text
stringlengths 66
377k
|
|---|
A novel long non-coding RNA AC073352.1 promotes metastasis and angiogenesis via interacting with YBX1 in breast cancer
Breast cancer is the major cause of cancer death worldwide in women. Patients with metastasis have poor prognosis and the mechanisms of breast cancer metastasis are not completely understood. Long non-coding RNAs (lncRNAs) have been shown to have crucial roles in breast cancer development and progression. However, the underlying mechanisms by which lncRNA-driven breast cancer metastasis are unknown. The main objective of this paper is to explore a functional lncRNA and its mechanisms in breast cancer. Here we identified a novel lncRNA AC073352.1 that was significantly upregulated in breast cancer tissues and was associated with advanced TNM stages and poor prognosis in breast cancer patients. In addition, AC073352.1 was found to promote the migration and invasion of breast cancer cells in vitro and enhance breast cancer metastasis in vivo. Mechanistically, we elucidated that AC073352.1 interacted with YBX1 and stabilized its protein expression. Knock down of YBX1 reduced breast cancer cell migration and invasion and could partially reverse the stimulative effects of AC073352.1 overexpressed on breast cancer metastasis. Moreover, AC073352.1 might be packaged into exosomes by binding to YBX1 in breast cancer cells resulting in angiogenesis. Collectively, our results demonstrated that AC073352.1 promoted breast cancer metastasis and angiogenesis via binding YBX1, and it could serve as a promising, novel biomarker for prognosis and a therapeutic target in breast cancer.Cell Death and Disease (2021) 12:670 ; https://doi.
# Introduction
Breast cancer (BC) is the most common malignancy and the leading cause of cancer-related death in women worldwide [bib_ref] Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for..., Sung [/bib_ref]. Of late, the age-standardized morbidity and mortality rates in those with BC have been rising in China [bib_ref] Breast cancer in China, Fan [/bib_ref]. Although BC is a molecularly heterogeneous multi-factorial disease [bib_ref] Breast cancer risk factors and mammographic density among high-risk women in urban..., Sung [/bib_ref] , studies have indicated that mortality levels are closely related to metastasis, which occurs in >90% of deaths [bib_ref] A perspective on cancer cell metastasis, Chaffer [/bib_ref]. It is known that tumor metastasis is a complex and multistep process, mainly involving migration, invasion, epithelial-mesenchymal transition, and angiogenesis, as well as other processes [bib_ref] Emerging biological principles of metastasis, Lambert [/bib_ref] [bib_ref] Tumor microenvironment participates in metastasis of pancreatic cancer, Ren [/bib_ref]. However, the mechanisms of BC metastasis are still not fully understood. Therefore, it is of great importance that the underlying molecular mechanisms of pathogenesis in BC metastasis are studied and that novel therapeutic targets are identified to improve the prognosis of BC.
Long non-coding RNAs (lncRNAs) are a class of non-coding RNAs that are >200 nucleotides in length [bib_ref] The bright side of dark matter: lncRNAs in cancer, Evans [/bib_ref] [bib_ref] Molecular mechanisms of long noncoding RNAs, Wang [/bib_ref]. New studies have shown that lncRNAs play a critical role in a wide range of cellular processes that could regulate multilevel gene expression, including transcription, posttranscriptional alternative splicing, protein translation and modification, protein transport, localization, and more [bib_ref] Unique features of long non-coding RNA biogenesis and function, Quinn [/bib_ref] [bib_ref] Noncoding RNA NORAD regulates genomic stability by sequestering PUMILIO proteins, Lee [/bib_ref] [bib_ref] Long noncoding RNAs: functional surprises from the RNA world, Wilusz [/bib_ref] [bib_ref] Adam12 and lnc015192 act as ceRNAs in breast cancer by regulating miR-34a, Huang [/bib_ref] [bib_ref] The STAT3-binding long noncoding RNA lnc-DC controls human dendritic cell differentiation, Wang [/bib_ref]. Abnormal expression of lncRNAs is closely related to the occurrence, development, metastasis, and drug resistance of malignant tumors, including BC [bib_ref] LncRNA UCA1 promotes tumor metastasis by inducing miR-203/ZEB2 axis in gastric cancer, Gong [/bib_ref] [bib_ref] lncRNA HOTAIR contributes to 5FU resistance through suppressing miR-218 and activating NF-kappaB/TS..., Li [/bib_ref] [bib_ref] The molecular mechanism of LncRNA34a-mediated regulation of bone metastasis in hepatocellular carcinoma, Zhang [/bib_ref]. For instance, found that a lncRNA, RAB11B-AS1, promoted hypoxia-mediated angiogenesis and BC metastasis driven by hypoxia-inducible factor-2 [bib_ref] HIF2-induced long noncoding RNA RAB11B-AS1 promotes hypoxia-mediated angiogenesis and breast cancer metastasis, Niu [/bib_ref]. Furthermore, Liang et al. demonstrated that a novel lncRNA, BCRT1, promoted BC progression by activating the miR-1303/PTBP3 axis [bib_ref] LncRNA BCRT1 promotes breast cancer progression by targeting miR-1303/PTBP3 axis, Liang [/bib_ref]. Moreover, recent reports have revealed that lncRNAs can also be packaged into exosomes to participate in intercellular communication in BC metastasis and progression [bib_ref] Ovarian cancer cell-secreted exosomal miR-205 promotes metastasis by inducing angiogenesis, He [/bib_ref] [bib_ref] Exosomal long noncoding RNA LNMAT2 promotes lymphatic metastasis in bladder cancer, Chen [/bib_ref]. Thus, such studies suggested that lncRNA may be critical in the pathogenesis of BC and provide new molecular targets for BC treatment and prognosis.
In this present study, we intended to investigate function and related mechanisms of lncRNAs in BC metastasis. We identified that a novel lncRNA, AC073352.1, was significantly upregulated in BC tissues. High expression levels of AC073352.1 were associated with poor prognosis in patients with BC. Functionally, we found that AC073352.1 promoted BC metastasis and invasion in vitro and in vivo. We further explored the potential molecular mechanism and revealed that AC073352.1 directly bound to the YBX1 to stabilize its protein, which led to the induction of BC metastasis. Additionally, we found that AC073352.1 could be packaged into exosomes and induce angiogenesis. Together, these findings demonstrated that AC073352.1 exerted tumorigenic potential and may be a useful prognosis biomarker and therapeutic target in BC.
## Materials and methods bc patient samples, cell lines, and cell culture
All human samples were obtained surgically from BC patients from whom we received informed consent at the Qilu Hospital of Shandong University. The selected patients had not received preoperative chemotherapy or radiotherapy. Fresh breast tumor and adjacent normal tissues were frozen in liquid nitrogen and stored at −80°C before being used for microarray analysis. Our study was approved by the ethics committee of Qilu Hospital of Shandong University.
Human BC cell lines (MDA-MB-231, MCF-7, BT549, MDA-MB-468, and HCC1937), a normal human mammary epithelial cell line (MCF-10A), HEK293T, and human umbilical vein endothelial cells (HUVECs) were purchased from the Cell Bank, Type Culture Collection Committee, Chinese Academy of Sciences (Shanghai, China). Cell lines were maintained using standard media and conditions. MDA-MB-231, MCF-7, MDA-MB-468, HEK293T, and HUVECs were cultured in Dulbecco's modified Eagle's medium (DMEM, Gibco) supplemented with 10% fetal bovine serum (FBS, Gibco). BT549 and HCC1937 were maintained in Roswell Park Memorial Institute (RPMI, Gibco) 1640 medium supplemented with 10% FBS. Finally, MCF-10A was grown in DMEM/F12 (Macgene) medium containing 5% horse serum, 100 ng/ml cholera toxin (Macgene), 20 ng/ml epidermal growth factor, 0.5 µg/ml hydrocortisone (Macgene), and 10 μg/ml insulin (Macgene). All cell lines were grown at 37°C in a 5% CO 2 cell culture incubator.
# Lncrna microarray analysis
Total RNA of six paired sets of tissue from BC and adjacent normal tissues were extracted using TRIzol (Life Technologies, USA) and purified using the RNeasy Mini Kit (Qiagen, Germany). RNA quality and quantity were measured by an Agilent Bioanalyzer 2100. The microarray analysis was carried out by Agilent technologies Inc. at Sinotech Genomics Corporation. Briefly, total RNA was checked for RNA integration by the Agilent Bioanalyzer 2100 (Agilent technologies, USA), and RNA samples were synthesized to biotinylated cRNAs for the Sino Human ceRNA array V3.0. Next, the biotinylated cRNAs were hybridized with the slides and the processed slides were scanned with an Agilent Microarray Scanner (Agilent technologies, USA). The acquired array images were analyzed by the Feature Extraction software 10.7 (Agilent technologies, USA). Finally, raw data and heatmaps were produced using the R software package, and the aberrant genes were selected by fold change >2 or <0.5.
## Tissue microarray (tma) and in situ hybridization (ish)
The expression level of AC073352.1 in tissues was detected by ISH using a specific digoxigenin-labeled AC073352.1 probe on TMAs, which contained 137 paraffin-embedded BC samples and 67 paired adjacent normal samples. The tissues of BC patients were obtained with detailed clinicopathologic features, including age, gender, tumor-node-metastasis stage, etc. (Qutdo Biotech, Shanghai, China). All of these tissues were sliced at a thickness of 4 μm, and the diameter of each tissue core is 1.5 mm. Further, the quantitative scanning approach was taken to analyze the staining and expression of AC073352.1 using a Nikon microscope. The ISH score was calculated by multiplying the value of intensity of positive staining by the proportion of positively stained cells. An ISH score of <0.5 represented low expression, while ≥0.5 indicated high expression. The sequences for the probing of AC073352.1 are listed in .
## Small interfering rna (sirna) transfection
Three individual siRNAs were used to knock down AC073352.1 and YBX1, these were designed and synthesized by Gene Pharma (Shanghai, China), and the two with the best efficiency were selected for use intransient knockdown experiments. And siNC was used as a negative control. The siRNAs were transfected using lipofectamine 2000 (Invitrogen, USA) according to the manufacturer's instructions. The sequences of the siRNAs are listed in .
## Lentivirus packaging and infection
For overexpression, full AC073352.1 cDNA was inserted into the lentivirus expression vector pLent-EF1a-FH-CMV-GFP-P2A-puro, and an empty vector was used for a transfection control group. For knockdown, a lentivirus short hairpin RNA expression vector targeting the same sequence as siAC073352.1-1 was constructed by Obio Technology (Shanghai, China). Next, HEK293T packaging cells were transfected using a retroviral mechanism. Culture supernatants were harvested at 72 h after transfection and infected of BC cells. Stably transfected cells were established using 1 μg/ml puromycin treatment for 2 weeks and were bulk cultured for subsequent assays.
Quantitative reverse transcription-polymerase chain reaction (qRT-PCR)
Total RNA was isolated from cultured cells using TRIzol (Invitrogen, USA) according to the manufacturer's instructions. First-strand cDNA was synthesized using the PrimeScript RT Reagent Kit (Takara, Dalian, China). Real-time PCR analyses were performed with SYBR Premix Ex Taq (Takara). The sequences for the gene-specific primers used are listed in Supplementary . β-Actin was used as an internal control.
## Cell migration and invasion assays
Cell migration and invasion assays were performed using a 24-well plate and 8-μm pore size chamber inserts (Corning, USA), which were precoated with or without 50 μl Matrigel (BD Biosciences, USA). In both assays, cells harvested in 200 μl serum-free media were seeded into the upper chamber and 600 μl medium supplemented with 20% FBS was added to the lower chamber. For migration assays, 5 × 10 4 cells were seeded into the upper chamber without Matrigel. In addition, for invasion assays, 1 × 10 5 cells were suspended in the upper chamber with Matrigel. After 36 h of incubation at 37°C and 5% CO 2 , the cells on the lower surface of the membrane were fixed with 4% paraformaldehyde and stained by Giemsa's stain. The images were captured with an inversion microscope (Zeiss, Germany) and counting was undertaken using the ImageJ software.
## Wound-healing assay
To investigate their metastatic ability, cells (2 × 10 5 per well) were seeded in 24-well plates and cultured for 12 h until the confluency of monolayers reached 90-100%. Further, a 200-μl sterile pipette tip was used to scratch across the center of each well before the medium was replaced by new medium supplemented with 2% FBS. After incubating for 0 and 24 h, images of cells in the wells were captured by an inversion microscope (Zeiss, Germany). The wound area was calculated to estimate the cell's migration efficacy.
## Cell viability assay
Cell proliferation efficacy was determined using the Cell Counting Kit 8 (CCK-8; Bestbio, Shanghai, China). A total of 3000 cells with corresponding treatment were seeded in 96-well plates. Cell viability was measured every 24 h, whereby 10 μl of CCK-8 solution was used for incubation for 2 h at 37°C and the spectrophotometric absorbance at OD450 was determined by the SpectraMax i3X (Molecular Devices, USA).
## Edu (5-ethynyl-2′-deoxyuridine) incorporation assay
The transfected BC cell lines were incubated in 96-well plates for 24 h, and the EdU Kit (RiboBio, Guangzhou, China) was used following the manufacturer's instructions. DAPI (4′,6-diamidino-2-phenylindole) was applied to stain cell nuclei (Solarbio, China), and images were obtained to identify positively proliferative cells by use of a fluorescence microscope (Zeiss, Germany).
## Rna fluorescence in situ hybridization
An AC073352.1 probe was obtained from Shanghai Qutdo Biotech. RNA fluorescence in situ hybridization (FISH) for cells on coverslips was performed following the manufacturer's protocol. Briefly, the cells were washed, fixed, and treated by 0.5% Triton. Next, cells on coverslips were incubated with each specific probe overnight before the slides were stained with DAPI. Finally, fluorescence images were captured by a confocal microscope (Zeiss, Germany). The sequences of the AC073352.1 probe for FISH is listed in using the Enhanced BCA Protein Assay Kit (Beyotime). Proteins were denatured at 100°C for 10 min, subjected to sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE), and transferred to 0.22 mm polyvinylidene fluoride membranes (Merck-Millipore, Darmstadt, Germany). The membranes were blocked using 5% non-fat milk and 1% Tween 20 in phosphate-buffered saline (PBS) for 1 h and incubated with primary antibodies at 4°C overnight. The membranes were subsequently washed and incubated with the appropriate secondary antibodies at room temperature. Finally, the signals were detected by standard analysis of HRPO-induced chemiluminescence. Details of the antibodies used in western blot are listed in .
## Rna pull-down assay
In vitro, full-length sequences of AC073352.1 were transcribed from their corresponding plasmids using Riboprobe Combination Systems (Promega, USA) before being purified using the RNAclean Kit (Qiagen, DP412) and labeled with desthiobiotinylation overnight (Magnetic RNA-Protein Pull-Down Kit, Thermo). Next, 1 mg protein extracted from MDA-MB-231 cells was incubated with biotin-labeled RNA and streptavidin magnetic beads for 6-12 h at 4°C. After elution of lncRNA-associated proteins, they were separated by SDS-PAGE and visualized by silver staining. Furthermore, the retrieved proteins were subjected to mass spectrometry (MS) analysis (Novogene, Tianjin, China).
## Rna immunoprecipitation (rip) assay
A RIP assay was performed by using the Imprint RNA Immunoprecipitation Kit (Millipore, USA) according to the manufacturer's instructions. In brief, 1 × 10 7 MDA-MB-231 cells were lysed with RIP lysis buffer, and the lysates were incubated with magnetic beads conjugated to 5 μg anti-YBX1 (Abcam), anti-SNRP70 (positive controls) or anti-lgG (negative controls) for 3-6 h at 4°C. Immunoprecipitated RNA was eluted, purified, and dissolved in RNase-free water, which was further measured through qRT-PCR analysis of YBX1. Specific primers are listed in .
## Exosomes isolation, characterization, and treatment
Exosomes were isolated from BC cell supernatants by ultracentrifugation. BC cells were initially cultured in DMEM medium supplemented with 10% exosome-free FBS. Cell supernatants were collected after 60 h and centrifuged at 300 × g for 10 min, 2000 × g for 10 min, and 10,000 × g for 30 min, all at 4°C. The obtained supernatant was filtered through 0.22-μm filter (Millipore) followed by ultracentrifugation at 110,000 × g for 90 min at 4°C. The supernatant was discarded, and exosomes pellets were washed with PBS before a second ultracentrifugation was performed at 110,000 × g for 90 min at 4°C.
The purified exosomes were subjected to the following experiments: Morphological images of exosomes were analyzed using transmission election microscopy (TEM) at 100 keV, and the size distribution of exosomes was analyzed by the nanoparticle tracking analysis (NTA) software.
To monitor exosome trafficking, a PKH67 Green Fluorescent Cell Linker Kit (Sigma Aldrich, St Louis, USA) was used to label exosomes according to the manufacturer's instruction. The staining was terminated by addition of 1% bovine serum albumin. The PKH67-labeled exosomes were washed in PBS, precipitated by ultracentrifugation, and resuspended in PBS. Next, exosomes were added to HUVECs and incubated for 12 h. The nuclei were stained with DAPI, and images were obtained by using a Zeiss confocal microscope system.
## Endothelial tube-formation assay
HUVECs were treated with exosomes (MDA-MB-231-vector or MDA-MB-231-AC073352.1) for 12 h. Then pretreated HUVECs were seeded onto a 96well plate coated with reduced growth factor Matrigel (BD Biosciences) and incubated at 37°C in 5% CO 2 for 6 h. Tube-like structures formed by the HUVECs were imaged using a light microscope, and total junction numbers were calculated automatically using the angiogenesis analyzer plugin in the ImageJ software.
## Xenograft model
Five-week-old female BALB/c nude mice were purchased from Weitonglihua (Peking, China) and housed in pathogen-free conditions. For experimental lung metastases assays in the xenograft model, MDA-MB-468 cells were stably transfected with shAC073352.1 or shNC lentiviruses, and a total of 1 × 10 6 cells were suspended in 0.1 ml sterile PBS before being injected into mice via the lateral tail vein (n = 7 in each group). After 6 weeks, these mice were sacrificed by cervical dislocation and the resected lung tissues were harvested, fixed, and embedded in paraffin. Subsequently, the number of metastatic foci in the lung was determined using hematoxylin and eosin (HE) staining. All animal experiments were performed in accordance with safe guidelines and regulations, and protocols were approved by the Institutional Animal Care and the Committee of the Second Hospital of Shandong University.
# Statistical analysis
All the measurement data analyses were performed using GraphPad Prism V6.0 (GraphPad prism, Inc., La Jolla, CA, United States) and SPSS V19.0 (SPSS, Chicago, IL, United States). The significance of differences between groups were assessed using Student's t test or χ 2 test as appropriate. The results are presented as mean ± standard error of mean (SEM) of three independent experiments. A p value < 0.05 was considered to be statistically significant.
# Results
## Ac073352.1 expression is upregulated in bc and associated with poor outcomes
To identify differentially expressed lncRNAs in BC, we performed a lncRNA microarray analysis of six BC tissues and paired adjacent normal breast tissues and analyzed The Cancer Genome Atlas (TCGA) database. Twenty-eight overlapping differentially expressed lncRNAs were found, according to the criteria of a fold change >2 or <0.5 and a p < 0.001. Among them, ten candidate lncRNAs were significantly associated with the outcome of BC patients . Especially, patients with a high AC073352.1 expression had a markedly poorer overall survival compared to those with low AC073352.1 expression; from this, AC073352.1 was selected for further experiments [fig_ref] Figure 1: AC073352 [/fig_ref].
To further validate the clinical significance of AC073352.1 expression, AC073352.1 expression level was detected in 137 BC samples and 67 adjacent normal tissues using ISH. The results showed that AC073352.1 was more highly expressed in BC tissues compared to adjacent normal tissues [fig_ref] Figure 1: AC073352 [/fig_ref]. The relationship between AC073352.1 expression and clinical pathological factors was further explored in 137 BC patients. As shown in Supplementary , greater AC073352.1 expression positively correlated with a high pathological tumor stage (p = 0.026) and increased lymph node (LN) metastasis (p = 0.006). In summary, this data suggested that AC073352.1 is highly expressed in BC and may serve as a potential prognostic biomarker in the clinic.
To characterize AC073352.1, its sequence and coding potential were comprehensively studied. AC073352.1 (ENSG00000272662.1) is located on chromosome 3q33.13 and it has one 504-bp length transcript (for UCSC Genome browser: http://genome.ucsc.edu/; [fig_ref] Figure 1: AC073352 [/fig_ref]. The ORF Finder (>300) and Conserved Domain Database from the National Center for Biotechnology Information failed to predict a protein for AC073352.1 (Supplementary [fig_ref] Figure 1: AC073352 [/fig_ref]. Further investigation using the Coding Potential Assessment Tool and Coding Potential Calculator revealed that AC073352.1 had no coding capability (http://lilab. research.bcm.edu/cpat/index.php and http://cpc2.cbi.pku.edu.cn/; [fig_ref] Figure 1: AC073352 [/fig_ref]. [fig_ref] Figure 2: AC073352 [/fig_ref]. The results of transwell assays showed that AC073352.1 knockdown significantly inhibited the migration and invasion b Kaplan-Meier analyses of the correlation between AC073352.1 RNA levels and the overall survival rate of patients with BC, using TCGA data. Patients were stratified for the analysis using the median. c-e ISH analysis detecting AC073352.1 expression (red) in BC tissue (n = 137) and adjacent normal tissues (n = 67). Representative images from c two BC cases and d adjacent normal tissue are shown (magnification: ×200). e Statistical analysis was assessed by t test. ***p value < 0.001.
## Ac073352.1 enhances the migration and invasion capabilities of bc cells in vitro
capabilities of MDA-MB-468 and MCF-7 cells [fig_ref] Figure 2: AC073352 [/fig_ref] ; in contrast, the overexpression of AC073352.1 substantially enhanced the migration and invasion capabilities of MDA-MB-231 and BT549 cells [fig_ref] Figure 2: AC073352 [/fig_ref]. Similarly, the migration capability of MDA-MB-468 and MCF-7 cells in wound healing assays was markedly decreased after the downregulation of AC073352.1 [fig_ref] Figure 2: AC073352 [/fig_ref] and Supplementary [fig_ref] Figure 2: AC073352 [/fig_ref] , the opposite results were found in MDA-MB-231 and BT549 cells that overexpressed AC073352.1 [fig_ref] Figure 2: AC073352 [/fig_ref] and 1s. e, g Transwell assays were used to investigate the changes in the migratory and invasive capabilities of MDA-MB-231 and BT549 cells after transfection with pcDNA-AC073352.1. h, i Wound healing assay assessing BC cells' metastasis. The data represent the mean ± SEM from three independent experiments. Values are expressed as mean ± SEM, n = 3. *p < 0.05, **p < 0.01, ***p < 0.001, and ns not significant. [fig_ref] Figure 2: AC073352 [/fig_ref]. Neither the CCK-8 assay nor the EdU assay showed that the knockdown and overexpression of AC073352.1 had a significant effect on proliferation of MCF-7 or MDA-MB-231 cells, respectively. [fig_ref] Figure 2: AC073352 [/fig_ref].
The above results indicated that AC073352.1 expression improves the migration and invasion capabilities of BC cells in vitro.
## Silencing ac073352.1 inhibits bc metastasis in vivo
To further assess the role of AC073352.1 in BC metastasis in vivo, MDA-MB-468 cells were stably transferred with shAC073352.1 or shNC and injected into mouse tail veins to build a metastatic mice model. Metastatic nodules in mice injected with MDA-MB-468 cells were dissected out for HE analysis [fig_ref] Figure 3: AC073352 [/fig_ref]. Compared to the control group, the knockdown of AC073352.1 significantly decreased the number of pulmonary metastatic nodules [fig_ref] Figure 3: AC073352 [/fig_ref]. Metastatic rates were also reduced in the shAC073352.1 group [fig_ref] Figure 3: AC073352 [/fig_ref]. Therefore, this data showed that silencing AC073352.1 significantly impairs the metastasis of BC cells to the lung in vivo.
## Ac073352.1 directly binds to ybx1
LncRNAs have been reported to exert their function by interacting with proteins during cancer progression [bib_ref] JAK2-binding long noncoding RNA promotes breast cancer brain metastasis, Wang [/bib_ref]. It was hypothesized that AC073352.1 promotes BC metastasis through protein binding. To confirm this, an RNA FISH assay was performed on BC cells using an AC073352.1 probe [fig_ref] Figure 4: AC073352 [/fig_ref]. Consistent with the ISH [fig_ref] Figure 1: AC073352 [/fig_ref] , the results of this experiment found that the lncRNA exists both in the nucleus and cytoplasm of BC cells. It was mainly distributed in the cytoplasm of the BC cells.
Next, the endogenous binding proteins of AC073352.1 in BC cells were identified using RNA pull-down assays; bound protein was revealed by sliver staining and MS [fig_ref] Figure 4: AC073352 [/fig_ref]. This analysis found 329 proteins that interacted with AC073352.1, including splicing factors and RNA-binding proteins. Among these binding proteins, proteins that correspond to the top ten unique peptides were selected; from this, it was found that YBX1 had the highest levels of interaction with this highest score. This result indicated that AC073352.1 may bind to the transcriptional activator YBX1 (Supplementary and [fig_ref] Figure 3: AC073352 [/fig_ref]. Following this, three independent RNA pull-down complexes were performed, and the results were detected by a western blot analysis using an YBX1 antibody. The results showed that a strong YBX1 signal was observed with the labeled AC073352.1 RNA; this result was not seen when using antisense RNA [fig_ref] Figure 4: AC073352 [/fig_ref]. A RIP assay using the YBX1 antibody also detected the RNA-YBX1 complex and confirmed that YBX1 had significant interaction with AC073352.1 [fig_ref] Figure 4: AC073352 [/fig_ref]. Furthermore, FISH and immunostaining showed that AC073352.1 and YBX1 co-located in BC cells [fig_ref] Figure 4: AC073352 [/fig_ref] , demonstrating that AC073352.1 interacted with the YBX1. These results together indicated that YBX1 specifically interacts with AC073352.1 in BC cells.
## Ac073352.1 stabilizes the protein level of ybx1
Next, qRT-PCR and western blot assays were used to detect whether AC073352.1 regulated the expression of YBX1. Interestingly, it was found that AC077352.1 did not change YBX1 mRNA levels but significantly affected the protein levels. AC073352.1 depletion dramatically reduced the protein expression of YBX1, whereas ectopic overexpression of AC073352.1 elevated YBX1 protein levels in BC cells [fig_ref] Figure 5: AC073352 [/fig_ref]. In addition, we performed FISH and immunostaining experiments to assess [fig_ref] Figure 5: AC073352 [/fig_ref]. To further investigate how AC073352.1 affects YBX1 protein stability, YBX1 protein levels were examined in MDA-MB-231 cells overexpressing AC073352.1; these cells had been treated with the protein synthesis inhibitor cycloheximide. As shown in [fig_ref] Figure 5: AC073352 [/fig_ref] , e, the protein stability of YBX1 was increased c Western blot analysis of the specific association between YBX1 and biotinylated AC073352.1; results obtained from three independent streptavidin RNA pull-down assays. d RIP assays were performed using an anti-YBX1 antibody; lgG was used as negative control. Specific primers were used to detect AC073352.1 (n = 3). e The co-localization of AC073352.1 and YBX1 was assessed by FISH and immunofluorescence. Scale bar: 10 μm. ***p < 0.001. and the half-life was prolonged when AC073352.1 was upregulated in MDA-MB-231 cells. These results indicated that AC073352.1 affects YBX1 expression by increasing its protein stability.
## Ac073352.1 promotes migration and invasion of bc cells via ybx1
Existing studies have demonstrated that YBX1 may contribute to the progression of multiple cancers, especially in the context of promoting tumor metastasis [bib_ref] Long noncoding RNA HULC modulates the phosphorylation of YB-1 through serving as..., Li [/bib_ref] [bib_ref] YBX1 gene silencing inhibits migratory and invasive potential via CORO1C in breast..., Lim [/bib_ref] [bib_ref] Silencing Y-box binding protein-1 inhibits triple-negative breast cancer cell invasiveness via regulation..., Lim [/bib_ref]. Therefore, it was explored whether YBX1 may positively regulate this biological function in BC. We first detected YBX1 expression levels in BC cell lines; the results showed that YBX1 expression were higher in BC cell lines than normal cell line (MCF-10A) [fig_ref] Figure 3: AC073352 [/fig_ref]. When YBX1 was effectively silenced [fig_ref] Figure 3: AC073352 [/fig_ref] , the migration and invasion capabilities of BC cells was reduced [fig_ref] Figure 6: AC073352 [/fig_ref]. Moreover, rescue experiments were used to detect whether the interaction between YBX1 and AC073352.1 contributed to BC metastasis. Unsurprisingly, the knockdown of YBX1 partly reversed the AC073352.1-mediated increase of BC cell migration and invasion [fig_ref] Figure 6: AC073352 [/fig_ref]. Meanwhile, the expression of YBX1 was detected in rescue experiments [fig_ref] Figure 3: AC073352 [/fig_ref]. These results further verified that AC073352.1 enhances BC metastasis by interacting with YBX1.
Exosomal AC073352.1 regulates the angiogenesis of HUVECs via binding YBX1 As exosomes are key mediators of cell-cell communication in the promotion of cancer progression, this experiment aimed to explore the exosome-based mechanism of BC progression. To examine whether AC073352.1 exists in BC cells derived from exosomes, BC-derived exosomes were isolated and characterized. Exosomes purified from the cell culture supernatant of BC cells [fig_ref] Figure 7: Exosomal AC073352 [/fig_ref] exhibited a round-shaped morphology and a size ranging from 30 to 150 nm according to TEM and NTA [fig_ref] Figure 7: Exosomal AC073352 [/fig_ref]. A western blot assay further verified the presence of welldefined exosome markers CD63 and CD9 [fig_ref] Figure 7: Exosomal AC073352 [/fig_ref]. These results indicated that the exosomes were isolated successfully.
Additionally, to explore whether the expression of AC073352.1 in exosomes from stable BC cell lines was altered, qRT-PCR analysis was performed. As expected, when compared to the control group, the expression level of exosomal AC073352.1 was significantly higher in MDA-MB-231 cells overexpressing AC073352.1; these levels were significantly lower in MDA-MB-468 cells treated with shAC073352.1 [fig_ref] Figure 7: Exosomal AC073352 [/fig_ref]. Previous studies have shown that YBX1 could participate in the packaging of RNAs into exosomes [bib_ref] Broad role for YBX1 in defining the small noncoding RNA composition of..., Shurtleff [/bib_ref]. To clarify whether YBX1 is essential for the loading of AC073352.1 into exosomes, qRT-PCR was performed and it was found that silencing YBX1 decreased the expression levels of exosomal AC073352.1 [fig_ref] Figure 7: Exosomal AC073352 [/fig_ref].
Several studies have shown that tumor-derived exosomes can be internalized by recipient cells to achieve their function. Therefore, this experiment focused on "tumor cell-endothelial cell" communication, and it was hypothesized that exosomes derived from BC cells may be internalized by HUVECs, as well as playing a role in the angiogenesis of HUVECs. As expected, HUVECs exhibited an uptake of PKH67-labeled exosomes derived from MDA-MB-231 cells [fig_ref] Figure 7: Exosomal AC073352 [/fig_ref] , suggesting that exosomes derived from BC cells could be effectively internalized by HUVECs. Furthermore, tube-formation assays revealed that exosomes derived from AC073352.1-overexpressing MDA-MB-231 cells increased HUVEC angiogenesis compared to control group exosomes [fig_ref] Figure 7: Exosomal AC073352 [/fig_ref]. Taken together, these results suggested that lncRNA AC073352.1 is loaded into exosomes via YBX1 and promotes the angiogenic ability of HUVECs.
# Discussion
With the improvement of diagnosis and treatment methods, the overall survival rate of BC has been improved [bib_ref] Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for..., Sung [/bib_ref]. However, the prognosis for patients with metastases is still poor, with metastasis the major cause of death among BC patients [bib_ref] Targeting metastasis, Steeg [/bib_ref]. Hence, there remains a considerable need to explore the underlying molecular mechanisms of BC metastasis and identify novel molecular therapeutic markers. So far, a large number of studies have found that, just like protein-coding genes, lncRNAs act as important players in the development of disease; this has become especially evident in tumor metastasis, including BC. For example, Xiu et al. revealed that the lncRNA LINC02273 could promote BC metastasis via the hnRNPL-AGR2 axis [bib_ref] LINC02273 drives breast cancer metastasis by epigenetically increasing AGR2 transcription, Xiu [/bib_ref]. Zheng et al. demonstrated that a novel triple-negative BC LN metastasis-associated lncRNA, HUMT, can activate FOXK1 transcription to induce BC proliferation and metastasis [bib_ref] Long non-coding RNA HUMT hypomethylation promotes lymphangiogenesis and metastasis via activating FOXK1..., Zheng [/bib_ref]. In this study, AC073352.1, located at chromosome 3q33.13, was identified in a microarray analysis and TCGA database as one of the several different and novel lncRNAs in BC. The lncRNA AC073352.1 was significantly overexpressed in BC tissues and displayed a positive correlation with lymphatic metastasis and advanced disease in BC. It was important that the higher levels of AC073352.1 associated with a lower overall survival rate in BC patients. Additionally, with GTEx data, it was found that AC073352.1 expression was lower in mammary tissues, which further confirmed that AC073352.1 may play a critical role in BC progression [fig_ref] Figure 1: AC073352 [/fig_ref]. In support of this notion, further functional experiments were conducted in vitro and in vivo. The results showed that AC073352.1 promoted BC cell migration and invasion in vitro; in addition to this, silencing AC073352.1 markedly inhibited BC cell lung metastasis in vivo. Taken together, these results provide key evidence that AC073352.1 plays a critical role in BC metastasis and serves as a potential prognostic and therapeutic target for BC metastatic.
Multiple investigations have demonstrated that lncRNAs can perform their biological functions via protein interactions [bib_ref] LncRNA ANCR promotes hepatocellular carcinoma metastasis through upregulating HNRNPA1 expression, Wen [/bib_ref] [bib_ref] Revealing protein-lncRNA interaction, Ferre [/bib_ref]. In this study, RNA pull-down was utilized to identify that transcription factor YBX1 binds to AC073352.1 in BC. RIP and FISH assays further confirmed the binding of AC073352.1 and YBX1. Subsequently, a series of AC073352.1 deletion mapping analysis was detected to further map the specific YBX1-binding region. And we found that the YBX1-binding site may be mainly distributed in the 252-272 nucleotide region of AC073352.1. [fig_ref] Figure 3: AC073352 [/fig_ref]. More work is needed to prove this in the future. In addition, this study investigated the subcellular fractionation location of AC073352.1, with the results showing that AC073352.1 exists both in the nucleus and cytoplasm of BC cells and it was mainly distributed in BC cell cytoplasm. Growing evidence has indicated that cytoplasmic lncRNAs may regulate protein stability and modification [bib_ref] Long noncoding RNA GMAN, upregulated in gastric cancer tissues, is associated with..., Zhuo [/bib_ref] [bib_ref] A TRIM71 binding long noncoding RNA Trincr1 represses FGF/ERK signaling in embryonic..., Li [/bib_ref]. In consistency with this, it was found that AC073352.1 affected YBX1 protein expression without influencing mRNA levels. In addition, the overexpression of AC073352.1 increased the stability of YBX1 at a protein level. These results suggest that AC073352.1 interacts with YBX1 and stabilizes the protein. Previous studies have shown that lncRNAs regulate the stability of protein via diverse mechanisms, such as ubiquitination, phosphorylation, and other methods at a posttranscriptional level [bib_ref] Long noncoding RNA AGPG regulates PFKFB3-mediated tumor glycolytic reprogramming, Liu [/bib_ref] [bib_ref] MetaLnc9 facilitates lung cancer metastasis via a PGK1-Activated AKT/mTOR pathway, Yu [/bib_ref]. However, how AC073352.1 affects the protein stability of YBX1 needs to be further explored.
YBX1, a key transcription factor, is located on chromosome 1 (1p34), and as a multi-functional RNA-binding protein, it is highly overexpressed in a variety of tumors; it is established to play a role in several cellular processes, including tumor metastasis [bib_ref] Silencing of long noncoding RNA MIR22HG triggers cell survival/death signaling via oncogenes..., Su [/bib_ref] [bib_ref] Endogenous tRNAderived fragments suppress breast cancer progression via YBX1 displacement, Goodarzi [/bib_ref] [bib_ref] The long noncoding RNA LINC00312 induces lung adenocarcinoma migration and vasculogenic mimicry..., Peng [/bib_ref] [bib_ref] Nuclear expression of Y-box binding protein-1 is associated with poor prognosis in..., Shinkai [/bib_ref] [bib_ref] The RNA-binding protein YBX1 regulates epidermal progenitors at a posttranscriptional level, Kwon [/bib_ref]. To explore whether YBX1 affects cell metastasis in BC, the YBX1 protein was knocked down using siRNAs. The results of this experiment found that silencing YBX1 inhibits BC cell migration and invasion; this may partially rescue AC073352.1-induced BC metastasis. Thus, our results indicated that AC073352.1 mediates cell metastasis by interacting with YBX1 in BC. In recent years, researchers paid special attention that microcalcifications presented in the breast microenvironment and osteoclast-like cells played an important role in BC bone metastases [bib_ref] Breast cancer and microcalcifications: an osteoimmunological disorder?, Clemenceau [/bib_ref] [bib_ref] Microcalcifications in breast cancer: novel insights into the molecular mechanism and functional..., Cox [/bib_ref]. Previous studies have reported that dysregulated YBX1 is associated with chordoma, osteosarcoma, synovial sarcoma, and other bone diseases [bib_ref] Y-box binding protein-1 promotes tumorigenesis and progression via the epidermal growth factor..., Liang [/bib_ref] [bib_ref] The Y-box binding protein YB-1 is associated with progressive disease and mediates..., Chatterjee [/bib_ref] [bib_ref] Nuclear expression of Y-box-binding protein-1 correlates with P-glycoprotein and topoisomerase II alpha..., Oda [/bib_ref] [bib_ref] Nuclear expression of YB-1 protein correlates with P-glycoprotein expression in human osteosarcoma, Oda [/bib_ref]. Hence, it is a suppose that AC073352.1 may create a favorable condition for bone metastasis by YBX1 in BC. Further studies are needed to identify the detailed mechanisms. It was noted that lncRNA may be directly and indirectly related to gene regulation in BC; in addition to YBX1, other genes could participate in AC073352.1associated biological function; however, more evidence is needed to confirm this.
Exosomes are a class of small macrovesicles ranging from 30 to 150 nm and are a critical means of exchanging intercellular information that have aroused great research interest [bib_ref] Exosome-mediated secretion of LOXL4 promotes hepatocellular carcinoma cell invasion and metastasis, Li [/bib_ref] [bib_ref] CircLONP2 enhances colorectal carcinoma invasion and metastasis through modulating the maturation and..., Han [/bib_ref]. Interestingly, several studies suggest that YBX1 is an important RNA-binding protein that can be secreted into the extracellular matrix, including tumor cell exosomes [bib_ref] YBX1 at the crossroads of non-coding transcriptome, exosomal, and cytoplasmic granular signaling, Suresh [/bib_ref] [bib_ref] Y-box protein 1 is required to sort microRNAs into exosomes in cells..., Shurtleff [/bib_ref]. Therefore, this investigation tried to explore whether AC073352.1 is incorporated into exosomes by binding YBX1. The results showed that knocking down YBX1 decreased the level of AC073352.1, which indicated that YBX1 may aid in the loading of AC073352.1 into exosomes. Further results showed that exosomes isolated from BC cell supernatant could be internalized by HUVECs and exosomal AC073352.1 promoted angiogenesis. As a point of great interest, this may be another key way that AC073352.1 promotes BC metastasis via YBX1 binding. This data strongly suggests that AC073352.1 may act not only as an intracellular lncRNA in BC metastasis but also as an exosomal lncRNA; this will aid in the identification of novel therapeutic targets for BC.
In conclusion, this study identified and characterized a novel lncRNA, AC073352.1, which was significantly overexpressed in BC tissues and associated with poor survival in BC patients. It was also shown that AC073352.1 promotes BC cell metastasis by physically interacting with YBX1 and stabilizing YBX1 at a protein level in BC cells. Moreover, AC073352.1 was transferred into exosomes via YBX1 binding, while exogenous AC073352.1 promotes angiogenesis. These findings illustrate that AC073352.1 promotes BC cell metastasis by functioning within cells and changing the tumor microenvironment . Collectively, this data offers a mechanistic focus on the oncogenic roles of AC073352.1 and it may serve as a prognostic and therapeutic biomarker for BC. A schematic diagram of the proposed mechanisms. A schematic illustration for the mechanisms by which AC073352.1 interacts with YBX1 to promote BC metastasis and how extracellular AC073352.1 from breast cancer cells was packaged into exosomes and induced HUVEC angiogenesis.
[fig] Figure 1: AC073352.1 expression is upregulated and associated with poor outcomes in BC. a A flowchart showing the selection of lncRNA AC073352.1 from the LncRNA microarray and TCGA database in BC. A heatmap representation of 10 overlapping lncRNA for gene expression profiles of 6 BC tissues and paired adjacent normal tissues in LncRNA microarray (fold change >2 or <0.5, p < 0.001). [/fig]
[fig] Figure 2: AC073352.1 regulates Breast Cancer cells metastasis in vitro. a Expression of AC0733521.1 in the normal breast epithelium cell line (MCF-10A) and BC cell lines was detected by qRT-PCR and normalized to β-actin. b, c qRT-PCR analysis of AC073352.1 expression in control, siAC073352.-1-, siAC073352.-2-, siAC073352.-3-, and pcDNA-AC073352.1-treated BC cells. d, f Transwell assays were used to investigate the changes in the migratory and invasive capabilities of MDA-MB-468 and MCF-7 cells after transfection with siAC073352. [/fig]
[fig] Figure 3: AC073352.1 accelerates breast cancer metastasis in vivo. a The representative images of lung metastasis of mice. Hematoxylin and eosin (HE) staining of lungs were removed from mice injected with shNC or shAC073352.1 stable cell line. b Metastasis nodules were counted and analyzed using Student t test. Values are expressed as mean ± SEM. c The ratio of lung metastasis was calculated for the indicated groups. The p value was determined using χ 2 test. *p < 0.05 and **p < 0.01.the expression of YBX1 in BC cells. Consistently, the results showed that the expression level of YBX1 was significantly downregulated after AC073352.1 siRNA treatment in MDA-MB-468 cells, and YBX1 expression was upregulated in MDA-MB-231 cells with AC073352.1 overexpression [/fig]
[fig] Figure 4: AC073352.1 binds to the transcriptional activator YBX1. a AC073352.1 expression in MDA-MB-231 and MDA-MB-468 cells was detected by RNA FISH. Scale bars: 20 μm. b AC073352.1 sense and antisense RNAs were used for in vitro transcribed and pull-down assays in MDA-MB-231 cells. After electrophoresis and silver staining, the different bans located at 45-60 kDa. [/fig]
[fig] Figure 5: AC073352.1 affects the protein level of YBX1. a Western blots analysis showing the protein levels of YBX1 following the knockdown or overexpression of AC073352.1 in BC cells. b qRT-PCR analysis showing the mRNA levels of YBX1 mRNA following the knockdown or overexpression of AC073352.1 in BC cells. c Expression of YBX1 and AC073352.1 in MDA-MB-468 cells transfected with siNC or AC073352.1 siRNAs and MDA-MB-231 cells transfected with vector or pcDNA-AC073352.1 was analyzed by FISH and immunofluorescence. Scale bar: 20 μm. d, e The stable overexpression of AC073352.1 in MDA-MB-231 cells treated with cycloheximide (CHX, 200 μg/ml) for the indicated times. d Western blot showing YBX1 levels in whole-cell extracts. e Densitometry analysis of YBX1 protein levels: the relative fold of the level at 0 h. [/fig]
[fig] Figure 6: AC073352.1 interacts with the transcriptional activator YBX1 to affect BC metastasis. a, b Transwell and wound-healing assays on MDA-MB-231 cells transfected with YBX1 siRNA. c Rescue assays following transwell assays revealed that the depletion of YBX1 partly reversed the effects of AC073352.1 overexpression in MDA-MB-231 cells. Values are expressed as mean ± SEM, n = 3. **p < 0.01, ***p < 0.001. [/fig]
[fig] Figure 7: Exosomal AC073352.1 regulates the angiogenesis of HUVECs via binding YBX1. a Flow chart for the exosome purification procedure based on differential ultracentrifugation. b TEM scanning images of purified exosomes derived from MDA-MB-231 cells overexpressing AC073352.1. Scale bars: 100 nm. c Characterization of purified exosomes derived from MDA-MB-231 cells overexpressing AC073352.1. The size, number, and distribution of exosomes was analyzed by NTA. d Exosomal protein marker (CD63 and CD9) detection by western blot from purified exosomes and exosome-depleted cell extracts. e qRT-PCR analysis of AC073352.1 levels in exosomes from AC073352.1overexpressing MDA-MB-231 cells and shAC073352.1-treated MDA-MB-468 cells. f qRT-PCR analysis of AC073352.1 levels: AC073352.1 levels were decreased in exosomes after MDA-MB-231 cells were transfected with a YBX1 siRNA. g HUVECs were incubated with PKH67-lableled exosomes for intercellular trafficking. Scale bars: 10 μm. h, i Representative images and quantification of tube-formation assays by HUVECs treated with MDA-MB-231 vector or MDA-MB-231 AC073352.1 . Scale bars: 100 μm. *p < 0.05 and **p < 0.01. [/fig]
|
Management Options for Extensor Mechanism Discontinuity in Patients With Total Knee Arthroplasty
# Introduction and background
The extensor mechanism of the knee joint is one of the most important structures encountered during total knee arthroplasty. It consists of the quadriceps tendon, patella, and patellar tendon. The importance of extensor mechanism is well-established in the field of knee replacement, not only in terms of minimizing intraoperative complications but also in early postoperative recovery and in attaining long-term goals. The structural and functional integrity of the extensor mechanism is crucial for optimum biomechanics of both the native and prosthetic knee. Restoration of the optimum Q-angle is necessary to obtain desirable patellar tracking and to minimize the chances of anterior knee pain.
One of the major steps during any knee replacement surgery is to retract the extensor mechanism in a way that minimizes damage to the structural and functional integrity of the extensor mechanism and, at the same time, to obtain adequate exposure to performing all the necessary procedures. All the contemporary approaches to the knee joint revolve around how to handle the extensor mechanism during the surgery and how to ensure the optimum function of the extensor mechanism postoperatively. The rationale presented by surgeons using subvastus and midvastus approaches is that they are less damaging to the extensor mechanism and hence help in faster recovery.
For revision cases, the biggest obstacle in attaining proper exposure is stiff and sometimes deficient extensor mechanism. Extensive approaches like the rectus snip, V-Y quadricepsplasty, and tibial tubercle osteotomy have been described in the literature for such cases. These extensile approaches help the surgeon to obtain adequate exposure in performing such complex procedures.
Extensor mechanism dysfunction in the recipients of total knee arthroplasty can result from various causes ranging from iatrogenic to post-traumatic disruption. This dysfunction can be acute, subacute, or chronic and can involve any of the anatomical components of the extensor mechanism. Sometimes extensor mechanism disruption is combined with other complications, such as infection, periprosthetic fracture, loose implants, and failed previous reconstructions. These coexisting conditions have profound implications on the management and outcomes of extensor mechanism repair. Adequate repair or reconstruction of the extensor mechanism is considered paramount for the success of knee replacement in such cases.
Based on all the aforementioned factors, there are numerous management strategies described in the literature to treat this rare but devastating complication . To date, there is no consensus amongst orthopedists regarding ideal treatment for extensor mechanism deficiency, especially in chronic cases. Treatment decisions made by surgeons largely depend upon individual training, experience, and preferences rather than evidence-based guidelines.
The purpose of this review is to summarize the existing literature regarding extensor mechanism reconstruction and provide the readers an evidence-based summary of etiology, diagnosis, and techniques for extensor mechanism reconstruction and outcomes of various strategies to treat this catastrophic condition.
## Figure 1: anatomical classification of extensor mechanism disruption
Patellar tendon disruptions are the most common and the most devastating form of extensor mechanism dysfunction, mainly because of the poor soft tissue coverage, lack of adequate blood supply, and resultant extremely high incidence of failures and considerable reoperation rates.
## Etiology
The etiology of the dysfunction has a great influence on the management and outcomes. Acute traumatic disruptions are relatively easier to treat since scarring and retraction of the extensor mechanism is minimal and primary repair is still an option in many such cases, while chronic or failed reconstructions are often associated with infection or loosening of components. These cases are less amenable to primary repair, and failure rates and complications are higher. Suprapatellar disruptions result from either a quadriceps tendon rupture or avulsion from the superior pole of the patella. It is a rare complication in comparison to patellar tendon rupture or avulsion. The estimated incidence is 0.1% amongst total knee replacement recipients. While trauma is often considered as a trigger event, there are some predisposing factors that increase the risk of rupture in individuals. Periprosthetic fracture of the patella is the second most common periprosthetic fracture around the knee joint after distal femur fracture. The estimated average incidence is 1.19%. It is one of the most difficult fractures to treat. It is worth noting that only about 12% of these fractures are associated with trauma. Most of these present at routine follow-up with anterior knee pain with or without extensor lag. Most of these fractures are the result of osteonecrosis of the patella. The etiology of almost all of these fractures is iatrogenic.
## Systemic risk factors local risk factors
There are some risk factors associated with periprosthetic patella fracture. The fact that 99% of these fractures happen in a resurfaced patella shows that resurfacing of the patella is a prominent risk factor. This phenomenon challenges the wisdom behind the practice of resurfacing every patella routinely during primary TKA. The residual thickness of the patella is a known contributor to the risk of fracture. Traditionally, at least 12 mm of the thickness of the residual patella is considered adequate to resurface the patella. This signifies the importance of measuring the thickness of the patella before proceeding with a bony resection. A lateral release to improve patella tracking is detrimental to the blood supply to the patella. This becomes more important in the setting of arthroplasty where the medial parapatellar arthrotomy has already jeopardized the medial blood supply. An additional release laterally can render the patella completely avascular and results in osteonecrosis of the patella and resultant stress fracture. All care must be taken to identify and protect the superior lateral genicular artery while performing lateral release. For reasons not clearly understood, cementless components are associated with an increased incidence of fractures. The components with a large central peg create a stress riser and increase the risk of fracture. Component malalignment alters the biomechanics of the knee joint, increases the stress on the prosthetic patellar component, and indirectly increases the chances of fracture.
These fractures most commonly occur in the first two years after surgery with an average time period being 18.5 months.
Ortiguerra and Berry classified periprosthetic patella fracture in three types. Their classification is based on the integrity of the extensor mechanism, the presence or absence of loosening of the component, and the remaining bone stock. This classification, with the relative incidence of individual types of fractures, is depicted in.
## Type description
Type I Intact extensor mechanism and stable implant (25.3%)
Type II Disruption of extensor mechanism with or without implant in place (20%)Most of the patellar tendon avulsions or ruptures can be considered iatrogenic, resulting from either direct injury during the surgery or component malalignment putting excessive stress on the patellar tendon during successive ambulation. Patellar tendon rupture or avulsion is one of the most dreaded complications associated with total knee arthroplasty. Regardless of the treatment modality used, the outcomes are unsatisfactory. To prevent these injuries, it is important to prevent excessive stress on the patellar tendon and, if needed, to perform adequate arthrolysis or to switch to extensile approaches, like V-Y quadricepsplasty or rectus snip, at the appropriate time to prevent rupture or avulsion of the patellar tendon. In some instances, nonunion of the tibial tubercle osteotomy used as an extensile approach can lead to patellar tendon insufficiency.
## Diagnosis
Regardless of the etiology and anatomic location, the diagnosis is mainly clinical. In many cases, the diagnosis is delayed or missed altogether due to a lack of suspicion. Patients with extensor mechanism dysfunction typically present with loss of active extension, instability, and quadriceps minus gait. Patients have difficulty in climbing stairs. A history of trauma is commonly present but should not be overemphasized since, in many cases, it cannot be directly linked to the clinical picture. Anterior knee pain is a common presentation in patients with patella fractures with or without extensor lag, depending upon the degree of displacement and continuity of the extensor mechanism. In many cases, especially with complete disruption of the extensor mechanism, a palpable defect virtually confirms the diagnosis .
## Figure 2: palpable defect in the infrapatellar region at the site of patellar tendon disruption
In cases of a complete patellar tendon disruption, the patella alta is also an important clinical finding.
Apart from the degree of extensor lag and instability, a thorough history and clinical examination also help in determining the future surgical approach, availability of local soft tissue and flaps, the blood supply to the region, the presence or absence of infection, the necessity of additional procedures (e.g., revision of one or more components), medical comorbidities, and probable outcomes.
The clinical diagnosis can be further validated by appropriate radiological studies. A plain x-ray is usually sufficient in this regard. Direct evidence of extensor mechanism disruption can be seen on plain x-ray in the form of patella alta (patellar tendon disruption), tibial tuberosity avulsion, patellar fracture, nonunion or displacement of tibial tubercle osteotomy, or an anteriorly displaced patella (quadriceps tendon disruption). Imaging studies confirm the diagnosis and provide important information about component alignment, fixation of the component, and remaining bone stock. Magnetic resonance imaging (MRI) with special sequences or ultrasound may be helpful in some selected cases but are not routinely required.
The optimum treatment strategy is devised based on all the inputs from the patient's history, physical examination, and imaging studies. The final treatment strategy not only depends upon location and degree of extensor mechanism disruption but also on preexisting medical conditions, previous attempts at reconstructions, and expected overall outcomes.
## Management
## Quadriceps tendon rupture or avulsion
Incomplete ruptures can be treated non-operatively with acceptable results. In patients with less than a 20-degree extensor lag, incomplete ruptures can be treated with immobilization of the knee in extension for four to six weeks, either in a cast or in a brace, followed by gradual mobilization and physical therapy. A study done by Dobbs et al. showed that 85% of cases with partial quadriceps tendon ruptures yielded satisfactory outcomes with nonoperative treatment. This is in contrast to the operative treatment of partial ruptures, which yields satisfactory outcomes in about 75% of cases. The same study shows satisfactory outcomes in only 40% of cases with complete ruptures. These results signify the importance of adequately protecting partial ruptures and thus preventing them from being converted to complete ruptures.
When extensor lag is more than 20 degrees, operative intervention is necessary. For acute ruptures, primary repair following thorough debridement of the severed ends is recommended. Typically, Krackow sutures are placed in the quadriceps tendon and it is attached to the patella using either drill holes or suture anchor. Biomechanically suture anchor repairs are considered more robust. For chronic ruptures, the primary repair is often impossible owing to the contracture of the surrounding tissue and retraction of the extensor mechanism. In such cases, autologous augmentation using the vastus medialis, vastus lateralis, and medial head of the gastrocnemius can be used. In most cases, local augmentation is not sufficient because of the compromised soft tissue envelope and vascularity. However, it is proven that when a repair is augmented with the medial head of the gastrocnemius with or without the medial part of the soleus extensor lag is less pronounced in comparison to vastus medialis and lateralis flap advancement alone. Results reported by Dobbs et al. are discouraging with a 35% reoperation rate and 12% infection rate. Forty percent of the primary suture repairs had reruptured. These results are in stark contrast with the generally good results reported in patients without knee replacement. These differences can probably be explained by a compromised soft tissue envelope, higher infection rates, comorbidities, and a relatively older patient population in knee replacement recipients. Chronic and severe deficiency of the quadriceps tendon can be managed by an Achilles tendon allograft with calcaneal tuberosity bone block, extensor mechanism allograft, or patellar tendon allograft. Recently, the use of Marlex mesh has also been popular, which is discussed in detail in the section of patellar tendon disruptions. Results of operative interventions of chronic rupture and resulting insufficiency are far less than satisfactory.
Even after surgical repair, immobilization in extension for a period of four to six weeks, followed by gradual mobilization, is recommended to protect the repair.
## Patellar fracture
While considering various treatment options, it should be clearly understood that these fractures are different than traumatic patellar fractures in a native knee. Outcomes of open reduction and internal fixation (ORIF) are poor in these patients. Reasons for this discrepancy in results lie in the fact that these fractures occur in the background of poor vascularity, low residual bone stock, and altered biomechanics .
## Figure 5: periprosthetic patellar fracture, lateral view
Patients with acceptable quadriceps function and well-fixed components can be treated conservatively in the same manner as a quadriceps tendon rupture. Non-operative treatment yields satisfactory results in most patients as per Keating et al.. Mean flexion achieved was around 120 degrees with less than 5 degrees of extensor lag and minimal pain in most of their patients.
Operative treatment is associated with a guarded prognosis. ORIF is associated with an 88% failure rate [10].
## Figure 6: failed open reduction and internal fixation
performed for periprosthetic patellar fracture (lateral view)
## Figure 7: anteroposterior view
This fact signifies the basic difference between a periprosthetic fracture of the patella and traumatic fractures of the native patella. This high failure rate is related to poor bone stock and vascularity. The high failure rate, in addition to the potential of hardware-related complications like breakage and prominent hardware, makes ORIF an inappropriate choice. Partial excision of the patellar fragment with or without repair of the extensor mechanism, as needed, is the mainstay of the operative treatment of patients with periprosthetic patellar who are symptomatic. In cases of a resurfaced patella, integrity and tracking of the patellar prosthesis should be addressed in addition to the extensor mechanism reconstruction. Whenever the patellar component is loose, it should be removed. Further reconstruction is based on available bone stock and the quality of the remaining bone. If an adequate bone stock is available (at least 12 mm), patellar resurfacing may be a viable option. Nowadays, with some newer designs (e.g., trabecular metal components), resurfacing may be considered in selected cases with lower bone stock.
## Patellar tendon rupture or avulsion
Partial disruptions with less than 30 degrees of extensor lag can be treated non-surgically in the same line as described in previous sections. Patients with more significant extensor lag should be treated surgically, provided there are no contraindications.
Surgical repair of this condition is extremely challenging because of the constant pull on the repair by the extensor mechanism and inadequate soft tissue coverage available locally.
Direct repair almost always fails and leads to unacceptable outcomes. The main reason for failure is the paucity of soft tissue to achieve adequate strength to withstand the forces during ambulation, which are concentrated on a relatively small area. The vascularity of the tendon is usually poor due to repetitive parapatellar arthrotomies performed in the past. In many cases, there is component malalignment responsible for excessive stress on the extensor mechanism. For these reasons, direct repair is rarely performed nowadays for chronic disruptions and some sort of augmentation is usually considered. Acute patellar tendon avulsions can be treated with either soft tissue repair or direct repair to the bone using drill holes or suture anchors. Acute mid-substance tears can be treated with end-to-end repair, usually augmented by the semimembranosus, gracilis, or Achilles tendon.depicts the studies related to local flap augmentation for extensor mechanism disruptions.
## Study
## Technique n results conclusion
Whiteside et al.Vastus medialis (VM) and vastus lateralis (VL) flap reconstruction with or without soleus or gastrocnemius (GN) Chronic patellar tendon ruptures usually require augmentation in the form of allograft, autograft, or various synthetic materials. Bone-patellar tendon-bone grafts, Achilles tendon allografts, and synthetic materials, such as polypropylene mash, have been tried to augment the repair with variable success.show a technique for repair of chronic patellar tendon rupture in a patient with distal femur replacement using bone-patellar tendon allograft. Synthetic material appears to be a potential alternative to allografts since they do not depend upon the availability of local muscle flaps, is more cost-effective, and yields similar results in an early study ( A chronic retracted patellar tendon with extensive scarring and soft tissue or bone loss can be treated with an extensor mechanism allograft. This procedure is technically challenging and is associated with high failure and complication rates. Some of the studies are listed inand available literature in this area. In a published study, eight cases of allograft failures were reoperated using allografts. Two out of the eight cases failed due to infection and the remaining six showed no clinical improvement. Repetitive surgeries further compromise the soft tissue envelope and increase the risk of infection. In refractory cases, knee fusion or amputation may be appropriate choices. The decision should be made only after due consideration of the patient's expectations, physiological condition, and anticipated outcomes.
# Conclusions
There are many techniques described in the literature to treat extensor mechanism discontinuity. Lack of unanimity between experts probably shows the complexity of this problem and the guarded prognosis of the condition.
In summary, non-operative treatment for partial disruptions at any anatomical location results in acceptable outcomes. For complete disruptions, treatment depends upon the location, extent, and chronicity of the disruption, the patient's overall health condition and expectations, and the experience and expertise of the operating surgeon.
Every surgeon should adopt the technique that works well in his or her hands. It is also important to discuss all the aspects of the condition (e.g., treatment options, nature of the treatment, rehabilitation, and possible outcomes) in detail with the patient to bring mutual expectations to a pragmatic level. |
The Interplay between PolyQ and Protein Context Delays Aggregation by Forming a Reservoir of Protofibrils
Polyglutamine (polyQ) diseases are inherited neurodegenerative disorders caused by the expansion of CAG codon repeats, which code for polyQ in the corresponding gene products. These diseases are associated with the presence of amyloid-like protein aggregates, induced by polyQ expansion. It has been suggested that the soluble aggregates rather than the mature fibrillar aggregates are the toxic species, and that the aggregation properties of polyQ can be strongly modulated by the surrounding protein context. To assess the importance of the protein carrier in polyQ aggregation, we have studied the misfolding pathway and the kinetics of aggregation of polyQ of lengths above (Q41) and below (Q22) the pathological threshold fused to the well-characterized protein carrier glutathione S-transferase (GST). This protein, chosen as a model system, is per se able to misfold and aggregate irreversibly, thus mimicking the behaviour of domains of naturally occurring polyQ proteins. We prove that, while it is generally accepted that the aggregation kinetics of polyQ depend on its length and are faster for longer polyQ tracts, the presence of GST alters the polyQ aggregation pathway and reverses this trend. Aggregation occurs through formation of a reservoir of soluble intermediates whose populations and kinetic stabilities increase with polyQ length. Our results provide a new model that explains the toxicity of expanded polyQ proteins, in which the interplay between polyQ regions and other aggregation-prone domains plays a key role in determining the aggregation pathway.
# Introduction
Polyglutamine (polyQ) diseases are caused by the expansion of CAG codon repeats resulting in extended polyQ tracts in the expressed proteins [bib_ref] Glutamine repeats and neurodegeneration, Zoghbi [/bib_ref]. This family of inherited neurodegenerative disorders includes Huntington's chorea, spinobulbar muscular atrophy, dentatorubral-pallidoluysian atrophy, and spinocerebellar ataxias (SCAs) 1, 2, 3, 6, 7, and 17. The polyQ region is the only common feature of the proteins associated to these diseases, that are otherwise totally unrelated [bib_ref] Molecular genetics: unmasking polyglutamine triggers in neurodegenerative disease, Gusella [/bib_ref]. In affected individuals, the polyQ tract is expanded above a threshold of ca. 35 consecutive glutamines, resulting in the aggregation of the mutant protein and the consequent formation of intranuclear inclusions [bib_ref] Fourteen and counting: unraveling trinucleotide repeat diseases, Cummings [/bib_ref].
Although the role of aggregation and fibre formation of expanded polyQ proteins has not yet been established clearly, protein misfolding and aggregation are accepted to be central issues for understanding the molecular mechanisms of these pathologies [bib_ref] Polyglutamine diseases: protein cleavage and aggregation, Zoghbi [/bib_ref]. In vitro studies have shown that polyQ aggregation depends on protein concentration, repeat-length, and time and that it occurs with a nucleation-dependent mechanism [bib_ref] Huntingtin-encoded polyglutamine expansions form amyloid-like protein aggregates in vitro and in vivo, Scherzinger [/bib_ref] [bib_ref] Selfassembly of polyglutamine-containing huntingtin fragments into amyloid-like fibrils: Implications for Huntington's disease..., Scherzinger [/bib_ref] [bib_ref] Polyglutamine aggregation behavior in vitro supports a recruitment mechanism of cytotoxicity, Chen [/bib_ref]. A conformational transition from random coil to b-sheet, which share most of the features typical of amyloids, takes place during the process of fibre formation [bib_ref] Amyloid-like features of polyglutamine aggregates and their assembly kinetics, Chen [/bib_ref] [bib_ref] Huntington's disease age-of-onset linked to polyglutamine aggregation nucleation, Chen [/bib_ref] [bib_ref] Huntingtin spheroids and protofibrils as precursors in polyglutamine fibrilization, Poirier [/bib_ref]. However, detailed structural information on polyQ aggregates is still unavailable and the steps leading to the assembly of mature fibres are not yet fully understood.
Kinetic studies of polyQ protein aggregation in vitro have shown that formation of amyloid or amyloid-like fibres generally occurs via fibrous intermediates that can have distinct morphologies [bib_ref] Huntingtin spheroids and protofibrils as precursors in polyglutamine fibrilization, Poirier [/bib_ref] [bib_ref] Expansion of polyglutamine induces the formation of quasi-aggregate in the early stage..., Tanaka [/bib_ref] [bib_ref] Hsp70 and Hsp40 attenuate formation of spherical and annular polyglutamine oligomers by..., Wacker [/bib_ref] [bib_ref] Formation of morphologically similar globular aggregates from diverse aggregation-prone proteins in mammalian..., Mukai [/bib_ref]. Cell biology studies have suggested that these early aggregates or proto-fibres rather than the insoluble aggregates are the main cytotoxic species, with mature fibres having a beneficial role for neuronal cells [bib_ref] Common structure of soluble amyloid oligomers implies common mechanism of pathogenesis, Kayed [/bib_ref] [bib_ref] Permeabilization of lipid bilayers is a common conformation-dependent activity of soluble amyloid..., Kayed [/bib_ref] [bib_ref] Calcium dysregulation and membrane disruption as a ubiquitous neurotoxic mechanism of soluble..., Demuro [/bib_ref] [bib_ref] Targeting protein aggregation in neurodegeneration-lessons from polyglutamine disorders, Weydt [/bib_ref]. This hypothesis has been formulated also for other neurodegenerative diseases related to protein misfolding and aggregation, such as Parkinson's and Alzheimer's diseases and the transmissible spongiform encephalopathies. Increasing evi-dence suggests that soluble aggregates-mediated toxicity might be a common pathogenesis mechanism for these disorders [bib_ref] Fresh and nonfibrillar amyloid beta protein(1-40) induces rapid cellular degeneration in aged..., Zhu [/bib_ref] [bib_ref] Inherent toxicity of aggregates implies a common mechanism for protein misfolding diseases, Bucciantini [/bib_ref] [bib_ref] Protofibrils, Pores, Fibrils, and Neurodegeneration: Separating the Responsible Protein Aggregates from the..., Caughey [/bib_ref] [bib_ref] Protein aggregation and aggregate toxicity: new insights into protein folding, misfolding diseases..., Stefani [/bib_ref] [bib_ref] Prefibrillar amyloid protein aggregates share common features of cytotoxicity, Bucciantini [/bib_ref] [bib_ref] Toxicity of recombinant beta-amyloid prefibrillar oligomers on the morphogenesis of the sea..., Carrotta [/bib_ref]. The characterisation of the early phases of fibrillation is therefore critical for understanding the molecular causes of pathogenesis.
Another central issue is the relationship between the polyQ tracts and other regions of the proteins that host them. Although polyQ expansion is certainly the main factor responsible for protein aggregation, various studies have demonstrated that the protein context plays an important role in determining the stability and solubility of polyQ peptides and may modify and contribute to the aggregation process [bib_ref] Amino acid sequences flanking polyglutamine stretches influence their potential for aggregate formation, Nozaki [/bib_ref] [bib_ref] Solution structure of polyglutamine tracts in GST-polyglutamine fusion proteins, Masino [/bib_ref] [bib_ref] Polyglutamines placed into context, Spada [/bib_ref] [bib_ref] Characterization of the structure and the amyloidogenic properties of the Josephin domain..., Masino [/bib_ref]. It is therefore important to investigate the properties of polyQ when flanked by different protein sequences in order to mimic their effect on aggregation.
In this work, we have studied the aggregation properties of a model system consisting of two polyQ peptides of different lengths, one below (22 glutamines) and one above (41 glutamines) the pathological threshold, fused to glutathione S-transferase (GST). GST was chosen because it is a well characterised protein of known structure, which can be used to mimic the environment surrounding the polyQ region in full-length proteins, thus providing an ideal model to assess the effects of protein context. GST-polyQ proteins have been instrumental in establishing that, prior to aggregation, polyQ is disordered regardless of its length, and in assessing the hypothesis of a structural threshold between short and long polyQ sequences [bib_ref] Solution structure of polyglutamine tracts in GST-polyglutamine fusion proteins, Masino [/bib_ref] [bib_ref] Pathogenic and non-pathogenic polyglutamine tracts have similar structural properties: towards a length..., Klein [/bib_ref].
Here, we have investigated the aggregation pathway of these fusion proteins and compared their polymerization kinetics with those of GST. Using complementary biophysical techniques, such as optical spectroscopy methods and both dynamic and static light scattering, we have been able to characterize the structural properties of the proteins during the process of aggregation, to estimate the timescale of aggregation, to assess the presence of different species and measure their size and relative populations. Our results show that the presence of a carrier may change significantly the aggregation pathway of polyQ and that protofibril formation may compete with bigger insoluble aggregates.
# Results
The presence of the polyQ tail influences the aggregate population
The aggregation state of GST, GST-Q22 and GST-Q41 was first characterized by dynamic light scattering (DLS) to obtain information about the size of the species present in solution.
Analysis of the correlation function shows a bimodal distribution, for all three samples which therefore contain two distinct populations [fig_ref] Figure 1: Comparison of the intensity-weighted size distribution of GST [/fig_ref]. The hydrodynamic radii of the dominant species are 4-7 nm. GST is well known to form stable dimers in the whole range of concentrations considered in this study, having a hydrodynamic radius of 3.7 nm (gyration radius of 2.9 nm) as calculated with the assumption of a globular shape of the molecule. The observed experimental radius measured for GST (4.260.5 nm) is therefore in excellent agreement with the presence of the dimer as the minimal unit. Marginally larger radii were observed for GST-Q22 and GST-Q41 (5.460.5 and 6.36 0.5 nm). These values can be explained by the progressive failure of the assumption of an isotropic globular shape, which is expected for the two samples in which the polyQ tail is unstructured [bib_ref] Solution structure of polyglutamine tracts in GST-polyglutamine fusion proteins, Masino [/bib_ref].
The second population corresponds to species with much larger molecular weights. Their hydrodynamic radius in the GST sample is 80610 nm which corresponds to a soluble aggregate of ca. 3500 dimeric units, but their population is so small as to be practically negligible. We estimated the numeric ratio between the soluble aggregate and the dimer to be ca. 1:7000000, which accounts for ca. 0.1% of the total population by mass. In GST-Q22 and GST-Q41, the soluble aggregates have averaged radii of 9164 and 9664 nm respectively, which correspond to ca. 5000-7000 dimeric units. The aggregate populations are also small in these samples. This explains why these species were not detected by NMR or CD in previous studies [bib_ref] Solution structure of polyglutamine tracts in GST-polyglutamine fusion proteins, Masino [/bib_ref] but can be observed by a technique as sensitive to aggregation as DLS: large aggregates scatter light much more than small species and therefore even small populations of them will be detectable (with the same argument, we cannot exclude the presence of minute populations of species with radii close to that of the dimer). Despite the minor populations of the soluble aggregates, a clear trend was observed in all samples studied (from three different protein batches): the numerical ratios between the larger and the smaller species are of about 1 to 140000635000 and of 1 to 3200068000, which correspond to ca. 3% and 11% of the total populations by mass of GST-Q22 and GST-Q41, respectively, suggesting that the relative populations of the soluble aggregates correlate with the length of the polyQ tract. No significant time dependence of these ratios was observed over a period of several months and the range of concentrations considered (6-20 mM).
These results strongly indicate an intrinsic tendency of polyQ to promote aggregation. While not affecting the size, which seems to be determined by the carrier, the length of the polyQ tract influences the relative populations of the aggregates.
## The presence of a reducing agent discriminates between different types of aggregation
We then explored the nature of the aggregation. GST has four cysteines that are not involved in intramolecular sulphur bridges and can thus bond intermolecularly and promote covalent aggregation [bib_ref] Conformational stability of pGEX-expressed Schistosoma japonicum glutathione S-transferase: a detoxification enzyme and..., Kaplan [/bib_ref]. This phenomenon would be unwanted in the present study since it would depend on the specific choice of the protein carrier and not on the polyQ tract. To make sure that the presence of DTT effectively inhibits the formation of disulphide bridges, and to estimate the extent of their interference with the formation of non-covalent polyQ aggregates, we compared measurements in the presence and in the absence of a reducing agent [fig_ref] Figure 2: Effect of an antioxidant on aggregation [/fig_ref]. In experiments carried out in the absence of DTT, the correlation function of GST showed a distribution which corresponds to a species with a weighted average radius of 7.5 nm, which corresponds roughly to a dimer of the dimeric form, and only minor traces of a species of 50 nm. This is much smaller than the one observed in the presence of DTT. The size distribution of GST-Q22 and GST-Q41 is bimodal with average radii of 7.2 and 74 nm, and 10 and 86 nm, respectively. We did not observe a dependence of their relative populations on the protein concentration, at least in the range of concentrations explored . When an excess of freshly prepared DTT (1 mM) was added to the samples without DTT, a progressive decrease of scattered light was seen and followed until a steady state was reached [fig_ref] Figure 2: Effect of an antioxidant on aggregation [/fig_ref] , indicating a decrease of the average dimensions of the sample species in response to the anti-oxidant effect of DTT. The final state is indistinguishable from that observed in samples always treated with DTT.
These results indicate that covalent aggregation of GST can compete with the non-covalent aggregation observed under reducing conditions producing different, smaller soluble aggregates. The effect can, however, be effectively reversed by addition of DTT. All the following studies were carried out in the presence of DTT.
## Thermally-induced aggregation occurs at higher temperatures in gst-polyq proteins
To study further the nature and the mechanism of aggregation, the thermal stabilities of GST, GST-Q22 and GST-Q41 were assessed by DLS, recording thermal unfolding profiles. Heat should promote aggregation if this is mainly hydrophobic. Thermal denaturation of GST causes a temperature-dependent increase of scattered light intensity, concomitant with a size increase of the aggregates [fig_ref] Figure 3: Effect of temperature on aggregation [/fig_ref]. The temperature at which deviation of the signal from the baseline starts appearing depends on the sample concentration. For GST, it is in the range of 38-48uC for 6-20 mM protein concentrations (data not shown). The increase continues up to ca. 55-58uC. Above this temperature, the signals decrease, concurring with the appearance of a visible precipitate at the bottom of the cuvette. This behaviour reflects the opposing trends of an increase of scattered light due to the increasing dimensions of the species present in solution and their loss from the same solution by precipitation. The event is irreversible, as previously described for GST [bib_ref] Conformational stability of pGEX-expressed Schistosoma japonicum glutathione S-transferase: a detoxification enzyme and..., Kaplan [/bib_ref].
DLS temperature scans measured for GST-Q22 and GST-Q41 show similar features [fig_ref] Figure 3: Effect of temperature on aggregation [/fig_ref]. Thermal unfolding is irreversible also for these samples, and the starting temperature of aggregation has an appreciable dependence on concentration (data not shown). However, a clear difference is observed between the three samples: at the same protein concentration, aggregation initiates at higher temperature for the polyQ fusion proteins than for GST. Interestingly, GST-Q41 starts aggregating at a higher temperature than does GST-Q22. The extent of these differences depends on the history of the samples and on the scanning rates, but the qualitative behaviour is consistent and reproducible.
Our results confirm that thermally induced destabilization of GST causes irreversible aggregation, and indicate that the same is true for GST-polyQ proteins. The temperature of this transition is influenced both by the presence and by the length of the polyQ tract. This suggests that, although undergoing the same process, the kinetics of aggregation of polyQ fusion proteins are slower than those of GST.
## Gst-polyq fusion proteins form a reservoir of soluble aggregates
To follow in more details the early steps which precede the final catastrophic events after which precipitation starts to occur, the behaviour of the individual species in solution was monitored by a three dimensional representation of the intensity-weighted size distribution during the temperature scan monitored by DLS [fig_ref] Figure 4: Analysis of the species size during the temperature scan [/fig_ref]. The plot clearly shows the simultaneous presence of the minimal dimeric form and of the soluble aggregates. The dominant and almost exclusive species present in the GST sample remains the minimal dimeric unit (of ca. 4 nm radius) up to 38uC, with only a minor contribution of a larger species. The populations of the soluble aggregates of GST-Q22 and GST-Q41 are more appreciable at all temperatures. For all three samples, there is no significant appearance of species of dimensions intermediate between the dimer and the soluble aggregates of 80-96 nm radii. At high temperature, the size of the large species increases exponentially (note that the x axes in [fig_ref] Figure 4: Analysis of the species size during the temperature scan [/fig_ref] are logarithmic), suggesting that there is a progressive and direct scavenging of the dimer by the soluble aggregates. Eventually, the signal from the large species disappears completely as they grow into an insoluble large aggregate and fall out of solution. The temperatures at which the size starts to increase and at which the signal disappears are different for the three samples and are progressively higher as a function of the polyQ length. It is also clear from the plots that the size growth of the soluble aggregate at the expense of the dimer observed at high temperatures is larger for GST than for GST-Q22 and GST-Q41, which increase more slowly.
A two-dimensional plot of the ratios of the average dimensions of the two main species present in solution (R 2 /R 1 , where species 1 refers to the minimal dimeric species, of ca. 4-7 nm radius, and species 2 is the soluble aggregate, originally of ca. 80-90 nm radius) vs. the ratio of the respective contributions to the correlation function amplitude (A 2 /A 1 ) provides a complementary description [fig_ref] Figure 4: Analysis of the species size during the temperature scan [/fig_ref]. For GST, there is a close to linear increase of the dimensions of the aggregates at increasing A 2 /A 1 ratios in the whole range of temperatures up to signal disappearance. For GST-Q22 and GST-Q41 instead, the amplitude increment is not paralleled by an increase of the radii ratios, which remain almost constant in the range 35uC-42uC, although the overall scattering intensities increase exponentially over the whole range of temperature (cf. [fig_ref] Figure 3: Effect of temperature on aggregation [/fig_ref]. This strongly suggests that GST aggregation proceeds through intermediate formation of soluble species which grow immediately into large insoluble aggregates. Their population does not however increase and the large aggregates grow so much that they start to precipitate. On the contrary, over the same temperature range, GST-Q22 and GST-Q41 convert first the minimal dimeric unit into the soluble aggregate, whose population rather than size increases. Only when this intermediate of aggregation is significantly populated does the process of aggregation proceed to formation of large aggregates.
These results indicate that aggregation occurs via an intermediate which behaves as a seed (or nucleus) for further aggregation. However, the mechanism of aggregation of GST is different from that of GST-Q22 and GST-Q41, which form a reservoir of soluble aggregates before proceeding to the formation of insoluble large species.
## Temperature induced aggregation is associated with a conformational transition
The process was monitored independently using far-UV CD to follow the conformational state of the samples during thermally induced aggregation. The CD spectrum of GST at 20uC is typical of an a/b protein, with a predominance of the a-helical signal, and in agreement with literature data [fig_ref] Figure 5: The effect of temperature on the far-UV CD spectra [/fig_ref] [bib_ref] Conformational stability of pGEX-expressed Schistosoma japonicum glutathione S-transferase: a detoxification enzyme and..., Kaplan [/bib_ref]. At 20uC, the spectra of GST-Q22 and GST-Q41 are very similar to that of GST, but have a slightly higher content of random coil conformation due to the presence of the polyQ tail, which has been proven to be highly flexible and not to interact with the carrier protein [fig_ref] Figure 5: The effect of temperature on the far-UV CD spectra [/fig_ref] , inset) [bib_ref] Solution structure of polyglutamine tracts in GST-polyglutamine fusion proteins, Masino [/bib_ref]. For all three proteins, spectra measured at higher temperatures showed that heating induces a secondary structure transition from an a/b structure to a conformation with a higher b-sheet content. Such a transition has been described for several proteins that are known to form fibres and are associated with misfolding diseases [bib_ref] Unfolding the role of protein misfolding in neurodegenerative diseases, Soto [/bib_ref]. At the same time, the overall signal decreases irreversibly. Visual inspection and electron microscopy images of the samples after the scan (see below) show that this irreversible process is aggregation, in agreement with LS measurements. The CD signal decrease therefore arises from two interconnected phenomena: a secondary structure variation and protein aggregation, which eventually causes precipitation and loss of the signal.
When thermal scans monitored at a fixed wavelength (222 nm) were registered, we observed the same phenomenon as in DLS measurements: there is a clear difference in the transition midpoints of the three samples, with GST having a transition at lower temperature [fig_ref] Figure 5: The effect of temperature on the far-UV CD spectra [/fig_ref]. The amplitude of the apparent transition is smaller for GST-Q41 than for GST-Q22 and, even more so, for GST. A final signal intensity similar to that of GST is reached also by GST-Q22 and GST-Q41, but only after incubation at high temperature for several hours (data not shown).
Quantitatively, the temperature at which formation of the insoluble aggregates starts being observable by DLS appears lower than that at which we observe a conformational transition by CD. This difference is partly a consequence of the slower scanning rate imposed by DLS measurements, which lead to work at quasiequilibrium conditions. However, even taking this effect into account, we observed consistently that the structural transition occurs at a temperature at which the dimer ceases to influence the DLS measurements. This delay suggests that the two techniques record different albeit interconnected events. The b-rich signal does not arise from the dimer but becomes detectable only when the aggregates become the dominant species in solution. We cannot, on the other hand, infer anything on the structure of the aggregates while they are in co-presence with the dimers, since their populations are too small to influence the CD signal.
These observations provide direct information about the structural changes occurring upon aggregation and confirm the presence of different kinetics of aggregation due to the presence of the polyQ tail.
## Studying further the kinetics of aggregation
To quantify the differences in behaviour of GST, GST-Q22, and GST-Q41, time scans were recorded at fixed temperatures using both LS and CD. The far-UV CD data, recorded at temperatures close to the beginning of the thermally induced transitions, showed that the predominantly helical signal converts, as a function of time, into a b-rich spectrum, as observed during temperature scans (data not shown). The signal intensity also decreases with time, in concomitance with the formation of an insoluble precipitate [fig_ref] Figure 6: Kinetics of aggregation [/fig_ref]. The kinetics of the a to b transition and aggregation are temperature dependent and are slower at lower temperature, as expected for a hydrophobic process. Conformational changes and aggregation occur with faster kinetics for GST, followed by GST-Q22 and then GST-Q41. The kinetics are characterised by a lag phase, followed by an exponential decay phase, typical of nucleation-dependent polymerization [bib_ref] Protein folding and misfolding, Dobson [/bib_ref]. At 50uC, the lag phase measured for GST is approximately 3 min, whereas GST-Q22 and GST-Q41 have lag phases of approximately 19 and 26 min, respectively. The estimated average values of the apparent transition half lives at 50uC are 31, 40, and 67 (+/23) min for 4 mM samples of GST, GST-Q22, and GST-Q41, respectively [fig_ref] Table 1: Half lives [/fig_ref].
When the aggregation kinetics were followed by DLS, the scattered light intensity was recorded at 37uC, which is the temperature at which the exponential increase of scattered light starts. This temperature is significantly lower than the one used to detect conformational changes by CD, but we wanted to make sure we could follow with each technique the early stages of aggregation. A longer persistence of the soluble aggregates of GST-Q41 is very clear in a three-dimensional representation [fig_ref] Figure 6: Kinetics of aggregation [/fig_ref]. Intensity plots measured at different time intervals show that GST aggregates faster, followed by GST-Q22, and finally GST-Q41, with half lives of 53 min, 70 min and 390 min [fig_ref] Figure 6: Kinetics of aggregation [/fig_ref]. These values are in excellent qualitative agreement with those obtained by CD, which, having been recorded at higher temperatures, are smaller. The conversion of the GST soluble aggregates into larger species is also much faster, over the same time interval, than that of the polyQ fusion proteins. Measure-ments performed with sample concentrations in the range 6-30 mM indicate a clear dependence of aggregation kinetics on concentration (data not shown).
Taken together, the CD and LS data strongly suggest that, although the presence of the polyQ tail does not affect the secondary structure of GST, it has a striking effect on the time scales of formation of insoluble aggregates. The polyQ tracts seem to delay the aggregation process of GST in a polyQ lengthdependent manner.
## The tendency to form fibrils increases with the polyq length
A different behaviour of the three samples could also be deduced from direct visual inspection of the samples. Immediately after the kinetic studies, only GST was cloudy, whereas the other two samples showed physical precipitation at the bottom of the cell or became opalescent only after several hours at high temperature (data not shown). Inspection of the samples by EM showed the presence of fibrillar aggregates only for GST-Q41 [fig_ref] Figure 7: Visualization of the large insoluble aggregates as observed by EM [/fig_ref]. They were typically 10-50 nm long and narrower than 5 nm. GST-Q22 and GST showed amorphous aggregates that were usually too big even to be visualised in the EM grids. While we cannot exclude that also these samples could form fibres if the correct conditions were found, we must conclude that longer polyQ tails facilitate fibre formation in GST fusion proteins, in agreement with what has been observed for isolated polyQ peptides [bib_ref] Pathogenic and non-pathogenic polyglutamine tracts have similar structural properties: towards a length..., Klein [/bib_ref].
# Discussion
An increasing number of diseases that result in neuronal death have been associated with protein misfolding and aggregate formation. Accumulating evidence strongly supports the view that cytotoxicity arises from the presence of soluble aggregates and/or prefibrillar species, rather than from the fibrillar species which form the insoluble aggregates accumulated in the cell either as amyloid fibres or as amorphous deposits. In this newer model, the insoluble species would serve only as an escape route for smaller aggregates. It is therefore essential to characterize the species involved and the misfolding pathways that relate them in order to understand the mechanism of aggregation. Here, we have adopted a model system which makes use of GST as a carrier to investigate the role of protein context in the aggregation pathway of polyQ. We have shown that the presence of polyQ tails results in the trapping and stabilization of a small but well defined population of soluble aggregates. Interestingly, the size of these species is defined by the carrier protein and not by the polyQ tract, as judged from their presence already in the GST sample. They must reflect an intrinsic tendency of this protein to misfold during over-expression and/or purification, a behaviour possible also in proteins as soluble as GST, which is known to undergo irreversible transitions upon stress conditions [bib_ref] Conformational stability of pGEX-expressed Schistosoma japonicum glutathione S-transferase: a detoxification enzyme and..., Kaplan [/bib_ref]. When the polyQ tail is added, the population of aggregation intermediates increases by 1-2 orders of magnitude (depending on the polyQ length) already at room temperature, without the appearance of additional species. Although not necessarily structurally and morphologically similar, the soluble aggregates observed in the three samples seem to have similar hydrodynamic radii. This suggests that the polyQ tract increases the probability of misfolding but the carrier protein acts as the template for aggregation formation. We do not observe significant dimer-toaggregate inter-conversion at room temperature, in agreement with the observation that no aggregation was detected by CD and NMR for the same constructs over a period of months [bib_ref] Solution structure of polyglutamine tracts in GST-polyglutamine fusion proteins, Masino [/bib_ref].
The pathway of thermal aggregation of the three samples was followed by CD, LS and both optical and electron microscopy. These techniques provided complementary information. LS, being highly sensitive even to minute populations of aggregates whose capacity to scatter light increases with molecular size [bib_ref] Protofibril formation of amyloid beta-protein at low pH via a non-cooperative elongation..., Carrotta [/bib_ref] , helped us to characterize the size of the species present in solution. CD provided a description of the secondary structure of the populations dominant in solution, whereas optical and electron microscopies gave us information on the morphology of the insoluble aggregates. The model which comes out from LS data is that aggregation proceeds through a two-step irreversible conversion of the three species, the minimal dimeric units, a soluble high molecular weight species and large insoluble aggregates of dimensions too large to be detected even by LS (i.e. roughly$10 mm). Before disappearing into the large insoluble aggregates, an appreciable population of dimer is trapped in the intermediate state which becomes increasingly more populated. The soluble aggregates are thus an integral part of the aggregation pathway: they appear to behave like 'sticky glue balls' which scavenge dimers remaining in solution, and thus act as the foci for further aggregation. The process does not consist of a simple unfolding event but, as observed by CD, is a complex transition in which aggregation is tightly associated with an irreversible structural transition which results in a significant enrichment of b content.
The effect is, however, not the same for all the samples tested, as already reported before [bib_ref] Solution structure of polyglutamine tracts in GST-polyglutamine fusion proteins, Masino [/bib_ref]. Irreversible aggregation of GST occurs either at lower temperatures or, when following the process at constant temperature, with faster aggregation kinetics than those of GST-Q22 and GST-Q41. This suggests that the two pathways dimer-to-soluble aggregate and from the soluble-toinsoluble aggregate have a similar efficiency only for GST, whereas one of them is disfavoured in GST-Q22 and GST-Q41 with consequent, if transient, formation of a reservoir of soluble high molecular weight intermediates or protofibrils. These results could look somewhat counterintuitive and in direct conflict with what is observed when the polyQ tail is proteolytically cleaved from the carrier: the cleaved polyQ peptides have aggregation kinetics which depend on the polyQ length so that, even at room temperature, Q41 aggregates must faster than Q22 [bib_ref] Pathogenic and non-pathogenic polyglutamine tracts have similar structural properties: towards a length..., Klein [/bib_ref]. A similar behaviour has been observed for polyQ tracts fused to CRABP I, a highly soluble protein with a reversible unfolding pathway [bib_ref] Extended polyglutamine tracts cause aggregation and structural perturbation of an adjacent Beta..., Ignatova [/bib_ref]. The main difference between these examples and our GST model system is that, under destabilizing conditions, unmodified GST is able to undergo irreversible aggregation on its own. This strongly suggests that, when together, the two distinct elements GST and polyQ, each with intrinsic tendencies to aggregation, do not behave independently but mutually affect each other's behaviour.
The importance of protein context in modulating the behaviour of polyQ is largely supported by independent evidence [bib_ref] Amino acid sequences flanking polyglutamine stretches influence their potential for aggregate formation, Nozaki [/bib_ref] [bib_ref] Polyglutamines placed into context, Spada [/bib_ref] [bib_ref] Characterization of the structure and the amyloidogenic properties of the Josephin domain..., Masino [/bib_ref] [bib_ref] Extended polyglutamine tracts cause aggregation and structural perturbation of an adjacent Beta..., Ignatova [/bib_ref] [bib_ref] Expansion of amino acid homo-sequences in proteins: Insights into the role of..., Menon [/bib_ref]. The effect seems to work both ways. Soluble carriers are known to solubilize the insoluble polyQ and to make it stable in solution for several months. Well known examples are myoglobin, CRABP I and GST itself. Addition of a proline-rich extension to a polyQ tract has also been shown to decrease its tendency to aggregate [bib_ref] Oligoproline effects on polyglutamine conformation and aggregation, Bhattacharyya [/bib_ref]. Conversely, protein domains outside the polyQ tract have been shown to increase the tendency of polyQ to misfolding: cellular studies of the aggregation propensity of expanded and non-expanded ataxin-1 and ataxin-3 have shown that, while promoted by polyQ expansion, aggregation can be noticeably reduced by deletion of such domains or their replacement with sequences with no known tendency to aggregate [bib_ref] Expansion of amino acid homo-sequences in proteins: Insights into the role of..., Menon [/bib_ref] [bib_ref] Polyglutamine is not all: the functional role of the AXH domain in..., De Chiara [/bib_ref]. Incorporation of a polyQ tract into a loop of the stably folded chymotrypsin inhibitor 2 (CI2) has also been shown to lead to formation of misfolded dimeric and trimeric species [bib_ref] Crystal structure of a dimeric chymotrypsin inhibitor 2 mutant containing an inserted..., Chen [/bib_ref]. It is therefore not entirely surprising that, in addition to increasing the probability of misfolding and causing structural destabilisation, polyQ can alter the kinetics of aggregation of its carrier proteins, thus having an effect on the time scale of aggregation.
The effect can be due to different causes. The flexible polyQ tail, which, before aggregation, fluctuates freely in solution, could mask the surface which promotes GST aggregation, decreasing the probability of effective collisions of this region with other molecules. This would disfavour further transition to the insoluble aggregates, thus slowing the kinetics. The polyQ tail could also increase the stability of the GST aggregation intermediate by transiently interacting with the carrier. It might be more difficult to promote transition of both the polyQ tail and the globular GST to the b-rich structure of the large aggregate for steric hindrance reasons. Recent theoretical studies also indicate that merely adding a large mass to the terminus of a protein tends to stabilise the fold (DJT, unpublished data).
The presence of larger reservoirs of soluble aggregates or protofibrils along the aggregation pathway observed for the polyQ fusion proteins could be correlated with their increasing tendency to form ordered fibres, a feature that is easily observed for polyQ peptides but not for GST. The trapped intermediates, which kinetically prevent the immediate appearance of large insoluble aggregates, could lead more easily to well-ordered fibrillar structures, as for crystal formation, where slower growth conditions are generally beneficial once nucleation has occurred.
How can our observations be related to 'real' polyQ proteins? The results presented here are coherent both internally and with what is known about the behaviour of polyQ proteins. GST, which was originally chosen mainly because it is a well characterised globular protein for which both the structure and the unfolding properties are known, turned out a posteriori to be an ideal model for naturally occurring polyQ proteins. Its ability to aggregate irreversibly through formation of a b rich conformation mimics the behaviour of at least two of the nine known proteins linked to polyQ diseases. Both ataxin-1 and ataxin-3 contain globular motifs, the AXH and the Josephin domains, which have strong intrinsic tendencies to form non-covalent aggregates [bib_ref] Characterization of the structure and the amyloidogenic properties of the Josephin domain..., Masino [/bib_ref] [bib_ref] Expansion of amino acid homo-sequences in proteins: Insights into the role of..., Menon [/bib_ref]. It is tempting to suggest that the tendency of protein context to stabilize larger populations of soluble aggregates or proto-fibrils is an important and more general behaviour, which could be at the basis of polyQ pathologies. If, as now widely believed, the toxic species is the soluble aggregates and/or the protofibrils rather than the insoluble aggregates, this model could help to explain why longer polyQ tracts are more toxic than short ones. More work will be needed to extend this model to more specific examples.
# Materials and methods
## Samples preparation
The DNA sequences coding for 22 and 41 glutamines were cloned into a pGEX-4T1 plasmid vector containing Schistosoma japonicum GST with a 21-residue linker at the C-terminus, as described before [bib_ref] Solution structure of polyglutamine tracts in GST-polyglutamine fusion proteins, Masino [/bib_ref]. The GST-fusion proteins were expressed in E. coli strain BL21 and purified by affinity chromatography using glutathione-agarose beads (Amersham Pharmacia). The purity of the samples was assessed by SDS-PAGE and mass spectrometry. Protein concentrations were determined using UV absorption, with calculated extinction coefficient at 280 nm of 40920. The buffer used was 40 mM sodium phosphate, pH 6.5, and 1 mM dithiothreitol (DTT) to prevent cysteine oxidation. The experiments performed under non-reducing conditions were performed in the same buffer, in the absence of DTT.
## Static and dynamic light scattering
Before each measurement, the samples were filtered using 0.2 mm pore diameter membranes (Sartorius), put into dust-free optical cells and placed into a thermostated cell compartment of a Brookhaven Instruments BI200-SM goniometer to carry out the measurements. The temperature was controlled within 0.1uC using a thermostated recirculating bath. The light scattered intensity and time autocorrelation function were measured by using a Brookhaven BI-9000 correlator and a 100 mW Argon laser (Melles Griot) tuned at l = 514.5 nm. The spatial resolution is defined by the scattering vector q = 4pnl 0 21 sin(h/2), where n is the refraction index of the solution, l 0 is the wavelength of the incident light, and h is the scattering angle. Static light scattering data were corrected for the background scattering of the solvent and normalized by using toluene as calibration liquid. In DLS experiments, the correlator was operated in the multi-channel mode. To assess reproducibility, each experiment was repeated at least three times using independent batches of proteins. The size and relative populations indicated throughout the manuscript are averaged over the measurements.
# Data analysis
The field autocorrelation function, g (1) (t), was obtained by measuring the intensity correlation function and analyzed by using CONTIN [bib_ref] A constrained regularization method for inverting data represented by linear algebraic or..., Provencher [/bib_ref] , in order to determine the distribution of relaxation times according to:
[formula] g 1 ð Þ t ð Þ~ð A C ð Þ exp { C t ð ÞdC [/formula]
where A(C) denotes the contribution amplitude of the mode with characteristic time C 21 . The latter is related to the diffusion coefficient by:
[formula] C~Dq 2 [/formula]
The hydrodynamic radius is obtained by the Stoke-Einstein relationship D = k B T/6pgR H . The simpler cumulative analysis, which gives the average value and the width of the size distribution [bib_ref] Analysis of macromolecular polydispersity in intensity correlation spectroscopy: the method of cumulants, Koppel [/bib_ref] , can be strictly applied only for GST samples under nonaggregating conditions, where large species are negligible and a modal distribution is observed. For GST-Q22 and GST-Q41 or even for GST samples under aggregating conditions, a contribution of two different particle populations was observed always.
## Circular dichroism
CD measurements were performed on a Jasco J-715 spectropolarimeter equipped with a PTC-348 Peltier temperature control system, which allows a maximal error of 0.1uC. CD spectra were recorded using quartz cuvettes (Hellma) with pathlengths of 1 mm. Protein samples were in 40 mM phosphate pH 6.5, 1 mM DTT, with protein concentrations of 4-20 mM. CD intensities are presented as the CD absorption coefficient calculated using the molar concentration of the proteins (De M ). Thermal scans were measured by increasing temperature from 20 to 90uC at 1uC/min or at 10uC/hour. Reversibility was assessed by cooling to 20uC using the same rate. Temperature and time scans were recorded by monitoring the CD signal at 222 nm.
## Electron microscopy
The samples used for CD temperature or time scans were directly analysed by EM, applied to carbon coated grids and stained with 1% sodium silico-tungstate (pH 7). The grids were viewed under minimal dose, accurate defocus conditions with a Jeol 1200EX operated at 100 kV.
[fig] Figure 1: Comparison of the intensity-weighted size distribution of GST (black line), GST-Q22 (red line) and GST-Q41 (green line) at 90u scattering angle. The measurements were carried out on 20 mM protein samples and at 20uC. doi:10.1371/journal.pone.0000111.g001 [/fig]
[fig] Figure 2: Effect of an antioxidant on aggregation. A) Comparison between the intensity-weighted size distributions obtained in the absence and in the presence of DTT. B) Time course of light scattered intensity at 90u angle upon addition of DTT to fresh protein samples. GST, GST-Q22 and GST-Q41 are shown using black, red and green lines, respectively. The measurements shown were carried out at 20uC using 20 mM samples. doi:10.1371/journal.pone.0000111.g002 [/fig]
[fig] Figure 3: Effect of temperature on aggregation. A) Light scattered intensity at 90u angle registered during the temperature scan. Data relative to different proteins are normalized by dividing for the respective initial values of scattered intensity. B) Weight-average radius of the largest species, as obtained by CONTIN analysis[38], followed during the temperature scan. doi:10.1371/journal.pone.0000111.g003 [/fig]
[fig] Figure 4: Analysis of the species size during the temperature scan. A) Contour plots of intensity-weighted size distribution during the temperature scan. B) Size ratio of large to small species, R 2 /R 1 , plotted versus the ratio of their relative light scattering contribution, A 2 /A 1 during the temperature scan. The corresponding temperatures are indicated. C) Weight-averaged radius of the largest species, as obtained by CONTIN analysis[38], in the course of temperature scan. doi:10.1371/journal.pone.0000111.g004 [/fig]
[fig] Figure 5: The effect of temperature on the far-UV CD spectra. A) Far-UV CD spectra of GST-Q41 recorded at different temperatures: 20uC (continuous black line), sample incubated at 51uC for 30 minutes (dashed line), and then incubated at 68uC for 1 hour and 20 minutes (grey line). Protein concentration was 4 mM. Inset: Far-UV CD spectra of GST (dashed line), GST-Q22 (dotted line), and GST-Q41 (continuous line) recorded at 20uC. B) Far-UV CD thermal scans of GST (continuous black line), GST-Q22 (dotted line), and GST-Q41 (grey line) recorded at 222 nm with a heating rate of 1uC/ min. The curves were normalized to the intensity of the GST sample at 20uC for comparison purposes. The cooling profile of GST (dashed line) was recorded with the same rate and is added as an example of irreversibility. A similar behaviour was observed for the other two proteins. Protein concentration was 20 mM in all samples. doi:10.1371/journal.pone.0000111.g005 [/fig]
[fig] Figure 6: Kinetics of aggregation. A) Far-UV CD time scans of GST (black line), GST-Q22 (red line), and GST-Q41 (green line) recorded at 50uC. The signal changes, monitored at 222 nm, arise from both protein aggregation and the associated conformational transition. The curves were normalized to the intensity of GST for comparison purposes. Protein concentration was 4 mM in all samples. B) Contour plots of intensity-weighted size distribution of samples quenched at 37uC. C) Time course of light scattered intensity at 90u angle of samples quenched at 37uC. doi:10.1371/journal.pone.0000111.g006 [/fig]
[fig] Figure 7: Visualization of the large insoluble aggregates as observed by EM. GST (left) and GST-Q41 (right) samples were analysed immediately after the kinetics at 50uC. doi:10.1371/journal.pone.0000111.g007 [/fig]
[table] Table 1: Half lives (min) of GST, GST-Q22 and GST-Q41 as estimated from LS and CD data recorded at different temperature. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [/table]
|
Hypothalamic glioma masquerading as craniopharyngioma
Hypothalamic glioma account for 10-15% of supratentorial tumors in children. They usually present earlier (first 5 years of age) than craniopharyngioma. Hypothalamic glioma poses a diagnostic dilemma with craniopharyngioma and other hypothalamic region tumors, when they present with atypical clinical or imaging patterns. Neuroimaging modalities especially MRI plays a very important role in scrutinizing the lesions in the hypothalamic region. We report a case of a hypothalamic glioma masquerading as a craniopharyngioma on imaging along with brief review of both the tumors.
# Introduction
The hypothalamic region may be involved by a large variety of disease processes including tumors, inflammatory, and developmental disorders. The clinical manifestations include vision loss, behavioral changes, and endocrine abnormalities. Magnetic resonance imaging (MRI) and computed tomography (CT) play a pivotal role in evaluation of such patients by narrowing the list of differential diagnosis based on the imaging features. However, an accurate diagnosis may not be possible in all cases.
## Case report
A 25-year-old female patient presented with a history of headache, vomiting, and blurring of vision of 2-month duration. On examination the patient was conscious and cooperative. Visual acuity was 6/18 bilaterally. Fundus was normal. Reduction of the left temporal visual field was seen. Pupils were bilaterally symmetrical and reactive to light. CT and contrast-enhanced MRI of the brain [ was performed which revealed a solid cystic mass lesion involving the hypothalamus and the thalamus displacing and compressing the third ventricle. Inferiorly the lesion was extending into the interpeduncular cistern. The pituitary stalk and optic chiasm were not visualized separately from the lesion. The pituitary gland was normal in location and signal intensity. The solid component of the lesion showed enhancement on contrast administration. A provisional diagnosis of a suprasellar craniopharyngioma was given. The patient was subsequently operated upon and a right frontoparietal craniotomy with subtotal excision was performed. On histopathology the lesion was diagnosed as a hypothalamic glioma [ ]. Post-surgery the patient developed central diabetes insipidus and is currently E4V4M4 status with left hemiplegia.
# Discussion
Craniopharyngiomas arise from the remnants of the craniopharyngeal duct. [bib_ref] Lesions of the Hypo-thalamus: MR imaging diagnostic features, Saleem [/bib_ref] They are the most common nonglial brain tumors in children and account for half of all suprasellar masses in this age group. [bib_ref] Aggressive surgical management of craniopharyngiomas in children, Hoffman [/bib_ref] More than half of craniopharyngiomas occur in children and young adults. Intrasellar craniopharyngiomas are seen in 4% of the cases, 21% are both sellar and suprasellar, and 75% of the cases are limited to the suprasellar region alone, often with extension up into the third ventricle. [bib_ref] Neuroimaging of childhood craniopharyngioma, Nash [/bib_ref] Histologically craniopharyngioma can be divided into papillary and adamantinomatous subtypes. Adamantinomatous craniopharyngioma have a bimodal distribution, with the first peak between 5 and 15 years old and a second peak between 45 and 60 years old. The papillary subtype occurs in adults at a mean age of 40-55. [bib_ref] The descriptive epidemiology of craniopharyngioma, Bunin [/bib_ref] Males are more commonly affected. [bib_ref] MR differentiation of adamantinous and squamous-papillary craniopharyngiomas, Sartoretti-Schefer [/bib_ref] Symptoms include headaches, visual field defects, and hypothalamic dysfunction. Adamantinous craniopharyngioma are usually suprasellar and are predominantly cystic. T1 hyperintense cysts are typical; however, T1 hypointense cysts are also possible. The tumor is mostly lobulated and may encase subarachnoid arterial vessel. Squamous-Papillary variant is spherical in shape, generally intrasellar/suprasellar or suprasellar. It is predominantly solid with hypointense cysts on T1. On CT large cysts with CSF density and soft tissue density enhancing solid component are seen. Calcification is seen in 90% and is stippled and often peripheral in location.
Hypothalamic chiasmatic gliomas, on the other hand, form 10-15% of supratentorial tumors in children. The most common age of presentation is 2-4 years. The patients usually present with loss of visual acuity. Growth hormone dysfunction is seen in 20% of the cases. [bib_ref] Lesions of the Hypo-thalamus: MR imaging diagnostic features, Saleem [/bib_ref] Hypothalamic gliomas are associated with a family history of neurofibromatosis type 1 in 20-50% of the cases. [bib_ref] Lesions of the Hypo-thalamus: MR imaging diagnostic features, Saleem [/bib_ref] In such cases the glioma may have a more indolent course and may even regress spontaneously. [bib_ref] Magnetic resonance analysis of suprasellar tumors of childhood, Kollias [/bib_ref] [bib_ref] Initial management of children with hypothalamic and thalamic tumors and the modifying..., Allen [/bib_ref] These gliomas are hypointense on T1-weighted sequences and hyperintense on T2-weighted and FLAIR sequences. Larger tumors tend to be heterogeneous with both cystic and solid components. The solid component shows significant contrast enhancement. [bib_ref] Magnetic resonance analysis of suprasellar tumors of childhood, Kollias [/bib_ref] A study by showed that in non-NF patients, the chiasm and hypothalamus were the most common sites of involvement, the tumor was mass-like, and cystic components were frequently seen, as was extension beyond the optic pathways. [bib_ref] Optic pathway glioma: Correlation of imaging findings with the presence of neurofibromatosis, Kornreich [/bib_ref] In our case the lesion had both solid and cystic component with enhancement and did not show any calcification; however, the age of the patient was not in favor of a diagnosis of hypothalamic glioma. Bisson et al., have reported two similar cases of OCHGs appearing like craniopharyngiomas on neuroimaging. [bib_ref] Hypothalamic-opticochiasmatic gliomas mimicking craniopharyngiomas, Bisson [/bib_ref] In another study by Bommakanti et al. nine out of 18 cases with a pre-op diagnosis of optico-chiasmatic-hypothalamic gliomas had mixed solid and cystic component. Of these histological examination revealed craniopharyngioma in two patients. [bib_ref] Optic chiasmatic-hypothalamic gliomas tissue diagnosis essential, Bommakanti [/bib_ref] To conclude, the main differentiating features between craniopharyngioma and hypothalamic glioma are the presence of mixed intensity cysts on T1 and calcification in craniopharyngioma and the relative young age of presentation in hypothalamic gliomas. Yet, it is not always possible to differentiate craniopharyngiomas from hypothalamic glioma. Thus, obtaining a tissue diagnosis via biopsy may be the right course of action in planning further management, whenever diagnosis is in doubt.
[fig] Figure 1, Figure 2, Figure 3, Figure 4: CT (a) and MR (b) axial image showing suprasellar multicystic lesion b a Coronal T2-weighted MR images showing multicystic suprasellar mass lesion, with peripheral small solid component (arrow), which is showing hyperintense signal. The lesion is superiorly compressing the third ventricle extending upto the lateral ventricles. Optic chiasma is not seen separately from the lesion a b Sagittal post-gadolinium T1-weighted MR images showing solid cystic suprasellar mass lesion which is showing peripheral enhancements of the cysts and homogenous enhancement of the solid component. Pituitary gland is seen separately from the lesion (arrow) Pilocytic astrocytoma composed of glial cells with small cytoplasmic processes forming loose fibrillary matrix (H and E, ×100) [/fig]
|
Prehabilitation with wearables versus standard of care before major abdominal cancer surgery: a randomised controlled pilot study (trial registration: NCT04047524)
Background Prehabilitation aims to improve post-operative outcomes by enhancing pre-operative fitness but is labourintensive. This pilot study aimed to assess the efficacy of a tri-modal prehabilitation programme delivered by smartwatches for improving functional fitness prior to major abdominal cancer surgery. Methods A single-centre, randomised controlled pilot study, in which 22 patients were randomised to: (a) a prehabilitation group (n = 11), comprising of home-based exercise, nutritional, and dietary advice delivered using a wrist-worn smartwatch connected to a smartphone application; or (b) a control group (n = 11) receiving usual care, with patients given a smartwatch as a placebo. Eligible participants had over two weeks until planned surgery. The primary outcome was pre-operative physical activity including 6-min walk test (6MWT) distance, with secondary outcomes including change in body weight and hospital anxiety and depression score (HADS). Results Recruitment was 67% of eligible patients, with groups matched for baseline characteristics. The prehabilitation group engaged in more daily minutes of moderate [25.1 min (95% CI 9.79-40.44) vs 13.1 min (95% CI 5.97-20.31), p = 0.063] and vigorous physical activity [36.1 min (95% CI 21.24-50.90) vs 17.5 min (95% CI 5.18-29.73), p = 0.022] compared to controls. They also had significantly greater improvements in 6MWT distance compared to controls [+ 85.6 m (95% CI, + 18.06 to + 153.21) vs + 13.23 m (95% CI − 6.78 to 33.23), p = 0.014]. HADS scores remained unchanged from baseline in both groups. Conclusion Prehabilitation in the colorectal cancer care setting can be delivered using smartwatches and mobile applications. Furthermore, this study provides early indicative evidence that such technologies can improve functional capacity prior to surgery Trial registration NCT04047524.In the UK, patients undergoing major colorectal cancer surgery have 3% 90-day mortality and an 11% unplanned readmission rate from post-operative complications [1]. Tri-modal prehabilitation aims to enhance post-operative recovery by improving patients' pre-operative functional capacity, nutritional status and psychological readiness for surgery via exercise, dietary, and psychological support[2]. Delivery of prehabilitation before major abdominal cancer surgery has ranged from unsupervised home-based interventions (exercise booklets, and CDs) and computer programs, to supervised hospital-based programmes using cycle ergometers and exercise equipment[3,4]. Unsupervised home-based programmes are less labour-intensive and more cost-effective but are reported to have lower compliance rates (16-87%)[4,5].The impact of prehabilitation on functional capacity can be assessed using the 6-min walk test (6MWT)[6]. A clinically meaningful increase in 6MWT distance is reported to Supplementary Information The online version contains supplementary material available at https:// doi.
be 20 m [bib_ref] Meaningful change and responsiveness in common physical performance measures in older adults, Perera [/bib_ref] [bib_ref] Interpreting small differences in functional status: the six minute walk test in..., Redelmeier [/bib_ref] [bib_ref] Minimal clinically important differences in the six-minute walk test and the incremental..., Wise [/bib_ref] following 3-6 weeks of home-based prehabilitation [bib_ref] Prehabilitation versus rehabilitation: a randomized control trial in patients undergoing colorectal resection..., Gillis [/bib_ref] [bib_ref] Impact of a trimodal prehabilitation program on functional recovery after colorectal cancer..., Li [/bib_ref]. A recent pooled analysis of comparative studies suggests this may result a 5-year disease free survival in stage III colorectal cancer patients from 50.9% without prehabilitation versus 73.4% with (p = 0.044) [bib_ref] Improved disease-free survival after prehabilitation for colorectal cancer surgery, Trepanier [/bib_ref]. Delivery of prehabilitation across a healthcare system however, remains a challenge [bib_ref] Implementing a system-wide cancer prehabilitation programme: the journey of Greater Manchester's 'Prehab4cancer, Moore [/bib_ref]. Wearable technologies linked to smartphones through software applications can promote health behaviour change including physical activity, rehabilitation, weight loss, and may be ideally suited to delivering prehabilitation [bib_ref] The efficacy of wearable activity tracking technology as part of a weight..., Cheatham [/bib_ref] [bib_ref] Effect of a home-based exercise intervention of wearable technology and telephone coaching..., Mcdermott [/bib_ref] [bib_ref] A lot of action, but not in the right direction: systematic review..., Bender [/bib_ref]. This prospective randomised controlled pilot study aimed to determine whether tri-modal prehabilitation delivered via a wearable technology can successfully increase pre-operative physical activity levels and improve pre-operative functional capacity in patients undergoing major abdominal cancer surgery.
# Materials and methods
This was a single-centre, parallel-arm randomised controlled pilot study (NCT04047524). During the study period consecutive patients undergoing major abdominal cancer surgery at a tertiary cancer centre (The Christie NHS Foundation Trust, Manchester, UK) were randomised (1:1) to the intervention or control group. A sample size of 15 in each group was calculated based on the ability to detect an increase in the primary outcome 6MWT distance of 20 m in the prehabilitation versus the standard care group. The intervention group received a prehabilitation programme delivered using a Fitbit Smartwatch (Fitbit Inc, San Francisco, CA, USA) with a digital display and Smartphone Application (App). The control group received a Fitbit smartwatch with no display or patient feedback as a placebo. This meant comparable activity data was collected in both groups. Clinical aspects of care were not altered for either group. The study was approved by Health and Social Care REC A on 25/02/2019 and Health Research Authority on 05/03/2019. Written informed consent was obtained from all participants prior to enrolment.
## Inclusion criteria
Participants had to be aged 18 years or over, able to consent to participate in the prehabilitation programme, be undergoing major abdominal cancer surgery, have at least 2 weeks to their operation date, and be able to understand written and spoken English.
## Exclusion criteria
Participants were excluded if they had any health conditions which prevented them from safely taking part in a home-based exercise programme. These included patients who had within the past 3 months had a myocardial infarction or stroke. Participants were also excluded if they were already active users of an activity monitoring smartwatch. Patients were not excluded if they were already completing the UK national recommendation of 150 min physical activity per week [17] to avoid selection bias.
## Randomisation
A random sequence with 15 participants in each study group was generated using the online randomisation software "GraphPad" (GraphPad Software Inc, San Diego, CA, USA). This sequence was placed into sequentially numbered sealed opaque envelopes. After participants gave their fully informed consent and completed baseline questionnaires and observations, the appropriate sequential envelope was opened to obtain group allocation. All participants then completed baseline 6MWT.
## Blinding
All patients approached were informed that the study was investigating whether prehabilitation could have a role to play in improving pre-operative fitness. It was clearly explained that allocation to the control or prehabilitation arm of the study would not knowingly advantage or disadvantage them in any way. Whilst both groups were given wearable devices (placebo in the control group as outlined below), the inherent differences between the intervention and control groups made blinding challenging for study participants. The baseline 6MWT distance was used by the physiotherapist to determine baseline fitness in the prehabilitation group participants in order to help tailor the intensity of their physical activity.
## Wearable activity monitors
A Fitbit Charge 2 smartwatch was provided to each participant in the prehabilitation group. This has a screen that allows the wearer to self-monitor real-time physical activity levels. A Fitbit Flex 2 (placebo device) was provided to each participant in the control group. This is a band without a screen that can be set to not provide feedback on physical activity levels to the wearer. Both smartwatches record daily steps and duration and intensity of physical activity, which can be viewed via the Fitbit App (https:// www. fitbit. com), which was downloaded onto participants' smartphones (participants without a smartphone were loaned one). All participants were instructed to wear their device all days of the pre-operative period and received instructions on how to upload smartwatch data to the Fitbit App daily using Bluetooth wireless technology. In the control group the App was only used to collect data from their Fitbit with no feedback given. Data from the control group were used as an objective measure of daily physical activity levels in a sample of patients not provided with a prehabilitation programme, for comparison with the same data from the intervention group.
## Tri-modal prehabilitation group intervention
Prehabilitation group participants were shown how to use the Fitbit Charge 2 device and Fitbit App as a motivational tool for increasing daily physical activity levels and were shown how to use the Fitbit App food log to support dietary behaviour change. A separate mindfulness app was provided for stress management. Standardised structured weekly phone calls were provided to allow reporting of technical issues and provide tailored prehabilitation support.
## Structured exercise and physical activity
Participants in the prehabilitation group were assessed by a physiotherapist and given an individualised structured exercise and physical activity programme. This comprised aerobic exercise (3 × per week), resistance exercise using a resistance band consisting of 8-10 repetitions in two sets (2 × per week), and encouragement to engage in additional physical activity (e.g. increasing daily walking, use of stairs, etc.) or structured exercise of their choice for 30 min on two other days. The intensity of activity was tailored to each participant's individual fitness level, as shown in , with the physiotherapist also assessing baseline strength and providing a resistance band of appropriate resistance. Each participant's target heart rate was calculated using the Karvonen formula and participants were instructed to aim for this target heart rate during the prescribed exercise using the heart rate monitor on the Fitbit Charge 2. Participants were considered to be completing moderate-intensity activity when achieving 50-70% of their predicted maximum heart rate (based on 220-age) [bib_ref] The effects of training on heart rate; a longitudinal study, Karvonen [/bib_ref]. Participants were also provided with a BORG Perceived Exertion Scale and instructed to work at a level of perceived exertion between 12 and 16 indicating "somewhat hard" to "hard" physical activity.
## Nutrition
Increased protein intake through Oral Nutritional Supplementation has been demonstrated to reduce length of stay by 2 days following colorectal surgery [bib_ref] Effects of nutritional prehabilitation, with and without exercise, on outcomes of patients..., Gillis [/bib_ref] and poor preoperative nutritional status in cancer patients is associated with sarcopenia and myopenia which increase post-operative complications [bib_ref] Pre-operative nutrition and the elective surgical patient: why, how and what?, Gillis [/bib_ref] [bib_ref] Influence of preoperative nutritional status on clinical outcomes after pancreatoduodenectomy, Kim [/bib_ref]. Therefore, participants in the The intensity of exercise programmes provided to the participants prehabilitation group were provided with written dietary advice and watched a presentation on pre-operative nutrition which emphasised the importance of avoiding unintentional weight loss and increasing protein consumption to maintain muscle mass prior to surgery. This advice is presented in Online Appendix. They were shown how to use the Fitbit App food log and encouraged to use this to monitor their daily dietary intake and increase protein intake.
## Psychosocial support
The peri-operative period is associated with increased levels of anxiety and depression in cancer patients [bib_ref] Prevalence of depression in cancer patients: a meta-analysis of diagnostic interviews and..., Krebber [/bib_ref] [bib_ref] Prevalence of depression and anxiety in colorectal cancer patients: a literature review, Peng [/bib_ref]. Preoperative psychological support may relieve these symptoms and improve Quality of Life for patients awaiting cancer surgery and mindfulness apps have previously been shown to reduce symptoms of anxiety and depression in cancer patients [bib_ref] Quality of life among women diagnosed with breast cancer: a randomized waitlist..., Rosen [/bib_ref] [bib_ref] Feasibility of the mobile mindfulness-based stress reduction for breast cancer (mMBSR(BC)) program..., Lengacher [/bib_ref]. Participants in the prehabilitation group were therefore instructed to complete one guided meditation per day using a mindfulness app which provides stress management and relaxation techniques (Smiling Mind Pty Ltd-https:// www. smili ngmind. com. au/ smili ng-mind-app).
## Outcome measures
The primary outcome was quantitative data collected via a Fitbit aimed at determining the number of participants whose physical activity increased and by how much during the study. This included levels of physical activity (light, moderate, and vigorous) and functional walking capacity as measured by the 6MWT which was conducted according to standard operating procedures at baseline and on the day before surgery. Secondary outcomes included: change in body weight, and change in psychological well-being measured using the Hospital Anxiety and Depression (HADS) questionnaire from baseline to the day before surgery, daily step counts, and participant satisfaction [bib_ref] The validity of the hospital anxiety and depression scale. an updated literature..., Bjelland [/bib_ref]. Daily step counts were determined by the smartwatch accelerometers and active minutes were calculated using estimates of metabolic equivalents after 10 continuous minutes of physical activity. Patient recruitment and retention rates and adverse events were also recorded.
Both the Fitbit Flex 2 and Charge 2 use the same 3-axis accelerometer technology to record activity levels and Fitbit devices have previously demonstrated accuracy and interdevice reliability [bib_ref] Validity of Fitbit's active minutes as compared with a research-grade accelerometer and..., Brewer [/bib_ref] [bib_ref] Systematic review of the validity and reliability of consumer-wearable activity trackers, Evenson [/bib_ref]. Physical activity data were synchronised to an online study dashboard "Fitabase" (Fitabase, San Diego, CA, USA) which allowed the research team to perform retrospective analysis. An Information Governance officer thoroughly reviewed the Fitabase privacy notice (Available from: www. fitab ase. com/ resou rces/ knowl edgebase/ worki ng-with-the-irb/ data-secur ity-priva cy/) to ensure compliance to data security and ethical guidelines and had no concerns relating to the security of the de-identified data. Fitabase is under the Privacy shield framework, and they do not collect IP addresses or GPS data.
# Statistical analysis
Data analysis was undertaken using StatsDirect (StatsDirect Ltd, Merseyside, UK). Statistical significance was defined as a p value of < 0.05. A Shapiro-Wilk test was used to determine whether data were normally distributed. Normally distributed data were compared using an independent groups t test. Non-normally distributed data were compared using a Mann-Whitney U test. Within-group analyses were conducted using a paired T test. Qualitative data was generated from end-of-study questionnaires.
# Results
This study was open to recruitment between 14/05/2019 and 23/09/2019. A total of 33 participants were identified, of which 11 were either declined or were ineligible, leaving 22 participants randomised during the enrolment period [fig_ref] Figure 2: CONSORT flow diagram of participants through the study [/fig_ref]. The overall recruitment rate was 67% (rising to 79% when excluding patients who did not meet the inclusion criteria). Of note, no patient was excluded because of pre-existing medical conditions. Retention on the study was 100%.
## Baseline characteristics
Baseline characteristics of the prehabilitation and control groups are reported in [fig_ref] Table 1: Baseline characteristics and surgical details of the prehabilitation and control groups Data... [/fig_ref]. The mean duration from enrolment to the day before surgery was greater in the prehabilitation group at 30.5 (95% CI 18.7-42.2) days versus 20.8 (95% CI, 12.7-29.0) days in the control group (p = 0.072), although this was not statistically significant. This difference in days before surgery between the two groups occurred as two prehabilitation group participants had their scheduled surgery date unexpectedly delayed following enrolment. There was no significant difference between the two groups for baseline 6MWT distance or HADS [fig_ref] Table 2: Mean baseline HADS Scores and change in mean 6-min walk test distance... [/fig_ref].
## Patient compliance
Prehabilitation group participants wore the Fitbit Charge 2 on 98.9% of days and control group participants wore the placebo Fitbit Flex 2 on 95.8% of days between enrolment and surgery. Compliance with the exercise component was high, with prehabilitation group participants achieving 30 min of moderate-intensity activity on average 59.9% (95% CI, 41.3-78.5) of the days during the pre-operative period, compared to 42.6% (95% CI 18.4-66.7) of days in Usability of the Fitbit food log was also high, with prehabilitation group participants logging food on average 82.9% of the days. Participants on average only listened to the mindfulness App on 15% of days. Reasons for noncompliance to the mindfulness intervention included participants reporting good baseline mental health, finding the App unhelpful, and only using the App when they felt they needed to. Most participants (73%) already owned a suitable smartphone or tablet and the other 27% (six participants) were loaned a smartphone due to incompatibility of their own device with the FitBit, connection issues, or not owning a smartphone. One technical issue was reported during the study (a Fitbit Flex 2, which stopped charging). No adverse events were reported. [fig_ref] Table 2: Mean baseline HADS Scores and change in mean 6-min walk test distance... [/fig_ref] presents the 6MWT distance at baseline and on the day before surgery for both study groups. The mean change in 6MWT distance for the prehabilitation group during the pre-operative period was + 85.6 m (95% CI 18.1-153.2 m) compared to + 13.2 m (95% CI − 6.8 to 33.2 m) in the control group (p = 0.014). All participants in the prehabilitation group improved their 6MWT distance, with nine participants (82%) increasing by 20 m or more. In the control group, eight participants increased their 6MWT distance, with only three (27%) participants improving by more than 20 m. In addition, three control group participants (27%) declined in 6MWT performance, with one participant (9%) declining by greater than 20 m.
## Physical activity levels
## Functional capacity
## Body weight
Changes in body weight during the pre-operative period were minimal for both study groups. Mean change in weight in the prehabilitation group was + 0.46 kg (range − 1.0 to + 2.5 kg) versus − 1.06 kg (range − 3.85 to + 0.70 kg) in the control group. Qualitative analysis of the nutritional intervention identified that all prehabilitation participants used the food log on most days. Participants stated that the food log "gave me increased focus on my calorie control" and "helped to increase my food intake, but protein targets were difficult to stick to", although two participants stated the app-based food diary was confusing and difficult to use.
## Participant satisfaction
Following completion of the prehabilitation programme, participants completed an end-of-study questionnaire.
Responses to this questionnaire are presented in [fig_ref] Table 3: Prehabilitation group participant's responses to the End-of-Study Questionnaire [/fig_ref].
Overall, 100% of participants rated both the prehabilitation programme overall and the exercise component "Good" or "Excellent". Ten participants (90%) responded "Strongly Agree" or "Agree" to the statement "The Fitbit motivated me to do the physical activity that was part of the prehabilitation programme". In addition, participants in the control group were asked "The Fitbit motivated me to increase my physical activity although I was not asked to complete an exercise programme", to which five (45%) participants stated they "Strongly Agree" and four (36%) participants stated they "Agree", with two (18%) responding "Neutral".
# Discussion
Prehabilitation aims to improve pre-operative functional capacity and "cardiopulmonary reserve" through preoperative exercise, nutritional optimisation, and enhancing psychological readiness for surgery. The results of this study suggest that a programme of remotely supported tri-modal prehabilitation may increase pre-operative functional capacity before major abdominal cancer surgery. Smartwatches and associated technologies were used as the mainstay support for a prescribed programme of structured exercise/physical activity, optimal nutrition, and stress management. Participants allocated to prehabilitation increased their pre-operative 6MWT distance by an average of 85.6 m, with nine participants achieving an increase of > 20 m, suggesting a clinically important improvement in functional capacity. A previous RCT has demonstrated marginally greater increases in 6MWT distance in an exercise prehabilitation group compared to a control group (+ 31 m vs + 27 m) following home-based prehabilitation; however, this study required a physical therapist to make regular home visits (on six occasions over 4 weeks) to ensure compliance to the exercise programme [bib_ref] Responsive measures to prehabilitation in patients undergoing bowel resection surgery, Kim [/bib_ref]. Conversely, Carli and colleagues found a reduction in 6MWT distance in a bike/strengthening group (− 6.8 m) following home-based prehabilitation How would you rate the prehabilitation programme overall? 5 (45%) 6 (55%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) How would you rate the exercise component of the prehabilitation programme? 5 (45%) 6 (55%) 0 (0%) 0 (0%) 0 (0%) 0 (0%)
How would you rate the dietary component of the prehabilitation programme?
[formula] 3 (27%) 5 (45%) 3 (27%) 0 (0%) 0 (0%) 0 (0%) [/formula]
How would you rate the mental well-being component of the prehabilitation programme?
[formula] 3 (27%) 3 (27%) 2 (18%) 0 (0%) 1 (9%) 2 (18%) [/formula]
Strongly agree Agree Neutral Disagree Strongly disagree Not answered
The Fitbit Charge 2 was easy to use 9 (81%) 0 (0%) 1 (9%) 0 (0%) 0 (0%) 1 (9%) The Fitbit Charge 2 was comfortable to wear 10 (90%) 1 (9%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) The Fitbit motivated me to do the physical activity that was part of the prehabilitation programme 5 (45%) 5 (45%) 1 (9%) 0 (0%) 0 (0%) 0 (0%)
The tailored exercise regime helped me to increase my exercise levels 5 (45%) 6 (55%) 0 (0%) 0 (0%) 0 (0%) 0 (0%)
Taking part in this programme has encouraged me to increase the amount of exercise I do once I recover from surgery 8 (72%) 2 (18%) 1 (9%) 0 (0%) 0 (0%) 0 (0%) and compliance rate of only 16% despite home visits and weekly phone calls [bib_ref] Randomized clinical trial of prehabilitation in colorectal surgery, Carli [/bib_ref]. Our study has demonstrated that functional walking capacity may be increased to a desired threshold with prehabilitation delivered through a wearable technology and smartphone application. The acceptability and compliance with this device and its associated exercise intervention was high with patients reaching 30 min moderate-intensity activity on 84% of days they were encouraged to do so. This compares to compliance rates of 16-87% in previously reported pilot studies of home-based prehabilitation [bib_ref] A systematic review of prehabilitation programs in abdominal cancer surgery, Hijazi [/bib_ref]. Daily activity levels were higher in the prehabilitation compared to the control group, with the latter completing significantly more minutes of vigorous daily exercise and increased minutes of moderate daily exercise. Feedback from end-of-study questionnaire suggests that the Fitbit wearable devices were an independent factor responsible for encouraging participants to complete the prescribed exercise by providing easily visible feedback on progress towards activity goals.
The nutritional intervention aimed to prevent unintentional weight loss prior to surgery which is particularly important in cancer patients. The relatively high baseline BMI of our groups (27-30 kg/m 2 ) reflects the nutritional and body mass status of patients in our UK population, although it should be noted that BMI is itself a crude measure of nutrition. One of the limitations of prehabilitation before cancer surgery is there are not more than 2-3 weeks in which to deliver a nutritional intervention. Changes in body weight from baseline to the day before surgery were minimal in both groups but this intervention may be more relevant for frailer patients to prevent further weight loss before surgery and warrants further review. Qualitative comments suggest that the protein targets were challenging for patients, suggesting the use of a food log App may need further development to improve efficacy in this setting. Oral Nutritional Supplementation has previously been shown to reduce post-operative length of stay following colorectal surgery [bib_ref] Effects of nutritional prehabilitation, with and without exercise, on outcomes of patients..., Gillis [/bib_ref] and could be used alongside dietary advice in frailer patients to enhance the dietary component of our prehabilitation programme. The incorporation of a nutritional status questionnaire, such as the Nutrition Risk Screening Tool 2002 [bib_ref] validation of nutritional risk screening-2002 in a hospitalized adult population, Bolayir [/bib_ref] , into baseline assessments could identify those patients at greatest risk for weight loss and malnutrition and ONS could be provided to these at-risk patients.
The psychological intervention aimed to reduce preoperative stress and anxiety by teaching mindfulness stress management techniques. In-group analysis revealed that reductions in anxiety and depression in the prehabilitation group were not statistically significant. This may reflect poor compliance with the psychological intervention or poor efficacy of the mindfulness app in this context, suggesting further development of the mindfulness component may also be needed.
Several limitations of this pilot study should be noted. Firstly, participants could not be completely blinded to their allocated study group. This may have generated bias as participants in the control group were aware that we were investigating the impact of pre-operative exercise and may have altered their exercise levels accordingly. The control group were provided with a placebo Fitbit Flex 2 (no display screen or feedback), with our end-of-study questionnaire suggesting 82% of control group participants stated that they increased physical activity just because they were given a placebo device to wear. This demonstrates how even just wearing a smartwatch can increase motivation for physical activity. It is important to note that we did not exclude patients who were already completing the national recommendation of 150 min physical activity per week, but even these patients reported enjoying participation and that it "gave them something to focus on", with a beneficial effect on their mental health. Finally the sample size in this study was small with a risk of Type 2 error.
To our knowledge, this is the first pilot study to demonstrate the efficacy of delivering home-based prehabilitation prior to major abdominal cancer surgery using wearable smart technologies, with clinically and statistically significant improvements in functional walking capacity and higher levels of vigorous physical activity in the prehabilitation group when compared to controls. This study has demonstrated that the delivery of home-based prehabilitation using smartwatches and mobile applications is well-accepted in patients awaiting major abdominal cancer surgery, and that the primary outcomes of 6-min walk test (6MWT) distance and levels of activity as measured by wearable devices may be reliably used in future trials. We feel that further optimisation of the nutritional and psychological components is required, perhaps through further integration into a single smartphone application. A fully powered RCT is warranted to investigate the effects of a prehabilitation programme on post-operative outcome measures such as length of hospital stay and complication rates in this setting.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.
[fig] Figure 2: CONSORT flow diagram of participants through the study [/fig]
[table] Table 1: Baseline characteristics and surgical details of the prehabilitation and control groups Data are expressed as mean (95% CI), n (%), or male:female AP abdominoperineal, BMI body mass index, CRS with HIPEC cytoreductive surgery and hyperthermic intraperitoneal chemotherapy, ECOG Eastern Cooperative Group [/table]
[table] Table 2: Mean baseline HADS Scores and change in mean 6-min walk test distance from baseline until the day before surgery (pre-operative) Data are presented as mean (95% CI) or n (%) 6MWT 6-min walk test, HADS hospital anxiety and depression scale a "No change" represents participants whose pre-operative 6MWT distance was within ± 20 m of their baseline 6MWT distance [/table]
[table] Table 3: Prehabilitation group participant's responses to the End-of-Study Questionnaire [/table]
|
End-of-life discussions reduce the utilization of life-sustaining treatments during the last three months of life in cancer patients
Studies to examine the impact of end-of-life (EOL) discussions on the utilization of life-sustaining treatments near death are limited and have inconsistent findings. This nationwide population-based cohort study determined the impact of EOL discussions on the utilization of life-sustaining treatments in the last three months of life in Taiwanese cancer patients. From 2012 to 2018, this cohort study included adult cancer patients, which were confirmed by pathohistological reports. Life-sustaining treatments during the last three months of life included cardiopulmonary resuscitation, intubation, and defibrillation. EOL discussions in cancer patients were confirmed by their medical records. Association of EOL discussions with utilization of life-sustaining treatments were assessed using multiple logistic regression. Of 381,207 patients, the mean age was 70.5 years and 19.4% of the subjects received life-sustaining treatments during the last three months of life. After adjusting for other covariates, those who underwent EOL discussions were less likely to receive life-sustaining treatments during the last three months of life compared to those who did not (Adjusted odds ratio [AOR] 0.87; 95% confidence interval [CI] 0.85-0.89). Considering the type of treatments, EOL discussions correlated with a lower likelihood of receiving cardiopulmonary resuscitation (AOR = 0.45, 95% CI 0.43-0.47), endotracheal intubation (AOR = 0.92, 95%CI 0.90-0.95), and defibrillation (AOR = 0.54, 95%CI 0.49-0.59).Since EOL discussions are associated with less aggressive care, our study supports the importance of providing these discussions to cancer patients during the EOL treatment.Cancer is the second leading cause of death worldwide, and was responsible for an estimated 10 million deaths in 2020 1 . Patients often experience pain, dyspnea, and distress during the end-of-life (EOL) care. Aggressive EOL treatment has predicted lower quality of life 2,3 and greater regret about treatment decisions 4 .
www.nature.com/scientificreports/ Advance care planning (ACP) is the process of discussing patient preferences regarding life-sustaining treatments and place of death during EOL care, and involves healthcare professionals, patients, and their family [bib_ref] The effects of advance care planning on end-of-life care: A systematic review, Brinkman-Stoppelenburg [/bib_ref]. It is considered a means for helping patients die in their preferred treatment and is emphasised as an end-of-life care strategy [bib_ref] The effects of advance care planning on end-of-life care: A systematic review, Brinkman-Stoppelenburg [/bib_ref]. A previous report showed that patients engaging in ACP have more stable EOL preferences compared to those who do not [bib_ref] Stability of end-of-life preferences: A systematic review of the evidence, Auriemma [/bib_ref].
Comprehensive EOL discussions with physicians during the ACP consultation could allow cancer patients to confront the limitations of medical treatments and consider their preferences regarding life-sustaining treatments through the EOL treatment. A meta-analysis review showed that these discussions predicted an increased use of hospice care and better quality of death [bib_ref] Associations among end-of-life discussions, health-care utilization, and costs in persons with advanced..., Starr [/bib_ref]. Although cancer patients may avoid discussing their preference of life-sustaining treatments during EOL discussions [bib_ref] Associations between end-of-life discussions, patient mental health, medical care near death, and..., Wright [/bib_ref] [bib_ref] Oncologists' attitudes toward and practices in giving bad news: An exploratory study, Baile [/bib_ref] , healthcare providers are allowed-or morally obligated-to decide whether life-sustaining treatments are viable for terminal cancer patients [bib_ref] Universal do-not-resuscitate orders, social worth, and life-years, Geijteman [/bib_ref].
Although EOL discussions offer patients the opportunity to define their expectations for life-sustaining treatments that they want to receive during the EOL treatment, the impact of EOL discussions on the use of lifesustaining treatments near death has not been extensively studied. A US cohort study followed up 332 patients with advanced cancer and found that patients who participated in EOL discussions were less likely to receive ventilation (adjusted odds ratio [AOR] 0.26; 95% confidence interval [CI] 0.08-0.83) and resuscitation (AOR 0.16, 95% CI 0.03-0.80) during the EOL treatment compared to those who did not [bib_ref] Associations between end-of-life discussions, patient mental health, medical care near death, and..., Wright [/bib_ref]. In another observational study in the US involving 145 patients with advanced cancer, EOL discussions lowered the rates of ventilation and resuscitations during the EOL treatment 2 . Previous studies on the association between EOL discussions and the use of life-sustaining treatment, however, have used small sample sizes (n = 145) 2 and inadequately controlled for potential confounders such as dementia [bib_ref] Health care costs in the last week of life: Associations with end-of-life..., Zhang [/bib_ref] [bib_ref] Associations between end-of-life discussions, patient mental health, medical care near death, and..., Wright [/bib_ref].
EOL discussions serve to maximize patients' quality of life while considering their goals and expectations regarding EOL treatment. It is important to understand the impact of these discussions on the administration of life-sustaining treatments among terminally ill patients. We therefore conducted a nationwide population-based cohort study to determine the impact of EOL discussions on the utilization of life-sustaining treatments during the last three months of life among Taiwanese cancer patients, from 2012 to 2018.
# Methods
## Promotion of end-of-life discussions in terminally ill patients in taiwan. the taiwan national
Health Insurance Administration, since 2012, has initiated an EOL discussions program that encourages healthcare providers to provide these discussions to terminally ill patients. While patients were admitted to the hospital, healthcare providers offered EOL discussions to those with life-limiting diseases. During the discussions, healthcare providers discuss patients' goals and preferences concerning medical care and life-sustaining treatments towards the end of their lives. Each EOL discussion with a patient awards the healthcare provider with the equivalent of US$75. There should be a record of all discussion content in the patient's medical chart and each session must be at least an hour long 11 . Study population. This nationwide cohort study used the Taiwan National Health Insurance Research Database (NHIRD), which contains healthcare data from more than 99% of the population in Taiwan 12 . It is a large-scale database derived from the national health insurance system, which consists of registration files and original claims data. The database de-identified and scrambled the patients' identification codes before releasing the data for research purposes.
This cohort study selected subjects aged 18 years or older who had received a cancer diagnosis and died between January 1, 2012 and December 31, 2018. We identified these patients from the Registry for Catastrophic Illness. In Taiwan, the Registry for Catastrophic Illness requires a peer review of pathohistological reports before a cancer diagnosis is reported [bib_ref] Latent tuberculosis infection and the risk of subsequent cancer, Su [/bib_ref]. This registry was also linked to Taiwan's death certificate database to confirm the demise of cancer patients 15 . The Institutional Review Board of Taipei City Hospital (no. TCHIRB-10709107-W) approved this study. The informed consents for study participants were waived by the Institutional Review Board of Taipei City Hospital. All methods in this study were performed in accordance with relevant guidelines and regulations. This study is also in accordance with the Declaration of Helsinki.
Outcome variable. The outcome was life-sustaining treatments during the last three months of life in cancer patients. Life-sustaining treatments included cardiopulmonary resuscitation, intubation, and defibrillation Main explanatory variable. The main explanatory variable was EOL discussions with physicians, which was determined by patients' medical records.
Controlling variables. The controlling variables included sociodemographic characteristics, type of cancer, and comorbidities. Sociodemographic factors included age, sex, urbanization, and income level. Urbanization described whether subjects resided in urban, suburban, or rural areas 17 . We calculated the income level from the average monthly income of the insured person and grouped it into three categories: low (≤ 19,200 New Taiwan Dollars [NTD]), intermediate [bib_ref] Advance directives and life-sustaining treatment: Attitudes of Hong Kong Chinese elders with..., Ting [/bib_ref] NTD to < 40,000 NTD), and high (≥ 40,000 NTD). Type of cancer and the comorbidities in the study subjects were determined according to the International Classification of Diseases, Ninth and Tenth Revision, Clinical Modification (ICD-9-CM and ICD-10-CM) code. The type of cancer included solid tumors and hematologic malignancies. Subjects' comorbidities include diabetes, chronic kidney disease, congestive heart failure, coronary heart disease, liver cirrhosis, chronic obstructive pulmonary disease, dementia, cerebrovascular disease, and depressive disorder (Supplementary Table S1). A person was considered to have a comorbidity only if the condition occurred in an inpatient setting or in three or more outpatient visits [bib_ref] Association of pulmonary tuberculosis and ethambutol with incident depressive disorder: A nationwide,..., Yen [/bib_ref] www.nature.com/scientificreports/ Statistical analysis. First, we analyzed the subjects' demographic data. We then analyzed categorical data using the Pearson χ 2 test where appropriate. We presented continuous data as mean ± standard deviation (SD), and conducted a two-sample t-test to compare outcomes between patients who underwent EOL discussions and those who did not.
We assessed the crude associations of EOL discussions and other covariates with the outcome (utilization of life-sustaining treatments during the last three months of life) by computing the odds ratios (ORs) and corresponding 95% confidence intervals (CIs). We then performed a multivariate logistic regression to estimate the association between EOL discussions and the utilization of life-sustaining treatments after adjusting for potential confounders. A variable with p < 0.05 was defined as a significant factor associated with the utilization of lifesustaining treatments in the multivariate analysis. Adjusted odds ratios (AOR) with 95% confidence intervals (CI) indicated the strength and direction of these associations.
We conducted subgroup and sensitivity analyses to examine the associations between EOL discussions and the utilization of life-sustaining treatments, after stratifying participants by age and sex. To examine the robustness of the main findings, propensity score analysis was conducted to evaluate the associations between hospice care services and utilization of life-sustaining treatments during the last three months of life. Logistic regression was conducted to calculate the probability of enrolling in a hospice care program in cancer patients according to their age, sex, income level, type of cancer, and comorbidities. A multivariate logistic regression was used to estimate the association between hospice care services and utilization of life-sustaining treatments after adjusting for patients' propensity score. We performed all data management and analyses using the SAS® v9.4 statistical software package (SAS Institute, Cary, NC, USA).
# Results
Participant selection. This cohort study involved the raw data of 383,128 deceased cancer patients from 1 January 2012 to 31 December 2018. After excluding those younger than 18 years (n = 850) and those with missing data (n = 1071), 381,207 patients were eligible for the purposes of the study. The overall mean (SD) age was 70.5 (14.3) years; 65.6% of the subjects were male; and 14.1% had EOL discussions with healthcare providers towards the end of their lives. The median duration from the EOL discussion to death for the 53,783 patients completing EOL discussions with healthcare providers was 34 days (interquartile range 16-79 days); of these patients, 7382 (13.7%) had such discussions after utilizing life-sustaining treatment. [fig_ref] Table 1: Characteristics of deceased cancer patients by end-of-life discussions [/fig_ref] shows the characteristics of cancer patients with and without EOL discussions before death. The patients who participated in EOL discussions tended to be younger and were more likely to be female. The proportion of patients completing EOL discussions with healthcare providers significantly increased from 3. shows the univariate and multivariate analyses of factors associated with life-sustaining treatments during the last three months of life among the cancer patients. After adjusting for the sociodemographic factors and comorbidities, those who underwent EOL discussions were less likely to receive life-sustaining treatments during the last three months of life, compared to those who did not (AOR = 0.87, 95% CI 0.85-0.89). [fig_ref] Table 3: Multivariate [/fig_ref] shows the multivariate analyses for the association of EOL discussions with cardiopulmonary resuscitation, endotracheal intubation, and defibrillation in cancer patients. After adjusting for the sociodemographic factors and co-morbidities, patients who underwent EOL discussions were less likely to receive cardiopulmonary resuscitation (AOR = 0.45, 95% CI 0.43-0.47), endotracheal intubation (AOR = 0.92, 95% CI 0.90-0.95), and defibrillation (AOR = 0.54, 95% CI 0.49-0.59) during their last three months of life, compared to those who did not. shows the results of subgroup analyses of the association between EOL discussions and life-sustaining treatments after stratifying participants by age and sex. In all patient subgroups, the discussions significantly predicted a lower likelihood of receiving life-sustaining treatments towards the end-of-life.
## Characteristics of patients by end-of-life discussions.
## Association of end-of-life discussions with utilization of life-sustaining treatments.
## Association of end-of-life discussions with cardiopulmonary resuscitation, endotracheal intubation, and defibrillation.
## Subgroup and sensitivity analysis for the association between end-of-life discussions and life-sustaining treatments.
Propensity score analysis for the association between end-of-life discussions and utilization of life-sustaining treatments. Propensity score analysis showed that cancer patients with EOL discussions were less likely to receive life-sustaining treatments during the last three months of life than those without the discussion (AOR = 0.87, 95% CI 0.85-0.89), including cardiopulmonary resuscitation (AOR = 0.45, 95% CI 0.43-0.47), endotracheal intubation (AOR = 0.92, 95% CI 0.90-0.94), and defibrillation (AOR = 0.54, 95% CI 0.49-0.59). Our study revealed robust associations between EOL discussions and life-sustaining treatments after stratifying patients by age and sex. In all patient subgroups, the discussions significantly lowered the likelihood of receiving life-sustaining treatments near death.
This study found that 14.1% of cancer patients in Taiwan participated in EOL discussions, which was lower than 37.0% among cancer patients in the US 8 . The culture difference regarding death-related issues between Asia and Western countries may explain the lower rate of completing EOL discussions among Taiwanese cancer patients. In traditional Chinese culture, death-related topics are thought to be taboo, and mentioning it is sacrilegious and to be avoided [bib_ref] Advance directives and life-sustaining treatment: Attitudes of Hong Kong Chinese elders with..., Ting [/bib_ref]. A previous survey in Hong Kong reported that 73% of elderly patients with chronic diseases have never received discussions regarding their EOL treatment [bib_ref] Advance directives and life-sustaining treatment: Attitudes of Hong Kong Chinese elders with..., Ting [/bib_ref]. Since EOL discussions provide patients www.nature.com/scientificreports/ the opportunity to define their goals and expectations regarding the EOL treatment 20 , our study suggests that it is imperative to provide such discussions for cancer patients nearing the end of their lives. This study found that 19.4% of cancer patients in Taiwan received life-sustaining treatments in the last three months of life. A previous report showed that the proportion of cancer patients receiving intensive care in the last six months of life was 40.3% in the US, 17.5% in Belgium, 15.2% in Canada, and 8.2% in Germany 21 . Since intensive care in patients with terminal diseases is associated with lower quality of life 3 , our study suggests that healthcare professionals should proactively provide EOL discussions for terminally ill patients to discuss their preferred treatment during EOL care.
The present cohort study found that cancer patients who underwent EOL discussions had 18% lower rates of receiving life-sustaining treatments in the last three months of life, than those who did not. EOL discussions could allow terminally ill patients to consider their preferences regarding life-sustaining treatments during the EOL care. A previous report showed that EOL discussions significantly increased the rates of do-not-resuscitate order completion among advanced cancer patients [bib_ref] Impact of end-of-life discussions on the reduction of Latino/non-Latino disparities in do-not-resuscitate..., Shen [/bib_ref]. Another cohort study in the US found that EOL discussions correlated with lower rates of ventilation and resuscitation near death [bib_ref] Associations between end-of-life discussions, patient mental health, medical care near death, and..., Wright [/bib_ref]. Moreover, a study in the US analyzed 145 cancer patients and found that patients who had undergone EOL discussions had lower rates of receiving ventilation and resuscitations in the last week of life 2 . Our cohort study followed up 381,207 cancer patients and found that EOL discussions associated with a lower utilization of life-sustaining treatments during the last three months of life. With regard to treatment types, cancer patients who had EOL discussions had a lower likelihood of receiving cardiopulmonary resuscitation, endotracheal intubation, and defibrillation during EOL care. Thus the findings of our study suggest that EOL discussions could reduce the utilization of life-sustaining treatments in cancer patients near death.
The EOL conversations with physicians improve the patients' understanding of the futility of aggressive therapies during EOL care 7 . This may account for the lower rate of receiving life-sustaining treatments at that stage. While cancer patients had EOL discussions with physicians, the benefit and harm of life-sustaining treatments during the EOL care would be discussed and emphasized. Through the EOL discussions with healthcare providers, cancer patients could consider and document their treatment preferences during EOL care. A prior study showed that EOL discussions significantly improved patients' understanding of their terminal illness and subsequently reduced the likelihood of receiving aggressive treatments during EOL care [bib_ref] Socialization to dying: Social determinants of death acknowledgement and treatment among terminally..., Prigerson [/bib_ref]. As EOL discussions are associated with less aggressive care 7 , the findings of our study suggest that it is important to provide them for cancer patients during EOL treatment.
This study showed that cancer patients receiving life-sustaining treatments during EOL care had higher medical expenditures than those not. A previous study involving 314 veterans in the US showed that aggressive EOL treatment was associated with significantly higher inpatient costs during a terminal hospitalization [bib_ref] Cost and utilization outcomes of patients receiving hospital-based palliative care consultation, Penrod [/bib_ref]. Another US study using administrative data found that EOL discussion was associated with significant hospital cost savings among patients during EOL care [bib_ref] Cost savings associated with US hospital palliative care consultation programs, Morrison [/bib_ref]. Since aggressive EOL treatment is associated with higher health care cost and poor quality of EOL care in terminally ill patients 2 , our study suggests that EOL discussions should be provided for cancer patients to discuss their treatment preference during EOL care.
This nationwide cohort study has several strengths. First, this study is the largest cohort study to determine the impact of EOL discussions on the utilization of life-sustaining treatments during the last three months of life among cancer patients. The research design enhances the validity of our findings by including unbiased subject selection and strict cancer diagnostic criteria. Moreover, since the Taiwan National Health Insurance covers all medical care of cancer patients, this nationwide population-based study traced all cancer patients, thus minimizing referral bias. Furthermore, the study's large sample size sufficiently detected the real, albeit subtle, difference between cancer patients who had and those who had not received EOL discussions towards the end of their lives.
This cohort study has several limitations. First, there may be important factors (e.g., patients' religion, functional status, and educational level) associated with the decision of receiving life-sustaining treatments, which the National Health Insurance Research Database did not record. Second, of the 53,783 patients completing EOL discussions with healthcare providers, 7382 (13.7%) had such discussions after utilizing life-sustaining treatment, which would underestimate the impact of EOL discussions on the utilization of life-sustaining treatments during the last three months of life among cancer patients. Finally, since all our enrollees were Taiwanese, the external validity of our findings may be a concern. Therefore, the generalizability of our results to other non-Asian ethnic groups requires further verification. However, our findings suggest new avenues for future research.
In summary, this nationwide, cohort study found that cancer patients who had EOL discussions were less likely to receive life-sustaining treatments during the last three months of life compared to those who did not have EOL discussions. With regard to treatment types, EOL discussions correlated with a lower likelihood of receiving cardiopulmonary resuscitation, endotracheal intubation, and defibrillation during the last three months of life. As end-of-life discussions are associated with less aggressive EOL care, it is important to provide them for cancer patients during EOL treatment.
[table] Table 1: Characteristics of deceased cancer patients by end-of-life discussions. EOL, end-of-life; SD, standard deviation; COPD, chronic obstructive pulmonary disease. a Unless stated otherwise. P value Total N = 381207 Patients with EOL discussions, n = 53,783 Patients without EOL discussions, n = 327,424 Table 2. Univariate and multivariate analysis of factors associated with utilization of life-sustaining treatments during the last 3 months of life among deceased cancer patients. AOR, adjusted odds ratio; CI, confident interval; EOL, end-of-life; COPD, chronic obstructive pulmonary disease. a During the last 3 months of life. [/table]
[table] Table 3: Multivariate [/table]
|
The Origin of Variation in Primary Care Process and Outcome Indicators
Healthcare providers are often evaluated by studying variability in their indicators. However, the usefulness of this analysis may be limited if we do not distinguish the variability attributable to health professionals and organizations from that associated with their patients.Our objectives are to describe the main process and outcome indicators of primary healthcare services, analyzing the contribution to variability in these indicators from different levels: individual, health professional, health center, and health district. This is a cross-sectional study that includes all the individuals covered by the public Basque Health Service (children [age 0-13], n ¼ 247,493; adults [!14 years old], n ¼ 1,959,682) over a 12-month period.We calculated the number of visits to primary care doctors, number of referrals, prescription costs, and potentially avoidable hospitalizations for ambulatory care sensitive conditions (ACSCs). Using multilevel analysis, we determined the percentage of variance attributable to each level.After adjusting for the characteristics of patients (demographic, socioeconomic, and morbidity), doctors (panel size), health center (size, staff satisfaction, demographic structure of the community), and health district, the variance in the indicators was mainly attributable to differences between patients, independently of the attending health professional, the center, or the healthcare organization, both in children (94.21% for visits to the doctor; 96.66% for referrals; 98.57% for prescription costs; 90.02% for potentially avoidable hospitalizations for ACSCs) and in adults (88.10%; 96.26%; 97.92%; and 93.77%, respectively).The limited contribution of health professionals and organizations to variability in indicators should be taken into account when performing evaluations and planning quality improvement strategies. (Medicine 94(31):e1314) Abbreviations: ACG = Johns Hopkins Adjusted Clinical Groups, ACSCs = ambulatory care sensitive conditions, ADGs = Aggregated Diagnosis Groups, ANOVA = analysis of variance, EHRs = electronic health records, GLMM = generalized linear mixed model, ICC = intraclass correlation coefficient, ICD9-CM = International Classification of Diseases, Ninth Revision, Clinical Modification, MEDEA = Mortalidad en áreas pequeñas Españolas y Desigualdades socio-Económicas y Ambientales (Mortality in small Spanish areas and socio-economic and environmental inequalities), OLS = ordinary least square.
# Introduction
V ariation in medical practice is a topic of interest in primary as well as other levels of care. Unexplained differences in healthcare may influence equity, outcomes of care provided and efficiency in the use of resources. As a consequence, numerous initiatives have been developed to measure and analyze such differences and to understand their causes.
Both in industrialized 1 and developing countries, [bib_ref] Provider-specific report cards: a tool for health sector accountability in developing countries, Mcnamara [/bib_ref] it is common to compare indicators obtained in populations under the care of primary care professionals. The results of this profiling have been used as a training and motivational tool for doctors (telling them their position with respect to their colleagues), to give financial rewards to those achieving good outcomes, and to guide patients toward the best-performing providers. [bib_ref] Performance profiling in primary care: does the choice of statistical model matter?, Eijkenaar [/bib_ref] In this context, since 2005, in the Basque Country (Spain) the Basque Health Service (Osakidetza) has used an annual evaluation system that assesses primary care doctors, health centers, and healthcare districts in terms of efficiency in the use of healthcare resources and quality of care. In order to compare apples to apples, this assessment is adjusted for age, sex, and morbidity in the populations seen, using the Adjusted Clinical Group (ACG) Case-Mix System developed at Johns Hopkins University. The logic behind this is that, once we have adjusted for differences between patients, variations in outcomes are attributable to medical practice, and hence, said outcomes may be modified directly, by acting on the way health professionals work. However, the usefulness of this approach may be limited, unless we take into account the hierarchical nature of data.Patients are not isolated units, independent of one another; rather they are individuals grouped in doctors' lists, and in turn, doctors are based in particular health centers and healthcare organizations. Although some studies have indicated the need to use hierarchical statistical models for making comparisons between hospitals, [bib_ref] Variation in the risk of readmission among hospitals: the relative contribution of..., Singh [/bib_ref] health centers, [bib_ref] The unreliability of individual physician ''report cards'' for assessing the costs and..., Hofer [/bib_ref] and doctors, [bib_ref] Performance profiling in primary care: does the choice of statistical model matter?, Eijkenaar [/bib_ref] [bib_ref] The unreliability of individual physician ''report cards'' for assessing the costs and..., Hofer [/bib_ref] [bib_ref] Meaningful variation in performance: a systematic literature review, Fung [/bib_ref] very few publications to date have analyzed the contribution of different levels of care to variability in primary care indicators.
In this study, we used the data from all Osakidetza primary care physicians providing healthcare to more than 2 million people, almost the entire population of the geographical area under study (Basque Country, Spain). Our objective was to describe the main indicators for measuring resource use (visits to doctors, referrals, and prescriptions), healthcare outcomes (potentially avoidable hospitalizations) and their relationship with factors related to patients, doctors, and health centers, as well as to ascertain the contribution of each of these levels to the variability observed in these indicators.
# Methods
This was a cross-sectional study, analyzing healthcare process and outcome indicators in the public health service in the Basque Country over a period of 1 year. The protocol was approved by the Clinical Research Ethics Committee of the Basque Country (reference number: 11/2012).
## Setting
Osakidetza is a public organization funded by taxes. It provides near-universal coverage for the population living in the Basque Country. Care is free at the point of delivery except for prescriptions, for which there is copayment, this varying with the type of disease and the patient's social and employment status.
At the time of this study, in 2008, primary care health services were organized into 7 health districts, corresponding to geographical areas. Each health district was composed of 9 to 22 health centers. Funded by annual contracts with the Department of Health of the Government of the Basque Country, health districts are organizationally independent and accountable in economic terms.
Primary care physicians are public sector workers, receiving a salary and a capitation payment, and work in multidisciplinary care teams. Every inhabitant is included on the list of a doctor, who is a pediatrician or a family doctor depending on the patient's age (<14 vs !14 years old). These primary care doctors act as gatekeepers to other levels of care. Electronic health records (EHRs) started to be introduced in 1990 and are now used by all primary care doctors.
## Study population and period
The observation period was set at 1 year, from September 1, 2007 to . In such period, the total population covered by the Basque public healthcare system was 2,265,058 individuals. We considered that the primary care doctor was responsible of the care of a patient if such person was registered in his/her list most part of the year of study (ie, 185 or more days), independently whether they had any type of contact. Due to several reasons (migration, dying, newborn, and multiple changes of physician or administrative modifications) 57,883 people (2.6%) did not complete such 6-month period with the same doctor and were excluded from our analyses. So, the study population included almost the entire population of the Basque Country.
In this study, we analyzed the data from across the public primary care network: 130 health centers, 1193 family doctors, and 286 pediatricians. The total number of registered inhabitants was 2,207,175 of whom 247,493 (11.2%) were children (age<14 years).
The extent of the relationship between providers and patients can influence the quality of care and its outcomes. Inclusion of individuals with very short enrollment periods can affect the results of some providers and, therefore, the Basque Health Service has established 185 days as the minimum period to consider that a provider (physician, health center, or health district) is accountable for the care provided to a patient in their annual evaluations. Accordingly, we kept such criteria in our study. However, there is not a unanimous opinion between researchers about the decision of include/exclude patients with partial enrollment. So that, we repeated all our analyses limiting the study population to the 1,901,927 people that remained the full 12-month period with the same physician. Such results were very similar to the ones presented in this article and are available on demand for interested readers (please contact the authors for further information).
## Sources of data
We use the following sources of data:
Primary care EHRs of the Basque public health system, containing demographic, administrative, and clinical data including diagnoses, prescriptions, ancillary examinations, and referrals generated in each patient visit. The minimum basic data set, gathering information regarding all the patient discharges from all hospitals across the Basque public health network, including data on patient characteristics, hospitalization episodes, diagnoses, and procedures.
## Variables
The explanatory variables used at the patient level were demographic characteristics (age and sex), morbidity (primary and secondary diagnoses), and socioeconomic status (censusbased deprivation index).
In order to include a manageable number of disease categories, all the patient diagnoses (ICD-9-CM codes used by the primary care doctors during the year of study) were classified into Aggregated Diagnosis Groups (ADGs). The ADG system assigns ICD-9-CM codes to 1 of 32 categories, as a function of clinical criteria, expected resource use and the type of care likely to be required for each health problem. It is part of the Johns Hopkins population-based ACG Case-Mix adjustment system. 4 In our study, we applied a ''lenient'' diagnostic certainty option: any single diagnosis included in an ADG was enough to accept the corresponding ADG for such patient.
The census-based deprivation index, used as a proxy for the socioeconomic status, was developed for the MEDEA project (Mortalidad en áreas pequeñas Españolas y Desigualdades socio-Económicas y Ambientales-Mortality in small Spanish areas and socio-economic and environmental inequalities). [bib_ref] Construcción de un índice de privación a partir de datos censales en..., Domínguez-Berjón [/bib_ref] In Spain, census tracts are the smallest geographical units into which census data are divided; they are mainly defined by criteria related to population size but also reflect natural and man-made features. Although the number of people living in a tract is variable, the median population per census tract is around 1200 people.
This index is constructed from variables related to employment (rates of unemployment, and of manual and short-term employment) and education (low rates of educational attainment among young people and overall). In this study, deprivation index scores were categorized into quintiles (the 5th corresponding to areas with the greatest deprivation, and the 1st to the least deprived areas). It is a measure of the ecological effect on individuals of dwelling in a census tract and provides an indicator of the level of access to economic and material resources in a community. It has been shown to be correlated with mortality rates 10 and prevalence of morbidity. [bib_ref] Prevalence of multimorbidity according to the deprivation level among the elderly in..., Orueta [/bib_ref] [bib_ref] Socioeconomic variation in the burden of chronic conditions and health care provision-analyzing..., Orueta [/bib_ref] At the doctor level, we considered the number of patients on a doctor's list to estimate his/her workload.
To analyze the characteristics of the primary care centers we included variables relative to demographic characteristics of the area (percentage of the population above 65 years of age and percentage of immigrants in the corresponding geographical areas),size of the center (number of primary care doctors), and satisfaction of the staff with their working environment. For the last of these variables, we used the overall satisfaction scores for health centers (on a scale of 0-10), taken from internal surveys that Osakidetza regularly carries out in all its organizations. 14 As for the doctors, the health centers were categorized into quintiles. Categorization of these variables was necessary because some of the statistical models did not achieved convergence when we used them as continuous.
The outcome variables were the number of visits to primary care doctors, number of forms for referrals to specialized care issued to users by these doctors, and cost of medications prescribed by primary care physicians and paid for by the Department of Health during the year of study. In addition, we identified patients with hospitalizations for ambulatory care sensitive conditions (ACSCs) that might have been avoided had they received appropriate and timely outpatient care. For ACSCs, we used the list developed by Casanova Matutano et al [bib_ref] Hospitalización pediátrica evitable en la Comunidad Valenciana y Cataluña, Casanova Matutano [/bib_ref] for the pediatric population and that of Caminal et al [bib_ref] The role of primary care in preventing ambulatory care sensitive conditions, Caminal [/bib_ref] for the rest of the population.
# Statistical analysis
The study population was divided into 2 groups that were analyzed separately: individuals on pediatricians' lists and those on family doctors' lists. The methodology for the analysis was identical for the 2 groups. All analyses were carried out using SAS (SAS/STAT, version 9.2; SAS Institute, Inc., Cary, NC, USA).
Four-level mixed effect models 5 were used to assess the relationship between the various different outcome variables and the characteristics of patients, doctors, and health centers. Taking into account the hierarchical nature of the data, these models include each of the explanatory variables as fixed effects and random intercepts for each of the levels (the patient, doctor, health center, and health district).
For prescription costs, we built generalized linear mixed models (GLMMs) with gamma distribution and log link. [bib_ref] Issues for the next generation of health care cost analyses, Basu [/bib_ref] As healthcare cost data are typically nonnormally distributed with a skew toward the right, Gamma regression is a better modeling approach to deal with this skewness than ordinary least square. [bib_ref] Estimating marginal and incremental effects on health outcomes using flexible link and..., Basu [/bib_ref] [bib_ref] Family structure and childhood obesity, early childhood longitudinal study-Kindergarten Cohort, Chen [/bib_ref] In the case of visits to the doctor and referrals, although Poisson regression is often employed for this type of outcomes, we used GLMMs with negative binomial distribution as our data showed signs of over-dispersion.Using likelihood ratio tests we checked that the negative binomial regression models provided a better fit than did the Poisson models.Finally, the existence of an ACSC hospitalization is a binary variable (yes/no), and therefore was modeled using logistics GLMMs.We rescaled the fixed and random effects using the method developed by Hox, 23 because in multilevel logistic models the variance of the residual variance (at the individual level) is fixed at a constant.
All the models were developed using PROC GLIMMIX with the Laplace method. Effect sizes of associations were exponentiated and reported as ratios. We consider an association as significant when the P value was below 0.05.
For each outcome variable we calculated the intraclass correlation coefficients (ICCs) and percentages of variance at each level, both for empty models (no explanatory variables, random effects analysis of variance [ANOVA]) and for adjusted models (full set of explanatory variables).
In the case of prescriptions cost, visits to the doctor and referrals we used the method of normal approximation treating the outcomes as if they were normally distributed variables and analyzing them employing linear mixed models (PROC MIXED). [bib_ref] Partitioning variation in multilevel models, Goldstein [/bib_ref] Using the variance components obtained with these models, the calculation of the percentages of variance and ICCs at each level is straightforward. In a multilevel model, the ICCs are estimated as the ratio between the variance at the level (which include the variance at its higher levels) and the total variance. The ICCs can be interpreted as the correlation among observations within the same cluster (eg, correlation of 2 patients in the same health center). For ACSC we used logistic mixed models, which provide the variance components for the nonindividual levels (doctor, health center, and health district), and the formula of Snijders,based on a latent continuous variable that considers that the variance attributable to individuals is approximately 3.29.
# Results
Patient, physician, and health center level characteristics are presented for adult and pediatric patients in [fig_ref] TABLE 1: Patient, Physician, and Health Center Level Characteristics [/fig_ref]. [fig_ref] TABLE 2: Unadjusted OutcomesPrescription costs in Euros, visits to the physician and referrals [/fig_ref] shows the unadjusted outcomes: visits to the physician, referrals, cost of prescriptions, and ACSC.
By applying multilevel analysis, the demographic characteristics of patients had an influence on resource use and hospitalization for ACSCs, although not homogeneously. In the adult population [fig_ref] TABLE 3: Multilevel Analysis [/fig_ref] , the number of visits to the doctor and prescription costs increased with age, until 84 years of age, at which point they decreased slightly; while in the case of referrals to a specialist, such reduction was observed at 65 years of age and beyond; we found a bimodal distribution in ACSC hospitalizations with a peak among younger patients and another in the most elderly ones. Among the children , the rates of visits to primary care pediatricians and of hospitalization for ACSCs decreased with age, while the opposite trend was observed for referral rates and prescription costs. With respect to sex, both in adults and children, prescription costs per patient were higher among male patients, and they visit their primary care doctor less often than females. Nevertheless, adult men were less often referred to specialists and were more likely to be hospitalized for ACSCs than women; in children, there were no significant differences in referrals and the opposite pattern was found in hospitalizations.
Although the differences were small, the least disadvantaged socioeconomic groups presented less frequent visits to the primary care doctor and referrals; for prescriptions, there were not differences in adults and the rates of hospitalizations for ACSCs were smaller among individuals living in richer areas; in pediatric patients, although being statistically significant their differences, there was no clear gradient for prescriptions and ACSC hospitalizations.
In adults, diagnoses in any of the 32 diagnostic groups (ADGs) were associated with more visits to the primary care doctor and referrals (see Supplementary Materials, http:// links.lww.com/MD/A371). This pattern was also observed for children, except 1 group in visits and 2 in referrals for which the differences were not statistically significant. Regarding prescriptions, we found a higher use of drugs in 30 of the ADGs in both populations. The risk of potentially avoidable hospitalization in adults was higher for 20 ADGs and lower for 5 ADGs, while in children, the risk was higher for 14 ADGs and lower for 3.
Concerning variables related to the doctor (panel size) and center (size, staff satisfaction, percentages of immigrants, and elderly people in the population), we did not find any statistically significant trends [fig_ref] TABLE 3: Multilevel Analysis [/fig_ref] , with 2 exceptions in adults: patients in small panels receive less prescriptions and the percentage of aged population that achieve statistical signification for referrals but does not show a defined gradient.
In the empty models, the variance observed was almost exclusively due to differences between patients (children or adults). Specifically, the percentage of variance associated with health professionals, and hence, health centers and districts, was <3% for prescriptions, 5% for referrals, and 9% for visits to the primary care doctor, being around 7.4% and 3.8% for potentially avoidable hospitalizations in children and adults, respectively [fig_ref] TABLE 6: Proportion of Variability Attributed to Patients, Physicians, Primary Healthcare Centers, and Districts [/fig_ref]. After adjusting for all the explanatory variables, there was a significant decrease in total variance: in adults, by more than 50% for visits to the primary care doctor and by around 30% for the other dependent variables; and in children, by 58% for visits, 23% for potentially avoidable hospitalizations, 16% for referrals, and 14% for prescriptions. However, we did not find relevant changes in the percentage of the unexplained variances attributable to levels higher the patient, and, in fact, their sum only reaches figures of over 5% for visits to the primary care doctor (12.2% in adults; 6.2% in children) and for potentially avoidable hospitalizations (children 10.3%; adults: 6.1%).
# Discussion
Our results demonstrate that only a small proportion of variability in costs and in healthcare outcomes can be attributed to doctors or health centers. Even using a sophisticated and widely validated population-adjustment system, a significant percentage of the overall variance remains to be explained; it is attributable to patients but cannot be explained by their morbidity or demographic characteristics.
Other authors have also observed that patient characteristics determine most of the variability observed both between doctors and between levels of care. Hofer et al 7 estimated that the effect attributable to doctors' way of working accounted for only 4% of the variance in visits of patients with diabetes Comparing hospital indicators, other authors found that the percentage of the variance attributable to health centers did not even reach 1% in the rates of readmission within 30 days [bib_ref] Variation in the risk of readmission among hospitals: the relative contribution of..., Singh [/bib_ref] and was around 10% in terms of use of intensive care resources. [bib_ref] Hospital-level variation in the use of intensive care, Seymour [/bib_ref] In a systematic review, Fung et al 8 analyzed 21 studies that provided estimates of the percentage of variance explained by different levels of care, using hierarchical models; although results were mixed (in part attributable to methodological differences), the total variability attributable to the levels higher than the patient (doctors' list, center, hospital, or healthcare plan) was low, in almost all cases, when analyzing process or outcome measures, and higher considering patient satisfaction. More recently, it has been reported that 85% of the variance in outcomes in terms of overall technical quality in primary care centers [bib_ref] Characteristics of primary care practices associated with high quality of care, Beaulieu [/bib_ref] and 92% of differences in the adequate prescription of beta blockers after a heart attack 27 are attributable to patientlevel factors.
In our analysis, we did not find any associations of family doctor's panel size or work satisfaction, or health center size with outcomes in terms of healthcare resource use or ACSC hospitalizations in pediatric patients; in adults, we only observed more prescriptions in doctors with large panels, as well as differences in referrals to the specialists among health centers according to the percentage of elderly people in the geographic area, but without a defined gradient. Although some researchers have found an association between the satisfaction of primary care doctors and the provision of better quality care [bib_ref] Outcomes of physician job satisfaction: a narrative review, implications, and directions for..., Williams [/bib_ref] ; this has not been confirmed in other studies. [bib_ref] Working conditions in primary care: physician reactions and care quality, Linzer [/bib_ref] [bib_ref] Is the job satisfaction of primary care team members associated with patient..., Szecsenyi [/bib_ref] Further, there is no consensus on what is the most suitable size for a primary care health center to achieve the best healthcare outcomes. [bib_ref] Characteristics of primary care practices associated with high quality of care, Beaulieu [/bib_ref] [bib_ref] Does practice size matter? Review of effects on quality of care in..., Ng [/bib_ref] In healthcare organizations, it is important to record indicators of healthcare providers and analyze their variability, as they can help monitor the quality of the healthcare provided. In particular, it is unquestionable that, in this way, we may be able to detect inefficiencies in medical practice, clinical errors, and healthcare gaps, but this task is not straightforward and often variability can be attributed to the adaptation of the healthcare provided to the individual needs of patients or to the participation of patients in decision making. [bib_ref] Improving productivity in the NHS, Mulley [/bib_ref] Wennberg classified clinical care into 3 categories: effective, patient preference sensitive, and supply sensitive, largely due to the capacity of the local health services.He defined unwarranted variation as care that is not consistent with patient preferences or related to patient's underlying diseases.However, as other authors recognize, discriminating between warranted and unwarranted variation can be challenging and interventions designed to reduce the observed variability risk eliminating variations that exist for good reason. [bib_ref] Medical practice variations: what the literature tells us (or does not) about..., Mercuri [/bib_ref] In relation to this, primary care practice reports providing data on the results individual doctors has been criticized, because, among other issues, they only take into account very specific aspects of healthcare and do not provide a real measure of the complexity of the work of a primary care doctor. [bib_ref] Quality measures and the individual physician, Ofri [/bib_ref] On the other hand, publishing such reports does not, in itself, ensure improvements in quality or motivate doctors to put in place measures to solve problems related to quality; in fact, an excessive emphasis by managers on reducing variability may encourage physician to ''de-select'' complex patients, declining to see the most sick patients, those whose treatments have failed or who do not adhere to treatment plans. [bib_ref] The unreliability of individual physician ''report cards'' for assessing the costs and..., Hofer [/bib_ref] Consequently, we believe that our findings help to put these indicators in context, since we show that only a small proportion of the variability can be attributed to healthcare professionals or the organization of health centers. However, a small proportion does not equate to no effect at all. If we translate the differences observed into costs related to prescriptions or visits to the doctor, we find significant absolute values that justify managerial interventions. Moreover, even recognizing that the greater part of the variability is attributable to patient-level factors, it may be easier and more efficient for healthcare organizations to design actions focused on their staff rather than on patients. Further, it has been suggested that doctors may be able to influence some important factors at the patient level. [bib_ref] Meaningful variation in performance: a systematic literature review, Fung [/bib_ref] The present study has some limitations. First, the data analyzed are taken from the healthcare information systems of the Basque Country, and may contain some incomplete or inaccurate information, as is always the case when using data from administrative databases or notes in EHRs. On the other hand, factors other than those analyzed may contribute to patient variability, such as those related to the family environment, psychosocial factors, or health-related habits, 37 the same being true of the other levels studied. Besides, the socioeconomic variable used (deprivation index) may reduce the individual contribution to socioeconomic characteristics, given its ecological nature. Finally, the variability in other process or outcome indicators may be different to that of those analyzed in this study.
To conclude, our study found that doctors and health centers make only a modest contribution to variability in resource use and healthcare outcomes. This fact should be taken into account when evaluating healthcare professionals, trying to modify their behavior, and attempting to improve the efficiency and quality of organizations. Compared to traditional analysis, complex statistical methods are more costly, require more time, and are more difficult for clinicians and managers to understand. Nevertheless, a greater understanding of the factors involved in variability and a greater reliability in the estimates will help the evaluations of healthcare providers to become accepted. Given all this, strategies for improving clinical practice should include multilevel interventions that go beyond the health professionals and include individuals themselves and the organization of healthcare services. [bib_ref] The PCORI perspective on patientcentered outcomes research, Frank [/bib_ref]
[table] TABLE 2: Unadjusted OutcomesPrescription costs in Euros, visits to the physician and referrals: mean per patient; number and percentage of patients with 1þ admissions due to ACSC. ACSC ¼ ambulatory care sensitive conditions; CI ¼ confidence interval. [/table]
[table] TABLE 1: Patient, Physician, and Health Center Level Characteristics [/table]
[table] TABLE 3: Multilevel Analysis: Fixed Effects Adult Patients (Age 14þ) [/table]
[table] TABLE 4 TABLE 5: Multilevel Analysis: Fixed Effects Pediatric Patients (Age <14) Impact of the characteristics of the pediatric patients (age <14), primary care physicians and healthcare centers on the analyzed variables. ACSC Proportion of Variability Attributed to Patients, Physicians, Primary Healthcare Centers, and Districts: Adult Patients (Age 14þ) ACSC ¼ ambulatory care sensitive conditions; ICC ¼ intraclass correlation coefficient. [/table]
[table] TABLE 6: Proportion of Variability Attributed to Patients, Physicians, Primary Healthcare Centers, and Districts: Pediatric Patients (Age <14) ACSC ¼ ambulatory care sensitive conditions; ICC ¼ intraclass correlation coefficient. [/table]
|
Global hypomyelination of the brain white and gray matter in schizophrenia: quantitative imaging using macromolecular proton fraction
Myelin deficiency is commonly recognized as an important pathological feature of brain tissues in schizophrenia (SZ). In this pilot study, global myelin content abnormalities in white matter (WM) and gray matter (GM) of SZ patients were non-invasively investigated using a novel clinically-targeted quantitative myelin imaging technique, fast macromolecular proton fraction (MPF) mapping. MPF maps were obtained from 23 healthy subjects and 31 SZ patients using a clinical 1.5T magnetic resonance imaging (MRI) scanner. Mean MPF in WM and GM was compared between the healthy control subjects and SZ patients with positive and negative leading symptoms using the multivariate analysis of covariance. The SZ patients had significantly reduced MPF in GM (p < 0.001) and WM (p = 0.02) with the corresponding relative decrease of 5% and 3%, respectively. The effect sizes for the myelin content loss in SZ relative to the control group were 1.0 and 1.5 for WM and GM, respectively. The SZ patients with leading negative symptoms had significantly lower MPF in GM (p < 0.001) and WM (p = 0.003) as compared to the controls and showed a significant MPF decrease in WM (p = 0.03) relative to the patients with leading positive symptoms. MPF in WM significantly negatively correlated with the disease duration in SZ patients (Pearson's r = −0.51; p = 0.004). This study demonstrates that chronic SZ is characterized by global microscopic brain hypomyelination of both WM and GM, which is associated with the disease duration and negative symptoms. Myelin deficiency in SZ can be detected and quantified by the fast MPF mapping method.Translational Psychiatry (2021) 11:365 ; https://doi.
# Introduction
Myelin abnormalities in schizophrenia (SZ) have been attracting substantial interest over the past two decades. A number of post-mortem investigations of the SZ brain confirmed various aspects of myelin disorganization and dysfunction associated with this disease. Electron microscopy studies demonstrated abnormalities in the myelin sheath ultrastructure associated with degeneration of oligodendroglia in both gray matter (GM) and white matter (WM) of SZ patients. Post-mortem genetic transcriptome studies (detailed review can be found in ref.identified downregulation of a series of genes encoding key proteins related to the oligodendrocyte function and myelin biosyntheses, such as myelin basic protein (MBP), myelin oligodendrocyte glycoprotein, and transferrinin various anatomic regions of the brain. Histological studies reported a reduced myelin density in the cortex according to MBP immunostainingand luxol-fast-blue staining, as well as a decreased amount of oligodendrocytes in the hippocampus, cortical GM, and subcortical WM. Impaired production of myelinating oligodendrocytes from oligodendrocyte precursor cells also has been found in SZ. While the etiology of oligodendroglial dysfunction and related dysmyelination in SZ is still poorly understood, it is commonly recognized to date that oligodendrocyte and myelin damage represents a fundamental aspect of the molecular and cellular brain pathology in this disease and may become a target for future therapeutic interventions.
Subtle structural and microstructural brain abnormalities in SZ were detected in numerous non-invasive magnetic resonance imaging (MRI) studies. A number of morphological studies demonstrated regional signs of accelerated GM and WM atrophy as compared to the normal population and aberrant cortical gyrification. Certain microstructural features of brain tissues in SZ reported in the literature appeared consistent with abnormal myelination. The vast majority of such findings have been based on diffusion tensor imaging (DTI), which produces a series of indices reflecting the integrity of WM fiber tracts. The common observations in SZ typically linked to hypomyelination include reduced fractional anisotropyand increased radial diffusivity. However, these DTI parameters strongly depend on the diameter, density, and spatial coherence of axons, and, therefore, cannot be used as myelin biomarkers in the presence of a diverse axonal population. Another widely used SZ research technique is magnetization transfer (MT) imaging, which produces a semi-quantitative index, MT ratio (MTR) that has been commonly interpreted as a marker of brain tissue myelination. A number of studies (reviewed infound a reduced MTR in SZ, though some reports indicated no effect or an increase in this parameter However, MTR was shown to have low specificity and sensitivity to myelin because of its dependence on both cross-relaxation and longitudinal relaxation in tissues. More specific myelin imaging measures, such as myelin water fraction (MWF) and T1/ T2 signal ratio were applied in a few studies of SZ. MWF demonstrated regional reductions in WM of SZ patients relative to the normal control population 35,36 but did not show changes in the first-episode SZ 37 . T1/T2 signal ratio showed a global decrease in WM and GM in SZ in one study 38 and opposite regional trends of either decrease or increase in different cortical areas of the firstepisode SZ patients in another study. It also should be pointed out that both MWF and T1/T2 signal ratio relies on magnetic relaxation as the source of quantitative information, and, therefore, are influenced by any factors affecting T1 and T2 relaxation times, such as concentrations of paramagnetic ions including iron. In brain diseases, alterations of the iron content are a common cause of T1 and T2 changes 42 , and the corresponding confounding effect cannot be ignored in the applications for the above techniques. In summary, a general limitation of the previous neuroimaging studies of myelin abnormalities in SZ is the use of imaging biomarkers, which do not provide sufficient sensitivity and specificity to enable unambiguous interpretation of the results in terms of the myelin content changes. A more accurate and precise myelin imaging technique is required for the prospective use as both a measure of the disease progression and a potential surrogate marker of the effect of therapies targeted at the treatment of oligodendroglial dysfunction.
Several more specific myelin imaging techniques based on the MT effect, such as relaxation-independent MT saturation 43 , inhomogeneous MT 44 , and macromolecular proton fraction (MPF)mapping, were recently developed but have not been applied in SZ research. One of these methods, fast MPF mappingyields the maps of the fraction of macromolecular protons involved in cross-relaxation with free water protons. During past years, this technique showed promise as a clinically targeted tool for the assessment of myelin damage and development in the human brain. MPF has been validated as a myelin biomarker in a number of animal model studies and demonstrated a close linear relationship with the myelin content in both WM and GM. A recent meta-analysis 55 indicated high specificity of MPF to myelin and its metrological superiority as compared to other myelin imaging biomarkers. In the pilot clinical studies, MPF mapping enabled the quantitative assessment of microscopic demyelination in normal-appearing WM and GM caused by multiple sclerosisand mild traumatic brain injury. The method also showed the capability to quantify subtle changes in the myelin content at the earliest stages of prenatal brain developmentand in the course of brain maturation during adolescence. In the recent synthetic-reference design, MPF mapping utilizes only three source images, thus enabling clinically reasonable acquisition time. In the quantitative aspects, MPF mapping provides high reproducibility, independence of magnetic field strength, and insensitivity to iron deposition 47 , axonal organization, and changes in the content of nonmyelin cellular components of neural tissues, such as glia and neurons. In view of the above technical advantages, it would be reasonable to investigate the capability of MPF mapping to assess myelination abnormalities in SZ.
The objectives of this pilot study were to test the feasibility of the detection and quantitation of myelin deficiency in SZ at the level of global brain WM and GM measurements using the fast clinically-targeted MPF mapping method and explore potential clinical correlations of the myelin content in the SZ population.
## Materials and methods subjects
Thirty-one patients with schizophrenia were recruited to study. The inclusion criteria were the following: the diagnosis of schizophrenia (F20.0, F20.6), in accordance with the International Statistical Classification of Diseases and Related Health Problems, 10th Revision (ICD-10)and the Structured Clinical Interview for DSM-IV Axis I Disorders (SCID)), age from 18 to 60 years.
The exclusion criteria were the following: the presence of acute and chronic infectious, inflammatory, autoimmune, or neurological diseases, other organic mental disorders, and mental retardation. The sample was formed from outpatients visiting the clinics of the TNIMTs from August till October 2019 with a continuous observation method. At the time of inclusion in the study, patients were receiving anti-relapse therapy. Most of the patients (81%) have taken atypical antipsychotics in the last 6 months. MRI scan of the brain was performed during the period of therapeutic remission. In order to analyze the MPF maps of the brain, taking into account the clinical characteristics of schizophrenia, the study included patients with different leading symptoms and duration of the disease. The patients were stratified into 2 groups: with positive and negative symptoms, using a clinical interview and the Positive and Negative Syndrome Scale (PANSS) for schizophrenia.
The control group consisted of 23 healthy subjects. These control subjects were mentally and somatically healthy individuals and were recruited from volunteers from Tomsk State University. They did not have significant differences in sex and age with the group of patients with schizophrenia (p = 0.98). The excluding criteria for the controls were the presence of acute and chronic infectious, inflammatory, autoimmune, or neurological diseases and organic, including symptomatic, mental disorders. A subgroup of 8 control participants underwent two repeated MRI examinations with the scan-rescan interval from 1 to 7 days to assess the reproducibility of MPF measurements. All individuals included in the study gave written informed consent. Ethical approval was granted (protocol N 78/1.2019) by the Local Bioethics Committee of the Mental Health Research Institute in accordance with Helsinki ethics committee guidelines. None of the participants were compromised in their capacity/ability to consent; thus, consent from the next-of-kin was not necessary, and it was not recommended by the local ethics committee. The MRI scanning procedure was well tolerated by all participants.
## Mri data acquisition
MRI data were acquired on a 1.5T clinical scanner (Magnetom Essenza; Siemens, Erlangen, Germany). The fast MPF mapping protocol was implemented using a standard manufacturer's 3D spoiled gradient-echo sequence according to the single-point synthetic reference method. The method is based on the reconstruction of MPF maps from three source images with MT, T1, and proton density (PD) contrast weightings by voxelwise iterative solution of the pulsed MT equationwith a calculated synthetic reference image for data normalization. The protocol implementation was similar in concept to those used in the earlier fast MPF mapping studies with 1.5T MRI equipmentand utilized the shortest available repetition time (TR) in the MT-weighted sequence. Acquisition parameters for the source images were as follows: Image processing and analysis MPF maps were reconstructed using custom C++ language software (available at https://www.macromolecularmri.org/) based on the singlepoint synthetic reference algorithm. The two-pool model parameter constraints in the single-point algorithm were set according to the previous implementation for 1.5T data 49 as follows: cross-relaxation rate constant R = 19 s −1 ; T2 of bound macromolecular protons, T 2 B = 10 μs; and the product of the relaxation rate R1 and T2 of free water protons, R 1 T 2 F = 0.055. The last constraint was used to account for direct saturation of water protons based on the Lorentzian lineshape approximation as detailed earlier. No correction of B0 and B1 field inhomogeneity was performed due to the absence of corresponding field mapping sequences on the used MRI scanner. According to the literature 56 , the impact of B0 and B1 field non-uniformity on the accuracy of MPF measurement in segmented brain tissues at 1.5T is expected to be very small with residual relative errors <1%.
Further processing of the 3D MPF maps was carried out according to the earlier described pipelineusing FSL software (FMRIB Software Library v. 6.0; Oxford Centre for Functional Magnetic Resonance Imaging of the Brain, University of Oxford, Oxford, England; available at www.fmrib.ox.ac. uk/fsl/). In brief, MPF maps were skull stripped using the brain-extraction tool BET 61 and then segmented using the automated segmentation tool FAST 62 into the four tissue classes. The segmented brain tissues included pure WM, pure GM, mixed voxels containing partial volumes from white and gray matter (PVWGM), and mixed voxels containing a partial volume from cerebrospinal fluid (PVCSF). This segmentation technique was specifically designed for the analysis of MPF mapsto alleviate the partial volume effects by separating pure (WM and GM) and mixed (PVWGM and PVCSF) tissue classes and thus enable accurate MPF measurements in WM and GM. The segmentation procedure was initialized with the tissue priors of 13%, 6%, 9%, and 1% for WM, GM, PVWGM, and PVCSF, respectively. The PVWGM and PVSCF tissue classes were used to separate MPF measurements in WM and GM from a potential confounding effect of partial volume averaging. The PVCSF tissue was also used to exclude the residual contribution from cerebrospinal fluid after applying the MPF threshold (MPF > 1%) and to correct for brain masking imperfectness since this tissue class typically absorbs residual non-brain voxels. The PVCSF data were not used in subsequent analyses similar to the earlier studies. The resulting binary segmentation masks were applied to calculate mean MPF in WM, GM, and PVWGM tissues.
# Statistical analysis
The normality of continuous data distributions was assessed by the Shapiro-Wilk test within each subject group. Parametric analyses were subsequently used since no significant departures from the normal distribution were identified. The gender distribution and mean age were compared between SZ patients and controls using the Pearson χ 2 test and independent-samples Student t-test, respectively. Clinical variables were compared between the patients with positive and negative symptoms using the independentsamples t-test. To compare mean MPF values in segmented brain tissues between the participant groups, the multivariate analysis of covariance (MANCOVA) was used with age as a covariate (to adjust for age) and either two (patients and controls) or three (patients with negative and positive symptoms and controls) levels of the group factor. Overall MANCOVA significance was assessed using the Wilks Λ-test. Tukey honest significant difference post-hoc pairwise tests were used to adjust for multiple comparisons. Distinctions in MPF between patients and controls were also characterized by the percentage differences and effect sizes (Cohen's d) calculated as the ratio of the mean difference to the pooled standard deviation. Pearson's correlation coefficient (r) was used to test relationships between MPF and clinical variables. For significant correlations, partial correlation coefficients with adjustment for age were also calculated. Scanrescan reproducibility of MPF measurements in a subgroup of healthy controls was assessed using Bland-Altman plots, paired Student t-test, and withinsubject coefficient of variation (CoV), which were calculated for each segmented brain tissue class. Two-tailed tests were used in all analyses with a significance level of P < 0.05. All statistical analyses were carried out in Statistica 12 for Windows software (StatSoft, Tulsa, OK, USA).
# Results
An example MPF map and the results of brain segmentation for an SZ patient are presented in . Mean MPF values in segmented brain tissues and the results of group comparisons are summarized in. In comparison between the SZ patients and controls, MANCOVA revealed the significant global effect of the disease (Λ = 0.67, F(3,49) = 7.9, p < 0.001) and insignificant influence of the age covariate (Λ = 0.89, F(3,49) = 2.1, p = 0.11). The SZ patients had significantly reduced MPF in all brain tissues. When the SZ patients with positive and negative symptoms were compared to controls, the group factor remained significant (Λ = 0.60, F(6,96) = 4.7, p < 0.001), while the age effect was not (Λ = 0.93, F(3,48) = 1.2, p = 0.31). Pairwise tests revealed no significant differences between the positive SZ patients and controls, though a trend of an MPF decrease was observed for GM. SZ patients with negative symptoms had significantly lower MPF in all tissues as compared to controls. SZ patients with negative symptoms also showed a trend of an MPF reduction relative to the patients with positive symptoms in all tissues, while the significant difference was observed for WM only.
Percentage changes in MRI parameters between the subject groups and the corresponding effect sizes (Cohen's d) are presented in. MPF in GM demonstrated the largest decrease in SZ patients relative to controls as compared to other tissue classes. The percentage changes and effect sizes for MPF in the mixed PVWGM tissue showed intermediate values between those in pure WM and GM. In all tissues, a decrease in MPF was substantially larger for patients with leading negative symptoms as compared to the patients with positive symptoms.
Correlations between MPF in brain tissues and clinical variables for SZ patients are presented in. In the control group, MPF significantly negatively correlated with age in GM (r = −0.45, P = 0.03) and showed a trend of an age-related decrease in PVWGM (r = −0.40, P = 0.06) and WM (r = −0.34, P = 0.11). In SZ patients, a marginally significant negative correlation between MPF in WM and age was found. Among the clinical variables, MPF significantly negatively correlated with the disease duration only, particularly in WM and PVWGM with a stronger relationship (r = −0.51, P = 0.004) identified for WM. These correlations retained significance after adjustment for age (partial r = −0.39, P = 0.04 for PVWGM and partial r = −0.38, P = 0.04 for WM). Example three-dimensional MPF macromolecular proton fraction map and binary tissue segmentation masks obtained from a 33year old male patient with schizophrenia. Segmentation masks correspond to the following tissue classes: white matter (WM), gray matter (GM), and mixed WM and GM (PVWGM). Results of the statistical analysis of scan-rescan repeatability of MPF measurements in brain tissues for a subgroup of control participants are summarized in, and the corresponding Bland-Altman plot is presented in . Repeated MPF measurements demonstrated high reproducibility with remarkably small CoV (1% for GM, 0.9% for PVWGM, and 0.8% for WM) and the absence of significant biases.
# Discussion
This study provides the first report of quantitative myelin assessment in SZ using the fast MPF mapping method. Our results indicate the global myelin loss in both WM and GM of SZ patients, which is associated with the disease duration and dominant-negative symptoms. Our findings are in agreement with the evidence of myelin damage in SZand corroborate the results of earlier studies obtained using different myelin-sensitive MRI methods. Our results also add confidence to the understanding of DTI and MTR abnormalities in SZas at least partially related to myelin, though other pathological factors, such as axonal remodeling and degeneration, inflammation, and iron content changes need to be taken into account in the interpretation of non-specific to myelin imaging measures.
The results of this study suggest that myelin loss in GM and WM of SZ patients may have different pathological mechanisms. Both the percentage MPF changes and effect sizes indicate that the relative reduction in the myelin content is substantially larger for GM as compared to WM. The observed effects are corroborated by the high reproducibility of MPF measurements according to the scan-rescan data presented in the Supplementary material. The distinctions between WM and GM appear even more dramatic in terms of the absolute myelin content changes. The amount of myelin in brain tissues can be estimated from the histologically validated linear relationship between MPF and myelin density according to the formula MPF = 0.21 × MD + 3.9, where MD is the myelin density expressed in percentage units based on histological staining. While the exact coefficients in the above equation may vary depending on the pathological material, MPF mapping technique, and histology quantitation approach, this transformation provides an advantage of eliminating the non-myelin contribution (or a substantial portion of it) from MPF, thus resulting in a more realistic representation of the relative myelin content in WM and GM. A similar approach was recently used to approximate myelin content changes in WM and GM during adolescent brain development. Application of this estimation technique to the data fromresults in the myelin densities in GM and WM of SZ patients of 10.6% and 41.6% and the corresponding values for controls of 12.1% and 43.5%, respectively. Accordingly, the relative myelin loss in SZ in GM is approximately 13% in contrast to only an about 5% reduction in WM. Despite this large difference, WM myelination showed a much stronger negative correlation with the disease duration as compared to GM. We believe that such distinctions may be indicative of different temporal trajectories of GM and WM dysmyelination in SZ. Particularly, the myelin deficiency in GM may represent either a pre-existing trait associated with the developmental SZ etiology or a result of rapid progression in the early disease stage. On the other hand, myelin loss in WM could be described by a slow progressive pattern associated with such pathological processes as neurodegeneration, inflammatory demyelination, and excitotoxic damage to oligodendrocytes 1-7 . These findings suggest an intriguing hypothesis of SZ as a two-stage disease in terms of myelin damage, where GM dysmyelination reflects the disrupted intracortical connectivity in the early disease manifestation and follows by the gradual breakdown of long-range connectivity associated with WM demyelination in chronic disease.
It is known that negative symptoms of SZ have an adverse impact on psychosocial functioning and disease outcome and show less response to antipsychotic therapy. Our results indicate that patients with leading negative symptoms have a more pronounced decrease in myelin density as compared to the patients with positive symptoms. The results of this study are consistent with the literature data on a decrease in the volume of white and gray matter in negative SZ. The meta-analysis by the ENIGMA-Schizophrenia consortium indicates an association between cortical thinning in prefrontal regions and the severity of negative symptoms in SZ. In smaller samples of patients, an inverse correlation was found between the volume of white matter and negative symptoms of SZ. The voxel-wise analysis demonstrated significant fractional anisotropy reductions in the WM tracts in chronic negative SZ which indicates more pronounced microstructural changes. Proteomic analysis of the area-specific for cognitive impairments in SZ revealed dysfunction of pathways associated with the regulation of synaptic processes in the caudate nucleus, cerebellum, and cingulate gyrus. Associations of leading negative symptoms and impairment in the prefrontal cortex, hippocampus 67 , amygdala, as well as a metabolic disturbance in the prefrontal-thalamic-parietal networkwere found. As such, more extensive myelin damage in SZ with negative symptoms appears a biologically plausible finding, which is in line with a general view of the greater severity of this disease phenotype and its associations with anatomical and functional brain abnormalities.
Correlation analysis revealed significant negative correlations between the disease duration and MPF in WM and PVWGM, which remained significant after adjustment for age. As such, a decrease in myelin density in chronic SZ patients cannot be explained by changes in myelination associated with normal aging. At the same time, negative correlations between MPF and age in healthy controls were found, being in agreement with the age-related trends for the adult population. Such an age effect seems to be suppressed by disease-related factors in SZ, as indicated by weaker correlations between MPF and age found in patients. The loss of myelin-associated with the disease duration is in line with the disease-related brain atrophyand a decrease of WM integrityreported in earlier SZ studies.
This study has several limitations. First, the group of patients with leading negative symptoms had a longer duration of the disease as compared to the group with positive symptoms. However, the fact that negative symptoms predominate in long-term illness is well known. As such, it is difficult or even impossible to avoid this confounder within the study design with consecutive patient recruitment. Second, a relatively small number of patients with positive symptoms was recruited for this study. Such patients frequently present with acute psychosis, which makes it difficult to achieve a sufficient level of cooperation for participation in an imaging research study. Third, the cross-sectional design of this pilot study limits the mechanistic interpretation of inferred temporal changes, particularly in relationships to such factors as age and disease duration. While our results indicate substantial disconnection between WM and GM myelination in SZ, this construal should be taken with caution and tested in future longitudinal studies. Fourth, a possible effect of antipsychotic therapy on the myelin content cannot be ruled out, although we did not find significant correlations with the antipsychotic dose. The question about the effect of antipsychotics on brain morphology and microstructure remains controversial. Future studies with the inclusion of drug-naïve patients are warranted to address this issue. Finally, our interpretation of an MPF decreases solely in terms of myelin loss may be oversimplified, as MPF could be also affected to some extent by the tissue water content. An increased water content theoretically may be related to inflammation, which is considered one of the possible mechanisms of microstructural tissue damage in the SZ brain. However, an earlier study 53 did not find correlations between MPF and postischemic microglial activation in an animal stroke model, suggesting that myelin remains the main determinant of MPF in brain tissues even in the presence of severe inflammation and L.P. Smirnova et al. edema. Furthermore, oligodendroglial injury, demyelination, and neuroinflammation are commonly viewed as interrelated phenomena in SZ and other brain diseases. A recent modification of the MPF mapping technique enabling correction of MPF values for water content changes 76,77 may be helpful to investigate this aspect of SZ pathology in more detail in the future.
In conclusion, this study demonstrated that chronic SZ is characterized by global microscopic brain hypomyelination of both WM and GM, which can be detected and quantified by the fast MPF mapping method. A decrease in the myelin content in the SZ brain is associated with the disease duration and negative symptoms, thus suggesting the clinical relevance of myelin damage in this disease. Our results underscore the importance of myelin pathology for understanding the biological mechanisms of SZ development and substantiate the interpretation of earlier microstructural neuroimaging findings in terms of myelin deficiency. In the methodological aspect, this study provides technical background for prospective applications of the fast MPF mapping method in clinical psychiatric research. |
Patients’ and Caregivers’ Experiences with the Multi-Payer Advanced Primary Care Practice Demonstration
# Introduction
In recent years, Centers for Medicare & Medicaid Services (CMS) and other payers have invested significant resources in testing the patient-centered medical home (PCMH) model as a means to improve the organization and delivery of primary health care and reduce health care expenditures. Goals of the PCMH model are to improve patient access to care, improve coordination and quality of care, and increase patient participation in health-related decisionmaking and self-management.Understanding how the PCMH model affects patient experience with care is essential to assessing the model's overall impact and identifying areas in need of improvement. Quantitative studies to date have shown mixed but promising results. [bib_ref] Differences in patient ratings of medical home domains among adults with diabetes:..., Hall [/bib_ref] [bib_ref] Effects of patient-centered medical home transformation on child patient experience, Harder [/bib_ref] [bib_ref] Patient-centered medical home transformation with payment reform: Patient experience outcomes, Heyworth [/bib_ref] [bib_ref] Patient experience over time in patient-centered medical homes, Kern [/bib_ref] [bib_ref] Patient experience with the patient-centered medical home in Michigan's statewide multi-payer demonstration:..., Sarinopoulos [/bib_ref] Qualitative research allows for more in-depth exploration and deeper understanding of patients' experiences of care, but little work has been done in this area.
The Multi-Payer Advanced Primary Care Practice (MAPCP) Demonstration provides an opportunity to explore the effects of PCMH transformation across a broad variety of populations, health care systems, and geographic areas. The MAPCP Demonstration began in 2011 when the CMS joined ongoing PCMH initiatives in eight states (Maine, Michigan, Minnesota, New York, North Carolina, Pennsylvania, Rhode Island, and Vermont). Through CMS's involvement, Medicare partnered with Medicaid and commercial payers to make payments to participating primary care practices to support PCMH transformation activities, including extending office hours, staffing care teams, coordinating care, and enhancing electronic health record (EHR) capabilities. By the end of 2014, 857 advanced primary care practices were participating in the demonstration.
This paper describes the perspectives of patients and patient caregivers from practices participating in the MAPCP Demonstration regarding their experiences with aspects of their care that are expectations for a PCMH (i.e., enhanced access to care, high-quality and coordinated care, support for shared decision-making and patient self-management, and solicitation of patient input) and how their care had changed since their practices' PCMH transformation.
# Methods
## Sampling and recruitment
The study design was to hold 12 focus groups in each of the eight MAPCP Demonstration states-one group for each of six categories in two separate geographical regions, including both rural and urban areas. The six categories were Medicare high-risk (defined as having a Hierarchical Condition Category score ≥ 1.22), Medicare low-risk, Medicaid, Medicare/ Medicaid dually eligible beneficiaries, caregivers of Medicaid or Medicare beneficiaries, and caregivers of Medicaid children (in Vermont, groups were conducted with participants of the Support and Services at Home [SASH] program, which provided support services and care coordination to Medicare beneficiaries living in subsidized housing and the surrounding communities, instead of caregivers of Medicaid children). We aimed to recruit 10 people per group, with the goal of having 6-8 participants.
We recruited participants by mailing letters to Medicare, Medicaid, and dually eligible beneficiaries inviting them or their caregivers to participate. To identify Medicare and dually eligible beneficiaries, we selected six MAPCP Demonstration practices in each of two regions of each state and then used the Medicare enrollment database to select a random sample of beneficiaries attributed to those practices who met the following criteria: age 18 or older, had been assigned to a MAPCP Demonstration practice for more than 1 year, and had visited the practice at least twice in the prior 12 months. To identify Medicaid beneficiaries and the caregivers of children with Medicaid, four practices in each state generated a random sample of Medicaid beneficiaries who had received care at those practices over the prior 12 months and mailed the recruitment letter to the sample on the study's behalf.
The recruitment letter asked beneficiaries to call to be screened for eligibility. To be eligible to participate, beneficiaries had to be proficient in English and not have participated in a focus group in the prior 12 months; they also had to confirm that they had either Medicare or Medicaid insurance and that they received their primary care from a practice participating in the MAPCP Demonstration. To ensure that beneficiaries and caregivers had sufficient experience with the practice to be able to speak knowledgably about it and that they would be able to address questions about coordination with specialists, they also had to have seen a specialist at least once in the prior year and have a chronic condition.
Because contact information for beneficiaries' caregivers was not available, the recruitment letters to beneficiaries also invited caregivers to call and be screened for eligibility. A caregiver was defined as the main person responsible for the beneficiary's health care who usually or always took the beneficiary to appointments at his or her primary care practice. To capture a broader variety of experiences, we did not recruit both a beneficiary and their caregiver.
If we did not receive a sufficient number of incoming calls to recruit 10 participants for any of the Medicare and dually eligible groups, we called beneficiaries from among those who received recruitment letters to identify additional participants. We were unable to supplement the Medicaid groups in this way because we did not have contact information for the Medicaid beneficiaries and caregivers.
## Data collection and analysis
An experienced focus group moderator facilitated the groups between July and November 2014. All participants read and signed an informed consent form. The moderator followed a discussion guide that explored participants' awareness of the PCMH concept and their experiences with care related to key PCMH constructs. For example, related to access to care, participants were asked about the ease or difficulty of getting an appointment when they needed one; how they could schedule appointments and their satisfaction with those methods; wait times; if they had used a patient portal and if so, their experience with it; and whether the practice had taken any steps to help them reduce use of the emergency room. Participants were also asked whether they had experienced any changes in any of these areas in the previous few years. Groups lasted 1.5 h on average, and participants were given a $50 Visa gift card for their participation. Each group was audio-recorded and transcribed.
To guide our analysis of the data, we developed a coding scheme based on a priori theoretical constructs as well as on themes that emerged from a review of the focus group transcripts. A team of six coders used NVivo qualitative data analysis software to code the transcripts. To refine the coding scheme and ensure that all of the coders were interpreting the codes in the same way, all six coders initially coded the same two transcripts and discussed any discrepancies. Once the team had refined the codebook and reached consensus on how to apply it, they divided the remaining transcripts, coded them, and prepared coding reports. For each state, one team member reviewed the coded reports, analyzed the data to identify patterns and themes, and prepared a report summarizing findings for the state. Though some distinctions were found across the eight states, our analysis focuses on the comprehensive findings from all the states, with differences noted by focus group type when applicable. When summarizing the frequency with which specific views were expressed, we use "few" for fewer than 10% of participants; "some" for 10 to 25%; "many" for 25 to 50%; and "most" for more than 50%.
# Results
We conducted 81 focus groups with 490 individuals, of whom 80 were caregivers and 410 were beneficiaries [fig_ref] Table 1: Number of Focus Groups and Participants by CategoryIn Vermont, groups were conducted... [/fig_ref]. This is 15 fewer groups than intended: The shortfall is primarily due to the fact that most of the participating practices had very few pediatric patients, so we were able to recruit enough participants for just 1 of the planned 14 focus groups with caregivers of Medicaid children (and even that group had only 3 participants). Twelve people participated in the two focus groups of participants in the SASH program in Vermont. Most participants described themselves (or, in the case of caregivers, the person they cared for) as being in very good (25%), good (35%), or fair (27%) health, with few on either extreme [fig_ref] Table 2: Characteristics of Focus Group Participants [/fig_ref]. Almost two-thirds (64%) were 60 years of age or older, and only 9% were under age 40. Sixty percent of participants were female, and a large majority (83%) were non-Hispanic White. Nearly one-third (31%) of participants had a high school education or less, half (52%) had some college or a college degree, and 17% had more than a college degree. Compared with Medicare beneficiaries, Medicaid and dually eligible beneficiaries were, on average, in worse health, younger, more likely to be female, and had lower levels of education; Medicaid beneficiaries also were more likely to be non-Hispanic Black.
Results from the focus group discussions are presented in the sections that follow and are summarized in [fig_ref] Table 3: Ways in Which Patients' Experiences Were or Were Not Consistent with PCMH... [/fig_ref].
## Access to care
Most participants reported convenient access to care during office hours, and some thought that it had improved in recent years: wait times were short and they could typically get a same-day appointment when needed (although usually not with their primary care providers [PCPs]). After-hours access was more problematic, however: most participants said that their practices' hours were limited to weekdays during the day, forcing participants to visit an urgent care facility or the emergency department when they had an urgent health care need after hours.
The biggest recent change in access described by participants was the addition of patient portals. Participants who had used the portal were enthusiastic about its ability to help them make appointments, check test results, and communicate with their PCPs. As one said, "I like that I'm able to see my test results beforehand, especially if there's something that we've been following for a period For the caregiver focus groups, overall health and age were reported by the caregivers for the beneficiaries for whom they care. Sex, race/ethnicity, and education are reported for the caregivers of time" (low-risk Medicare, New York). However, most participants had not used the portal, and many were not even aware of it. Some were interested in trying it, but others were not because the current process worked for them, they were "technology averse," they were worried about privacy issues, or they did not have a computer or Internet service. In the words of one participant, "I like to talk to people when I make appointments, not hit buttons on a computer"(dually eligible, Vermont).
## Coordination of care
Participants' experiences with coordination of care varied widely. Most said that information was readily shared between their PCPs, specialists, and hospitals-for example, their PCP knew if they had been in the hospital and was able to access test results from their specialists. Many said that their PCP either visited them in the hospital or called to follow up after they were discharged. Some participants, however, said that information did not readily transfer between the PCP and the hospital or specialists, particularly if they were in different systems. Some said their PCP did not know when they had been in the hospital, or that the PCP did not have access to lab results from the hospital and would order the same tests again.
Most participants thought that the transfer of information had improved in recent years, particularly with the introduction of EHRs. However, some participants noted that the transfer of information did not necessarily mean that their PCPs were coordinating their care for them: they were not sure to what extent their PCPs actually digested the information, and some caregivers, in particular, said that they still felt that they were the ones who had to take the lead in coordinating the care for their loved ones: "The information is there, but I always feel like I have to be the advocate for my mom to say, 'Well, she had this and the results came back that way, so does that mean we need to do something?'" (caregiver, New York).
Some participants mentioned that their practices were providing more assistance in scheduling appointments with specialists than they had in the past. Some Medicaid and dually eligible participants reported, however, that the specialists their PCPs referred them to often would not accept their insurance, so they ended up having to find specialists on their own. A few participants mentioned that their PCPs seemed more willing to provide referrals to specialists than in the past; some considered this an improvement, but some wished the PCP would provide more care themselves, rather than sending them to specialists.
A final change in coordination of care noted by some participants was that during hospital stays, care was increasingly provided by hospitalists, and their PCPs, who knew them best, were not involved with their care. They felt that this practice reduced coordination of care.
## Patient-centered care
Most participants thought that their PCPs communicated well with them: listened carefully, explained things thoroughly and in terms they understood, and spent as much time as needed to address all their concerns. On the other hand, some participants (most of whom were in Medicaid or dually eligible groups) said that their PCPs rushed them, allowed them to discuss only one or two concerns per appointment, did not address their emotional or mental health needs, or made assumptions about their needs.
Most participants were also pleased with the office staff at their practices, describing them as friendly, helpful, professional, courteous, and efficient, and a few thought that the staff had become more efficient and friendlier in recent years. A few participants had complaints, however, including that staff were rude, were inefficient, did not respect confidentiality, did not transmit messages to their PCPs, or made it difficult to reach the PCP. Again, most of the participants who had negative experiences were in Medicaid or dually eligible groups. As one said, "[The staff] act like they are doing you a favor by taking your .
The primary changes that participants noted in patientcentered care related to the introduction of EHRs. Participants said that the EHRs helped to ensure that their PCPs remembered their medical information, facilitated filling prescriptions, and shortened the wait time for getting test results. Many participants had started receiving a printout at the end of their appointments summarizing key information, which most appreciated. Some commented that the EHR did not always work well, especially when first implemented, but that it improved over time. One drawback of EHRs mentioned by some participants, however, was that their PCPs were now typing on the computer during their time with them, and they felt that this made for less personal communication "Sometimes I think they spend more time looking at the computer than looking at you" (low-risk Medicare, Pennsylvania).
A few participants mentioned other changes. Some said that their PCPs took more time to discuss health issues, took a more holistic approach, followed up more, and were more responsive. Others said that the practice seemed to be taking a more proactive approach to care, including administering screeners to assess depression or other health risks, calling patients to remind them about appointments or to tell them when they were due for a test or an appointment, and providing support for non-medical issues, such as transportation or food access. Finally, a few participants noticed a new teambased approach to care-for example, that a nurse would ask them questions before they saw their PCPs; some appreciated this because it made more efficient use of their time with the PCP, but others were frustrated that they had to repeat the same information.
## Shared decision-making
Most participants viewed their relationships with their PCPs as partnerships and said that their PCP respected their opinions and preferences and involved them in making decisions about their treatment. Some, however (primarily Medicaid and dually eligible beneficiaries), felt that their PCPs disregarded their perspectives by not focusing on the health concerns most important to them, not taking their health concerns seriously, or pushing them to have treatments or tests they did not want: "I just don't think they listen to me sometimes…. Something's wrong, and… I really wish it could be checked out, not just telling me to go home and relax" (Medicaid, New York). Some participants commented that they thought that, in general, patients need to advocate for themselves to make sure that their concerns are addressed. In the words of one participant, "It seems like I have to take the lead in asking [my PCP] questions about my… diabetes. I have to be very assertive… in getting information from her about diet or medication, things of that sort"(low-risk Medicare, Rhode Island).
A few participants noted some changes in shared decisionmaking. Some commented that their PCPs were now starting appointments by asking open-ended questions, such as "What's concerning you?" or "What are your goals?" Others noted a general shift in approach: "I think [they now] want you to be an active part of the team and not just sit back and be told what to do" (low-risk Medicare, Minnesota).
## Self-management support
Most participants said that their PCPs talked to them about things they could do to improve their health, but almost none reported having a written care plan and few indicated that they had set specific health-related goals with their PCPs. Moreover, a few participants said their PCPs talked to them about managing their health only if they, the patients, brought it up, and they wished that the PCPs spent more time discussing prevention. On the other hand, some participants said that their PCPs had given up talking to them about behavior change, because they knew that it would be "in one ear and out the other" (Medicaid, Michigan).
Participants reported receiving varying levels of support for managing their health. Only a few, all of whom were either Medicaid or dually eligible participants, had care managers through their PCPs, who helped with needs such as home care, housework, transportation, housing, employment, emotional support, reviewing medications, setting up appointments, and obtaining needed medical equipment. Nearly all participants who received these services found them to be very helpful. Some participants said that their PCPs had referred them to classes on topics such as smoking cessation, diabetes control, or weight loss. While most found the classes helpful, a few said that they were too basic or not relevant for their situation or that they were not able to attend the classes because they could not afford the cost of the class or transportation to get to the class. A few participants said that their PCPs had offered other supports, such as referring them to a dietitian or nutritionist to help support healthy eating or setting them up with a blood pressure cuff and logbooks to monitor their blood pressure. Some commented, though, that the only support their PCPs gave them to aid behavior change was written information such as a pamphlet, which they found inadequate: "Nothing's been said to either [me or my wife] about diet, especially me being diabetic. I was just handed a pamphlet" (dually eligible, North Carolina).
## Solicitation of patient input and awareness of medical home
The extent to which participants reported that their practices solicited feedback from them varied. Many said that their practice administered patient surveys or had other mechanisms for soliciting patient input; some participants said that this solicitation of feedback was new.
Almost no participants were familiar with the term "medical home." When the concept was explained to them, most agreed that their practices met the definition of a medical home and most thought a medical home would be a positive thing: "If everybody's involved in your situation, it's kind of like a brainstorm type thing, you know? If they're all working together, it could… be a lot more beneficial to everybody" (Medicaid, Rhode Island). Some participants expressed concerns about the medical home concept, however. Some thought that PCPs did not have time to fulfill the role envisioned under a medical home: "If doctors are so busy, how are they going to have time to look at records of everybody who don't necessarily have a problem?" (high-risk Medicare, Rhode Island). Others were concerned that the medical home concept could increase bureaucracy, restrict patients' access to providers, increase costs to patients, or threaten patient privacy. Some felt that they personally did not need a medical home, but it could benefit those with more health problems.
# Discussion
The PCMH paradigm is complex and primary care practices have many opportunities to integrate its diverse components into their daily activities. Our findings suggest that from the patient perspective, most of the practices participating in the MAPCP Demonstration were delivering care in a way that was largely consistent with PCMH principles. Most participants described having accessible, well-coordinated, and highquality care and appropriate engagement in shared decisionmaking. These findings are similar to earlier, smaller qualitative studies of patient perspectives on PCMH care. [bib_ref] Asking the patient about patient-centered medical homes: A qualitative analysis, Aysola [/bib_ref] [bib_ref] Veteran patient perspectives and experiences during implementation of a patient-centered medical home..., Tuepker [/bib_ref] Our findings also highlight several areas that may require enhanced transformation from the patient perspective. Relatively few participants used patient portals, indicating that practices' investments in new technologies may have limited effect when patients are not aware of their existence, cannot access them, or do not want to use them. Despite practices' efforts to expand access to care, participants still struggled to reach their PCPs and receive care when facing emergencies during nights and weekends, suggesting the need for additional effort in this area. Participants rarely received written plans or set specific goals with their physicians, which are critical tools for supporting patients' self-management of chronic conditions. Participants noted that although care was generally well-coordinated within a health care system, practices still need to improve coordination of care across systems, including the transfer EHR information and allowing for their PCPs to be engaged in their care when they were in the hospital. As primary care practices function within the realities of limited financial and staffing resources, adoption of PCMH components may require strategic prioritization. The need for prioritization is likely to escalate in the aftermath of COVID-19 pandemic, as many primary care practices may struggle to maintain PCMH functions that they have put in place and introduce new ones.
Patients' perspectives also revealed persistent disparities and the need to enhance the paradigm of patient-centered care. Some participants from Medicaid and dually eligible groups reported feeling stigmatized or that their health concerns were not taken seriously, and some experienced challenges with coordination of care because many specialists would not accept their insurance. These findings are consistent with previous research, which has shown that people who are uninsured or who have Medicaid insurance often feel that they are treated poorly by their health care providers [bib_ref] The role of stigma in access to health care for the poor, Allen [/bib_ref] [bib_ref] Reports of insurance-based discrimination in health care and its association with access..., Han [/bib_ref] and that many providers are unwilling to accept patients with Medicaid coverage. [bib_ref] Effects of changes in incomes and practice circumstances on physicians' decisions to..., Cunningham [/bib_ref] [bib_ref] nearly one-third of physicians said they would not accept new medicaid patients,..., Decker [/bib_ref] [bib_ref] No association found between The Medicaid Primary Care fee bump and physician-reported..., Decker [/bib_ref] These challenges suggest the need for enhanced focus on provider training in cultural competencies, reevaluation of effectiveness of cultural competency trainings across existing programs, and incentives to serve Medicaid beneficiaries and dually eligible.
As reported in other recent qualitative studies describing patient experiences with PCMHs, [bib_ref] Veteran patient perspectives and experiences during implementation of a patient-centered medical home..., Tuepker [/bib_ref] we also found that most participants did not know what a medical home was and were not aware that their practices were participating in an initiative seeking to enhance patient centeredness of care. In addition, few participants said that their practices had taken any steps to solicit their feedback. While practice improvement is a responsibility of practice staff, meaningful transformation cannot be achieved in the absence of patient engagement and buy-in. More recent iterations of CMS valued based care initiatives such as Accountable Care Organizations and the Comprehensive Primary Care Plus initiative require more structured patient engagement in organizational operations than those required in MAPCP, such as patient representation on organizational boards and advisory councils,which may help to improve practices' responsiveness to patient needs and concerns. While seeking patient input and feedback is an important and necessary step forward, it is the willingness of practices and providers to act on patient input and recommendations that will facilitate change.
This study is subject to several limitations. First, the selection criteria for the study (that participants be Medicaid, Medicare, or dually eligible beneficiaries or caregivers of beneficiaries in eight select states and speak English fluently) led to a sample that was primarily non-Hispanic White, over age 60, and English-speaking; results may not be generalizable to other demographic groups. We were also unable to summarize the experiences of children enrolled in Medicaid because we were unable to recruit a sufficient number of caregivers of Medicaid children. However, because participants were drawn from eight different states, including both urban and rural areas, the study represents a much greater breadth of experiences than previous studies, and the high number of Medicaid and dually eligible beneficiaries ensures that perspectives of low-income and disabled populations are reflected. Second, because we collected data only from patients of practices that were participating in the MAPCP Demonstration, we cannot assess the extent to which their experiences may differ from those whose practices were not participating in the MAPCP Demonstration. Third, because the focus groups were conducted at one point in time, approximately 2.5 years after the MAPCP Demonstration began, we can only indirectly assess the extent to which patient experiences changed after their practices began their PCMH transformation. Finally, because each focus group included participants from more than one practice, it was not possible to link the comments from participants in specific focus groups to individual practices. As a result, we were unable to assess which practices had been most successful in achieving PCMH transformation or to what extent patient perceptions of care aligned with practice transformation efforts.
Additional research is needed to further understand PCMH transformation from the patient perspective. Research that links patient experiences to specific practices, including characteristics of the practices (e.g., practice culture, leadership, and physician engagement) and the PCMH transformation strategies the practices have implemented, could help elucidate how practice characteristics or strategies influence patient experiences.
[table] Table 1: Number of Focus Groups and Participants by CategoryIn Vermont, groups were conducted with Support and Services at Home (SASH) program participants. SASH provided support services and care coordination to Medicare beneficiaries living in subsidized housing and the surrounding communities [/table]
[table] Table 2: Characteristics of Focus Group Participants [/table]
[table] Table 3: Ways in Which Patients' Experiences Were or Were Not Consistent with PCMH Principles [/table]
|
Clinical Application of Rapid Upper Limb Assessment and Nordic Musculoskeletal Questionnaire in Work-Related Musculoskeletal Disorders: A Bibliometric Study
Citation: Kakaraparthi, V.N.; Vishwanathan, K.; Gadhavi, B.; Reddy, R.S.; Tedla, J.S.; Alshahrani, M.S.; Dixit, S.; Gular, K.; Zaman, G.S.; Gannamaneni, V.K.; et al. ClinicalAbstract: Assessment of work-related musculoskeletal disorders (WMSDs) using the Rapid Upper Limb Assessment (RULA) and the Nordic Musculoskeletal Questionnaire (NMQ) has become widely accepted and reported in the literature. The objectives of this study are to (1) recognize and describe the topmost 50 cited scientific articles in WMSDs using the RULA and NMQ and (2) explore the factors that contribute to making an article influential. In this bibliometric study, we used the Web of Science and MEDLINE databases to identify the top 50 cited articles published from 1993 to 2022. The data collected were the title of the journal, number of citations, year of publication, type of the study, institution where the work was conducted, level of evidence, contribution of primary authors, and country of origin of the work. Our results showed that the top 50 cited articles were published between 1980 and 2010. The 2000s was the most valuable decade. Regarding journals, the Work journal had the highest number of articles concerning the use of RULA and NMQ in healthcare professionals. The maximum number of citations regarding RULA occurred in the Journal of Robotic Surgery (n = 50) and the maximum for NMQ occurred in the Journal of Safety Research (n = 106). Most articles originated from the United States, followed by England and the Netherlands. Eight authors had two publications published in the top 50 list. The majority of the topmost cited research articles were cross-sectional studies. Most of these studies were level III evidence. The bibliometric analysis from this study provides insights to researchers to choose the most appropriate and influential journal for submitting work on WMSDs.
# Introduction
Work-related musculoskeletal disorders are considered to be a major issue worldwide . Awkward non-physiological postures while working, hereditary causes, aging, and psychological issues could be the responsible factors causing work-related musculoskeletal disorders (WMSDs). These disorders include damage to various components of the locomotor system such as muscles, joints, nerves, and other connective tissues [bib_ref] Prevalence of upper limb musculo skeletal disorders among brass metal workers in..., Gangopadhyay [/bib_ref].
Healthcare specialists with prolonged interaction with patients had the greatest percentage of WMSDs. WMSDs are recognized as a major occupational hazard in various countries [bib_ref] Musculoskeletal symptoms among dentists in relation to work posture, Ratzon [/bib_ref] [bib_ref] Volvo award in clinical sciences. A new clinical model for the treatment..., Waddell [/bib_ref] [bib_ref] Work-break schedules for preventing musculoskeletal symptoms and disorders in healthy workers, Luger [/bib_ref] [bib_ref] Antecedents of work related musculoskeletal disorders in software professionals, Silvian [/bib_ref] since there is a considerable economic burden imposed by these conditions [bib_ref] Van der Beek, A. The economic evaluation of a participatory ergonomics programme..., Driessen [/bib_ref] [bib_ref] Costs of work-related musculoskeletal disorders (MSDs) in developing countries: Colombia case, Piedrahita [/bib_ref].
The use of a validated, uniform evaluation tool is recommended in order to diagnose and devise effective interventions to tackle WMSDs [bib_ref] Promoting work ability in a structured national rehabilitation program in patients with..., Stigmar [/bib_ref]. Two such important evaluation tools that have been utilized for many years are the Rapid Upper Limb Assessment (RULA) [bib_ref] RULA: A survey method for the investigation of work-related upper limb disorders, Mcatamney [/bib_ref] and the Nordic Musculoskeletal Questionnaire (NMQ) [bib_ref] Standardised Nordic questionnaires for the analysis of musculoskeletal symptoms, Kuorinka [/bib_ref]. Both these tools demonstrated greater reliability in evaluating ergonomic hazards in various healthcare practitioners [bib_ref] Musculoskeletal symptoms among dentists in relation to work posture, Ratzon [/bib_ref] [bib_ref] Evaluation of working posture and prevalence of musculoskeletal symptoms among medical laboratory..., Shreya [/bib_ref].
The RULA was developed by Corlett and McAtamney in 1993, primarily as an observational tool. It was utilized to assess the person's susceptibility to loads due to the posture of the upper extremity and spine and the assistance of the lower extremities required to cope with the extra load during one's work. The overall RULA score ranges from one to seven. Higher scores suggest a greater likelihood of a musculoskeletal problem. RULA scores either equal to or greater than five necessitate a change in posture during work [bib_ref] RULA: A survey method for the investigation of work-related upper limb disorders, Mcatamney [/bib_ref].
The Nordic Musculoskeletal Questionnaire (NMQ) was developed by [bib_ref] Standardised Nordic questionnaires for the analysis of musculoskeletal symptoms, Kuorinka [/bib_ref]. It is an easy, well-defined questionnaire comprising a body map describing nine functional sites revealing both sides' upper limbs, lower limbs, upper back, and lower back. It incorporates questions on symptoms felt by the individuals in the previous 12 months and the past seven days, along with constraints in activity levels in the previous 12 months. The responses to these questions are documented. NMQ can be administered either by a conversation method or can be self-administered [bib_ref] Standardised Nordic questionnaires for the analysis of musculoskeletal symptoms, Kuorinka [/bib_ref].
Numerous articles on WMSDs have been published previously, and there seems to be a steady increase in the number of publications on the subject. A bibliometric review on RULA by Gómez-Galan et al. [bib_ref] Musculoskeletal risks: RULA bibliometric review, Gómez-Galán [/bib_ref] only reported a country-wise, year-wise, and journal-wise analysis, while citations and other analyses were not reported. Despite their importance, the most influential research articles on the use of RULA and NMQ in healthcare practitioners have not been previously reported. The total number of citations of a research article is an indicator of the influence wielded in the area of interest [bib_ref] Top-cited articles in rehabilitation, Shadgan [/bib_ref]. A citation breakdown is a bibliometric analysis to calculate the comparative reputation of a scientific research paper by investigating the citations credited to that research paper [bib_ref] Citation analysis for measuring the value of scientific publications: Quality assessment tool..., Schoonbaert [/bib_ref]. In recent years, the highest-cited research articles have been detected by utilizing citation scrutiny in numerous medical fields [bib_ref] The Top 100 Most Cited Articles on Anterior Cruciate Ligament Reconstruction: A..., Tang [/bib_ref] [bib_ref] Analysis of 100 most cited articles on forensic odontology, Sengupta [/bib_ref] [bib_ref] Fifty top-cited classic papers in orthopedic elbow surgery: A bibliometric analysis, Huo [/bib_ref] [bib_ref] Top 100 cited articles on diabetes mellitus and Covid-19: A bibliometric analysis, Vishwanathan [/bib_ref].
Assessing bibliometrics is the arena of the medical field, which utilizes approaches such as citation analysis to estimate the implementation of research [bib_ref] Bibliometrics basics, Cooper [/bib_ref]. The total number of citations of formerly published work in scientific papers indicates its following identification in a particular area of study. It has an additional effect on the scientific community [bib_ref] Citation analysis as a tool in journal evaluation: Journals can be ranked..., Garfield [/bib_ref]. Bibliometric studies are mainly valuable for guiding healthcare specialists in creating evidence-based judgments in clinical practice. Along with this, the topmost cited papers in research are generally considered the most influential research within a particular field, raising innovative research ideas [bib_ref] The 50 most-cited articles in orthopaedic surgery from Mainland China, Jia [/bib_ref]. Therefore, a detailed bibliometric analysis of the topmost cited research articles can benefit the understanding of a specific discipline, which aids in generating useful guidelines in the field of biomedical research [bib_ref] A bibliometric analysis and visualization of the top-cited publications in mild traumatic..., Shi [/bib_ref].
Therefore, we conducted analyses on the citations, subject-wise contribution, quartilewise contribution, primary author contribution, institutional-wise, study type, and evidence level of the top 50 most-cited articles regarding the RULA and NMQ.
# Methods
## Collection of articles
We searched for all the relevant articles on the RULA and NMQ using Web of Science and MEDLINE records to identify the top 50 most commonly cited research articles in various healthcare-related specialties [bib_ref] The 100 top-cited classic papers in hand surgery, Joyce [/bib_ref] [bib_ref] Anatomy of success: The top 100 cited scientific reports focused on hypertension..., Oh [/bib_ref] [bib_ref] The 100 top-cited tuberculosis research studies, Chen [/bib_ref]. The following search terms were used: Ergonomics, assessment, workplace, musculoskeletal disorders, RULA, and NMQ. The search was executed on 20 June 2021 and yielded 429 results related to the RULA and 710 concerning the NMQ.
## Inclusion criteria
Only original articles reporting the use of the RULA and/or NMQ in healthcare professionals were included in this bibliometric analysis.
## Exclusion criteria
Original articles reporting the use of the RULA and/or NMQ in non-healthcare practitioners, narrative reviews, systematic reviews, meta-analyses, letters to the editor, short-communication-type articles, and abstracts of scientific meetings were omitted from the bibliometric analysis [bib_ref] Intraclass correlations: Uses in assessing rater reliability, Shrout [/bib_ref].
## Organization of articles
In total, 1139 studies were identified during the initial screening. Out of these, 376 studies were excluded for not having reported musculoskeletal outcomes. After a careful screening of the remaining 763 studies based on the inclusion standards, 544 studies were excluded due to inadequate data, failure to address the research question, dealing with non-healthcare practitioners, or duplicates, and the remaining 219 studies met the inclusion criteria. These articles were organized according to the number of citations, and the top 50 most-cited articles regarding the RULA and NMQ were incorporated into the final analysis [fig_ref] Figure 1: Flowchart illustrating the procedure of allocation of articles [/fig_ref].
# Methods
## Collection of articles
We searched for all the relevant articles on the RULA and NMQ using Web of Science and MEDLINE records to identify the top 50 most commonly cited research articles in various healthcare-related specialties [bib_ref] The 100 top-cited classic papers in hand surgery, Joyce [/bib_ref] [bib_ref] Anatomy of success: The top 100 cited scientific reports focused on hypertension..., Oh [/bib_ref] [bib_ref] The 100 top-cited tuberculosis research studies, Chen [/bib_ref]. The following search terms were used: Ergonomics, assessment, workplace, musculoskeletal disorders, RULA, and NMQ. The search was executed on 20 June 2021 and yielded 429 results related to the RULA and 710 concerning the NMQ.
## Inclusion criteria
Only original articles reporting the use of the RULA and/or NMQ in healthcare professionals were included in this bibliometric analysis.
## Exclusion criteria
Original articles reporting the use of the RULA and/or NMQ in non-healthcare practitioners, narrative reviews, systematic reviews, meta-analyses, letters to the editor, shortcommunication-type articles, and abstracts of scientific meetings were omitted from the bibliometric analysis [bib_ref] Intraclass correlations: Uses in assessing rater reliability, Shrout [/bib_ref].
## Organization of articles
In total, 1139 studies were identified during the initial screening. Out of these, 376 studies were excluded for not having reported musculoskeletal outcomes. After a careful screening of the remaining 763 studies based on the inclusion standards, 544 studies were excluded due to inadequate data, failure to address the research question, dealing with non-healthcare practitioners, or duplicates, and the remaining 219 studies met the inclusion criteria. These articles were organized according to the number of citations, and the top 50 most-cited articles regarding the RULA and NMQ were incorporated into the final analysis [fig_ref] Figure 1: Flowchart illustrating the procedure of allocation of articles [/fig_ref].
## Extraction of articles
Two autonomous assessors selected suitable studies based on the inclusion and exclusion criteria. Then, the reviewers identified the full texts of all relevant articles and selected the top 50 cited publications regarding the RULA and NMQ.
The following data were extracted from the included articles: Title of the paper, primary author's name, journal name, time of publication, impact feature of the journal in 2020, the total number of citations of the research paper, study design, article type, average citations per year, mentions, country, organization of basis, authorship, journal, and level of evidence. Citation statistics for all the selected research articles were achieved by exploring the SCOPUS records. The level of evidence of the articles was determined based on the Centre for Evidence-Based Medicine (CEBM) guidelines, Oxford, UK.
# Statistical analysis
Two reviewers individually agreed on the level of evidence for the research articles. The agreement was outstanding for the level of evidence, with an intraclass correlation coefficient of 0.90 [bib_ref] Intraclass correlations: Uses in assessing rater reliability, Shrout [/bib_ref]. Two assessors decided on all variances through conversation, and another assessor was involved when no agreement was reached. Data analysis was performed using SPSS software (version 24.0 for Windows; SPSS, Inc., Chicago, IL, USA).
# Results
A total of 1139 articles were found from the search, established on the citations, then the 50 most prominent articles concerning the RULA and NMQ regarding WMSDs were selected for the study.
## Citation breakdown
The 50 most commonly cited articles using the RULA and NMQ in regard to WMSDs have 2011 citations overall. These range from 1 to 50 concerning the RULA with a mean of 11.36 (SD = 11.32) mentions per article, with a total of 552 citations, as shown in. Regarding NMQ, the number of citations ranged from 13 to 106, with a mean of 30.66 (SD = 18.30) for each article, with a total of 1518 citations, as shown in.
## Title & year journal total citations
Design and ergonomic assessment of an infusion set connector tool used in nursing work. Otolaryngol. Head. Neck. Surg. 1 SOPEZ: study for the optimization of ergonomics in the dental practice-musculoskeletal disorders in dentists and dental assistants: a study protocol.
## Subject wise
Regarding the RULA, according to a subject-wise analysis, the majority of the research papers were published under the headings of Musculoskeletal disorders (n = 18; citations: 179), Work (n = 13; citations: 102), and Ergonomics (n = 8; citation: 110), followed by Risk factors (n = 5; citations: 45). In regard to the NMQ, according to the subject-wise analysis, the majority of the research papers were published under the heading of Musculoskeletal disorders (n = 41; citations: 1122); Work (n = 16; citations: 439), and Risk factors (n = 10; citation: 106), followed by Ergonomics (n = 2; citations: 37).
## Quartile wise
In regard to RULA, among all 35 journals, 10 journals (28.5%) were ranked in the first quartile (n = 225 citations), 1 (2.8%) in the second (n = 9 citations), 9 (25.7%) in third (n = 57 citations), 7 (20%) in the fourth (n = 112 citations), and 8 (22.8%) in the non-quartile category (n = 149 citations).
In regard to NMQ, among all 40 journals, 7 journals (17.5%) were ranked in the first quartile (n = 233 citations), 7 (17.5%) in the second (n =335 citations), 12 (30%) in third (n = 430 citations), 5 (12.5%) in the fourth (n = 267 citations), and 9 (22.5%) (n = 253 citations) in the non-quartile category.
The significance of dividing the journals into quartiles indicates where a journal's standing lies in a specific subject classification. These quartiles rank the journals from highest to lowest, established on their impact factor or impact index. A quartile is a category of scientific journals that shows their credibility, reliability, and quality. The quartile also reflects the demand for the journal by the scientific community.
## Oldest and latest articles
## Authors
Five authors published more than one research article related to the RULA. Two authors published more than one research article related to the NMQ.
## Authors
Five authors published more than one research article related to the RULA. Two authors published more than one research article related to the NMQ.
## Countries
The top-cited articles in WMSDs in healthcare practitioners in regard to the RULA originated from 13 countries. The country delivering the most articles was the USA (n = 18), followed by Iran (n = 11), India and Brazil (n = 4), Sweden (n = 3), and Germany and England (n = 2). The remaining articles were from other respective countries. Therefore, the geographic distribution of these most-cited articles is represented in.
The top-cited articles in WMSDs in healthcare practitioners in regard to the NMQ originated from 20 countries. The countries delivering the highest number of articles were Iran (n = 3), followed by Australia (n = 7), the USA (n = 5), France, Portugal, and Brazil (n = 3). The remaining articles were from the other respective countries. Therefore, the geographic distribution of these most-cited articles is represented in.
## Countries
The top-cited articles in WMSDs in healthcare practitioners in regard to the RULA originated from 13 countries. The country delivering the most articles was the USA (n = 18), followed by Iran (n = 11), India and Brazil (n = 4), Sweden (n = 3), and Germany and England (n = 2). The remaining articles were from other respective countries. Therefore, the geographic distribution of these most-cited articles is represented in.
## Institutions
A total of 50 institutions produced the top-cited articles in WMSDs concerning the RULA. Three institutions published three or more of the most commonly cited articles. Thirteen (26%) of these institutions were located in the USA, followed by 8 (16%) in Iran, 4 (8%) in Brazil and India, 3 (6%) in Sweden, and 2 (4%) in Germany, Canada, Turkey, Canada, South Korea, and England. The remaining countries contributed one article each. The institutions producing the most articles were Ohio State University, Tehran University, and the University of Pittsburgh (n = 3), followed by the University of North Carolina and Iran University of Medical Sciences (n = 2), and the remaining Universities contributed one article each. A total of 49 institutions were involved in producing the top-cited articles in WMSDS concerning the NMQ. Three institutions published three or more of the most commonly cited articles. Seven (14.2%) of these institutions were located in Iran, followed by six (12.2%) in Australia, five (10.2%) in the USA, three (6.1%) in Sweden, China, and Iran, and two (4%) in France, Saudi Arabia, and Egypt. The remaining countries contributed one article each. The institutions producing the most articles were Tehran University, Shiraz University, and Babol University (n = 3), followed by Universidade Nova de Lisboa, Queensland University, National Institute of Industrial Health, and King Abdulaziz University (n = 2), and the remaining universities contributed one article each.
Ranking institutions by country may help to build global brand visibility, forge strategic partnerships, and recruit international talent. It also allows for analyses according to academic reputation, employer reputation, research citations per paper, h-index, and the International Research Network.
## Study designs of articles
The main stream of the top 50 cited research articles associated with both the RULA and the NMQ were cross-sectional studies, followed by observational, comparative, experimental, and other types of studies [bib_ref] Overview of research designs, Clancy [/bib_ref] [fig_ref] Table 8: Study design of top-cited articles of RULA and NMQ in WMSD practitioners [/fig_ref]. [fig_ref] Table 8: Study design of top-cited articles of RULA and NMQ in WMSD practitioners [/fig_ref]. Study design of top-cited articles of RULA and NMQ in WMSDs in regard to healthcare practitioners.
## Study design articles
University, and Babol University (n = 3), followed by Universid Queensland University, National Institute of Industrial Health, an versity (n = 2), and the remaining universities contributed one arti Ranking institutions by country may help to build global bran tegic partnerships, and recruit international talent. It also allows fo academic reputation, employer reputation, research citations per p International Research Network.
## Study designs of articles
The main stream of the top 50 cited research articles associate and the NMQ were cross-sectional studies, followed by observati perimental, and other types of studies [bib_ref] Overview of research designs, Clancy [/bib_ref] [fig_ref] Table 8: Study design of top-cited articles of RULA and NMQ in WMSD practitioners [/fig_ref].
## Level of evidence
The top 50 most-cited articles could be classified into all evid The majority of the articles were within evidence level III (n = 70), level IV (n = 23), evidence level 1B (n = 4), and evidence level IIB (n
# Discussion
To the best of our knowledge, this is the first bibliometric analy related musculoskeletal disorders concerning the RULA and NMQ citations for each paper, the unique bibliometric indicator, is a valu the impact of publications. From the analysis of these top 50 most-c we attempted to determine what fundamentals of a research articl tive. The top-cited article in our list in relation to the RULA had 50 related to the NMQ had 106 citations [bib_ref] A detailed analysis of musculoskeletal disorder risk factors among Japanese nurses, Smith [/bib_ref]. This number is much lo in relation to non-healthcare specialties [bib_ref] A prospective study of work related factors and physical exercise as predictors..., Miranda [/bib_ref] [bib_ref] Number of pain sites is associated with demographic, lifestyle, and health-related factors..., Kamaleri [/bib_ref]. The citations varie predominantly reliant on the number of investigators in exact thera Identifying those classic articles is beneficial for better insight int gress of evaluation of WMSDs by using RULA and NMQ among h and also for planning future research.
## Level of evidence
The top 50 most-cited articles could be classified into all evidence levels [fig_ref] Table 9: Level of evidence of the top 50 cited articles in WMSDs in... [/fig_ref]. The majority of the articles were within evidence level III (n = 70), followed by evidence level IV (n = 23), evidence level 1B (n = 4), and evidence level IIB (n = 3) [bib_ref] Updating the assignment of levels of evidence, Marx [/bib_ref].
# Discussion
To the best of our knowledge, this is the first bibliometric analysis of papers on workrelated musculoskeletal disorders concerning the RULA and NMQ. The total number of citations for each paper, the unique bibliometric indicator, is a valuable tool to determine the impact of publications. From the analysis of these top 50 most-cited articles published, we attempted to determine what fundamentals of a research article make it vastly effective. The top-cited article in our list in relation to the RULA had 50 citations [bib_ref] Postural ergonomics during robotic and laparoscopic gastric bypass surgery: A pilot project, Lawson [/bib_ref] , and that related to the NMQ had 106 citations [bib_ref] A detailed analysis of musculoskeletal disorder risk factors among Japanese nurses, Smith [/bib_ref]. This number is much lower than the WMSDs in relation to non-healthcare specialties [bib_ref] A prospective study of work related factors and physical exercise as predictors..., Miranda [/bib_ref] [bib_ref] Number of pain sites is associated with demographic, lifestyle, and health-related factors..., Kamaleri [/bib_ref]. The citations varied between specialties, predominantly reliant on the number of investigators in exact therapeutic fields [bib_ref] Musculoskeletal symptoms among dentists in relation to work posture, Ratzon [/bib_ref]. Identifying those classic articles is beneficial for better insight into the history and progress of evaluation of WMSDs by using RULA and NMQ among healthcare practitioners and also for planning future research.
In relation to our present study, open access may be another significant factor in drawing citations [bib_ref] Ergonomics and human factors in endoscopic surgery: A comparison of manual vs..., Lee [/bib_ref] [bib_ref] Ergonomic deficits in robotic gynecologic oncology surgery: A need for intervention, Craven [/bib_ref]. It is widely known that open-access scientific papers are more easily accessible and cited compared to non-open scientific papers [bib_ref] Assessing work-related musculoskeletal symptoms among otolaryngology residents, Wong [/bib_ref] [bib_ref] Musculoskeletal disorders among a group of Iranian general dental practitioners, Tirgar [/bib_ref]. However, we also found that articles concerning otolaryngologists and neurosurgeons in open-access format did not obtain considerably more citations than non-open-access articles in the same specialties [bib_ref] Impact of Workspace Design on Radiation Therapist Technicians' Physical Stressors, Mental Workload,..., Mosaly [/bib_ref]. It can thereby be inferred that the extent to which open-access articles increase citations varies depending on the specialty.
All top 50 cited articles in WMSDs in relation to the RULA and NMQ were published in English. This establishes that English is the most commonly used scholarly language in WMSDs. Additionally, all of the most-cited papers occurred in 35 journals in relation to RULA and 40 in relation to NMQ. In relation to individual contributions of the journal published, the most productive journal was Work-A Journal of Prevention, Assessment & Rehabilitation, with ten articles (both RULA and NMQ) in the top list. This result suggests that this journal was the most preferred in regard to WMSDs.
In relation to quartile classifications, these were analyzed for individual journals in each theme classification, corresponding to which quartile of the scientific journal dominates in the impact factor division of that subject category [bib_ref] Features of top-rated gold open access journals: An analysis of the scopus..., Ennas [/bib_ref]. The total number of publications and/or the distribution of total publications in a particular quartile generally corresponded to the first quartile (Q1) in regard to the RULA and the third quartile (Q3) in regard to the NMQ. Moreover, in the present study, the overall greater number of cited articles occurred in the third quartile (Q3).
With relation to authors and institutions, good publication track records in WMSDs related to healthcare practitioners have been determined in the present study [bib_ref] Patient positioning during in-office otologic procedures impacts physician ergonomics, Govil [/bib_ref] [bib_ref] Predisposing factors for musculoskeletal symptoms in intensive care unit nurses, Sezgin [/bib_ref] [bib_ref] Effects of a PRECEDE-PROCEED model based ergonomic risk management programme to reduce..., Sezgin [/bib_ref] [bib_ref] Prevalence of work-related musculoskeletal symptoms among Iranian workforce and job groups, Choobineh [/bib_ref]. Tehran University has great priority in this field by contributing six publications overall. Musculoskeletal disorders are the most prevalent theme, with approximately 18 articles related to the RULA and 41 related to the NMQ, which emphasizes the significance of musculoskeletal disorders in subspecialties such as joints, bones, muscles, etc.
In regard to study type, we demonstrated that cross-sectional studies were the most popular study type in this study. This finding is consistent with many studies assessing WMSDs [bib_ref] Work-related musculoskeletal disorders among nurses in Ibadan, South-west Nigeria: A cross-sectional survey, Tinubu [/bib_ref] [bib_ref] Upper limb work-related musculoskeletal disorders in operating room nurses: A multicenter cross-sectional..., Clari [/bib_ref] [bib_ref] The prevalence, characteristics, and impact of work-related musculoskeletal disorders among physical therapists..., Kakaraparthi [/bib_ref]. Furthermore, most of the top cited articles were level III evidence. This result indicates that the level of evidence is certainly not a positive aspect of the total number of citations [bib_ref] Citations, citation indicators, and research quality: An overview of basic concepts and..., Aksnes [/bib_ref]. This is likely because original ideas and assessments are initially published as cross-sectional studies and still catch the attention of practitioners or researchers.
In relation to the sequential allocation of citations, the 2001-2010 time period had the highest number of mentions on our 50 top-cited research articles list related to WMSDs in healthcare practitioners; 33 articles were related to the RULA, and 37 were related to the NMQ. There are various reasons for this result in the present study. First, it could take approximately 10 years or further for prominent article citations to reach their peak, as acknowledged through bibliometric examination [bib_ref] Postural ergonomics during robotic and laparoscopic gastric bypass surgery: A pilot project, Lawson [/bib_ref] [bib_ref] A detailed analysis of musculoskeletal disorder risk factors among Japanese nurses, Smith [/bib_ref]. Second, researchers tend to underestimate the influence of the most recent studies compared to old studies [bib_ref] Core journals of rehabilitation: Identification through index analysis, Bohannon [/bib_ref]. Third, some research scholars believe that the accurate importance of the research papers cannot be considered until at least 10-20 years afterward the date of publication in the journal [bib_ref] Citation classics in anesthetic journals, Baltussen [/bib_ref]. Therefore, research articles published recently need additional time to gather citations to establish their importance.
Lastly, in relation to the geographical distribution, more than one-third of the topcited WMSD-related articles were from the United States. It is no wonder that the United States has benefited in the rankings, as various healthcare sectors have found similar results [bib_ref] Highly cited works in radiology: The top 100 cited articles in radiologic..., Pagni [/bib_ref] [bib_ref] One hundred citation classics in general surgical journals, Paladugu [/bib_ref] due to various aspects: (1) The American research and scientific community is substantially bigger than that of any other community in the world; (2) the availability of more funding for research; and (3) previous research studies have pointed out that there is a predisposition for researchers from the United States to cite local articles [bib_ref] National bias: A comparison of citation practices by health professionals, Campbell [/bib_ref]. This may result in the authors from the US having an essential advantage when trying to publish the most cited and dominant research articles.
## Limitations of the study
There are several limitations in the present study. First, we only utilized the Web of Science and MEDLINE for our analysis; consequently, some research papers indexed by Google Scholar or SCOPUS or basic medical journal databases may have been missed. Second, there was an association with the publication phase, with newly published research articles experiencing a disadvantage in relation to the citation calculation. With this issue, possibly more appropriate and prominent research articles published recently may not have been included in the top 50 cited articles list because they have not had sufficient time to gather the required number of citations. Lastly, possible factors that influence the citations cannot be exactly established in this analysis; these include researcher selfmentions, citations in workbooks, meeting notes, lectures, symposiums, and web-based writings [bib_ref] Highly cited works in radiology: The top 100 cited articles in radiologic..., Pagni [/bib_ref] [bib_ref] Fifty most cited articles in orthopedic shoulder surgery, Namdari [/bib_ref]. Nevertheless, the statistics delivered in the present work do provide insight into the most important investigation areas of WMSDs.
# Conclusions
The top-cited papers were all published in the English language, originated from institutions in the United States, and were primarily cross-sectional studies with level III evidence. Most articles were published in Work-A Journal of Prevention, Assessment & Rehabilitation. This top citation list offers valuable insight into the assessment of WMSDs using the RULA and NMQ, especially in the healthcare sector, and serves as a crucial basis of evidence for investigators and research scholars for future research on WMSDs.
[fig] Figure 1: Flowchart illustrating the procedure of allocation of articles. [/fig]
[fig] 2019: Appl. Ergon. 3 Study of musculoskeletal risks of the office-based surgeries. (2012) Work 3 Risk of musculoskeletal disorders in upper limbs in dental students: concordance of different methods for estimation of body angle. (2013) Indian J. Dent. Res. 3 Reliability, Construct Validity and Interpretability of the Brazilian version of the Rapid Upper Limb Assessment (RULA) and Strain Index (SI). (2018) Braz. J. Phys. Ther. 3 Estimation of surgeons' ergonomic dynamics with a structured light system during endoscopic surgery. (2019) Int. Forum. Allergy Rhinol. 3 Use of the Omaha System to identify musculoskeletal problems in intensive care unit nurses: a case study. (2019) Br. J. Nurs. 2 Ergonomic assessment of robotic general surgeons: a pilot study. (2020) J. Robot Surg. 2 Evaluating the efficacy of an educational ergonomics module for improving slit lamp positioning in ophthalmology residents. (2019) Can. J. Ophthalmol. 2 Ergonomic assessment of the first assistant during robot-assisted surgery. (2019) J. Robot Surg. 2 Risk factors for musculoskeletal disorders in an obstetrician-gynecologist and orthopedic surgeon. (2020) Work 1 Impact of Workspace Design on Radiation Therapist Technicians' Physical Stressors, Mental Workload, Situation Awareness, and Performance. (2021) Pract. Radiat. Oncol. 1 A 10-week exercise intervention can improve work posture but not neck/shoulder symptoms in dental health students: A pilot cohort study. (2020) Work 1 An intuitive surgical handle design for robotic neurosurgery. (2021) Int. J. Comput. Assist. Radiol. Surg. 1 Relationships between the postures of dentists and chairside dental assistants. (2020) J. Dent. Educ. 1 Quantitative Assessment of Surgical Ergonomics in Otolaryngology. (2020) [/fig]
[fig] Figure 2: All 50 highly cited articles on WMSDs concerning the RULA and NMQ were published in English over the past 28 years, from 1993 to 2021. The oldest article in relation to the RULA was published in the British Journal of Biomedical Science by Kilroy et al. (2000). The oldest article in relation to the NMQ was published in the Journal of Dental Hygiene by Oberg et al. (1993).It was shown that the length of time and paper citation count also depend on the length of the article, number of references, title length, type of affiliations, and number of authors. However, all those factors are independent of the scientific merit of publication.Year-wise distribution: The highest output of the top-cited articles in relation to the RULA was noted in the 2010s wherein 33 articles of the top 50 most-cited articles were published (Figure 2a). In regard to the NMQ, the highest output of the top cited articles was noted in the 2010s wherein 37 articles of the top 50 were published (Figure 2b). Int. J. Environ. Res. Public Health 2023, 20, x FOR PEER REVIEW 13 of 22 3.5. Oldest and Latest Articles All 50 highly cited articles on WMSDs concerning the RULA and NMQ were published in English over the past 28 years, from 1993 to 2021. The oldest article in relation to the RULA was published in the British Journal of Biomedical Science by Kilroy et al. (2000). The oldest article in relation to the NMQ was published in the Journal of Dental Hygiene by Oberg et al. (1993).It was shown that the length of time and paper citation count also depend on the length of the article, number of references, title length, type of affiliations, and number of authors. However, all those factors are independent of the scientific merit of publication.Year-wise distribution: The highest output of the top-cited articles in relation to the RULA was noted in the 2010s wherein 33 articles of the top 50 most-cited articles were published(Figure 2a). In regard to the NMQ, the highest output of the top cited articles was noted in the 2010s wherein 37 articles of the top 50 were published(Figure 2b). Cont. [/fig]
[fig] Figure 3: (a): Geographic distribution of the top 50 most-cited articles in WMSDS using RULA in regard to healthcare practitioners. (b): Geographic distribution of the top 50 most-cited articles in WMSDS using NMQ in regard to healthcare practitioners. [/fig]
[fig] Author: Contributions: Conceptualization, V.N.K. and K.V.; methodology, V.N.K. and R.S.R.; software, B.G., J.S.T. and R.S.R.; validation, V.N.K. and K.V.; formal analysis, V.N.K., K.V., B.G., R.S.R. and S.D.; investigation, K.G., G.N., S.D. and V.N.K.; resources, K.G. and V.N.K.; data curation, V.N.K., M.S.A., G.S.Z., V.K.G. and M.S.S.; writing-original draft preparation, V.N.K., K.V., V.K.G., R.S.R. and J.S.T.; writing-review and editing, V.N.K., K.V., V.K.G., R.S.R., J.S.T. and G.N.; supervision, V.N.K. and K.V.; project administration, V.N.K., K.V., B.G. and R.S.R.; funding acquisition, R.S.R. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by King Khalid University, grant number: RGP 1/6/43. Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. Data Availability Statement: All relevant data produced or evaluated during this study are included in the manuscript. [/fig]
[table] Table 1: (a): List of top 50 cited articles in WMSDS using RULA concerning healthcare practitioners. (b): List of top 50 cited articles in WMSDS using NMQ in regard to healthcare practitioners. [/table]
[table] Table 2: Cont. [/table]
[table] Table 3: Top Journals with their contribution to the 50 most-cited articles on WMSDs using NMQ related to healthcare practitioners. [/table]
[table] Table 4: (a) Number and citations of research articles as per subject-wise analysis concerning RULA, (b) number and citations of research articles as per subject-wise analysis in regard to NMQ. [/table]
[table] Table 5: (a): Number and citations of research articles as per quartile in regard to RULA. (b): Number and citations of research articles as per quartile in regard to NMQ. [/table]
[table] Table 6: (a): First authors of top-cited articles on WMSDs in relation to RULA. (b): First authors of top-cited articles on WMSDs in relation to NMQ. [/table]
[table] Table 7: (a): Top Institutional distribution of all top 50 cited articles in WMSDS related to healthcare practitioners concerning RULA. (b): Top Institutional distribution of all top 50 cited articles in WMSDS related to healthcare practitioners concerning NMQ. [/table]
[table] Table 8: Study design of top-cited articles of RULA and NMQ in WMSD practitioners. [/table]
[table] Table 9: Level of evidence of the top 50 cited articles in WMSDs in regard [/table]
[bib_ref] Analysis of 100 most cited articles on forensic odontology, Sengupta [/bib_ref] |
Obstetrics and Gynecology
OBSTETRICS AND GYNECOLOGY
M.DM.DOBSTETRICS AND GYNECOLOGY
The Time to Operate in Tubal Pregnancy.During the past year a prolonged discussion lias been going 011 in the American journals as to which is the best time for operative interference in cases of tubal pregnancy, and more particularly in those in which rupture has taken place with intra-peritoneal haemorrhage. I11
May of last year the American Gynecological Society discussed the subject of "Immediate versus Deferred Operation for Intra-peritoneal Hajmorrhage, due to Tubal Pregnancy," and papers were read by Frederick, Montgomery, Simpson, Manton, Lapthorn Smith, Boldt, Baer, King, Janvrin, and others, while Martin of Greifswald and Pfannenstiel of Kiel took part in the discussion (American Journal oj Obstetrics, July 1908). There were certain points 011 which there was practical unanimity among the speakers. All agreed that in many cases a diagnosis of tubal pregnancy could be made before rupture. Given a history of amenorrhoe t, of irregular crampy pains in the lower abdomen, with or without slight, intermittent vaginal haemorrhage, and the presence of a definite swelling to one side of a slightly enlarged uterus, the diagnosis of tubal pregnancy is practically certain. I11 such a case the treatment ought to be laparotomy and removal of the pregnant tube before symptoms of rupture appear. The operation can be done at the time of choice and under the best conditions as regards preparation of the patient and asepsis.
In the second class of case where there has been a partial rupture of the tube or an attempt at tubal abortion, as evidenced by severe spasms of abdominal pain accompanied by faintness and some pallor indicating internal haemorrhage, all were agreed that, while operation is the only rational treatment, there is no immediate hurry. The patient, provided she is under constant observation, may be allowed to recover from the immediate shock while suitable preparations are made for the operation either in a hospital or in her own home, which latter is preferable in most eases, as any moving of the patient may set up fresh bleeding. In such a case some of the operators delayed operation for weeks, but the majority preferred opening the abdomen within a day or two, for the reason that the operation is easier then, and involves less risk to the patient than when undertaken at a later period after adhesions have formed and possibly fresh haemorrhage has occurred.
The chief divergence of opinion occurred in those cases of complete rupture of the tube with large intra-peritoneal haemorrhages in which the symptoms occur suddenly, reducing the patient to a state of profound collapse and anaemia in a very short time. Ought such a case to be operated on immediately, or ought the operation to be deferred in the hope of the patient recovering from the initial shock and being in better condition later to stand the operation % Which constitutes the greater risk to the patient, the possibility of the haemorrhage continuing, or the addition of the shock of operation to that already present1? Statistics on the subject are contradictory. Frederick said that in his experience only 5 per cent, of tubal pregnancies are accompanied by haemorrhage to the danger of life at the time of primary rupture. Lapthorn Smith, on the other hand, quotes Schauta's statistics from Vienna showing that as the result of operative interference the mortality of tubal pregnancy has been reduced from 87 per cent, to 5 per cent. Some of the speakers quoted cases in which an initial haemorrhage had proved fatal, and most gynecologists have met with such. The advocates of the deferred operation held that the fatal result is due, not to haemorrhage, but to haemorrhage plus shock. Robb especially supported this view, holding that a patient cannot lose sufficient blood into the peritoneal cavity to cause death. The raised tension inside the abdomen and lowered tension in the circulation are sufficient to arrest bleeding before a sufficient quantity?3 per cent, of the body weight, according to physiologists?is lost. He based these conclusions largely on a series of experiments he performed in bitches (Amer. Journ. of Obst., October 1908), in which he cut the uterine and ovarian vessels and closed the abdomen without ligaturing them. The bleeding was spontaneously arrested and the animals recovered. The conditions in his experiments and those obtaining in ruptured tubal pregnancy can, however, hardly be held as parallel. The haemorrhage from a ruptured pregnant tube is to a great extent venous, and it is probable that the coagulability of the blood is interfered with by some action of the chorionic villi, otherwise we cannot explain the enormous quantity of blood which may be lost from even a pin-point aperture in the tube (Wade and Watson, Journ. of Obst. and Gyn., September 1908). The advocates of the deferred operation hold that the surgical principle in haemorrhage?" at once tie the bleeding point"?does not apply in intraperitoneal haemorrhage, clinical experience being against it. Stillwagen (Amer. Journ. of Obst., January 1909) points out that while the percentage of deaths from primary haemorrhage in unoperated cases of tubal haemorrhage is 5 per cent., the mortality from the immediate operation is from 30 per cent, to 40 per cent., and he advocates delay in every case, no matter how collapsed the patient be. If she is so collapsed and suffering from so much shock as not to be able to rally under expectant treatment, operation holds out no prospect of recovery. He quotes several cases in which he has adopted this expectant treatment, including cases where the patient was pulseless and in most profound collapse. In all, the circulation improved, the temperature rose, and he was able to operate under the most favourable circumstances some days, and, in one case, three weeks, after the primary haemorrhage, with perfect success. Baer (Amer. Journ. of Obst., January 1909) expresses the same views.
He refuses to operate when the patient is pulseless and in a profound collapse, having got better results by waiting for reaction and improvement in the circulation. The treatment to be adopted during this period of waiting is absolute rest, intra-venous transfusion of saline, and the administration of morphia gr. ^ (Baer). Most of the writers are sparing in the administration of stimulants such as strychnine, and some also condemn transfusion at this stage.
To sum up, it would appear that those operators who have tried both the immediate and deferred operation believe that they have obtained better results under the latter line of treatment. Those like Lapthorn Smith and Sturmdorf (Medical Record, January 1909), who advocate immediate opening of the abdomen and arrest of haemorrhage followed by the introduction of normal saline into the peritoneal cavity, are ecjually convinced of the soundness of their case.
The practical outcome of the discussion from the point of view of the practitioner is, that in a case of sudden intra-peritoneal haemorrhage there is a fair prospect of the patient rallying from the immediate shock and collapse, and that there is time to get assistance and arrange for the proper performance of the operation, the patient in most cases being in better condition a few hours after rupture than immediately after its occurrence. The ideal thing is to recognise the presence of tubal pregnancy early and to operate before rupture has taken place. |
Computational Modeling and Neuroimaging Techniques for Targeting during Deep Brain Stimulation
Accurate surgical localization of the varied targets for deep brain stimulation (DBS) is a process undergoing constant evolution, with increasingly sophisticated techniques to allow for highly precise targeting. However, despite the fastidious placement of electrodes into specific structures within the brain, there is increasing evidence to suggest that the clinical effects of DBS are likely due to the activation of widespread neuronal networks directly and indirectly influenced by the stimulation of a given target. Selective activation of these complex and inter-connected pathways may further improve the outcomes of currently treated diseases by targeting specific fiber tracts responsible for a particular symptom in a patient-specific manner. Moreover, the delivery of such focused stimulation may aid in the discovery of new targets for electrical stimulation to treat additional neurological, psychiatric, and even cognitive disorders. As such, advancements in surgical targeting, computational modeling, engineering designs, and neuroimaging techniques play a critical role in this process. This article reviews the progress of these applications, discussing the importance of target localization for DBS, and the role of computational modeling and novel neuroimaging in improving our understanding of the pathophysiology of diseases, and thus paving the way for improved selective target localization using DBS.
# Introduction
Deep brain stimulation (DBS) has been used to treat neurological disorders for several decades, but we still have little understanding of the mechanism by which it reverses pathological brain activity. This is in part due to the inherent difficulty in visualizing aberrant neural networks, thus challenging our ability to modulate such pathways via electrical stimulation. However, innovative developments in stereotaxy, computational modeling, engineering, and neuroimaging techniques have made it possible to identify complex connections of circuitry within the brain, improving our understanding of the pathophysiology of various diseases and facilitating our ability to influence these aberrant processes to produce therapeutic results. With continued advancements in engineering and neuroimaging, which enhance our targeting and delivery of stimulation, patients currently treated with DBS may experience improved, patient-specific outcomes, and those with other neurologic and psychiatric disorders may soon become candidates for surgical intervention using DBS.
## Mechanism of action of dbs
DBS involves the surgical implantation of electrodes into deep structures of the brain to modulate brain circuitry in an effort to restore normal physiological function. DBS has been used effectively for the treatment of movement disorders, including Parkinson's disease (PD), Essential tremor (ET), and dystonia, as well as for psychiatric disorders such as obsessive compulsive disorder (OCD). It is currently under investigation for the treatment of additional psychiatric disorders, cognitive dysfunction, epilepsy, and other potentially neurologicallymediated diseases. However, our lack of understanding of the exact mechanism of action of DBS remains a limiting factor in our ability to successfully treat such disease entities.
Given that the clinical effects of DBS produce outcomes comparable to those seen with lesioning techniques, the postulated mechanism of action of DBS was initially thought to be due to widespread neuronal inhibition [bib_ref] Deep brain stimulation, Kern [/bib_ref] [bib_ref] Mechanisms and targets of deep brain stimulation in movement disorders, Johnson [/bib_ref]. Such inhibitory effects were thought to occur from a depolarization blockade or a jamming of neural networks [bib_ref] Normalizing motor-related brain activity: subthalamic nucleus stimulation in Parkinson disease, Grafton [/bib_ref] [bib_ref] Deep brain stimulation, Kern [/bib_ref] [bib_ref] Deep brain stimulation of the subthalamic nucleus in Parkinson's disease: evaluation of..., Hamel [/bib_ref]. Interestingly, despite various studies demonstrating decreased activity of the targeted cells with high-frequency DBS, the output from these cells is not necessarily lessened [bib_ref] Mechanisms and targets of deep brain stimulation in movement disorders, Johnson [/bib_ref]. In fact, studies in which the activity of cells receiving input from a stimulated target were recorded suggest that the stimulated cells may release more neurotransmitters with electrical stimulation [bib_ref] Stimulation of the subthalamic nucleus changes the firing pattern of pallidal neurons, Hashimoto [/bib_ref]. Despite such seemingly contradictory evidence, it is now appreciated that the delivery of stimulation is vastly more complex, resulting in the activation of some elements and the inhibition of others, each of which results in a multitude of downstream effects [bib_ref] Deep brain stimulation for Parkinson's disease: disrupting the disruption, Lozano [/bib_ref] [bib_ref] Translational principles of deep brain stimulation, Kringelbach [/bib_ref] [bib_ref] Therapeutic subthalamic nucleus deep brain stimulation reverses corticothalamic coupling during voluntary movements..., Kahan [/bib_ref].
In addition, DBS may exert its influence via the correction of aberrant neuronal activity. For example, in the setting of DBS for the treatment of PD, the loss of dopamine, which is known to be largely responsible for the pathophysiology of the disease, results in changes in the underlying activity of cells within the basal ganglia [bib_ref] Alterations in neuronal activity in basal ganglia-thalamocortical circuits in the parkinsonian state, Galvan [/bib_ref]. Analyses of the electrical activity of neurons using local field potentials, electroencephalography, and electrocorticography have demonstrated alterations of interspike intervals, specific firing patterns, and oscillatory activities of neuronal cells in PD [bib_ref] Alterations in neuronal activity in basal ganglia-thalamocortical circuits in the parkinsonian state, Galvan [/bib_ref]. It is now believed that DBS acts to disrupt the abnormal activity of these cell populations, for instance, interfering with pathologically synchronous cell groups, thus further contributing to the clinical efficacy of stimulation [bib_ref] Deep brain stimulation for Parkinson's disease: disrupting the disruption, Lozano [/bib_ref] [bib_ref] Therapeutic subthalamic nucleus deep brain stimulation reverses corticothalamic coupling during voluntary movements..., Kahan [/bib_ref].
Finally, our notion of the target of stimulation has also changed over time. While it was once assumed that DBS affected only the regional gray matter structures in which the electrodes were placed, it is currently recognized that the effects of the electrical current are more far-reaching, modulating nearby white matter tracts as well as local and distant cortical and subcortical structures [bib_ref] Computational analysis of subthalamic nucleus and lenticular fasciculus activation during therapeutic deep..., Miocinovic [/bib_ref] [bib_ref] Therapeutic subthalamic nucleus deep brain stimulation reverses corticothalamic coupling during voluntary movements..., Kahan [/bib_ref] [bib_ref] Multitract orthogonal microelectrode localization of the subthalamic nucleus: description of a novel..., Sweet [/bib_ref]. Again looking at the model of DBS for PD, investigations have demonstrated similar motor outcomes with stimulation of either the globus pallidus interna (GPi) or the subthalamic nucleus (STN; [bib_ref] Comparison of pallidal and subthalamic nucleus deep brain stimulation for advanced Parkinson's..., Burchiel [/bib_ref] [bib_ref] Pallidal versus subthalamic deep-brain stimulation for Parkinson's disease, Follett [/bib_ref]. Neuroimaging studies performed on patients following DBS of the STN demonstrate that the therapeutic contact is frequently positioned at the dorsolateral border of the STN and overlaps with white matter tracts from the zona incerta and fields of Forel H2, which carry fibers from the GPi [bib_ref] Computational analysis of subthalamic nucleus and lenticular fasciculus activation during therapeutic deep..., Miocinovic [/bib_ref] [bib_ref] Deep brain stimulation of the subthalamic nucleus in Parkinson's disease: evaluation of..., Hamel [/bib_ref]. Thus, stimulation of adjacent white matter tracts, potentially affecting larger neuronal networks, may in part explain the similar effects of DBS with STN and GPi stimulation.
This concept of the activation of diffuse fiber pathways has been demonstrated in numerous imaging studies, in which widespread alterations in blood flow, glucose metabolism, and blood oxygenation level dependence (BOLD) across the brain as a result of DBS are shown, again suggesting its likely influence on the connectivity of networks [bib_ref] Limbic-cortical dysregulation: a proposed model of depression, Mayberg [/bib_ref] [bib_ref] Normalizing motor-related brain activity: subthalamic nucleus stimulation in Parkinson disease, Grafton [/bib_ref] [bib_ref] A phase I trial of deep brain stimulation of memory circuits in..., Laxton [/bib_ref] [bib_ref] Therapeutic subthalamic nucleus deep brain stimulation reverses corticothalamic coupling during voluntary movements..., Kahan [/bib_ref] [bib_ref] Neurosurgical targets for compulsivity: what can we learn from acquired lesions? Neurosci, Figee [/bib_ref] [bib_ref] Evidence from a rare case study for Hebbian-like changes in structural connectivity..., Van Hartevelt [/bib_ref]. In addition, computational models replicating the effects of DBS on cell body firing patterns and subsequent downstream axonal activation reveal that stimulation of the STN influences the firing of cells within the GPi, thus altering the output of pallidothalamic projection fibers, which in turn affect thalamic processing. Once again, this reinforces the idea that DBS modulates complex neuronal systems [bib_ref] Electric field stimulating influence generated by deep brain stimulation of the subthalamic..., Mcintyre [/bib_ref] [bib_ref] High frequency stimulation of the subthalamic nucleus eliminates pathological thalamic rhythmicity in..., Rubin [/bib_ref] [bib_ref] Computational analysis of subthalamic nucleus and lenticular fasciculus activation during therapeutic deep..., Miocinovic [/bib_ref] [bib_ref] Deep brain stimulation activation volumes and their association with neurophysiological mapping and..., Maks [/bib_ref] [bib_ref] Modeling shifts in the rate and pattern of subthalamopallidal network activity during..., Hahn [/bib_ref].
However, despite our incomplete understanding of the mechanism of action of DBS, there is substantial evidence to show its clinical efficacy for the treatment of movement disorders. In 1998, [bib_ref] Electrical stimulation of the subthalamic nucleus in advanced Parkinson's disease, Limousin [/bib_ref] assessed 24 patients with advanced motor symptoms of PD, who underwent bilateral STN DBS for 1 year. The patients had a 60% improvement in their motor scores of dopaminergic medications, and their daily activities also significantly improved. Moreover, the mean dose of medication was reduced by half and there was no change in their cognitive outcomes. Other studies have demonstrated similar results [bib_ref] Comparison of pallidal and subthalamic nucleus deep brain stimulation for advanced Parkinson's..., Burchiel [/bib_ref] [bib_ref] Safety and efficacy of pallidal or subthalamic nucleus stimulation in advanced PD, Volkmann [/bib_ref] [bib_ref] Surgical treatment for Parkinson's disease, Walter [/bib_ref] , and a long-term study of 42 PD patients with STN DBS assessed for 5 years showed sustained efficacy [bib_ref] Five-year follow-up of bilateral stimulation of the subthalamic nucleus in advanced Parkinson's..., Krack [/bib_ref]. In a randomized controlled trial in 2006, [bib_ref] A randomized trial of deep-brain stimulation for Parkinson's disease, Deuschl [/bib_ref] found a 25% improvement in quality of life and 41% improvement in motor symptoms in patients with PD, which was superior to the best medical management. Similarly, in a randomized open-label trial, DBS of the STN plus medical treatment had better outcomes medical treatment alone for PD [bib_ref] Deep brain stimulation plus best medical therapy versus best medical therapy alone..., Williams [/bib_ref].
These impressive outcomes are not unique to the treatment of PD, as such improvements in motor symptoms have also been shown in patients with ET and dystonia. In another randomized controlled trial, patients with severe tremor due to PD and ET were implanted with electrodes into the ventrointermediate (VIM) nucleus of the thalamus. The authors found almost complete cessation in tremor in 90% of patients [bib_ref] A comparison of continuous thalamic stimulation and thalamotomy for suppression of severe..., Schuurman [/bib_ref]. Moreover, in a follow-up study of the ET patients, up to 7 years after DBS, a significant reduction in tremor score was maintained [bib_ref] Thalamic deep brain stimulation in the treatment of essential tremor: a long-term..., Blomstedt [/bib_ref] [bib_ref] Current perspectives on deep brain stimulation for severe neurological and psychiatric disorders, Kocabicak [/bib_ref]. Similarly, prospective randomized controlled trials assessing GPi DBS for primary dystonias also reveal significant benefits in motor symptoms and quality of life measures [bib_ref] Bilateral deep-brain stimulation of the globus pallidus in primary generalized dystonia, Vidailhet [/bib_ref] [bib_ref] Pallidal deep-brain stimulation in primary generalized or segmental dystonia, Kupsch [/bib_ref] [bib_ref] Deep brain stimulation in DYT1 dystonia: a 10-year experience, Panov [/bib_ref] , with sustained and even improved effects many years following surgical intervention [bib_ref] Bilateral, pallidal, deep-brain stimulation in primary generalised dystonia: a prospective 3 years..., Vidailhet [/bib_ref] [bib_ref] Pallidal deep brain stimulation in patients with primary generalised or segmental dystonia:..., Volkmann [/bib_ref] [bib_ref] Deep brain stimulation in DYT1 dystonia: a 10-year experience, Panov [/bib_ref].
While there are fewer long-term studies investigating the effects of DBS for psychiatric disorders, the results are still promising. Several small studies evaluating DBS of the anterior limb of the internal capsule have been done for the treatment of OCD, based on data from capsulotomy lesions, demonstrating notable efficacy in treatment-refractory patients [bib_ref] Electrical stimulation in anterior limbs of internal capsules in patients with obsessive-compulsive..., Nuttin [/bib_ref] [bib_ref] Long-term electrical capsular stimulation in patients with obsessive-compulsive disorder, Nuttin [/bib_ref] [bib_ref] Three-year outcomes in deep brain stimulation for highly resistant obsessive-compulsive disorder, Greenberg [/bib_ref]. Other targets have also been studied for OCD including the nucleus accumbens and the STN [bib_ref] The nucleus accumbens: a target for deep brain stimulation in obsessive-compulsive-and anxiety-disorders, Sturm [/bib_ref] [bib_ref] Subthalamic nucleus stimulation in severe obsessive-compulsive disorder, Mallet [/bib_ref] [bib_ref] Unilateral deep brain stimulation of the nucleus accumbens in patients with treatment-resistant..., Huff [/bib_ref] [bib_ref] Deep brain stimulation for obsessive-compulsive disorder: systematic review and evidence-based guideline sponsored..., Hamani [/bib_ref] [bib_ref] Current perspectives on deep brain stimulation for severe neurological and psychiatric disorders, Kocabicak [/bib_ref]. Additionally, DBS for the treatment of psychiatric disorders, such as major depression (MD), have also been assessed. Targets for these various trials include the subgenual cingulate cortex, the nucleus accumbens, and the anterior limb of the internal capsule [bib_ref] Deep brain stimulation for treatment-resistant depression, Mayberg [/bib_ref] [bib_ref] Subcallosal cingulate gyrus deep brain stimulation for treatment-resistant depression, Lozano [/bib_ref] [bib_ref] Deep brain stimulation of the ventral capsule/ventral striatum for treatment-resistant depression, Malone [/bib_ref] [bib_ref] Deep brain stimulation for treatment-resistant depression: follow-up after 3 to 6 years, Kennedy [/bib_ref] [bib_ref] Longterm effects of nucleus accumbens deep brain stimulation in treatmentresistant depression: evidence..., Bewernick [/bib_ref] [bib_ref] Subcallosal cingulate deep brain stimulation for treatment-resistant unipolar and bipolar depression, Holtzheimer [/bib_ref]. The fact that each of these studies, targeting different structures within the brain, resulted in comparable outcomes, again suggests that a complex and inter-related network is influenced by electrode stimulation.
However, while the smaller studies yielded encouraging findings, these results could not be replicated in larger, multicenter trials [bib_ref] A randomized sham-controlled trial of deep brain stimulation of the ventral capsule/ventral..., Dougherty [/bib_ref]. This may in part be due to the heterogeneous patient population, high rates of concomitant psychiatric comorbidities among patients, or an overall still poor understanding of the involved neural circuitry [bib_ref] Crossspecies affective functions of the medial forebrain bundle-implications for the treatment of..., Coenen [/bib_ref]. Moreover, given the vulnerability of patients with psychiatric disease and the inconsistent outcomes shown with surgical interventions, ethical considerations must be taken into account prior to proceeding to surgery for such disease states. The potentially devastating consequences of poor clinical judgment is evidenced by historical events that transpired during the psychosurgery era. Led largely by Walter Freeman in the 1940s and 1950s, the rampant and unregulated practice of the frontal lobotomy, adapted from the frontal leucotomy procedure that was initially described in Europe by Egas Moniz, was performed indiscriminately in patients of all ages with diverse symptomatology across the United States. These surgeries resulted in unacceptably high complication rates with minimal evidence to support their efficacy [bib_ref] Surgery of the mind, mood and consciousnes: an idea in evolution, Robinson [/bib_ref]. This underscores the importance of careful patient selection and the need for stringent pre-surgical criteria.
Therefore, although the exact mechanism of DBS remains unclear, our understanding of the complex interplay of systems involved continues to grow. In addition, there is a great deal of evidence demonstrating the benefits of DBS in appropriately selected patients. However, if we are to further optimize observable outcomes by delivering patient-specific stimulation or treating novel neurological and psychiatric diseases in a successful and consistent manner, then utilization of sophisticated and innovative targeting, engineering, and imaging techniques is required. Advancements in computational modeling and neuroimaging that can better demonstrate neuronal activity within the context of widespread connection networks has largely contributed to our progress thus far. Such strategies and technologies are discussed in the duration of this review.
## Surgical targeting
## History of surgical targeting
As noted in the preceding section, the mechanism by which DBS exerts its clinical effects is incompletely understood. However, we do know that both the therapeutic benefits as well as many of the adverse symptoms resulting from DBS depend largely on the location of the electrode contacts within the brain. Thus surgical targeting is an essential factor contributing to the success of a DBS procedure.
Stereotaxis is the process by which deep regions of the brain may be surgically targeted via a minimally invasive approach using a three-dimensional coordinate system for spatial localization in reference to a targeting image [bib_ref] Moving forward: advances in the treatment of movement disorders with deep brain..., Schiefer [/bib_ref]. The role of stereotaxy for the treatment of neurological disorders dates to the 1940s when frame-based localization was used to aid in the precise targeting of intracranial structures [bib_ref] Stereotaxic apparatus for intra-cerebral surgery, Leksell [/bib_ref]. The necessity for such precise and accurate localization is easily understood when considering the size and location of commonly targeted structures. For instance, the STN measures approximately 6 mm by 4 mm by 5 mm in size and is obliquely oriented in all three planes [bib_ref] An assessment of current brain targets for deep brain stimulation surgery with..., Aboshch [/bib_ref] [bib_ref] Fiber tractography of the axonal pathways linking the basal ganglia and cerebellum..., Sweet [/bib_ref]. In addition, the STN has been postulated to be somatotopically organized in a tripartite manner with respect to functional territories, such that the desired location of the DBS electrode is in the dorsolateral aspect of the nucleus, believed to correspond to the motor region of the STN [fig_ref] FIGURE 1 |: Functional organization of subthalamic nucleus [/fig_ref] ; [bib_ref] The connections of the primate subthalamic nucleus: indirect pathways and the open-interconnected..., Joel [/bib_ref] [bib_ref] Confirmation of functional zones within the human subthalamic nucleus: patterns of connectivity..., Lambert [/bib_ref] [bib_ref] The striatum and subthalamic nucleus as independent and collaborative structures in motor..., Tewari [/bib_ref].
Historically, surgical targeting was done via a ventriculogram, to visualize the location of the anterior and posterior commissures which were used as reference points, and a standard stereotactic atlas, allowing for the determination of the desired target relative to its distance from these points [bib_ref] Surgical treatment for Parkinson's disease, Walter [/bib_ref]. The advent of computed tomography (CT) and magnetic resonance imaging (MRI), later replaced the use of the ventriculogram for indirect targeting. Yet, due to the potential error intrinsic to the process of indirect targeting, the practice of intraoperative microelectrode recording was developed. Microelectrode recording uses electrophysiology and somatosensory mapping to further confirm the precise location of the electrode within the brain [bib_ref] Surgical treatment for Parkinson's disease, Walter [/bib_ref]. Further improvements in MRI technology and software targeting platforms has led to the practice of ''direct'' targeting, either alone or in combination with indirect targeting, to facilitate DBS electrode placement. This involves directly visualizing the target, such as the STN, using highresolution MR images, rather than relying solely on the position of the target relative to the anterior and posterior commissures. Additional developments in neuroimaging, computational modeling, novel stimulation parameters, and improved electrode designs, are all drastically changing the way in which surgical targeting is performed. These developments aim to improve patient outcomes and aid in the understanding of both DBS and the pathophysiology of diseases.
## High-resolution mri and software targeting platforms
Improvements in neuroimaging modalities have revolutionized the process of DBS lead implantation. Advancements in MRI technology and the utilization of 3 tesla (T) MRI scanners rather than 1.5T MRI machines have made it possible to visualize deep structures of the brain to allow direct targeting of the structure with less reliance on indirect targeting based on standard landmarks [bib_ref] Direct visualization of deep brainstimulation targets in Parkinson disease with the use..., Cho [/bib_ref]. Such visualization is even more impressive with 7-T MR systems, even allowing for detection of the VIM for ET [bib_ref] An assessment of current brain targets for deep brain stimulation surgery with..., Aboshch [/bib_ref] [bib_ref] Direct visualization of deep brainstimulation targets in Parkinson disease with the use..., Cho [/bib_ref].
In 2007, [bib_ref] Surgical targeting accuracy analysis of six methods for subthalamic nucleus deep brain..., Guo [/bib_ref] looked at six methods of targeting, including a T2-weighted MRI-based approach, a brain atlas, T1 and T2 maps, an electrophysiological database, a collection of final surgical targets from previous patients, and the combination of all of these functional and anatomical data. These were compared against the ''gold standard,'' expert localization with intraoperative electrophysiologic confirmation. The displacement between the planned targets and the actual surgical targets compared to the gold standard ranged from 1.7 to 3 mm, with the technique combining anatomical and functional data being the most accurate [bib_ref] Surgical targeting accuracy analysis of six methods for subthalamic nucleus deep brain..., Guo [/bib_ref]. Additional advances enabling MRI-guided stereotaxy using a clinical workstation has been shown in animal models to be accurate to within 0.3-0.5 mm [bib_ref] MRI-guided stereotaxic targeting in pigs based on a stereotaxic localizer box fitted..., Bjarkam [/bib_ref].
Integration of imaging techniques with computer software platforms can further facilitate identification of brain structures and aid in targeting. Many such systems allow for the fusion of multiple imaging sequences while also superimposing standard brain atlases [bib_ref] Computer-aided placement of deep brain stimulators: from planning to intraoperative guidance, D'haese [/bib_ref] [bib_ref] A cross validation study of deep brain stimulation targeting: from experts to..., Castro [/bib_ref] [bib_ref] An active contour-based atlas registration model applied to automatic subthalamic nucleus targeting..., Duay [/bib_ref]. In addition, fusing different neuroimaging methodologies such as CT and MRI have also been found to ameliorate target localization [bib_ref] Targeting the subthalamic nucleus for deep brain stimulation-a comparative study between magnetic..., Chen [/bib_ref].
Such targeting platforms have also been integrated into both frame-based and frameless stereotaxy systems. Traditionally, the placement of stereotactic head frames followed by a volumetric head CT prior to DBS surgery has been used for determination of stereotactic coordinates within the frame space, and this has been shown to be a highly accurate method of stereotaxy [bib_ref] Safety of anterior commissure-posterior commissure-based target calculation of the subthalamic nucleus in..., Acar [/bib_ref] [bib_ref] Accuracy evaluation of microTargeting Platforms for deep-brain stimulation using virtual targets, Balachandran [/bib_ref]. Yet, frame placement can be uncomfortable for the patient, and requires repeated imaging once the frame is placed prior to surgery, therefore potentially adding to the length of surgery. As such, novel technologies have been developed in an effort to alleviate these concerns, allowing for the accurate targeting of deep structures without a frame. ''Frameless'' skull-mounted stereotactic platforms have been evaluated and found to have comparable accuracy with frame-based systems, although the use of these systems require operating through a narrow working field that can make neurophysiological recordings, if performed, more challenging [bib_ref] Nexframe frameless stereotaxy with multitract microrecording: accuracy evaluated by frame-based stereotactic X-ray, Fukaya [/bib_ref].
## Computational modeling techniques
## Whole-brain activity models
Computational modeling techniques have also contributed both to our understanding of and our progress in DBS. These techniques include a number of complex engineering technologies that allow us to better understand the pathophysiology of diseases treated with DBS, and the resultant effects of stimulation on the system as a whole. In particular, much work has been done looking at whole-brain network activity resulting from electrical stimulation. As previously mentioned, it is unclear as to how high-frequency DBS exerts its effects on surrounding structures. Thus, mathematical algorithms and computational models have been devised to help answer this question.
In 2004, [bib_ref] High frequency stimulation of the subthalamic nucleus eliminates pathological thalamic rhythmicity in..., Rubin [/bib_ref] created a computational network model to demonstrate how DBS of the STN results in the disruption of downstream pathological thalamic rhythms in PD. They did this by first creating mathematical equations and algorithms representative of the individual cell properties of each neuronal population with respect to the channels and molecular transporters comprising the cell membranes, the membrane capacitance, current flux across membranes, and additional factors in order to accurately model each cell type and its surrounding environment. They could then predict the various cell firing patterns under different circumstances depending on the variables changed. In this way, the authors evaluated the neuronal firing patterns of the STN, GP externa, GPi, and thalamus during simulation of three states: (1) normal conditions, in which GPi output to the thalamus is irregular with minimal effects on thalamic cells; (2) parkinsonian conditions, during which GPi cells, with output to the thalamus, fire in bursts at a tremor frequency, resulting in disruption of the processing capabilities of the thalamus; and (3) conditions of STN DBS in PD, causing subthalamopallidal projections that result in high-frequency tonic firing of the GPi, as opposed to the low-frequency bursting, which ultimately allows the thalamus to resume normal activity. From this model, the authors determined that STN DBS resulted in normalization of aberrant basal ganglia oscillations and thus abnormal basal ganglia output to the thalamus, ultimately restoring physiologic thalamic relay capabilities [bib_ref] Modeling shifts in the rate and pattern of subthalamopallidal network activity during..., Hahn [/bib_ref].
Additional models have since been developed to further elucidate the role that STN DBS has on GPi bursting and subsequent thalamic processing. In 2010, [bib_ref] Modeling shifts in the rate and pattern of subthalamopallidal network activity during..., Hahn [/bib_ref] looked at a model similar to that of Rubin and Terman, however, they attempted to parameterize a subthalamopallidal network model incorporating presumed cortical and striatal inputs thought to influence abnormal beta rhythms within the basal ganglia. They did this by training the model to fit in vivo microelectrode recordings from the basal ganglia of parkinsonian non-human primates [bib_ref] Modeling shifts in the rate and pattern of subthalamopallidal network activity during..., Hahn [/bib_ref]. As a result, the authors sought to more accurately predict the dynamic, multi-step cortico-striatal-thalamic network in a DBS model for PD. This complex analysis demonstrated that normalization of aberrant GPi bursting depends on the volume of STN tissue activated and may require a threshold reduction of bursting for therapeutic effects. It also reinforced the principle that extensive signaling pathways and networks, rather than isolated cell-to-cell interactions, play a large role in the etiology of disease states, as evidenced by the cortical and striatal inputs to the basal ganglia, which were reflected by the model. Therefore, using computational modeling systems, our understanding of the pathophysiology of diseases can be improved, which may in turn, allow us to more effectively treat diseases.
One further example of how computational modeling can be used to represent and predict multi-faceted systems within the brain is seen in a study by [bib_ref] Network effects of subthalamic deep brain stimulation drive a unique mixture of..., Humphries [/bib_ref]. They also utilized intricate models to assess the effects of STN DBS on the basal ganglia in PD [bib_ref] Network effects of subthalamic deep brain stimulation drive a unique mixture of..., Humphries [/bib_ref]. The authors hypothesized that STN DBS resulted in both increased and decreased firing rates within the basal ganglia output nuclei, contributing to the therapeutic effects of treatment. In their model, they replicated the diversification of responses from a primate STN DBS study, and showed that the mixed basal ganglia output responses underwent a step-wise change with therapeutic DBS frequencies (>100 Hz), which did not occur in optogenetic models of direct STN stimulation. They concluded that the efficacy of DBS was due to mixed effects of stimulation resulting in a combination of both stimulatory and inhibitory output from the basal ganglia in response to electrical STN stimulation unique to DBS at therapeutic thresholds. Once again, this highlights the complexity of processes concomitantly occurring as a result of DBS, leading to widespread network effects.
## Volume of tissue activated
Another computational modeling technique that has played a critical role in our understanding of the effects of DBS is the concept of visualizing the extent of tissue that is influenced by DBS electrode stimulation, or the volume of tissue activated ; [bib_ref] Multitract orthogonal microelectrode localization of the subthalamic nucleus: description of a novel..., Sweet [/bib_ref]. In 2006, [bib_ref] Computational analysis of subthalamic nucleus and lenticular fasciculus activation during therapeutic deep..., Miocinovic [/bib_ref] developed a computational model in parkinsonian macaques of STN DBS. This model predicted the pattern of axonal activation during electrode stimulation FIGURE 2 | Volume of Tissue Activated as determined from an anatomical model based on imaging data from subjects who have undergone deep brain stimulation (DBS) coupled with finite element modeling (FEM) to calculate the spread of electrical current through the tissue from the active electrode contact. (A) The active contact of each electrode was determined from postoperative programming sessions, and was defined as the cathodal electrode for either monopolar or bipolar stimulation. The active contact is shown on a postoperative computed tomography (CT) scan (left) and on image fusion on magnetic resonance imaging (MRI; right). (B) The volume of tissue activated was calculated from each patient's specific stimulation parameter settings by using a custom Python-based program from Python libraries. The volume of tissue activated is shown in dark pink alone (left) and in conjunction with the contored nuclei (right). The relevant nuclei were contored and include the thalami (yellow), the red nuclei (red), and the STN (green); the active contact and the volume of tissue activated were also contored (pink). Adapted from [bib_ref] Fiber tractography of the axonal pathways linking the basal ganglia and cerebellum..., Sweet [/bib_ref].
of the STN, and demonstrated that while both STN and GPi fibers were activated, only activation of the STN neurons correlated with clinically therapeutic stimulation [bib_ref] Computational analysis of subthalamic nucleus and lenticular fasciculus activation during therapeutic deep..., Miocinovic [/bib_ref].
The method by which such computational modeling systems determine the volume of tissue activated involves finite element modeling (FEM). FEM combines an anatomical model based on imaging data from subjects who have undergone DBS with an electrical model to calculate the electrical field from the DBS electrode. The FEM data can then be used to calculate the voltage spread of the electrical stimulation. This depends on the location of the electrode and the composition of the surrounding tissue, which is in turn evaluated by the electrode capacitance and impedance, as well as by the stimulation parameters used with programming of the electrode. Ultimately, the voltage distribution within the tissue is established from the FEM and the total amplitude of current is divided between the contacts of the electrode [bib_ref] Computational analysis of subthalamic nucleus and lenticular fasciculus activation during therapeutic deep..., Miocinovic [/bib_ref]. Axonal activation can then be predicted by again using an anatomical model to distribute axons in a grid around the DBS electrode and calculating the voltage spread produced by each active contact to find the threshold for generating an action potential within each surrounding axon. The volume of tissue activated is then predicted from the relationship between the electrical field and the evoked action potentials in the axon models [bib_ref] Current Steering to control the volume of tissue activated during deep brain..., Butson [/bib_ref].
Such modeling algorithms are not only able to accurately predict the volume of tissue activated, but can also determine its effect on clinical outcome. [bib_ref] Deep brain stimulation activation volumes and their association with neurophysiological mapping and..., Maks [/bib_ref] evaluated 10 patients with PD who underwent unilateral STN DBS using conventional preoperative imaging and intraoperative neurophysiology. The results were then superimposed to a brain atlas and fused with postoperative imaging of the electrodes and the volume of tissue activated calculated with a diffusion tensor stimulation model, followed by assessment of clinical outcomes. The authors found that the therapeutic benefits of stimulation were greatest with the active contacts near the dorsal border of the STN. Similarly, in 2011, [bib_ref] Patient-specific analysis of the relationship between the volume of tissue activated during..., Mikos [/bib_ref] used computational modeling to analyze of the association between the volume of tissue activated and verbal fluency in PD patients who underwent STN DBS, finding a subtle decline in verbal fluency with stimulation of non-motor regions of the STN. Thus the volume of tissue activated can be assessed relative to the surrounding anatomy in the brain to better determine which structures are actively being stimulated in an effort to improve patient outcomes.
## Limitations of computational modeling
It should be noted that there are limitations to computational modeling techniques. With respect to the modeling of network activity, numerous simplifications and assumptions are made and are therefore limitations of the model. For instance, the input and output connections to and from the basal ganglia likely ignore many fiber tracts in order to focus on the pathways under investigation. As a result, networks that are potentially involved in sensorimotor signals to the thalamus may also be neglected [bib_ref] High frequency stimulation of the subthalamic nucleus eliminates pathological thalamic rhythmicity in..., Rubin [/bib_ref]. In addition, field effects and variations due to electrode positioning are also ignored, and each cell is treated as a single compartment, again simplifying the model substantially [bib_ref] High frequency stimulation of the subthalamic nucleus eliminates pathological thalamic rhythmicity in..., Rubin [/bib_ref] [bib_ref] Modeling shifts in the rate and pattern of subthalamopallidal network activity during..., Hahn [/bib_ref]. Many of these models also look at an open-loop system, studying the effects of a unidirectional pathway or network. However, in actuality, there are likely closed-loop, feedback interactions that are not represented with such models [bib_ref] Modeling shifts in the rate and pattern of subthalamopallidal network activity during..., Hahn [/bib_ref].
Similarly, calculations for the determination of the volume of tissue activated are also based on several assumptions and simplifications regarding the properties of the activated neurons. Preliminary studies to calculate the volume of tissue activated were initially based on the properties of rat STN neurons. Attempts were then made to approximate in vivo conditions of Parkinson macaques. In addition, activated neuronal populations were presumed to be uniform in their membrane dynamics, myelination, morphology, and projection pattern. Yet we know such homogeneity does not exist in vivo. As a result, the true thresholds for producing action potentials and thus the overall volume of tissue activated could vary somewhat compared to the predicted model. Finally, studies investigating the volume of tissue activated looked at the white matter tracts surrounding the STN that were felt to be the most likely contributing tracts to the therapeutic effects of STN DBS, while many other nearby tracts were ignored. Also ignored were the potential effects of antridromic stimulation of afferent fibers projecting to the STN and the effects of adjacent glial cells that could influence the extracellular environment and thus the effects of current spread and axonal activation [bib_ref] Computational analysis of subthalamic nucleus and lenticular fasciculus activation during therapeutic deep..., Miocinovic [/bib_ref].
## Additional engineering developments
## Novel stimulation parameters
Many of the computational models predicting network activity with DBS have looked at the effects of high-frequency stimulation. In fact, as demonstrated by [bib_ref] Network effects of subthalamic deep brain stimulation drive a unique mixture of..., Humphries [/bib_ref] , the therapeutic effects of DBS are thought to occur with continuous stimulation at frequencies greater than 100 Hz. However, altering the parameters of the stimulation delivered could have significant results. Computational analyses using predictive models of stimulation have demonstrated that the volume of tissue activated is determined by both the stimulation parameters and the physical properties of the electrodes [bib_ref] Role of electrode design on the volume of tissue activated during deep..., Butson [/bib_ref]. This has substantial implications regarding the importance of modulating stimulation settings to influence patient outcomes. Studies exploring innovative stimulation parameters such as high-frequency or patterned stimulation may have additional benefits compared to conventional stimulation. Moreover, such novel patterns of stimulation may be more effective in disrupting abnormal beta band oscillations within the cortico-basal ganglia network that are thought to be involved in the symptomatology of PD [bib_ref] Dynamics of tremor-related oscillations in the human globus pallidus: a single case..., Hurtado [/bib_ref] [bib_ref] From Parkinsonian thalamic activity to restoring thalamic relay using deep brain stimulation:..., Meijer [/bib_ref] [bib_ref] Simulation of cortico-basal ganglia oscillations and their suppression by closed loop deep..., Grant [/bib_ref] [bib_ref] Interaction of oscillations and their suppression via deep brain stimulation, in a..., Kang [/bib_ref] [bib_ref] Computational modeling of deep brain stimulation, Mcintyre [/bib_ref].
## Closed-loop dbs and current steering
The delivery of stimulation can also be affected by the design of implanted DBS electrodes. New types of electrodes with recording capabilities have been studied such that changes in the environment surrounding the electrode can be assessed. Such changes include the recording of local field potentials, which could be used to correlate beta band oscillations with activity and posture in PD patients [bib_ref] Beta oscillations in freely moving Parkinson's subjects are attenuated during deep brain..., Quinn [/bib_ref] or even the characteristic chemical composition of the environment [bib_ref] A diamond-based electrode for detection of neurochemicals in the human brain, Bennet [/bib_ref]. Recording electrodes such as these would then transmit the information to a pulse generator capable of analyzing the data, which would in turn deliver stimulation in response to the recorded information [bib_ref] Beta oscillations in freely moving Parkinson's subjects are attenuated during deep brain..., Quinn [/bib_ref] [bib_ref] A diamond-based electrode for detection of neurochemicals in the human brain, Bennet [/bib_ref]. This closed-loop responsive form of stimulation, may ultimately be used to further impact patient outcomes.
The volume of tissue activated can also be manipulated with changes in electrode design. For instance, as discussed by [bib_ref] Current Steering to control the volume of tissue activated during deep brain..., Butson [/bib_ref] , shaping the volume of tissue activated by selectively activating specific fiber tracts via ''current-steering'' may better alleviate symptoms experienced by patients due to their disease. For example, in situations in which traditional STN DBS produces both therapeutic effects as well as concurrent side effects due to the undesired activation of adjacent fiber tracts like the internal capsule, current steering could be applied. This would enable directing the spread of current to avoid the activation of certain nearby structures, while activating the intended target. Currently utilized DBS electrodes have four cylindrical contacts, yet new electrode designs with more contacts of different shapes, capable of delivering stimulation in specific and variable configurations, are being developed [fig_ref] FIGURE 3 |: Left [/fig_ref] ; [bib_ref] A novel lead design enables selective deep brain stimulation of neural populations..., Van Dijk [/bib_ref]. These designs take advantage of the innumerable possible electrode orientations to maximize the efficacy of stimulation with surrounding anatomical structures [fig_ref] FIGURE 3 |: Left [/fig_ref] ; [bib_ref] Towards computer-assisted deep brain stimulation targeting with multiple active contacts, Beriault [/bib_ref] [bib_ref] A novel lead design enables selective deep brain stimulation of neural populations..., Van Dijk [/bib_ref]. This notion of current-steering has given rise to multiple investigations [bib_ref] Directional steering: a novel approach to deep brain stimulation, Contarino [/bib_ref] [bib_ref] Multiple-source current steering in subthalamic nucleus deep brain stimulation for Parkinson's disease..., Timmerman [/bib_ref] [bib_ref] Computational field shaping for deep brain stimulation with thousands of contacts in..., Willsie [/bib_ref] and combining these engineering designs with computational models, such as those looking at the volume of tissue activated, could also prove to be useful in surgical planning and subsequent electrode programming. However, additional studies with long-term follow-up will be necessary to better determine the efficacy of such technologies.
## Advanced neuroimaging technologies
## Diffusion-weighted imaging and tractography
The development of sophisticated imaging techniques has also enhanced our understanding of the complex circuitry of neural networks in the brain and the role of DBS in influencing such networks. As previously discussed, many of the effects of DBS are likely due to the involvement of white matter tracts comprising interconnected functional systems traveling between various intracranial structures. Neuroimaging modalities measuring metabolism and blood flow have demonstrated diffuse patterns of seemingly distant regions of the brain being simultaneously activated with DBS, thus supporting this notion of stimulation affecting widespread connectivity [bib_ref] Limbic-cortical dysregulation: a proposed model of depression, Mayberg [/bib_ref] [bib_ref] Normalizing motor-related brain activity: subthalamic nucleus stimulation in Parkinson disease, Grafton [/bib_ref] [bib_ref] A phase I trial of deep brain stimulation of memory circuits in..., Laxton [/bib_ref] [bib_ref] Therapeutic subthalamic nucleus deep brain stimulation reverses corticothalamic coupling during voluntary movements..., Kahan [/bib_ref] [bib_ref] Neurosurgical targets for compulsivity: what can we learn from acquired lesions? Neurosci, Figee [/bib_ref]. However, these studies do not directly show the actual white matter pathways thought to be involved.
Visualization of such fibers is now in part feasible with the use of diffusion-weighted imaging (DWI) MRI sequencing and tractography. DWI is a mode of MRI that illustrates the spontaneous diffusion of water within the brain. Since water diffuses easily along cell barriers rather than across them, one can assume that the diffusion of water throughout the brain represents the anatomical course of neuronal axons. Thus, by assessing water diffusivity in living tissue, the approximate location of axonal pathways can be estimated, and the likelihood of their connections to various structures can be established [bib_ref] Connectomicsurgery'': diffusion tensorimaging (DTI) tractography as a targeting modality for surgical modulation..., Henderson [/bib_ref] [bib_ref] Comparing a diffusion tensor and non-tensor approach to white matter fiber tractography..., Auriat [/bib_ref]. From DWI, a diffusion tensor can be established within a given voxel on MRI. A diffusion tensor can take a perfectly spherical shape when there is no water diffusion, such as the cerebrospinal fluid within a ventricle [fig_ref] FIGURE 4 |: Diffusion of water along cellular barriers in brain rather than across them [/fig_ref] ;. However, the tensor can also become cylindrically-shaped when the water is diffusing in a given direction, such as along an axon as part of a group of white matter fibers [fig_ref] FIGURE 4 |: Diffusion of water along cellular barriers in brain rather than across them [/fig_ref] ;. The more directions that are added to the DWI sequencing, the more specific the directionality of the diffusion tensor of a given voxel becomes. From the diffusion tensors, which are derived from DWI sequences, comes diffusion tensor imaging (DTI), which provides a voxel-to-voxel based estimation of the three-dimensional orientation of an axonal fiber [bib_ref] Comparing a diffusion tensor and non-tensor approach to white matter fiber tractography..., Auriat [/bib_ref].
However, DTI has numerous limitations. Firstly, each diffusion tensor is only an estimation of the approximate directionality of water within a given voxel. As such, each tensor is only representative of a small portion of a single axonal fiber, and thus is not a representation of the entire axon or a group of axons comprising a fiber tract. Therefore, DTI may not even depict the actual anatomical course of a neuronal axon. As a result, any interpretation of the course of an axon or group of axons from DTI is relatively subjective [bib_ref] Comparing a diffusion tensor and non-tensor approach to white matter fiber tractography..., Auriat [/bib_ref]. Furthermore, DTI is limited by its technical acquisition, including poor resolution of images, poor signalto-noise ratio, involuntary movements from subjects, and other related variables [bib_ref] Connectomicsurgery'': diffusion tensorimaging (DTI) tractography as a targeting modality for surgical modulation..., Henderson [/bib_ref].
A natural extension of DTI that attempts to correct for its limitations is fiber tractography. While DTI approximates the main nerve fiber direction within a given voxel, tractography attempts to assess all the voxels between two or more regions of interest in an effort to discern the entire course of axons comprising a white matter fiber tract. In order to achieve this, tractography uses the DWI data and a series of algorithms to produce a proposed fiber tract. By defining starting and ending seeds, or regions of interest, post-imaging analysis can be performed using various computer software platforms to determine the predominant fiber pathways coursing between the assigned seeds. However, it should be noted that tractography results are highly dependent upon the algorithms used, and larger white matter tracts are easier to visualize than smaller, unknown tracts [bib_ref] Diffusion tensor imaging and neuromodulation: DTI as key technology for deep brain..., Coenen [/bib_ref]. In addition, sharply decussating tracts may be more difficult to identify than tracts without sudden or significant changes in direction [bib_ref] Diffusion tensor imaging and neuromodulation: DTI as key technology for deep brain..., Coenen [/bib_ref].
Thus tractography provides an anatomical three-dimensional model of white matter fiber tracts. There are two types of tractography, deterministic and probabilistic. Deterministic tractography is the more straightforward of the two and involves a streamline algorithm based on thresholds for collinearity of primary diffusion direction among adjacent voxels to establish a putative axonal pathway. These algorithms, such as FACT (fiber assignment by continuous tracking) create continuous tubular structures that are representative of white matter tracts between two or more regions of interest [fig_ref] FIGURE 5 |: Regions of interest [/fig_ref] ; [bib_ref] Multitract orthogonal microelectrode localization of the subthalamic nucleus: description of a novel..., Sweet [/bib_ref]. However, this is an all-or-none calculation, and not all of the tracts identified using this technique will in fact exist, which again represents a substantial limitation. There is also, as noted above, particular ambiguity in areas with less anisotropy such as decussating fibers [bib_ref] Diffusion tensor imaging and neuromodulation: DTI as key technology for deep brain..., Coenen [/bib_ref]. Thus, deterministic tractography works well for highly anisotropic tensors, but may be quite inaccurate for areas with less anisotropy. While any potential error in estimation may be evident for well-established white matter tracts, such as the corticospinal tract, the degree of inaccuracy is indeterminable for tracts that are not well established. Therefore, adding to the above limitations of deterministic tractograpy, is the major disadvantage that the error or confidence in the estimation of a tracts is unknown [bib_ref] Modulation of the cerebello-thalamo-cortical network in thalamic deep brain stimulation for tremor:..., Coenen [/bib_ref].
In contrast, probabilistic tractography attempts to overcome the limitations of deterministic tractography by explicitly characterizing the confidence with which connections may be FIGURE 6 | Deterministic fiber tractography, demonstrating the dentatothalamaic tract (DTT; blue) in patients with tremor-predominant Parkinson's disease (PD). The relevant nuclei were contored and include the thalami (yellow), the red nuclei (red), and the dentate nuclei (peach). Data from from [bib_ref] Fiber tractography of the axonal pathways linking the basal ganglia and cerebellum..., Sweet [/bib_ref]. established through the diffusion data itself [bib_ref] Diffusion tensor imaging and neuromodulation: DTI as key technology for deep brain..., Coenen [/bib_ref]. Thus, probabilistic tractography uses advanced mathematical algorithms to produce an output of connection probability values to determine the statistical probability of a fiber tract running through a given set of voxels or between two anatomic structures based on the diffusion model [bib_ref] Diffusion tensor imaging and neuromodulation: DTI as key technology for deep brain..., Coenen [/bib_ref]. This often requires the use of computer software programs such as FSL, which calculate and produce a color scale to indicate the probability of a given streamline connecting a start point to other voxels within the brain [bib_ref] Modulation of the cerebello-thalamo-cortical network in thalamic deep brain stimulation for tremor:..., Coenen [/bib_ref]. The same predictability could only be done with deterministic tractography by performing the fiber tracking on multiple patients and combining the results to determine a probability [bib_ref] Modulation of the cerebello-thalamo-cortical network in thalamic deep brain stimulation for tremor:..., Coenen [/bib_ref]. Once again, the primary limitation of tractography is its intrinsic nature of estimating the course and location of a fiber tract from the diffusion imaging data. However, whereas deterministic tractography cannot assess the degree of estimation error, probabilistic tractography is able to more precisely predict the likelihood of accuracy of a given fiber tract.
From a DBS standpoint, tractography makes it possible not only to visualize white matter pathways in the brain, but also to postulate how modulation of these pathways might lead to improved treatment outcomes. For example, both of the common DBS targets for PD (STN and GPi) are gray matter structures, and stimulation of either target results in notable improvement in motor symptoms [bib_ref] Surgical treatment for Parkinson's disease, Walter [/bib_ref] [bib_ref] Pallidal versus subthalamic deep-brain stimulation for Parkinson's disease, Follett [/bib_ref]. Tractography has made it possible to view the axonal pathways surrounding each structure that may also be activated, such as subthalamopallidal and pallidothalamic pathways, as well as other connections to regions of the cortex, thalamus, and brainstem [bib_ref] Tractography patterns of subthalamic nucleus deep brain stimulation, Vanegas-Arroyave [/bib_ref] and this activation may be responsible to some degree for the therapeutic effects of DBS. Similarly, studies assessing white matter tracts associated with the VIM thalamus, such as the dentantothalamic tract, indicate changes in tremor patients compared to controls that may account for some of the efficacy of DBS of this target [bib_ref] Diffusion tensor imaging of white matter involvement in essential tremor, Klein [/bib_ref]. By identifying specific white matter pathways like the dentothalamic tract, which is thought to play a role in the pathogenesis of tremor, it is possible to determine whether there is any benefit to tract-specific targeting using DBS [bib_ref] Dentatorubrothalamic tract in human brain: diffusion tensor tractography study, Kwon [/bib_ref] [bib_ref] The tremor network targeted by successful VIM deep brain stimulation in humans, Klein [/bib_ref] [bib_ref] Modulation of the cerebello-thalamo-cortical network in thalamic deep brain stimulation for tremor:..., Coenen [/bib_ref] [bib_ref] Multitract orthogonal microelectrode localization of the subthalamic nucleus: description of a novel..., Sweet [/bib_ref] [bib_ref] Deep brain stimulation for essential tremor: targeting the dentato-rubro-thalamic tract?, Schlaier [/bib_ref]. In the same way, a better understanding of the involvement of white matter tracts in the pathophysiology of disease using tractography techniques can help direct future therapies. Using tractography to examine fiber networks in the brain of patients with psychiatric or cognitive disorders may allow identification of new targets in the brain for the treatment of various disorders. For example, [bib_ref] A tractography analysis of two deep brain stimulation white matter targets for..., Gutman [/bib_ref] used DTI and probabilistic tractography to investigate two DBS white matter targets for the treatment of depression, the subcallosal cingulum bundle and the anterior limb of the internal capsule, and concluded that these targets are involved in distinct neural networks that overlap, explaining the benefit of DBS of each target in patients with depression. Likewise, [bib_ref] Novel fingerprinting method characterizes the necessary and sufficient structural connectivity from deep..., Fernandes [/bib_ref] used advanced tractography combined with whole-brain anatomical parcellation to establish connectivity ''fingerprints'' of successful and unsuccessful DBS targets for chronic pain, a technique which offers promise for discovery of new neuromodulation targets for other neuropsychological disorders. [bib_ref] Tractography study of deep brain stimulation of the anterior cingulate cortex in..., Boccard [/bib_ref] used probabilistic tractography to demonstrate that connectivity with medial forebrain bundle and precuneus were associated with good and bad outcome, respectively, among patients treated with DBS of the anterior cingulate for chronic pain. The targeting of temporal fiber pathways for disorders of memory and cognition is also actively under investigation [bib_ref] Memory enhancement induced by hypothalamic/fornix deep brain stimulation, Hamani [/bib_ref] [bib_ref] A phase I trial of deep brain stimulation of memory circuits in..., Laxton [/bib_ref] [bib_ref] Memory enhancement and deep-brain stimulation of the entorhinal area, Suthana [/bib_ref] [bib_ref] Visual-spatial memory may be enhanced with theta burst deep brain stimulation of..., Miller [/bib_ref]. Incorporating tractography imaging techniques to better visualize neural pathways may prove to be invaluable in such targeting endeavors.
## Resting-state functional connectivity mri and connectonomics
It is also worth discussing additional imaging technologies that have been developed to further assess connectivity, such as resting-state functional-connectivity MRI (rs-fcMRI). This technique allows for the visualization of neural networks within the brain by assessing fluctuations in blood oxygenation [bib_ref] Resting-state networks link invasive and noninvasive brain stimulation across diverse psychiatric and..., Fox [/bib_ref]. Spontaneous alterations in the BOLD has been correlated with known functional networks within the brain, and may demonstrate continuous activity between various intracranial regions, thus representing connectivity [bib_ref] Atlas-based analysis of resting-state functional connectivity: evaluation for reproducibility and multi-modal anatomy-function..., Faria [/bib_ref]. Such networks have been shown to be active in the absence of sensorimotor or cognitive tasks in fMRI studies, and deactivate during such goal-directed tasks [bib_ref] Default mode network abnormalities in bipolar disorder and schizophrenia, Ongur [/bib_ref]. As a result, these pathways have been termed the default-mode network (DMN) and are implicated in ongoing, default functions of the brain [bib_ref] Default mode network abnormalities in bipolar disorder and schizophrenia, Ongur [/bib_ref]. Thus, rs-fcMRI allows us to assess the DMN in patients as a surrogate for connectivity. Moreover, rs-fcMRI also enables visualization of brain structures more easily than tractography, which is derived from DWI MRI sequences rather than structural MRI sequencing. This also allows for structural connectivity-based parcellation to be incorporated into the imaging analysis. The parcellated functional connectivity approach is a natural extension of rs-fcMRI due to the spatial dimensionality visible in rs-fcMRI. This entails identifying brain structures and clustering voxels around such structures [bib_ref] Atlas-based analysis of resting-state functional connectivity: evaluation for reproducibility and multi-modal anatomy-function..., Faria [/bib_ref]. This results in a more clearly depicted relationship between functional connectivity and structural anatomy. Such analysis can be taken a step further to include multi-modal analysis, combining DWI-based connectivity with rs-fcMRI connectivity using a parcel-to-parcel approach [bib_ref] Atlas-based analysis of resting-state functional connectivity: evaluation for reproducibility and multi-modal anatomy-function..., Faria [/bib_ref].
Data from rs-fcMRI pertaining to the DMN can be used to help draw conclusions regarding the wiring diagram of the brain, or the connectome. By modeling the DMN derived from BOLD signals on rs-fcMRI, lesions or disruptions of connections can be replicated to determine the resultant impact such changes would produce on the DMN [bib_ref] Graph theory analysis of complex brain networks: new concepts in brain mapping..., Hart [/bib_ref]. Thus by weakening specific projection fibers, as one might see in injuries induced by traumatic brain injury, one can predict the resultant effects this wound have on the connectome as a whole [bib_ref] Graph theory analysis of complex brain networks: new concepts in brain mapping..., Hart [/bib_ref]. Similarly, by strengthening certain connections, we can speculate as to the impact that DBS might have on the connectome or DMN.
## Limitations of connectivity analyses
However, like any novel modality, limitations in both tractography and rs-fcMRI do exist. With respect to tractography, as previously discussed, the visualization of fiber tracts is based on numerous assumptions regarding the diffusion properties of water within the brain. Firstly, the premise of tractography relies on the presumption that the spontaneous flow of water in the brain only occurs along the course of an axon. Moreover, the tensors are themselves estimations of the mean vector of water diffusivity within a given voxel. Lastly, the mathematical algorithms that generate proposed fiber tracts are at best educated estimations of the presence of a given tract.
In addition, limitations of rs-fcMRI should also be noted. While there is a demonstrable correlation between rs-fcMRI and white matter connections, rs-fcMRI is still a surrogate for connectivity based on a suspected DMN. In addition, the exact causal relationship between two or more regions with increased BOLD can only be postulated, and polysynaptic interactions may play a larger role then suspected. This could thus be a misrepresentation of a particular connection or relationship in the brain, which could be misleading if used for target guidance in DBS [bib_ref] Resting-state networks link invasive and noninvasive brain stimulation across diverse psychiatric and..., Fox [/bib_ref].
## Integration of novel technologies and future applications in target selection
Ultimately, the integration of computational modeling techniques and novel imaging technologies can perhaps be used to validate hypotheses regarding mechanisms of disease states and implicated networks, and thus aid in the discovery of new surgical targets. For example, in a study by [bib_ref] Multitract orthogonal microelectrode localization of the subthalamic nucleus: description of a novel..., Sweet [/bib_ref] , patients with advanced PD who underwent STN DBS were assessed using deterministic tractography from preoperative imaging to identify fiber pathways extending between the cerebellum and the basal ganglia believed to be associated with the pathophysiology of the tremor component of PD. The findings from the tractography were then combined with postoperative computational modeling to determine the volume of tissue activated based on the therapeutic electrode contacts. The study showed that in patients with tremor-predominant PD, the volume of tissue activated tended to involve the dentatothalamic tract ; [bib_ref] Multitract orthogonal microelectrode localization of the subthalamic nucleus: description of a novel..., Sweet [/bib_ref] , supporting the idea that this tract contributes to tremor in PD. The findings also suggest that specific targeting of the dentatothalamic tract may better alleviate symptoms in tremor-predominant patients, and may thus represent a form of patient-specific DBS targeting.
Other studies have similarly integrated computational modeling, tractography, and patient outcomes. Specifically, the dentatothalamic tract has been assessed in PD and ET patients by several groups, using both deterministic and probabilistic tractography [bib_ref] The tremor network targeted by successful VIM deep brain stimulation in humans, Klein [/bib_ref] [bib_ref] Modulation of the cerebello-thalamo-cortical network in thalamic deep brain stimulation for tremor:..., Coenen [/bib_ref] [bib_ref] Multitract orthogonal microelectrode localization of the subthalamic nucleus: description of a novel..., Sweet [/bib_ref] [bib_ref] Deep brain stimulation for essential tremor: targeting the dentato-rubro-thalamic tract?, Schlaier [/bib_ref] , and combined with computational models to further validate this concept of patients-specific targeting to treat tremor symptomatology.
Outcome data from patient-specific targeting techniques can also be combined with computational models to construct a probabilistic stimulation atlas that incorporates advanced computational modeling with clinical outcomes to determine the ideal target based on a patient's particular symptoms [bib_ref] Probabilistic analysis of activation volumes generated during deep brain stimulation, Butson [/bib_ref]. Though further studies are needed, tractography and predictive modeling modalities may be able to be used to rather directly target specific pathways in a patient-specific manner .
Finally, the combination of computational models of electric field patterns with tractography could help us to better predict which specific white matter tracts are being activated for the treatment of psychiatric, cognitive, and neurologically-mediated disorders. [bib_ref] Tractography-activation models applied to subcallosal cingulate deep brain stimulation, Lujan [/bib_ref] , combined probabilistic tractography and electrical field modeling to help predict connectivity in patients with depression. The authors looked at white matter tracts within the subcallosal cingulum bundle who underwent DBS to look at which electrodes would be the most useful in treatment of the depression based on the pattern of current delivery and activation of the associated fibers using computational modeling. Similarly, [bib_ref] Diffusion tensor imaging and neuromodulation: DTI as key technology for deep brain..., Coenen [/bib_ref] demonstrated the involvement of the medial forebrain bundle in depression using tractography, and utilized computational modeling to show how a new DBS target stimulating the medial forebrain bundle would potentially activate more white matter tracts associated with the pathophysiology of the disease than currently investigated targets. As more disease entities are studied for potential treatment using DBS, such as other psychiatric disorders, memory and cognitive dysfunction, and epilepsy [bib_ref] Electrical stimulation of the anterior nucleus of thalamus for treatment of refractory..., Fisher [/bib_ref] [bib_ref] Deep brain stimulation for disorders of memory and cognition, Sankar [/bib_ref] [bib_ref] Visual-spatial memory may be enhanced with theta burst deep brain stimulation of..., Miller [/bib_ref] , the FIGURE 7 | Integration of computer modeling techniques and tractography to demonstrate that the volume of tissue activated of the active STN electrode contact includes the DTT in patients with tremor-predominant PD. The relevant nuclei were contored and include the globus pallidus and putamen (blue), the thalami (yellow), the red nuclei (red), the STN (green), and the dentate nuclei (peach). Data from [bib_ref] Fiber tractography of the axonal pathways linking the basal ganglia and cerebellum..., Sweet [/bib_ref]. applications for these techniques will undoubtedly grow as well.
# Conclusion
Improvements in surgical targeting, computational modeling, engineering, and neuroimaging techniques have greatly enhanced our understanding of the pathophysiology of various diseases, allowing for the effective treatment of such conditions using DBS. In the future, it is likely that these novel technologies will have further applications in DBS surgery, including patientspecific targeting and the discovery of new targets for the treatment of neurological, psychiatric, and even cognitive disorders.
# Author contributions
JPM, JAS, JP, and FG wrote and edited the article.
[fig] FIGURE 1 |: Functional organization of subthalamic nucleus (STN) with cortical inputs into specific regions of the STN. Adapted fromTewari et al. (2016). [/fig]
[fig] FIGURE 3 |: Left: Conventional cylindrical four contact lead (CC) compared to high density lead design (HD). Right: An example of diffuse fiber activation with conventional DBS (A,B) vs. selective activation of fiber tracts with current steering (C,D). Reprinted with permissions from Van Dijk et al. (2015). [/fig]
[fig] FIGURE 4 |: Diffusion of water along cellular barriers in brain rather than across them (A), and examples of tensors with isotropic diffusion in CSF (represented by sphere) and anisotropic diffusion in callosal fibers (B). Adapted fromKlein (2013). [/fig]
[fig] FIGURE 5 |: Regions of interest (red boxes) shown on MRI in axial planes to include the right subthalamic nucleus (STN) and Red Nucleus and the contralateral cerebellar hemisphere for investigation of white matter tracts traversing between the two regions of interest. Adapted fromSweet et al. (2014a). [/fig]
|
Acute liver lobe torsion in a kitten
# Introduction
An acute abdomen is a common presentation seen at veterinary emergency hospitals caused by various etiologies. An acute abdomen secondary to a liver lobe torsion (LLT) is rarely reported in dogs, cats, horses and humans, but is relatively frequent in the rabbit. The exact cause of LLT is still unknown. There are several potential causes of LLT in dogs and cats, such as secondary to trauma, neoplasia, gastric dilatation and volvulus in dogs, abnormal hepatic ligaments or a diaphragmatic hernia.The purpose of this case report is to document the first peer-reviewed report of an acute LLT in a kitten and its successful management.
## Case description
A 10-week-old intact male domestic shorthair kitten was referred to an emergency animal hospital for an abdominal exploratory surgery for a suspected foreign body. The kitten initially presented to the referring veterinarian for an acute onset of lethargy, vomiting and anorexia of <12 h duration. Initial diagnostics performed by the referring veterinarian included a complete blood count and serum biochemical panel (CBC/CHEM; IDEXX Laboratories), feline leukemia virus and feline immunodeficiency virus point-of-care ELISA test (FeLV/FIV; IDEXX Laboratories) and abdominal radiographs. The CBC/CHEM abnormalities were a marked nonregenerative anemia with a hematocrit of 9.8% (reference interval 30.3-52.3%) and reticulocyte count of 34.5 K/µl (RI 3.0-50.0 K/µl), toxic neutrophils with suspected bands, mild lymphocytosis of 7.23 K/µl (RI 0.92-6.88 K/ul), moderate thrombocytopenia of 54 K/µl (RI 151-600 K/µl), mild hyperglycemia of 186 mg/dl (RI 77-153 mg/dl), increased blood urea nitrogen (BUN) of 57 mg/dl (RI 16-33 mg/dl), decreased creatinine of 0.3 mg/dl (RI 0.6-1.6 mg/dl), decreased total protein of 4.9 g/dl (RI 5.2-8.2 g/dl) and decreased globulin of 2.6 g/dl (RI 2.8-4.8 g/dl). The kitten's liver values were normal (alanine aminotransferase, aspartate aminotransferase, alkaline phosphatase, gamma-glutamyl transferase and total bilirubin). FeLV/FIV tests were both negative. The primary veterinarian's interpretation of the abdominal radiographs was abdominal distension with decreased serosal detail and radio-opaque grit-like material in the stomach. Given the severity of the kitten's condition, the kitten was referred to an emergency facility for further care.
On physical examination at the hospital, the kitten was laterally recumbent with a temperature of 94.1ºF (34.5°C), a heart rate of 190 beats/min, a respiratory rate of 40 breaths/min with increased respiratory effort, pale white mucous membranes, non-palpable femoral pulses, abdominal pain, and palpable cranial abdominal organomegaly or mass effect. An indirect systolic blood pressure of 138 mmHg was obtained via Doppler (Biomed Technologies). A point-of-care abdominal ultrasound (SonoScape: Universal Medical Systems) showed a mild amount of ascites in the cranial abdomen. An abdominocentesis on presentation was unsuccessful, but a sample was successfully obtained an hour later. There was only sufficient amount of fluid obtained to perform a microscopic evaluation. In-house cytological evaluation of the ascites revealed only red blood cells without evidence of microorganisms or white blood cells.
The Resources]) was given to address the kitten's anemia. Based on the clinical presentation and concern for an abdominal organ torsion, an abdominal exploratory surgery was pursued.
Within 4 h of presentation, the kitten was taken to surgery. An additional dose of hydromorphone (0.05 mg/kg IV [West-Ward]) was used as premedication. General anesthesia was induced with propofol (3.5 mg/kg IV given to effect [Zoetis]). A size 3 mm cuffed endotracheal tube was used for intubation, and the kitten was maintained on 2% sevoflurane (Zoetis) in oxygen at 1.5 l/min. The kitten was aseptically prepared for surgery. A ventral midline incision was made from the xiphoid process extending caudally to the pubis, approximately 7 cm in length. A moderate amount of hemorrhagic abdominal effusion was present. The entire caudate liver lobe was twisted on its axis and appeared congested, enlarged and necrotic. A hemoclip (Teleflex Medical) liver lobectomy of the caudate liver lobe was performed. The left lateral liver lobe appeared slightly congested, and a biopsy was obtained. The caudate liver lobe and left lateral liver lobe biopsy were submitted for histopathology review (Antech Diagnostics). The remainder of the abdomen was unremarkable, and no cause for this torsion was identified. The abdomen was lavaged copiously with warmed saline. The body wall was closed with a continuous pattern using 3-0 nylon suture (Nylon; J&J Healthcare Systems). The subcutaneous tissue was closed with a continuous pattern using 4-0 polydioxanone suture (PDS; J&J Healthcare Systems). The skin was closed using surgical staples.
Owing to the amount of blood loss during surgery, a PCV/total solids (TS) was performed 2 h postoperatively. The PCV and TS were 14% and 4.6 g/dl, respectively. The kitten was also exhibiting clinical signs of anemia (pale-pink mucous membranes, slight increased to increased respiratory effort, snappy femoral pulses, bradycardia and hypotension). Therefore, a second packed red blood cell transfusion (12.5 ml/kg IV over 4 h [Animal Blood Resources]) was administered. One dose of diphenhydramine hydrochloride (2.2 mg/kg IM [West-Ward]) was given towards the end of the transfusion owing to concern over transfusion reaction (tachycardia). The following day the kitten was cardiovascularly stable and eating well. The kitten was discharged 3 days postoperatively on buprenorphine (0.015 mg/kg PO [Roadrunner Pharmacy MFG]), amoxicillin trihydrate/ clavulanate potassium (Clavamox 13.9 mg/kg [Zoetis]), and a probiotic (Purina Pro Plan Veterinary Fortiflora; Nestlé Purina PetCare). Antibiotic therapy was continued postoperatively owing to concerns about a secondary infection from hypoxic injury given the changes viewed in the left lateral liver lobe.
At follow-up examination, 15 days postoperatively, the kitten was reported to be doing well. The recheck PCV/TS was 22%/6.4 g/dl. A fecal specimen for intestinal parasite screening was submitted, and results were positive for Toxocara cati. A 5-day course of fenbendazole was prescribed (50 mg/kg/day PO [Panacur; Intervet, Merck Animal Health]).
The histopathology findings (Antech Diagnostics) of the caudate liver lobe revealed mild sinusoidal and vascular dilation and congestion, mild diffuse coagulative necrosis with a variable infiltrate of degenerated neutrophils throughout the parenchyma, subcapsular hemorrhage and a multifocal capsular fibrinohemorrhagic exudate. The overall findings were diffuse, subacute, hepatic congestion with necrosis, hemorrhage and hemorrhagic exudate. The left lateral liver lobe biopsy revealed mostly diffuse panlobular coagulative necrosis with scattered degenerate neutrophils and lesser macrophages, dilated bile ductules with inspissated luminal bile pigment, a small amount of portal scattered inflammatory and hematolymphoid cells, and dilated portal lymphatic vessels and portal hemorrhage. There was no evidence of infectious organisms or neoplasia. The overall findings were marked subacute panlobular hepatic necrosis with bile duct ectasia and cholestasis. No culture or special stains were performed on the samples submitted.
# Discussion
To our knowledge, there are six published feline case reports of LLT (11-month-old, 5-, 6-, 10-, 13-and 15.7year-old cats).This is the first peer-reviewed report of LLT in a kitten. In the six previous reports, five were diagnosed at the time of surgery, while the remaining cat was diagnosed post mortem.The suspected causes for LLT in these cats were a pseudocyst, neoplasia, diaphragmatic hernia secondary to a motor vehicle accident, pectus excavatum and unknown etiology. 1-3,5,6 A report of an 11-month-old kitten diagnosed with torsion of the right medial and quadrate liver lobes along with the gallbladder was presented at the ECVS Annual Scientific Meeting in 2019.A pseudocyst and hemorrhagic infarction associated with a mucocele was discovered on histopathologic examination of the liver segments and gallbladder in this kitten.In the present case, the cause for LLT is unknown. There was no history of trauma, neoplasia, diaphragmatic hernia or congenital abnormalities predisposing to LLT. This kitten was later diagnosed with T cati, which is known to migrate from the gastrointestinal tract to the liver after ingestion of infected eggs.A parasitic infiltration of the liver resulting in alterations in liver architecture has been a proposed cause for torsion.However, no microorganisms were observed on histopathology.
It has been postulated that the ligaments supporting the liver lobes have a role in LLT.Increased laxity, aplasia, stretching or rupture of these ligaments can result in increased mobility of the liver lobe and possible torsion.There were no abnormalities reported at the time of surgery in these ligaments. For most cases, the cause for LLT is not identified, as in the case reported here.The clinical signs and physical examination findings in this case were of acute onset and non-specific for LLT, but were consistent with previous reports. This kitten exhibited lethargy, vomiting and anorexia. Previous cases in dogs and cats have reported abdominal pain, lethargy, vomiting, anorexia, collapse and diarrhea, all of which are non-specific for LLT.The most significant physical examination finding in this case suggestive of LLT was a palpable abdominal mass effect and abdominal pain. In previous cases in dogs and cats, abdominal pain, abdominal distension and a palpable abdominal mass were reported.This kitten exhibited signs of hypovolemic shock and severe decompensation on presentation. It is suggested that the severity of presentation depends on the degree of torsion and occlusion of hepatic blood flow, leading to hepatic congestion, arterial and venous thrombosis, and subsequent necrosis.A complete LLT resulting in complete occlusion or avulsion of hepatic blood flow results in a more severe and acute presentation. 2,7 A partial or intermittent LLT resulting in partial or intermittent complete occlusion results in a less severe and chronic presentation.This kitten had a complete LLT diagnosed at the time of surgery with subsequent hemorrhage and hypovolemic shock, resulting in a severe, acute presentation. Although a PCV/TS on the ascites was not performed, the cytological and gross examination of the ascites, histopathologic findings of the liver and low peripheral PCV/TS were suggestive of a hemoabdomen. The ascites may have also developed from increased hydrostatic pressure secondary to torsion or thrombosis of the hepatic vessels.In this case report, LLT was diagnosed at the time of surgery. The abdominal radiographs performed prior to surgery showed non-specific changes of abdominal distension with decreased serosal detail. Further diagnostics, such as an abdominal ultrasound with Doppler flow or CT scan, were not available at the time of presentation. Abdominal radiographs may reveal decreased serosal detail due to peritoneal effusion, mass effect in the cranial abdomen or displacement of the abdominal organs.Abdominal ultrasound may reveal changes in echogenicity, decreased hepatic blood flow with color-flow Doppler, peritoneal effusion, liver mass and/or segmental hepatomegaly.In a previous case series study in dogs, LLT was diagnosed on Doppler abdominal ultrasound findings in five cases. 7 CT may reveal enlargement or abnormal positioning of the liver lobe or a mass adjacent to the liver with low attenuation and reduced to absent contrast enhancement.There has also been a report of a 'whirl sign' on CT findings of a dog with LLT and secondary disseminated intravascular coagulation and multiorgan infarction.A 'whirl sign' is a CT finding described as the rotation of vessels around the point of torsion.Changes on imaging are usually non-specific, which creates a challenge for definitively diagnosing LLT.In this case, hematological and biochemical changes were also non-specific for LLT. Laboratory abnormalities reported in the veterinary literature include anemia, leukocytosis, elevated liver enzymes, thrombocytopenia, azotemia, hypoalbuminemia, hypoproteinemia, hyperlactatemia and elevated amylase and lipase. The significant laboratory findings in this kitten were a non-regenerative anemia, thrombocytopenia, leukocytosis, increased BUN, and decreased total protein and globulins. The non-regenerative anemia, thrombocytopenia and decrease in globulins were most likely due to acute hemorrhage. Owing to the acute nature of hemorrhage, the bone marrow had not yet generated a regenerative response.The leukocytosis was most likely a result of the kitten's inflammatory response to LLT. The elevated BUN was most likely pre-renal secondary to hypovolemia and dehydration. The liver values in this patient were normal at the time of presentation. It can take up to 12 h for alanine aminotransferase activity to peak after injury.Blood was obtained from this patient within 12 h of presentation, so it is possible that this may have changed if another chemistry panel was checked at a later time.
LLT has been reported in all liver lobes in the veterinary literature, with the left lateral lobe most commonly affected. It has been suggested that the left liver lobe is more mobile owing to its large size and location relative to the other liver lobes.In reported cases in dogs and cats from 2001 to 2020 (38 LLTs in 32 dogs and cats [in two cases, affected liver lobes were not recorded]), the left lateral liver lobe was most commonly affected (12/32 cases), with two of these cases also having the left medial lobe affected.The left medial liver lobe (7/32 cases: two of these cases also had the left lateral lobe affected), caudate liver lobe (3/32 cases), papillary process of caudate lobe (3/32 cases), right lateral lobe (1/32 cases), right medial liver lobe (5/32 cases, with four of these cases also having the quadrate liver lobe affected) and quadrate liver lobe (5/32 cases, with four of these cases also having the right medial liver lobe) have also been reported.In the cat, the right medial liver lobe was most frequently reported (two right medial and quadrate liver lobes, one right medial liver lobe, two left lateral liver lobes, one papillary process of the caudate lobe).In this study, the entire caudate liver lobe was affected, which, to our knowledge, has not previously been reported in the cat.
For an increased chance of survival, the treatment choice for both chronic and acute LLT for dogs and cats is prompt surgical intervention.In this case report the entire affected liver lobe was resected without untwisting the lobe back to its normal position. Untwisting the affected liver lobe can lead to ischemiareperfusion injury, thromboembolism and endotoxin release, which can lead to disseminated intravascular coagulation and septic shock.Therefore, it is not recommended to untwist the liver lobe prior to resection.
The histopathologic findings seen in this kitten are consistent with the histopathologic findings reported in previous cases of LLT.The histopathology examination of the caudate liver lobe showed necrosis, congestion, inflammation and hemorrhage. The histopathology examination of the left lateral lobe that was biopsied also showed significant changes of necrosis, inflammation, hemorrhage, and bile duct dilation and stasis. Although the reasons for these changes are unknown, we speculate that they may have been secondary to hypoxic injury from decreased perfusion secondary to shock and hypovolemia, or could have been evidence of intermittent LLT.
# Conclusions
This is the first peer-reviewed report of LLT in a kitten. The kitten described in this case report presented in acute hypovolemic shock with clinical signs consistent with acute LLT. An exploratory abdominal surgery was used to diagnose and successfully treat this kitten. It is up to the clinician to recommend surgery when diagnostic imaging is not available. LLT can carry a good prognosis if diagnosed and treated in a timely fashion.
## Conflict of interest
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
# Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
Ethical approval This work involved the use of nonexperimental animals only (including owned or unowned animals and data from prospective or retrospective studies). Established internationally recognized high standards ('best practice') of individual veterinary clinical patient care were followed. Ethical approval from a committee was therefore not specifically required for publication in JFMS Open Reports.
Informed consent Informed consent (either verbal or written) was obtained from the owner or legal custodian of all animal(s) described in this work (either experimental or nonexperimental animals) for the procedure(s) undertaken (either prospective or retrospective studies). No animals or humans are identifiable within this publication, and therefore additional informed consent for publication was not required. |
NF‐κB inhibition reverses acidic bile‐induced miR‐21, miR‐155, miR‐192, miR‐34a, miR‐375 and miR‐451a deregulations in human hypopharyngeal cells
We previously demonstrated that acidic bile activates NF-jB, deregulating the expression of oncogenic miRNA markers, in pre-malignant murine laryngopharyngeal mucosa.Here, we hypothesize that the in vitro exposure of human hypopharyngeal cells to acidic bile deregulates cancer-related miRNA markers that can be reversed by BAY 11-7082, a pharmacologic NF-jB inhibitor. We repetitively exposed normal human hypopharyngeal primary cells and human hypopharyngeal keratinocytes to bile fluid (400 lmol/L), at pH 4.0 and 7.0, with/without BAY 11-7082 (20 lmol/L). We centred our study on the transcriptional activation of oncogenic miR-21, miR-155, miR-192, miR-34a, miR-375, miR-451a and NF-jB-related genes, previously linked to acidic bile-induced pre-neoplastic events. Our novel findings in vitro are consistent with our hypothesis demonstrating that BAY 11-7082 significantly reverses the acidic bileinduced oncogenic miRNA phenotype, in normal hypopharyngeal cells. BAY 11-7082 strongly inhibits the acidic bile-induced up-regulation of miR-192 and down-regulation of miR-451a and significantly decreases the miR-21/375 ratios, previously related to poor prognosis in hypopharyngeal cancer. This is the first in vitro report that NF-jB inhibition reverses acidic bile-induced miR-21, miR-155, miR-192, miR-34a, miR-375 and miR-451a deregulations in normal human hypopharyngeal cells, suggesting that acidic bile-induced events are directly or indirectly dependent on NF-jB signalling.K E Y W O R D SBAY 11-7082, bile acids, hypopharyngeal cancer,
between acidic bile or GDF and early pre-neoplastic events in laryngopharyngeal mucosa. [bib_ref] In vitro model for gastroduodenal reflux-induced nuclear factor-kappaB activation and its role..., Sasaki [/bib_ref] [bib_ref] Gastro-duodenal fluid induced Nuclear Factor-jappaB activation and early pre-malignant alterations in murine..., Vageli [/bib_ref] [bib_ref] miR-21, miR-155, miR-192, and miR-375 deregulations related to NF-kappaB activation in gastroduodenal..., Sasaki [/bib_ref] Specifically, the combination of bile and acid (pH ≤4.0) constitutively activates NF-jB, up-regulating the expression of cancer-related genes and deregulating the expression of oncogenic miRNA markers, such as "oncomirs" miR-21, miR-192, miR-155 and "tumour suppressors" miR-34a, miR-375 and miR-451a, in pre-malignant lesions of murine laryngopharyngeal mucosa. [bib_ref] Gastro-duodenal fluid induced Nuclear Factor-jappaB activation and early pre-malignant alterations in murine..., Vageli [/bib_ref] [bib_ref] miR-21, miR-155, miR-192, and miR-375 deregulations related to NF-kappaB activation in gastroduodenal..., Sasaki [/bib_ref] Here, we describe an in vitro model exploring repetitive exposures of normal human hypopharyngeal cells to acidic bile, with and without BAY 11-7082, a pharmacologic inhibitor of NF-jB. [bib_ref] Novel inhibitors of cytokine-induced IkappaBalpha phosphorylation and endothelial cell adhesion molecule expression..., Pierce [/bib_ref] We hypothesize that NF-jB inhibitor is capable of preventing the acidic bile-induced up-regulation of "oncomirs" miR-21, miR-155 and miR-192 and down-regulation of "tumour suppressor" miR-34a, miR-375 and miR-451a, previously associated with laryngopharyngeal cancer, [bib_ref] Correlation between mir-21 expression and laryngeal carcinoma risks, Zhou [/bib_ref] [bib_ref] miR-21 and miR-375 microRNAs as candidate diagnostic biomarkers in squamous cell carcinoma..., Hu [/bib_ref] [bib_ref] MiR-21/miR-375 ratio is an independent prognostic factor in patients with laryngeal squamous..., Hu [/bib_ref] [bib_ref] MicroRNAs regulate epithelialmesenchymal transition of supraglottic laryngeal cancer, Tai [/bib_ref] [bib_ref] Overexpression of miR-155 promotes proliferation and invasion of human laryngeal squamous cell..., Zhao [/bib_ref] [bib_ref] MicroRNA-34a affects the occurrence of laryngeal squamous cell carcinoma by targeting the..., Shen [/bib_ref] providing insight into interactions of transcriptionally active NF-jB with cancer-related miRNA markers.
MicroRNA (miRNA) molecules have been considered to play an important role in both inflammation and cancer, [bib_ref] MicroRNAs play a central role in molecular dysfunctions linking inflammation with cancer, Tili [/bib_ref] modulating the expression of genes by causing target mRNA degradation or inhibiting their translation. [bib_ref] Mechanisms of posttranscriptional regulation by microRNAs: are the answers in sight?, Filipowicz [/bib_ref] Specifically, some miRNAs, such as "oncomirs"
and "tumour suppressor" miRNAs, show altered expression levels in tumour cells compared to normal cells (up-regulated or down-regulated) and are capable of contributing to carcinogenesis, demonstrating a significant regulating role in the multistep process of cancer initiation and progression. [bib_ref] OncomiR or tumor suppressor? the duplicity of MicroRNAs in cancer, Svoronos [/bib_ref] Previous studies have demonstrated that deregulation of "oncomirs" miR-21, miR-155, miR-192 and tumour suppressor miR-375, miR-451a and miR-34a is associated with laryngopharyngeal cancer. [bib_ref] Correlation between mir-21 expression and laryngeal carcinoma risks, Zhou [/bib_ref] [bib_ref] miR-21 and miR-375 microRNAs as candidate diagnostic biomarkers in squamous cell carcinoma..., Hu [/bib_ref] [bib_ref] MiR-21/miR-375 ratio is an independent prognostic factor in patients with laryngeal squamous..., Hu [/bib_ref] [bib_ref] MicroRNAs regulate epithelialmesenchymal transition of supraglottic laryngeal cancer, Tai [/bib_ref] [bib_ref] Overexpression of miR-155 promotes proliferation and invasion of human laryngeal squamous cell..., Zhao [/bib_ref] [bib_ref] MicroRNA-34a affects the occurrence of laryngeal squamous cell carcinoma by targeting the..., Shen [/bib_ref] Moreover, an independent association has been demonstrated between NF-jB activation and up-regulation of oncogenic miR-21 and/or down-regulation of tumour suppressor miR-34a and miR-451a. [bib_ref] Transcriptional activation of microRNA-34a by NF-kappa B in human esophageal cancer cells, Li [/bib_ref] [bib_ref] miR-451 inhibits cell proliferation in human hepatocellular carcinoma through direct suppression of..., Li [/bib_ref] [bib_ref] Unraveling regulatory programs for NF-kappaB, p53 and microRNAs in head and neck..., Yan [/bib_ref] BAY 11-7082 was selected as a reliable inhibitor of NF-jB pathway that has been widely used in many studies exploring the effect of NF-jB. [bib_ref] Novel inhibitors of cytokine-induced IkappaBalpha phosphorylation and endothelial cell adhesion molecule expression..., Pierce [/bib_ref] [bib_ref] BAY 11-7082, a nuclear factor-jB inhibitor, induces apoptosis and S phase arrest..., Chen [/bib_ref] [bib_ref] Nuclear factor-kappa B inhibition can enhance apoptosis of differentiated thyroid cancer cells..., Meng [/bib_ref] It has been suggested that BAY 11-7082 offers the most rapid and potent antitumour effect among other NF-jB inhibitors [bib_ref] BAY 11-7082, a nuclear factor-jB inhibitor, induces apoptosis and S phase arrest..., Chen [/bib_ref] and can possibly be used as a sensitizer of anticancer therapy, [bib_ref] Nuclear factor-kappaB inhibitors as sensitizers to anticancer drugs, Nakanishi [/bib_ref] [bib_ref] Targeting NF-jB signaling pathway suppresses tumor growth, angiogenesis, and metastasis of human..., Li [/bib_ref] increasing the intrinsic susceptibility of cancer cells to chemotherapeutic agents. [bib_ref] Potentiation of paclitaxel cytotoxicity in lung and esophageal cancer cells by pharmacologic..., Nguyen [/bib_ref] Evidence that inhibition of acidic bile-induced NF-jB activation effectively reverses the altered cancer-related miRNA phenotype will encourage the in vivo application of NF-jB inhibitors, as possible preventers of acidic bile effect in hypopharyngeal mucosa. telomerase-immortalized human hypopharyngeal keratinocytes (HHK) (4th passage) [bib_ref] In vitro model for gastroduodenal reflux-induced nuclear factor-kappaB activation and its role..., Sasaki [/bib_ref] to bile (pH 4.0 and pH 7.0) and corresponding controls (pH 4.0 and pH 7.0), for 10-15 minutes, 3 times per day, for 5 days, as previously described. [bib_ref] In vitro model for gastroduodenal reflux-induced nuclear factor-kappaB activation and its role..., Sasaki [/bib_ref] Bile fluid consisted of a mixture of conjugated bile salts (400 lmol/L) considered to be "physiologic" [bib_ref] Composition and concentration of bile acid reflux into the esophagus of patients..., Kauer [/bib_ref] [bib_ref] Systematic review: the role of bile acids in the pathogenesis of gastro-oesophageal..., Mcquaid [/bib_ref] (Supplementary Methods).
Experimental groups included HHPC and HHK repetitively exposed to (a) acidic bile at pH 4.0, the cut-off of reflux disease [bib_ref] Pharyngeal pH monitoring in patients with posterior laryngitis, Ulualp [/bib_ref] [bib_ref] Gastroesophageal reflux in patients with subglottic stenosis, Walner [/bib_ref] and (b) neutral bile, containing the same bile salts mixture, at pH 7.0.
Control groups included (a) acid control (pH 4.0), and (b) neutral control (pH 7.0) with identical media used in experimental groups (Supplementary Methods).
## Acidic bile + bay-11-7082 treatment
In parallel with acidic bile treatment, we performed an additional procedure of repetitive exposure of HHPC (2nd passage) and HHK At the end of treatment, media were removed and cells or cell extracts were analysed.
## | luciferase assay
We performed a luciferase assay in order to monitor the transcriptional activity of the NF-jB in HHPC exposed to acidic bile and corresponding controls, with or without the pharmacologic inhibitor of NF-jB, BAY 11-7082. We used Firefly | 2923 as ratios of relative NF-jB activity (with/without NF-jB inhibitor).
(Data were obtained from three independent experiments).
# | mirna analysis
We performed miRNA analysis in order to determine the expression levels of miR-21, -155, -192, -34a, -375 and -451a, previously characterized in laryngopharyngeal cancer, [bib_ref] Correlation between mir-21 expression and laryngeal carcinoma risks, Zhou [/bib_ref] [bib_ref] miR-21 and miR-375 microRNAs as candidate diagnostic biomarkers in squamous cell carcinoma..., Hu [/bib_ref] [bib_ref] MiR-21/miR-375 ratio is an independent prognostic factor in patients with laryngeal squamous..., Hu [/bib_ref] [bib_ref] MicroRNAs regulate epithelialmesenchymal transition of supraglottic laryngeal cancer, Tai [/bib_ref] [bib_ref] Overexpression of miR-155 promotes proliferation and invasion of human laryngeal squamous cell..., Zhao [/bib_ref] [bib_ref] MicroRNA-34a affects the occurrence of laryngeal squamous cell carcinoma by targeting the..., Shen [/bib_ref] in normal human hypopharyngeal cells, HHPC and HHK, exposed to acidic bile (pH 4.0), neutral bile (pH 7.0), acid (pH 4.0) and neutral control (pH 7.0) fluids, with or without pharmacologic inhibitor BAY 11-7082. [bib_ref] Novel inhibitors of cytokine-induced IkappaBalpha phosphorylation and endothelial cell adhesion molecule expression..., Pierce [/bib_ref] We estimated relative expression levels (target miRNA/RNU6B) for each specific miRNA marker, in each experimental and control group treated with or without NF-jB inhibitor (CFX96TM software; Bio-Rad, . (Data were obtained from three independent experiments).
[formula] CA, USA) (Supplementary Methods & [/formula]
We used the same pool of total RNA to determine, by qPCR, the effect of BAY 11-7082 on transcriptional levels of RELA(p65), TNFa, IL-1b, IL-6 and STAT3 in acidic bile-treated and control HHPC with or without BAY 11-7082, as previously described [bib_ref] Inhibition of NF-jB prevents the acidic bile-induced oncogenic mRNA phenotype, in human..., Vageli [/bib_ref]. These genes were selected because they demonstrated an increased transcriptional activity under acidic bile exposure of HHPC [bib_ref] In vitro model for gastroduodenal reflux-induced nuclear factor-kappaB activation and its role..., Sasaki [/bib_ref] [bib_ref] Gastro-duodenal fluid induced Nuclear Factor-jappaB activation and early pre-malignant alterations in murine..., Vageli [/bib_ref] that was prevented by BAY 11-7082 in our previous study. [bib_ref] Inhibition of NF-jB prevents the acidic bile-induced oncogenic mRNA phenotype, in human..., Vageli [/bib_ref]
[formula] (Supple- mentary Methods & [/formula]
## | immunofluorescence assay
We performed an immunofluorescence (IF) assay in HHPC, as previously described [bib_ref] In vitro model for gastroduodenal reflux-induced nuclear factor-kappaB activation and its role..., Sasaki [/bib_ref] [bib_ref] Gastro-duodenal fluid induced Nuclear Factor-jappaB activation and early pre-malignant alterations in murine..., Vageli [/bib_ref] [bib_ref] miR-21, miR-155, miR-192, and miR-375 deregulations related to NF-kappaB activation in gastroduodenal..., Sasaki [/bib_ref] (Supplementary Methods) to explore the effect of NF-jB inhibitor (20 lmol/L of BAY 11-7082) on the acidic bileinduced nuclear translocation of phospho-NF-jB (p65, Ser536) and phospho-STAT3 (Tyr705), previously shown to be up-regulated in 45-day acidic bile-treated murine laryngopharyngeal mucosa. [bib_ref] miR-21, miR-155, miR-192, and miR-375 deregulations related to NF-kappaB activation in gastroduodenal..., Sasaki [/bib_ref]
## | cell viability assay
We performed a cell viability assay, using CellTiter-Glo â Luminescent Cell Viability Assay (Promega) to monitor the effect of NF-jB inhibitor, BAY 11-7082, on viability of HHPC and HHK treated with bile at pH 4.0 and pH 7.0, and corresponding controls, as described in Supplementary Methods. We determined cell viability by comparing the mean values of cells exposed to NF-jB inhibitor against the mean value of cells that were not exposed to inhibitor, for each experimental and control group. Statistically significant difference in cell viability was determined using paired test and P value <.05 (Graph Pad Prism 6.0).
# | statistical analysis
We performed statistical analysis, using GraphPad Prism 6 software and one-way ANOVA (by Friedman and Dunn's multiple analysis test; P-values <.05) to compare expression changes of the analysed miRNA markers and the analysed genes induced by BAY 11-7082 (with/without NF-kB inhibitor) between different experimental and control groups. We also used t test analysis (multiple comparisons by Holm-Sidak) to reveal differential expression (P-values) for each analysed miRNA marker in treated cells, with and without NF-jB inhibitor.
Finally, we performed a Pearson correlation to estimate the correlation coefficient between BAY 11-7082-induced miRNA and mRNA expression levels, as well as between "oncomirs" and "tumour suppressor" miRNA levels, of different treated groups (P-values < .05). NF-jB in acidic bile-treated HHPC and HHK. We observed that cells exposed to neutral bile, acid or neutral control with NF-jB inhibitor also demonstrated a reduced transcriptional activity of NF-jB, compared to those treated without BAY 11-7082 . However, the acidic bile-treated group demonstrated the most reduced ratios of relative NF-jB transcriptional activity (NF-jB luciferase responsive element/control luciferase reporter) with/without BAY 11-7082, relative to neutral bile, acid or neutral control groups .
## | bay 11-7082 reverses the acidic bile-induced deregulation of cancer-related mirnas in normal human hypopharyngeal cells
We performed miRNA analysis, by qPCR, in normal human hypopharyngeal cells, HHPC and HHK, exposed to acidic bile (pH 4.0), neutral bile (pH 7.0) and corresponding controls, with and without NF-jB inhibitor (BAY 11-7082). We analysed specific miRNA markers, previously characterized as "oncomirs", such as miR-21, miR-155 and miR-192, or "tumour suppressors", such as miR-34a, miR-375 and miR-451a. , compared to controls (one-way ANOVA; by Friedman). In contrast, HHPC and HHK exposed to acidic bile with NF-jB inhibitor demonstrated lower levels of "oncomirs" and significantly higher levels of "tumour suppressor" miRNAs, compared to those cells exposed to acidic bile without NF-jB inhibitor (P < .05) (one-way ANOVA; by Friedman).
Human hypopharyngeal primary cells exposed to acidic bile with BAY 11-7082 demonstrated a significant decrease in "oncomirs" miR-21 (P = .006), miR-155 (P = .0035) and particularly of miR-192 levels (P < .00001) , as well as a significant increase in "tumour suppressor" miR-34a (P = .00023), miR-375 (P = .00134) and particularly of miR-451a levels (P < .000001) , compared to HHPC exposed to acidic bile without NF-jB inhibitor (t test analysis; multiple comparisons by Holm-Sidak). We also observed that HHPC exposed to acid with BAY 11-7082 demonstrated a significant decrease in miR-21 and miR-192 levels, compared to those exposed to acid without BAY 11-7082 (P = .038 and P = .007, respectively) Similarly, HHK exposed to acidic bile with BAY 11-7082 exhibited a significant decrease in "oncomir" miR-21 levels (P = .001265), miR-155 (P = .002) and miR-192 (0.003) , as well as a significant increase in "tumour suppressor" miR-451a levels (P = .000339) and miR-375 (P < .0001) , compared to HHK exposed to acidic bile without NF-jB inhibitor (t test analysis; multiple comparisons by Holm-Sidak). We also observed that HHK exposed to neutral bile with BAY 11-7082 demonstrated a significant increase in miR-34a (P = .0012) and miR-451a (P = .0029) levels, relative to HHK expose to neutral bile without NF-jB inhibitor . Finally, we observed that HHK exposed to acid with BAY 11-7082 showed a trending reduction, without statistical significance, in oncomirs miR-21 and miR-192, compared to HHK exposed to acid without BAY 11-7082 .
## | nf-jb inhibitor induces a reversed
cancer-related miRNA phenotype in acidic bile-treated normal human hypopharyngeal cells, relative to controls
We observed that NF-jB inhibitor induced an inverted miRNA phenotype of "oncomirs" and particularly of "tumour suppressor" miRNAs in between acidic bile and control groups (P = .0442 and P = .0139, respectively, by Friedman). The reversal of miRNA phenotypes, by NF-jB inhibitor, was particularly intense in acidic bile-treated HHPC, demonstrating a significant decrease in expression ratios of "oncomirs" [fig_ref] 3. 4 |: Correlations among BAY 11-7082-induced oncogenic miRNA expressions in treated normal human hypopharyngeal... [/fig_ref] and increase in expression ratios of "tumour suppressor" miRNAs (with/without BAY 11-7082), compared to control (P = .0269, one-way ANOVA; by Kruskal-Wallis) [fig_ref] 3. 4 |: Correlations among BAY 11-7082-induced oncogenic miRNA expressions in treated normal human hypopharyngeal... [/fig_ref]. We also observed an inverted miRNA phenotype, by NF-jB inhibitor, in acidic bile-treated HHK, demonstrating a significant decrease in expression ratios of "oncomirs" (P = .0114, one-way ANOVA; by Kruskal-Wallis) and an increase in expression ratios of "tumour suppressor" miRNAs (with/without BAY 11-7082), compared to control .
Specifically, we observed that each particular miRNA, including miR-21, miR-155, miR-192, miR-34a, miR-375 and miR-451a, was affected by NF-jB inhibitor in HHPC (P < .005, by Kruskal-Wallis) [fig_ref] 3. 4 |: Correlations among BAY 11-7082-induced oncogenic miRNA expressions in treated normal human hypopharyngeal... [/fig_ref]. On the other hand, miR-21 and miR-375 were the most affected by the application of NF-jB inhibitor in HHK, among the analysed miRNA markers, representing significant expression changes in acidic bile-treated normal hypopharyngeal cells, compared to control (P = .0022, by Kruskal-Wallis) .
We further showed HHPC treated with acidic bile without BAY 11-7082 demonstrated a significantly higher miR-21/375 ratio, compared to neutral control or neutral bile, (P = .0022, and P = .0415, respectively, by Kruskal-Wallis) [fig_ref] 3. 4 |: Correlations among BAY 11-7082-induced oncogenic miRNA expressions in treated normal human hypopharyngeal... [/fig_ref] , and similarly HHK treated with acidic bile without NF-jB inhibitor showed a significantly higher miR-21/375 ratio compared to neutral control or acid alone (P = .0022 and P = .0415, respectively, by Kruskal-Wallis). In contrast, we observed that BAY 11-7082 induced a significant reduction in miR-21/375 ratios in acidic bile-treated HHPC, compared to neutral control and neutral bile (P = .0022 and P = .0415, respectively) [fig_ref] 3. 4 |: Correlations among BAY 11-7082-induced oncogenic miRNA expressions in treated normal human hypopharyngeal... [/fig_ref] , as well as in HHK, compared to neutral control and acid alone (P = .0022 and P = .0415, respectively) .
These observations suggest that NF-jB inhibition is capable of decreasing the acidic bile-induced elevated miR-21/375 ratios in both HHPC and HHK.
3.3 | BAY 11-7082 preferentially affects the acidic bile rather than the neutral bile-induced cancer-related miRNA phenotypes
We observed that miRNA changes induced by NF-jB inhibitor included a significant difference between acidic bile-(pH 4.0) and neutral bile (pH 7.0)-treated cells . F I G U R E 2 NF-jB inhibitor (20 lmol/L ΒΑΥ 11-7082) reverses the acidic bile-induced deregulation of cancer-related miRNA markers, in normal human hypopharyngeal primary cells (HHPC). A, Acidic bile induces in HHPC (a) an up-regulation of the analysed "oncomirs" demonstrated by significantly higher miRNA levels, compared to controls, inverted by NF-jB inhibitor (BAY 11-7082). (b) Acidic bile induces a down-regulation of the analysed "tumour suppressor" miRNAs, demonstrated by significantly lower expression levels, compared to controls, that is also inverted by NF-jB inhibitor (BAY 11-7082) in HHPC (one-way ANOVA; by Friedman; *P < .05; **<P < .005; GraphPad Prism 6.0) B, Graphs depict the significantly (a) decreased expression levels of miR-21, miR-155 and miR-192 and (b) increased expression levels of miR-34a, miR-375 and miR-451a, in HHPC exposed to acidic bile with NF-jB inhibitor (BAY 11-7082), compared to HHPC exposed to acidic bile without NF-jB inhibitor (P values by t test; mean AE SD; multiple comparisons by Holm-Sidak; GraphPad Prism 6.0). (Normalization control: small RNA RNU6B). (Data were obtained from three independent experiments)
We observed that BAY 11-7082 induced significantly lower expression ratios (with/without BAY 11-7082) of "oncomirs", miR-21, miR-155 and miR-192, in acidic bile compared to neutral biletreated HHPC (P < .005) and HHK (P < .05) . On the other hand, we found that BAY 11-7082 induced significantly higher expression ratios (with/without BAY 11-7082) of "tumour suppressor" miR-34a, miR-375 and miR-451a, in acidic bile, compared to neutral bile-treated HHPC (P < .0005) and higher ratios of "tumour suppressor" miR-375 and miR-451a, in acidic bile, compared to neutral bile-treated HHK (P < .0005) . We also identified a significant positive correlation between BAY 11-7082-induced expression changes of "tumour suppressor" miR-34a and miR-375 (r = .99320292, P = .0007), miR-34a and miR-451a (r = .92967608, P = .0221) and between miR-375 and miR-451a (r = .88047938, P = .0487 in treated HHPC.
We found a strong inverse correlation between BAY 11-7082 induced expression changes of miR-21 and miR-375 (r = À.990359, P = .0096) of different treated groups of normal human hypopharyngeal cells (both HHK and HHPC) . We observed a strongly inverted correlation between BAY 11-7082-induced expression changes (with/without BAY 11-7082) of "oncomirs" miR-155 or miR-192 and "tumour suppressor" miR-451a in HHPC (r = À.914115213, P = .0298 and r = À.995811529, P = .0003, respectively), as well as miR-155 and miR-375 in HHK (r = À.989308022, P = .0107), as well as of miR-192 and miR-451a in HHK (r = À.996114636, P = .0039) .
## | correlations between bay 11-7082-induced changes of oncogenic mirna markers and nf-jb-related genes in treated human hypopharyngeal primary cells
To determine the correlations between BAY 11-7082-induced expression changes (with/without BAY 11-7082) of the analysed oncogenic miRNA markers and NF-jB-related genes, in the same treated groups, we performed qPCR analysis from the same pool of total RNA. F I G U R E 3 NF-jB inhibitor (20 lmol/L ΒΑΥ 11-7082) reverses the acidic bile-induced deregulation of cancer-related miRNA markers, in human hypopharyngeal keratinocytes (HHK). A, Acidic bile induces in HHK (a) an up-regulation of the analysed "oncomirs", demonstrated by significantly higher miRNA levels, compared to controls, inverted by NF-jB inhibitor (BAY 11-7082). (b) Acidic bile induces a down-regulation of the analysed "tumour suppressor" miRNAs, demonstrated by significantly lower expression levels, compared to controls, that is also inverted by NF-jB inhibitor (BAY 11-7082) in HHK (one-way ANOVA; by Friedman; *P < .05; **<P < .005; GraphPad Prism 6.0). B, Graphs depict the significantly (a) decreased expression levels of miR-21, miR-155 and miR-192 and (b) increased expression levels of miR-34a, miR-375 and miR-451a, in HHK exposed to acidic bile with NF-jB inhibitor (BAY 11-7082), compared to HHK exposed to acidic bile without NF-jB form of IL-6 up-regulation. [bib_ref] DNA damage induces NF-jB-dependent microRNA-21 up-regulation and promotes breast cancer cell invasion, Niu [/bib_ref] Our data from immunofluorescence assay demonstrated that the acidic bile-induced activated STAT3
and NF-jB were inhibited by NF-jB inhibitor in treated HHPC . Specifically, we observed that acidic bile-treated HHPC showed an intense nuclear staining for both phospho-NF-jB (p-p65 S556) and phospho-STAT3 (Tyr705)
# | discussion
Nuclear factor kappa B (NF-jB) is a key factor that mediates inflammatory and early tumorigenic events in epithelial cells, [bib_ref] The complexity of NF-jB signaling in inflammation and cancer, Hoesel [/bib_ref] that are related to activated NF-jB and that may contribute to an aggressive phenotype of head and neck cancer. [bib_ref] Unraveling regulatory programs for NF-kappaB, p53 and microRNAs in head and neck..., Yan [/bib_ref] [bib_ref] A novel nuclear factor-kappaB gene signature is differentially expressed in head and..., Lee [/bib_ref] Here, we present the first in vitro report that bile and acid com- Whereas the ratio of miR-21/375 has been considered a potential biomarker related to poor prognosis of supraglottic cancer, [bib_ref] miR-21 and miR-375 microRNAs as candidate diagnostic biomarkers in squamous cell carcinoma..., Hu [/bib_ref] [bib_ref] MiR-21/miR-375 ratio is an independent prognostic factor in patients with laryngeal squamous..., Hu [/bib_ref] our data demonstrate a significant increase in miR-21/375 ratios in cells exposed to acidic bile (pH 4.0), compared to controls, a relationship that is effectively inverted in the presence of NF-jB inhibitor, again strongly supporting the miR-21/375 ratio, as a potential biomarker in acidic bile-induced cancer-related molecular events in hypopharyngeal cells, mediated by NF-jB pathway.
Our novel findings also demonstrate that NF-jB inhibition significantly prevents the acidic bile-induced up-regulation of miR-155. Our current findings also demonstrate that miR-451a could be an important marker of acidic bile-related laryngopharyngeal carcinogenesis, in agreement with Fukumoto et al who previously suggested miR-451a is a tumour suppressor marker in hypopharyngeal SCC. [bib_ref] Identification of tumour suppressive microRNA-451a in hypopharyngeal squamous cell carcinoma based on..., Fukumoto [/bib_ref] This in vitro measure is in line with our 45-day in vivo model where a significant down-regulation of miR-451a occurred. [bib_ref] miR-21, miR-155, miR-192, and miR-375 deregulations related to NF-kappaB activation in gastroduodenal..., Sasaki [/bib_ref] We also showed a significant effect of NF-jB inhibition on "tu- suppressor and key regulator miRNA in HNSCC, [bib_ref] MicroRNA-34a affects the occurrence of laryngeal squamous cell carcinoma by targeting the..., Shen [/bib_ref] [bib_ref] Unraveling regulatory programs for NF-kappaB, p53 and microRNAs in head and neck..., Yan [/bib_ref] and our prior in vivo findings showed that acidic bile down-regulated miR-34a levels in treated laryngopharyngeal mucosa. Although previous studies reported NF-jB binding sites on the promoter of miR-34a, [bib_ref] Transcriptional activation of microRNA-34a by NF-kappa B in human esophageal cancer cells, Li [/bib_ref] [bib_ref] IL-6R/STAT3/miR-34a feedback loop promotes EMT-mediated colorectal cancer invasion and metastasis, Rokavec [/bib_ref] and an increase in miR-34a levels in oesophageal cells under NF-jB activation, [bib_ref] Transcriptional activation of microRNA-34a by NF-kappa B in human esophageal cancer cells, Li [/bib_ref] the exact mechanism of miR-34a regulation by NF-jB is not yet obvious. Alternatively, it has been shown that STAT3 can directly repress miR-34a, while an active IL-6R/STAT3/miR-34a loop was found necessary for EMT, invasion and metastasis of colorectal cancer cell line. [bib_ref] IL-6R/STAT3/miR-34a feedback loop promotes EMT-mediated colorectal cancer invasion and metastasis, Rokavec [/bib_ref] Our current data support the notion that NF-jB inhibition is capable of preventing down-regulation of miR-34a by acidic bile.
As would be expected, the individual response of HHPC and HHK was not always congruent. The ability of BAY 11-7082 to reverse acidic bile-induced phenotype of "oncomir" miR-192 was more prominent in HHPC compared to HHK. Likewise, the effect of NF-jB inhibition regarding "tumour suppressor" miRNAs miR-34a
and miR-451a was more intense in HHPC relative to HHK under acidic bile exposure. We are of the understanding that effects may be related to differences in maturational status. Primary cells HHPC, considered to be less mature and more sensitive to injurious external stimuli than immortalized keratinocytes HHK, may respond differently to NF-jB inhibition. The miRNA phenotype observed in HHPC was in fact similar to that observed in vivo, an observation in support of this view. [bib_ref] miR-21, miR-155, miR-192, and miR-375 deregulations related to NF-kappaB activation in gastroduodenal..., Sasaki [/bib_ref] Our novel data demonstrate strong inverted correlations among the BAY 11-7082-induced levels of the analysed tumour suppressors" miR-34a, miR-375 and miR451a, and NF-jB-related genes, such as RELA(p65), STAT3, TNF-a, IL-6 and IL-1b, that previous studies documented as crucial mediators of inflammatory and neoplastic events in head and neck cancer [bib_ref] The complexity of NF-jB signaling in inflammation and cancer, Hoesel [/bib_ref] [bib_ref] Aberrant IKKa and IKKb cooperatively activate NF-jB and induce EGFR/AP1 signaling to..., Nottingham [/bib_ref] [bib_ref] Chemoprevention of head and neck squamous cell carcinoma through inhibition of NF-jB..., Vander Broek [/bib_ref] [bib_ref] A novel nuclear factor-kappaB gene signature is differentially expressed in head and..., Lee [/bib_ref] [bib_ref] A signal network involving coactivated NF-kappaB and STAT3 and altered p53 modulates..., Lee [/bib_ref] [bib_ref] Genome-wide identification of novel expression signatures reveal distinct patterns and prevalence of..., Yan [/bib_ref] [bib_ref] Dysregulated molecular networks in head and neck carcinogenesis, Molinolo [/bib_ref] [bib_ref] Loss of the polarity protein PAR3 activates STAT3 signaling via an atypical..., Guyer [/bib_ref] [bib_ref] Molecular cross-talk between the NFkappaB and STAT3 signaling pathways in head and..., Squarize [/bib_ref] [bib_ref] Evidence that TNF-TNFR1-TRADD-TRAF2-RIP-TAK1-IKK pathway mediates constitutive NF-kappaB activation and proliferation in human..., Jackson-Bernitsas [/bib_ref] [bib_ref] Constitutively active NF-kappaB triggers systemic TNFalpha-dependent inflammation and localized TNFalpha-independent inflammatory disease, Dong [/bib_ref] [bib_ref] Adenoviral transgene delivery provides an approach to identifying important molecular processes in..., Foxwell [/bib_ref]. In line with our recent study, [bib_ref] Inhibition of NF-jB prevents the acidic bile-induced oncogenic mRNA phenotype, in human..., Vageli [/bib_ref] our current data suggest that NF-jB inhibition may reverse acidic bile-induced molecular events in normal human hypopharyngeal cells that are known to link inflammation to cancer, thereby in a sense shielding HHPC from the effects of bile-induced oncogenic molecular events.
We showed that BAY 11-7082 is also capable of preventing acid alone-induced deregulations of miR-21 and miR-192, but not miR-155 or "tumour suppressor" miRNAs in HHPC, suggesting that NF-jB inhibition could reverse a part of acid-induced miRNA phenotype in normal human hypopharyngeal cells. Although acid alone may upregulate selected "oncomirs", it was not capable of accelerating activation of oncogenic STAT3 or other cancer-related molecular events.
Similarly, bile at neutral pH seemed to contribute to deregulations of "tumour suppressor" miRNAs, such as miR-34a and miR-451a, but was not capable of accelerating activation of oncogenic STAT3 or other cancer-related molecules. Our data support the observation that acid and bile in combination but not acid or bile alone may contribute to cancer-related molecular events, mediated by NF-jB.
Our findings revealed that NF-jB inhibition resulted in a significant reduction in viable cells, particularly in acidic bile-treated groups. The identified decreased cell viability induced by 7082 is in line with previous studies. [bib_ref] Targeting NF-jB signaling pathway suppresses tumor growth, angiogenesis, and metastasis of human..., Li [/bib_ref] The preferential effect of BAY 11-7082 in acidic bile-treated groups is especially interesting because it suggests its effect on cell viability is not global but related to specific events, in those groups. Interestingly, other studies have shown that increased expression of miR-34a and miR-375 reduces cell viability of cancer cell lines. [bib_ref] Effect and mechanism of miR-34a on proliferation, apoptosis and invasion of laryngeal..., Wang [/bib_ref] [bib_ref] MicroRNA-375 targets PAX6 and inhibits the viability, migration and invasion of human..., Zou [/bib_ref] In this regard, our present data also showed that BAY 11-7082 decreases the levels of "oncomirs", miR-155, miR-192 and miR-21 levels, while increasing the levels of "tumour suppressors", miR-34a and miR-375, raising the view that specific cancer-related miRNAs interactions may be at least partially responsible for decreased cells viability. and -451a in treated hypopharyngeal mucosa.
# | conclusion
## Confli ct of interest
The authors whose names are listed in this article certify that they have no affiliations with or involvement in any organization or entity with any financial interest or non-financial interest in the subject matter or materials discussed in this manuscript.
## Author contri butions
DV, SD and CTS contributed to conceptualization and data curation.
[fig] |: BAY 11-7082 effectively reduces acidic bile-induced NF-jB transcriptional activity in normal human hypopharyngeal cells To investigate the effect of 20 lmol/L of BAY 11-7082 in NF-jB transcriptional activity of acidic bile-treated normal human hypopharyngeal cells, HHPC and HHK, we used an NF-jB luciferase assay (Figure 1), demonstrating that 20 lmol/L of pharmacologic inhibitor BAY 11-7082 effectively prevented the transcriptional activity of [/fig]
|
Renal Leiomyosarcoma, a Rare Presentation
Renal sarcomas are very rare malignant tumours with a very poor prognosis. Renal leiomyosarcoma, a malignant tumour of smooth muscle origin, is the most common histological type. The article reports a case of leiomyosarcoma of renal location, with a review of the literature. A 38-year-old female patient, with no previous pathological history, consulted the nephrology department of the Teaching Hospital of Lomé for abdominal pain that had been present for 4 years. Histology showed a tumour proliferation of fasciculated architecture, made of spindle cells arranged in long bundles, with cytonuclear atypia and cytoarchitectural abnormalities. Immunohistochemical examination showed positive staining for smooth muscle actin, h-caldesmone, desmin and CD34 and negative for pancytokeratin (AE1/AE3), myogenin and PS100. Renal leiomyosarcoma is an exceptional malignancy. It remains the most common renal sarcoma, the differential diagnosis of which is based on immunohistochemical findings.
# Introduction
Renal sarcomas represent 1-3% of malignant tumours of the kidney, frequently occurring in the fifth or sixth decade of life [bib_ref] Primary léiomyosarcoma of the kidney, Venkatesh [/bib_ref]. Leiomyosarcoma remains the most common histological type accounting for 50-60% of renal sarcomas [bib_ref] Leiomyosarcoma of the kidney: Case report, Martinez-Cornelio [/bib_ref]. It is a malignant tumour of smooth muscle origin that usually occurs in soft tissue and the uterus. However, a small percentage may originate from smooth muscle or vascular walls, most of which are venous in origin [bib_ref] Leiomyosarcoma of kidney, Niceta [/bib_ref]. Renal leiomyosarcomas can arise from the smooth muscle fibres of the renal pelvis and the renal capsule or the renal vessels, the latter being more common [bib_ref] High-grade primary renal leiomyosarcoma, Hakan [/bib_ref].
We report a case of renal leiomyosarcoma, describe the clinical presentation and diagnostic features and discuss the prognostic and therapeutic features of primary renal leiomyosarcomas.
## Case report
A 38-year-old woman with no previous pathological history consulted the nephrology department of the Teaching Hospital of Lomé for abdominal pain that had been present for 4 years and had been considered as renal colic.
The clinical examination revealed an axillary temperature of 37.4°C, a body weight of 56 kg for a height of 1.67 m. Blood pressure was 120/70 mmHg. The patient was in good general condition with well-stained conjunctiva. A painful mass was palpated on the left flank. The lymph nodes were free. The biological workup included a normal blood count, uremia, creatinine and blood glucose, all of which were returned normal. A computed tomography (CT) scan revealed a suspicious tissue mass measuring 9 × 3 cm, with homolateral hydrocalice and without any secondary thoraco-abdominal lesion [fig_ref] Figure 1: Abdominal CT scan showing a suspicious tissue mass in the left kidney [/fig_ref]. The paraclinical workup was completed by a uroscanner which revealed a spontaneously isodense process in the left kidney, enhanced after injection of contrast medium, without extension to the renal vein, resulting in stage III left ureterohydronephrosis with excretory disorder. The right kidney and intra-abdominal solid organs were without abnormalities. There was a large polymyomatous uterus with a normal bladder. An extended nephrectomy was performed and the specimen was sent to the pathology laboratory. Macroscopically, it was a nephrectomy specimen, without a ureter, weighing 640 g, measuring 15 × 12 × 5 cm, with a bumpy surface and an elastic consistency. On section, there was a well-limited whitish tumour lesion measuring 12 × 12 × 4 cm with haemorrhagic changes and a peripheral healthy portion of 3 cm without peri-renal fat infiltration [fig_ref] Figure 2: Macroscopic appearance of the sample after opening [/fig_ref].
The resected specimen was treated with 10% formaldehyde, followed by conventional dehydration, paraffin embedding, sectioning and hematoxylin and eosin (HE) staining. Histology showed a tumour proliferation with a fasciculated architecture, made of spindle cells arranged in long bundles, showing cytonuclear atypia and cytoarchitectural abnormalities. The cells had eosinophilic cytoplasm, an elongated nucleus, marked anisokaryosis and mitotic rate estimated at 9 mitoses/10 HPF in hotspot areas. No carcinomatous contingent or necrosis was found. Due to this, Immunohistochemistry reveals positive immunostaining for smooth muscle actin (SMA), h-caldesmone, desmin and vimentin. They are negative for epithelial markers (4).
## Treatment and prognosis
Renal leiomyosarcomas are generally biologically aggressive with a poor prognosis. Radical nephrectomy is the treatment of choice [bib_ref] Evaluation of biological potential of smooth muscle tumours, Miettinen [/bib_ref]. The major prognostic factor is total surgical resection (12). Lewis et al. reported that the presence of unresectable disease and incomplete surgical resection were the most important predictors of disease-specific death [bib_ref] Retroperitoneal soft-tissue sarcoma: Analysis of 500 patients treated and followed at a..., Lewis [/bib_ref].
Life expectancy is lower for renal sarcomas than for other urinary tract sarcomas. The 5-year survival is 82% for patients with retroperitoneal sarcoma, 73% for bladder sarcoma, 44% for prostate sarcoma and 39% for patients with renal sarcoma [bib_ref] Prognostic factors and clinical outcomes of urological soft tissue sarcomas, Geonseok [/bib_ref]. According to Geonseok et al., 5-year survival rate is 51.4% for all urogenital sarcomas [bib_ref] Prognostic factors and clinical outcomes of urological soft tissue sarcomas, Geonseok [/bib_ref].
Our patient underwent radical and extended nephrectomy. She is free of recurrence after a 16-month follow-up.
# Conclusion
Renal leiomyosarcomas are rare malignant mesenchymal tumours with a poor prognosis. They usually arise from the renal capsule, the smooth muscle tissue of the renal pelvis and the intra-renal vessels. They can only be differentiated from sarcomatoid renal cell carcinoma by immunohistochemical examination. Radical nephrectomy is the standard treatment and total surgical resection is the major prognostic factor.
# Statements
## Ethical approval and consent to participate
The study received approval from the head of the laboratory to be conducted. The manuscript was not submitted to more than one journal for simultaneous review and has not been previously published. Single study, not divided into multiple parts and no data were generated.
## Consent for publication
The Department of Pathology of the Teaching Hospital of Lomé has authorized the publication of this manuscript
## Conflicts of interest
The authors declare that they have no conflict of interest.
we suggested as diagnoses, a renal fibrosarcoma, an inclined renal carcinoma with a sarcomatoid component estimated at 100%, a renal synovialosarcoma and a renal leiomyosarcoma in the last instance. Immunohistochemical examination showed intense and diffuse labelling of actin, h-caldesmone and desmin and intense and focal labelling of CD34. There was an absence of immunostaining for pan-cytokeratin AE1/ AE3, myogenin and S100 protein. In view of this immunomorphological profile, we made the diagnosis of FNCLCC grade 2 renal leiomyosarcoma (3 + 1 + 0).
The patient has no evidence of recurrence after a 16-month follow-up.
# Discussion
## Epidemiology
Renal sarcomas represent 1-3% of adult renal malignancies (1). Leiomyosarcomas constitute 50-60% of renal sarcomas [bib_ref] Leiomyosarcoma of the kidney: A clinicopathologic study, Deyrup [/bib_ref]. These tumours usually arise from the renal capsule, the smooth muscle tissue of the renal pelvis and the intra-renal vessels [bib_ref] Leiomyosarcoma of the kidney: A clinicopathologic study, Deyrup [/bib_ref]. They occur predominantly in the fifth or sixth decade of life, with an incidence that increases with age [bib_ref] Case report: Good prognosis in leiomyosarcoma of the kidney, Demir [/bib_ref] [bib_ref] Apropos a case, M Rebassa Llull [/bib_ref]. Renal leiomyosarcoma is more common in women, most often occurring in the right kidney (6).
## Clinic and imaging
Renal leiomyosarcomas usually have an insidious clinical presentation with late-stage symptoms of abdominal pain, vomiting, weight loss and a palpable mass. Compression of neighbouring organs causes back pain and haematuria [bib_ref] Case report: Good prognosis in leiomyosarcoma of the kidney, Demir [/bib_ref]. Neither ultrasound, CT scan, nor MRI can differentiate leiomyosarcomas from renal cell carcinomas [bib_ref] Abdominal leiomyosarcomas: Radiologic appearances at various locations, Kurugoglu [/bib_ref].
## Histopathology
On microscopic examination, leiomyosarcomas have the characteristics of a smooth muscle tumour with a fasciculated architecture, made up of spindle-shaped cells arranged in long tangled bundles. The cells have sharp-ended, nontapered nuclei with eosinophilic cytoplasm. Indicators of malignancy are necrosis, nuclear pleomorphism with rare mitotic figures (1). It is extremely difficult to differentiate a renal leiomyosarcoma from a renal sarcomatoid carcinoma. Both tumours have similar clinical, radiographic and pathological features [bib_ref] High-grade primary renal leiomyosarcoma, Hakan [/bib_ref]. Only the absence of an epithelial contingent on morphological examination and the absence of cytokeratin expression on immunohistochemical examination can formally rule out a sarcomatoid carcinoma of the kidney [bib_ref] Leiomyosarcoma of the renal vein: Analysis of outcome and prognostic factors in..., Novak [/bib_ref] [bib_ref] Leiomyosarcoma of the renal vein mimicking a primitive renal cell carcinoma: Case..., Fekkar [/bib_ref]. Primary monophasic synovial sarcoma of the kidney also shows monophasic spindle cells. The spindle cells are plump with irregular cell borders. They tend to grow in sheets
## Availability of data and materials
All data supporting the conclusions of this article are included in the manuscript and its supporting documents.
# Funding
The authors did not receive funding from any source to conduct this study.
[fig] Figure 1: Abdominal CT scan showing a suspicious tissue mass in the left kidney. [/fig]
[fig] Figure 2: Macroscopic appearance of the sample after opening. [/fig]
[fig] Figure 3: (HE × 40). Normal kidney (left) and fasciculated tumour proliferation (right). [/fig]
[fig] Figure 4: (HE × 400). Spindle cell proliferation with clear anisokaryosis. and usually have trapped renal tubules in the form of cysts. However, these tumours show positivity for Bcl-2 (1). [/fig]
|
Clinical and research applications of a brain tumor tissue bank in the age of precision medicine
## A multidisciplinary team for successful tumor banking & precision medicine research
Much like the multidisciplinary teams that manage patient care, a multidisciplinary collaboration is necessary to generate and maintain a successful tumor bank [fig_ref] Figure 1: Contributions of the team [/fig_ref] , blue arrows) [bib_ref] Brain tumor biobanking in the precision medicine era: building a high-quality resource..., Ostrom [/bib_ref] [bib_ref] Challenges for quality management in implementation, maintenance, and sustainability of research tissue..., Schmitt [/bib_ref]. This same multidisciplinary team can also provide input regarding how to successfully conduct and translate precision medicine research [fig_ref] Figure 1: Contributions of the team [/fig_ref] , blue circle).
Depending on clinical practice variations, either the neuro-oncologist or surgeon is the team member most likely to introduce the patient to the tumor banking project and invite their participation. A level of rapport must be established to ensure patients are comfortable entrusting the research team with sensitive molecular data and derivatives generated from donated biospecimens [bib_ref] Issues in collecting, processing and storing human tissues and associated information to..., Grizzle [/bib_ref]. Informed consent to contribute biospecimens must be acquired from the patient or legally authorized representative by the clinical research team, prior to the surgical procedure. Blood specimens may also be collected at this time to enable correlative research and to provide a reference genome. Pairing the blood draw with informed consent acquisition ensures that these samples are obtained before any surgical interruption of the blood-brain barrier takes place. In research, this same patient-clinician connection will inform avenues of study and potentially facilitate patient-reported outcome studies. This is particularly true for patient-engaged research [bib_ref] Patients' and clinicians' research priorities, Stewart [/bib_ref] , which formulates patient advisory councils to provide input on the direction of research.
Surgeons and pathologists must be engaged with the tissue banking process in order to ensure successful specimen collection in the operating room. Standard operating procedures for specimen collection and handling should be collaboratively devised by the surgeon, pathologist, biobanking staff and research teams, with continuity across all participating sites. This ensures that the specimen is collected in an efficient manner, reducing ischemic time to preserve molecular stability and with minimal impact on the flow of patient care. Notably, the pathologist must have access to sufficient tissue for an accurate diagnosis and the surgeon must not discard excess tissue that could be banked. Regular training of the clinical staff may be necessary to maintain engagement as trainees and nursing teams rotate into the operating schedule, especially in academic facilities. In research, the clinical teamneuro-oncologist, oncologist, surgeon, pathologist -provides insight into current treatment practices and gaps in knowledge for the application of precision medicine. Close collaboration ensures that research programs maintain a path toward future clinical translation and implementation into daily practice.
The role of the neuropathologist is critical. Correct diagnosis according to current WHO guidelines is imperative for both treatment decisions and subsequent research analyses. Initial triaging of tissue can be challenging, particularly when only small amounts of tissue are available (e.g., in stereotactic biopsies). Brain tumors can also show histologic and molecular heterogeneity that may not be apparent at the time of intraoperative consultation. Although rendering a correct tissue diagnosis is the highest priority, pathologists should also strive to ensure that the banked tissue is representative of the tumor and that sufficient tissue is retained for the tissue bank. Separate maintenance of a research biobank from the clinical pathology archive preserves the maximum amount of tissue while ensuring clinical access. Although formalin-fixed and paraffin-embedded tissue is commonly used for pathology archives, due to its stability, in recent years, molecular testing from archived paraffin-embedded tissue has become widely available. Close cooperation between a tissue repository and the department of pathology opens opportunities to access and utilize both tissue collections.
Because tissue bank personnel play a critical role in specimen management, the development of standard operating procedures and appropriate staff training are vital. Preparation and storage of tissue and biofluids in aliquots minimizes deterioration from freeze-thaw cycles by preparing single-use quantities. Each aliquot must be labeled, tracked and stored with histology slides for pathology review; tissue bank personnel are key players in research, as specimens are located, retrieved and used. Generation of cell models or molecular extraction may also be performed upon entry of a new specimen into the tissue bank. Having a dedicated team for these banking future science group www.futuremedicine.com tasks, who are trained and experienced with the procedures of the standard operating protocol, provides consistency across banked samples and derivatives. The data science team includes statisticians, bioinformaticians, data analysts and clinical data abstractors. This team creates standardized data entry forms and oversees clinical data collection, curation of the database, tracking of sample utilization and data generation, and implements data quality control procedures. Routine assessment of the repository's clinical population relative to the overall treated or health system population may be used to characterize the patient sample or identify missed subgroups. For research, the statistician will provide input on study design, calculate the sample size to ensure an effective study without waste and perform data analyses. The bioinformatician will provide molecular data curation pipelines, analysis and classification schemes.
Molecular scientists and data scientists work in collaboration with the clinical teams to support and lead precision medicine research. Opportunities in bioinformatics and patient avatar models are described in more detail below. As end users of the biobanking resource, the research team should be involved in the development of protocols for biobanking. Their input on the collection and storage procedures -such as rapid freezing to ensure that protein or small molecules are not degraded, or generation of cell lines from fresh tissue -ensures maximum utility of the banked specimens for research.
Regulatory compliance is a responsibility of all members of the team. However, it is efficient to have the protocol for the biobanking and subsequent studies managed by a designated research staff member or project manager. This person assists with regular compliance audits of the tumor bank, generation and renewal of the tissue bank IRB protocol, and coordination of parties for data usage, both internally and with external collaborators. The project manager role may also include collection and monitoring of resource use requests to ensure timely review and access.
The HBTB has a multidisciplinary Tumor Bank Operations Committee that includes all the groups described above. The operations committee meets monthly to discuss general operation, strategize for planned updates and review usage requests. Training, audits and other regulatory affairs are also brought before this team. All team members in [fig_ref] Figure 1: Contributions of the team [/fig_ref] are represented in this committee. Although no patient advocates currently attend the operations committee meetings on a regular basis, the brain tumor center has an active patient advisory council which is available for consult on any new proposals.
## Use of a brain tumor tissue bank to advance precision medicine research
Precision medicine aims 'to establish a molecular taxonomy of disease through the creation of a multiparametric information network that can be used to refine diagnosis and treatment, and ultimately improve health outcomes'. Well-structured tumor banks, with clinical data associated with the acquired tissues, can contribute to the development of such precision medicine efforts. The HBTB has made significant contributions to foundational neuro-oncology molecular taxonomy efforts through TCGA [bib_ref] The Cancer Genome Atlas Research Network. Comprehensive genomic characterization defines human glioblastoma..., Institute [/bib_ref] [bib_ref] Comprehensive, integrative genomic analysis of diffuse lower-grade gliomas, Cancer Genome Atlas Research [/bib_ref] , the Glioma Longitudinal AnalySiS Consortium, the recently developed International Consortium on Meningiomas and other initiatives [bib_ref] Glioma through the looking GLASS: molecular evolution of diffuse gliomas and the..., Consortium [/bib_ref]. For example, the HBTB contributed nearly a quarter of the total glioma specimens and related clinical data to the federally funded TCGA Project, which has resulted in transformative research impacting the clinical care of brain tumor patients throughout the USA [bib_ref] Comprehensive, integrative genomic analysis of diffuse lower-grade gliomas, Cancer Genome Atlas Research [/bib_ref] [bib_ref] The somatic genomic landscape of glioblastoma, Brennan [/bib_ref].
## Development of molecular diagnostic, prognostic, & predictive biomarkers & classifiers through bioinformatics analysis
Translational research on banked tissues -including development of machine learning and comprehensive molecular data generation techniques -has rapidly increased our understanding of disease. The comprehensive molecular profiling generated by the TCGA Project and the integrative power of bioinformatics tools have already contributed to the development of 'signatures' that predict future tumor behavior as well as molecular markers that supported the 2016 update of the WHO classification of CNS tumors. More recently, a study showed the robustness of the machine learning methodology to classify CNS tumors into distinct DNA-methylation profile classes to improve the accuracy of clinical diagnosis. The authors profiled the DNA methylome of 2801 retrospective cases that represented a range of CNS tumors, and built a decision-tree classifier applying the random forest algorithm [bib_ref] Random Forests, Breiman [/bib_ref]. This classifier was able to identify methylation classes that correlate with established WHO types or subtypes and identify novel entity classes. The authors established a threshold value for the prediction of a matching class that can be used to classify new cases. This classifier was reproducible using different DNA methylation profiling techniques (EPIC array and whole-genome bisulfite sequencing data). A prospective application of the classifier to 1104 specimens showed a 76% concordance between pathological and methylation classifications [bib_ref] Practical implementation of DNA methylation and copy-number-based CNS tumor diagnostics: the Heidelberg..., Capper [/bib_ref]. This classification system is freely available online (www.molecularneuropathology.org).
Beyond diagnosis classification, the integration of clinical data has allowed the development of predictive signatures of tumor behavior. In one study, authors developed candidate signatures that predicted progression of IDH-mutant glioma to a more aggressive subtype using methylome analysis and various bioinformatics tools [bib_ref] A distinct DNA methylation shift in a subset of glioma CpG island..., De Souza [/bib_ref]. Such advances in neuro-oncology would not have been possible without high quality tissue specimens and clinical correlates from established biorepositories.
Recent data-sharing requirements placed on researchers submitting work to major journals or in applications for national funding have improved public access to large-scale molecular data, which expands translational research capacities. Although platforms such as Group on Earth Observations (GEO) and the Database of Genotypes and Phenotypes (dbGAP) have been available for over a decade, the disparate nature of the studies collected there reduces cross-study research [bib_ref] Gene expression Omnibus: NCBI gene expression and hybridization array data repository, Edgar [/bib_ref] [bib_ref] NCBI GEO: archive for functional genomics data sets -update, Barrett [/bib_ref]. A recent addition to data-sharing options, the National Cancer Institute's (NCI) Genomic Data Commons (GDC) is a data-sharing platform that promotes precision medicine in oncology through standardized processes [bib_ref] Toward a shared vision for cancer genomic data, Grossman [/bib_ref]. Seeding the platform with data from TCGA and Therapeutically Applicable Research to Generate Effective Treatments (TARGET)studies, the GDC aims to provide a common platform to capture and combine cancer genomics data from multiple studies. These extensive data bases are built upon high quality tissue repositories such as the HBTB. Quality standards required for upload, common informatics pipelines for data processing and common data elements for clinical curation are among the methods used by GDC provide a basis for translational research. These publicly available data repositories are based on the specimen collected in tissue banks and now can be exploited by researchers from around the world to make novel discoveries and/or validate important findings.
To facilitate such exchange and to enhance these findings, several important bioinformatics tools have been developed to harness cancer genomics data. The Bioconductor project is a common bioinformatics tool development platform; TCGAbiolinks [bib_ref] TCGAbiolinks: an R/Bioconductor package for integrative analysis of TCGA data, Colaprico [/bib_ref] is an example of a Bioconductor package that was developed to facilitate the analysis of TCGA/TARGET data by incorporating the query, download and processing steps within Bioconductor in order to apply statistical methodologies to biologically derived data. In addition, TCGAbiolinks provides integrative methodologies to perform several important downstream analyses (summarized inTable 2). A user-friendly, webbased graphical user interface (GUI) version is also available: the TCGAbiolinksGUI. This interface enables user access to the methodologies offered in TCGAbiolinks using a 'point-and-click' style analysis without the need to write specific programming codes. A sample analysis session is shown in [fig_ref] Figure 2: A volcano plot created in TCGA Biolinks GUI [/fig_ref]. Another benefit of TCGAbiolinks is that proprietary research data from a local tumor bank can be analyzed and integrated with GDC data, allowing the researcher to explore clinical questions. Researchers can use these tools to analyze and validate the genomic data of their local tissue banks and put it in the context of broader populations.
Improvements and incorporation of bioinformatics tools will help to bridge the knowledge gap between clinicians and scientists and will offer solutions to integrate the findings of publicly available data and local biorepositories. Efforts to integrate additional data analysis with the original data source (i.e., HBTB database) will magnify the return of scientific discovery and strengthen the primary resource. This is of importance for clinicians who participate in molecular tumor boards with intent to offer tailored treatment based on actionable molecular targets. Clinical repositories using an informatics architecture designed to harmonize clinical and genomic data to identify available therapies in combination with research bioinformatics tools will allow clinicians to interpret coding and noncoding variants as well as other drivers of disease development, and to apply this knowledge to precision medicine.
## Integration of clinical information & the use of database to advance clinical care & research
The clinical annotation of the HBTB uses a standardized process that complies with patient privacy regulations and addresses semantic and technical heterogeneity. Aggregating this data at an institutional level can be challengingbut this process is necessary to maintain data integrity and promote collaboration.
Molecularly annotated tissue specimen associated longitudinal drug exposure data have led to the discovery of predictive biomarkers that provide insight into a tumor's response to specific therapies -a hallmark of precision medicine [bib_ref] Randomized clinical trials with biomarkers: design issues, Freidlin [/bib_ref]. When a treatment-responsive cohort is identified, the archived biospecimens and clinical data can be interrogated to identify key molecular features; the factors or biomarkers can then be validated through prospective assessment in a clinical trial. This approach is attractive in glioma research as many patients that have participated in therapeutic clinical trials with targeted agents also contributed tumor specimens to the HBTB. A personalized approach to precision medicine uses a patient's own health data to direct treatment at an individual level and can be enriched by the utilization of patient-derived models, (described below) to test for individual drug responses. 'N-of-1' trials use a patient's own data to explore individual variability in response to a therapeutic intervention [bib_ref] The n-of-1 clinical trial: the ultimate strategy for individualizing medicine?, Lillie [/bib_ref]. Use of standardized N-of-1 methodology when clinical trials or standard therapy are not available may identify unique characteristics that contribute to improved treatment outcomes, identification of individual drug tolerability and permit investigation of unique end points not possible with larger clinical trial methodologies (i.e., a particular adverse event). This methodology may also lend to insights across clinical trials and beyond one individual patient's treatment course. The use of patient-derived animal avatars to develop an individualized therapy plan with N-of-1 methodology has further potential to improve health outcomes. Conditions that are extremely rare or excluded from standard clinical trials, such as CNS metastatic disease, may also benefit from an N-of-1 approach and other innovative trial designs that support limited datasets for early therapeutic discovery and disease treatment [bib_ref] Genomic characterization of brain metastases reveals branched evolution and potential therapeutic targets, Brastianos [/bib_ref]. When bolstered with a patient-derived, molecularly annotated biorepository, use of a standardized N-of-1 methodology when treating patients outside of a clinical trial may lead to discoveries in pharmacogenomics, therapeutic sequencing and predictive biomarkers.
Looking forward, the use of patient reported outcomes and new forms of individual health data generated by patients using wearable devices, such as smart watches, activity trackers and health monitors, will provide additional value in precision medicine. Digital phenotyping, which refers to the 'quantification of the individual-level human phenotype in situ using digital devices', [bib_ref] New tools for new research in psychiatry: a scalable and customizable platform..., Torous [/bib_ref] will provide further insight into the patient's environment and function. Once this data are standardized and validated, the HBTB will include structured digital phenotyping data in our neuro-oncology database. As data use capabilities evolve, we anticipate recording neuroanatomical, image-guided, surgical navigation correlates of harvested biospecimens during surgery into our database to capture intratumoral heterogeneity.
## Translational research: use of a tissue bank for the development of patient-derived models
Although no experimental model can fully capture the complexity of a high-grade brain tumor, patient-derived mouse xenografts (PDX) are unmatched in the faithful representation of the molecular heterogeneity of the original tumors [bib_ref] New cast for a new era: preclinical cancer drug development revisited, Herter-Sprie [/bib_ref]. Because primary brain tumor pathology is better reproduced in the orthotopic location, tumor specimens that are initially propagated subcutaneously in immunocompromised mice must be excised, dissociated, briefly cultured and implanted into the brain. The resulting orthotopic xenograft tumor recapitulates invasion and retains somatic mutations, including EGFR amplification, frequently lost in traditional cell lines [bib_ref] Establishment, maintenance and in vitro and in vivo applications of primary human..., Carlson [/bib_ref] [bib_ref] Patient tumor EGFR and PDGFRA gene amplifications retained in an invasive intracranial..., Giannini [/bib_ref] [bib_ref] Integrated genomic analysis identifies clinically relevant subtypes of glioblastoma characterized by abnormalities..., Verhaak [/bib_ref] [bib_ref] Use of an orthotopic xenograft model for assessing the effect of epidermal..., Sarkaria [/bib_ref]. Encouraging examples of correlation between clinical and PDX response to therapy are fueling the development of PDX panels for drug development [bib_ref] New insights into drug development for pediatric solid tumors: what preclinical data..., Houghton [/bib_ref].
There are different approaches to utilizing PDX in precision medicine. Most studies address the correlation of PDX and patient outcomes at a population level [bib_ref] High-throughput screening using patient-derived tumor xenografts to predict clinical trial drug response, Gao [/bib_ref] rather than assessing responses in individual patients. However, new so called 'co-clinical trials' that involve PDX models are already becoming more common in other cancers (e.g., PDX Modeling of Treatment Response for Triple Negative Breast Cancer; NCT02247037) and are expected to enroll increasing numbers of high-grade glioma patients. The expansion of the biorepository to include a live biobank through the propagation of tumor cells or tissue in the lab plays a transformative role in developing these PDX models for upcoming clinical trials.
Fresh tumor specimens from the HBTB are routinely dissociated and cultured in selective media for cancer stemlike cells (CSC); this allows us to create a comprehensive collection of renewable cryopreserved patient material for preclinical studies. Our research team has cultured more than 100 high-grade glioma specimens [bib_ref] Optimization of high grade glioma cell culture from surgical specimens for use..., Hasselbach [/bib_ref] , and has refined the procedure for intracranial implants in immunocompromised mice to achieve high throughput and consistency for orthotopic tumor growth in PDX, essential for preclinical studies [bib_ref] Optimization of glioblastoma mouse orthotopic xenograft models for translational research, Irtenkauf [/bib_ref]. A panel of GBM tumors and matched CSCs and PDX underwent comprehensive molecular characterization to ensure that they would represent all three glioblastoma transcriptional subtypes and frequent genomic abnormalities [bib_ref] Tumor evolution of glioma-intrinsic gene expression subtypes associates with immunological changes in..., Wang [/bib_ref]. The models presented remarkable heterogeneity and conservation of the original somatic genomic alterations including extrachromosomal oncogene amplification [bib_ref] Discordant inheritance of chromosomal and extrachromosomal DNA elements contributes to dynamic disease..., Decarvalho [/bib_ref]. The HBTB CSC and PDX models are used in basic research and preclinical studies focused on precision medicine in internal and collaborative projects involving academic, nonprofit and pharmaceutical industry studies [bib_ref] Copper-binding small molecule induces oxidative stress and cell-cycle arrest in glioblastoma-patient-derived cells, Shimada [/bib_ref] [bib_ref] The cyclin-like protein Spy1 regulates growth and division characteristics of the CD133+..., Lubanska [/bib_ref].
future science group www.futuremedicine.com
## Conclusion: opportunities & limitations for precision medicine from biobanks
Precision medicine relies on the concept of giving the right treatment to the right patient at the right time. The use of annotated biospecimens as well as advancements in bioinformatics has led to refined disease classifications and treatment paradigms [bib_ref] Towards precision medicine, Ashley [/bib_ref] [bib_ref] P4 medicine: how systems medicine will transform the healthcare sector and society, Flores [/bib_ref]. Research on banked biospecimens provides the opportunity to 'crowd-source' molecular and clinical data patterns needed to inform the presenting precision medicine decision. The more patients who are assessed in the bank, the more likely that a patient (or set of patients) with characteristics similar to those of the presenting patient has been observed. Due to the sensitive location of tumors within or adjacent to the brain, biobanking tumor specimens in neurooncology has unique challenges that may be less common for more accessible tumor sites. For example, the ability of a patient to provide informed consent to specimen banking prior to surgery may be impacted by the brain tumor itself. Changes to cognitive ability, vision, speech, language processing, emotional response and motor control are all potential presenting symptoms of a brain tumor, which introduce barriers to assuring informed consent. Although it may be easier to focus banking efforts on the most cognitively stable patients, the location of the tumor impacts the type and extent of impairment and tumor types tend to localize within the brain; exclusion due to cognitive symptoms may result in selection bias and failure to adequately sample patients with specific diagnoses. It is therefore critical to develop protocols and informed consent processes that incorporate both the patient's level of decision making and an authorized representative to ensure that patients and caregivers understand the benefits and limitations of tissue banks.
If a subset of patients or tumor types are systematically excluded, there is a risk that biobank specimens and resulting conclusions are not representative of all cohorts being potentially researched and treated. On an individual level, the patient at hand may carry a tumor type that is not well represented in the biobank and thus potentially not well characterized. It is important that treating physicians are aware of the population represented in the biobank so that they can understand the potential limitations in precision medicine applications. Maintenance of a diagnosis-level screening record of all patients in the bank will allow researchers and clinicians to determine whether there is significant bias in the clinical characteristics of those participating in the tumor bank and those who do not. In the HBTB, this screening log is performed monthly; however, other intervals may be appropriate. Regular assessment of this screening log will identify patient groups not well represented and may uncover opportunities to improve recruitment practices. Such monitoring should also spur innovative thinking about how to capture the molecular profiles of these nonbanked patients in other research endeavors.
Another challenge in using banked tissues to inform precision medicine is that the banked tissue represents only the portion of the disease that has been removed. It has been observed that brain tumors are molecularly heterogeneous; in glioma, clonal development shows pockets of unique molecular signatures within the larger tumor mass [bib_ref] Intratumor heterogeneity in human glioblastoma reflects cancer evolutionary dynamics, Sottoriva [/bib_ref]. As a result, multiple portions of tumor specimens should be banked and assessed when possible. Further, recent work has shown that gliomas demonstrate spatial differences (for example, between the central mass and the leading edge), and radiology studies indicate that the tumor field is larger still, with tissues affected beyond the leading edge. Moreover, the tissue and cells left behind are likely the most important for the subsequent treatment of disease. Additionally, the limited work available in longitudinal molecular profiling has revealed that tumors evolve with progression of disease and in response to therapy. Although some patients will undergo surgery for recurrent disease, many will not. Thus, the characteristics of the naive tumor are used to infer the likely progress of the tumor; this may become less reliable with time and subsequent treatments. Tissue specimens could be obtained postmortem to observe disease evolution; however, this tends to be incongruent with patient care where patients often spend end-of-life at home. Alternatively, a comprehensive tissue bank will prioritize longitudinal sampling of tissues and other biospecimens, such as blood or cerebral spinal fluid, to inform temporal and chronological heterogeneity. Although cerebral spinal fluid sampling may easily be obtained in some scenarios such as pediatric or metastatic cases receiving intrathecal therapy, the risk and burden of this moderately invasive procedure may deter other patients from participating in the biobank. Making it possible for patients to donate tissue at the time of surgery but opt-out of serial biofluid donations may allay those concerns without preventing tissue collection.
Despite these limitations, brain tissue repositories that integrate clinical as well as tissue-derived data such as molecular characteristics, cell line and PDX-derived information offer opportunities to refine therapeutic decisionmaking, foster multidisciplinary high-impact research, and ultimately improve patient outcomes.
## Future perspective
This article focused on tissue banking; however, the future of precision medicine will rely on the banking of more than tissues. Given the limited access to diseased tissues during the treatment of brain tumors, integration of external factors for the assessment and monitoring of disease is a natural evolution. Radiomic imaging studies are under development to align patterns in MR images with molecular and clinical features of disease. Identification of circulating free DNA or other biomarkers in the blood or cerebral spinal fluid that track with treatment response or disease progression will afford less-invasive clinical monitoring. Similarly, patient reported outcomes or physiological tracking will provide new guidance in patient care. Systematic capture, integration and analysis of these large data resources will create new challenges for the data science team. However, a full multidisciplinary team will be required to ensure that a translational focus is maintained. This does not indicate the end of tissue banking, but an expansion of biobanking capacity. Multifaceted data that includes tumor tissues as the source of biological truth will be necessary for the development of reliable precision medicine markers, preclinical or parallel avatar trials and clinical decision tools.
## Executive summary
- Progress in the field of neuro-oncology can be attributed to deepening analysis of systematically collected, quality-controlled and clinically annotated central nervous system (CNS) tumor banks, also called biorepositories. Hermelin brain tumor bank - The data warehouse is used to screen patients for clinical trials, manage multidisciplinary neuro-oncology tumor board and index clinical trial participation.
- An imaging repository of magnetic resonance imaging (MRI) studies and other imaging modalities is linked to annotated biospecimens such that molecular, tumor avatar, radiomic and treatment response data can be combined to yield comprehensive analytics on a personalized level. Multidisciplinary team - Critical ingredients for the establishment and maintenance of a CNS biorepository are the clinician-patient relationship and multidisciplinary collaboration.
- Necessary research components include banking personnel for specimen preparation, identification, and tracking and for initiation of cellular and animal disease models.
- The data science team is comprised of statisticians, bioinformaticists and clinical data abstractors and is key to database curation, veracity and data analysis.
- Molecular and data scientists work in collaboration with the clinical teams to optimize biobanking and tissue utilization and to support and lead precision medicine research.
## Biomarkers & bioinformatics
- Biorepositories that house clinically annotated tissue using standard protocols can be leveraged to augment central databases, which can incorporate multiple sites and thousands of specimens for impactful research.
- Predictive biomarkers have been identified through Hermelin Brain Tumor Center Tissue Bank participating consortia including The Cancer Genome Atlas, the Glioma Longitudinal AnalySiS Consortium, the International Consortium on Meningiomas and others.
- The National Cancer Institute hosts a Genomic Data Commons designed to promote precision medicine and the study of genomics in oncology through a public data sharing platform.
- Bioinformatics tools have been developed to harness genomics cancer data in a research-friendly format for database query, download and processing. Translational research - Cultured and stored cancer stem-like cells can serve as a cryopreserved renewable source of patient-derived material for preclinical studies.
- Patient-derived mouse xenografts provide the best preclinical representation of the molecular heterogeneity of individual tumors and their development can be an essential part of the repository.
- Comprehensive molecular characterization of matched tumor with patient derived cancer stem-like cells and patient-derived mouse xenografts models can serve as the foundation of a high-grade glioma molecular library and validate models of precision medicine. Conclusion - Brain tissue repositories that integrate clinical and tissue derived data offer extensive opportunities to refine therapeutic decision making, foster high impact research and improve patient outcomes.
future science group www.futuremedicine.com
## Financial & competing interests disclosure
This work was supported by the Henry Ford Health System Department of Neurosurgery and the Hermelin Brain Tumor Center.
Additionally, LMP, HN, ACD, and AM are supported by NCI R01CA222146 and HN, LMP, ACD, and AVC are supported by DOD CA170278. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.
No writing assistance was utilized in the production of this manuscript.
# Authors contributions
All authors contributed equally to this manuscript.
## Open access
This work is licensed under the Attribution-NonCommercial-NoDerivatives 4.0 Unported License. To view a copy of this license, visit http://creativecommons.org/licenses/by-nc-nd/4.0/
[fig] Figure 1: Contributions of the team. A multidisciplinary team is required for successful biobanking (blue arrows) and the sharing of knowledge across this team promotes successful research (blue circle). [/fig]
[fig] Figure 2: A volcano plot created in TCGA Biolinks GUI. The menu bar on the left shows different analyses available with the volcano plot item highlighted. The panel on the right shows the controls available for the current analysis; all aspects of the plot are configurable here. [/fig]
[table] Table 1: Clinical information and outcomes data assessed in the clinical database. [/table]
[table] Table 2: Feature comparison of bioinformatics tools specifically designed to analyze TCGA data.TCGA: The Cancer Genome Atlas. [/table]
|
Salvage therapies in relapsed and/or refractory myeloma: what is current and what is the future?
The treatment landscape for multiple myeloma (MM) is evolving with our understanding of its pathophysiology. However, given the inevitable cohort heterogeneity in salvage therapy, response to treatment and overall prognoses tend to vary widely, making meaningful conclusions about treatment efficacy difficult to derive. Despite the hurdles in current research, progress is underway toward more targeted therapeutic approaches. Several new drugs with novel mechanism of action and less toxic profile have been developed in the past decade, with the potential for use as single agents or in synergy with other treatment modalities in MM therapy. As our discovery of these emerging therapies progresses, so too does our need to reshape our knowledge on knowing how to apply them. This review highlights some of the recent landmark changes in MM management with specific emphasis on salvage drugs available for relapsed and refractory MM and also discusses some of the upcoming cutting-edge therapies that are currently in various stages of clinical development.
# Introduction
Multiple myeloma (MM) is a plasma cell disorder representing 1.5% of all cancers and up to 13% of all hematologic malignancies worldwide. [bib_ref] Multiple myeloma: charging toward a bright future, Katzel [/bib_ref] According to the American Cancer Society, ∼30,330 new cases of myeloma are expected to be diagnosed in 2016.Although myeloma is usually responsive to cytotoxic therapy in all stages, ie, in initial and relapsed, responses are often ephemeral, mandating the development of new therapeutic targets and more successful combination therapies. Over the years, notable progress has been made in autologous stem cell transplantation (ASCT) along with the introduction of several breakthrough drugs, including newer generation immunomodulators and proteasome inhibitors (PIs), which led to a significant increase in response rate of those affected as well as survival rate. [bib_ref] International Myeloma Working Group. Risk of progression and survival in multiple myeloma..., Kumar [/bib_ref] In fact, 5-year survival rates have almost doubled, increasing from 27% to 47% between 1989 and 2010, respectively. [bib_ref] Recent improvement in survival of patients with multiple myeloma: variation by ethnicity, Pulte [/bib_ref] Indeed, even with these enormous headways in the management of the disease, MM still remains a serious malady with many patients eventually developing treatment resistance. [bib_ref] Thalidomide and hematopoieticcell transplantation for multiple myeloma, Barlogie [/bib_ref] In addition, response time generally decreases with subsequent number of treatment lines. [bib_ref] Clinical course of patients with relapsed multiple myeloma, Kumar [/bib_ref] Overcoming this challenging nature of the disease remains a herculean task, with an increasing pressure to bring in other PIs and immunomodulatory drugs (IMiDs) as well as drugs with a niche mechanism of action which are effective even in progressed stages of myeloma. Currently available and investigational drugs for the treatment of MM are listed in [fig_ref] Table 1: Snapshot of current and upcoming therapies Abbreviations [/fig_ref]. This review highlights some of the landmark submit your manuscript | www.dovepress.com
## Dovepress
## Dovepress
## 4844
Thumallapally et al changes in the MM management with specific emphasis on salvage drugs available for relapsed and refractory MM (RRMM) and discusses some of the new and emerging drugs that are currently in various stages of clinical development.
## Definitions
Traditionally, active MM diagnosis required confirmation of end-organ damage using the CRAB criteria (hypercalcemia, renal insufficiency, anemia, bone lesion).Based on this assessment, patients who did not exhibit any signs of end-organ damage, but showed evidence of clonal plasma cell proliferation, were classified as having either monoclonal gammopathy of undetermined significance or smoldering MM. Because of slow progression and less aggressive nature of monoclonal gammopathy of undetermined significance and smoldering MM, most of the patients in this disease stage were not treated, due to the finite number of safe treatment options available and lack of curability at that time. Currently, given that bone marrow and myeloma pathogenesis microenvironments are clearly defined, and a number of safer therapies have become available, these limitations no longer apply. To enhance the early diagnosis of the condition, the International Myeloma Working Group carried out progressive revision of MM criteria to avoid early occurrence of end-organ damage. [bib_ref] International myeloma working group updated criteria for the diagnosis of multiple myeloma, Rajkumar [/bib_ref] The aim of this initiative was to include all patients who did not meet the CRAB criteria but had suspected early presence of clonal bone marrow plasma cells. The current criteria for diagnosing MM are summarized in [fig_ref] Table 2: Revised diagnostic criteria for MM [/fig_ref]. 8
## Risk stratification
The initial step in any treatment pertains to the assessment of patient eligibility. In particular, it is not possible to make even a speculative prognosis before conducting risk stratification. This assessment includes careful consideration of the disease stage, performance status, and patient's age. Detecting biological mutations by the use of fluorescence hybridization in situ and conventional karyotyping is also usually performed when stratifying the disease in terms of the risk to the patient. [bib_ref] Treatment of newly diagnosed multiple myeloma based on mayo stratification of myeloma..., Dispenzieri [/bib_ref] ,10 MM has various molecular subtypes that are all unique in response and presentation of More than 10% monoclonal plasma cell proliferation in bone marrow or biopsy-proven solitary bone plasmacytoma or extramedullary plasmacytoma including any one or more of the following (myeloma-defining events): 1) Evidence of end-organ damage - Hypercalcemia (serum Ca .1 mg/dL higher than the upper limit of normal)
- Renal insufficiency (creatinine clearance .2 mg/dL) - Anemia (value of Hb .20 g/L below the lower limit of normal)
- Bone lesions: one or more skeletal lytic lesion seen on PeT/CT, X-ray 2) One or more of biomarkers of neoplastic growth
- Clonal plasma cell proliferation in bone marrow $60%
- involved/uninvolved serum-free light chain ratio $100%
## 4845
Salvage therapies in relapsed and/or refractory myeloma disease. A good example of this distinctive presentation is the trisomic MM that responds appropriately to therapies that are lenalidomide based. [bib_ref] Interpretation of cytogenetic results in multiple myeloma for clinical practice, Rajan [/bib_ref] [bib_ref] Characteristics of exceptional responders to lenalidomide-based therapy in multiple myeloma, Vu [/bib_ref] At the same time, researchers have demonstrated that t(4;14) MM requires induction with bortezomib (BTZ) and suitable maintenance for successful outcomes. [bib_ref] Bortezomib induction and maintenance treatment in patients with newly diagnosed multiple myeloma:..., Sonneveld [/bib_ref] With respect to clinical presentation, t(4;MM has been found to display lower rates of predilection of bone disease during diagnosis. On the other hand, association between t(14;16) MM with high serum-free light chain levels has frequently been reported, and the available evidence suggests that it may also lead to increased risks of acute renal failure during the diagnosis. [bib_ref] Relationship between initial clinical presentation and the molecular cytogenetic classification of myeloma, Greenberg [/bib_ref] In addition to cytogenetics, in recent years, many useful technological advances have been made, allowing exploration of genomic variability of specific malignant cells using gene expression profiles. Among these, the most notable are myPRS, SKT-92, and M3P. [bib_ref] A gene expression signature for high-risk multiple myeloma, Kuiper [/bib_ref] [bib_ref] Translating a gene expression signature for multiple myeloma prognosis into a robust..., Van Laar [/bib_ref] [bib_ref] Targeted sequencing using a 47 gene multiple myeloma mutation panel (M(3) P)..., Kortüm [/bib_ref] [bib_ref] Longitudinal analysis of 25 sequential sample-pairs using a custom multiple myeloma mutation..., Kortüm [/bib_ref] However, due to their novelty, very little empirical evidence regarding their value in predicting prognosis and long-term survival of patients more accurately, presently exists.
## Response criteria
For the optimal understanding of the management of myeloma, it is also important to know the definitions of relapse and disease progression, set by the International Myeloma Workshop Consensus Panel, which are outlined in [fig_ref] Table 3: Defining myeloma [/fig_ref]. 20
## Current treatment options for rrmm
The current treatment standards for patients with RRMM include either rechallenging the disease with salvage chemotherapy or hematopoietic stem cell transplant. At present, the role of ASCT and post-transplant consolidation/ maintenance therapy remains unclear. Nonetheless, it is believed that these treatment options can be safely offered in specific cases. Historically, RRMM was treated by cytotoxic combination chemotherapy using melphalan, doxorubicin, or cyclophosphamide with dexamethasone or prednisone to reduce tumor burden prior to transplantation. However, high risk of long-term complications, such as secondary leukemia and myelodysplastic syndrome, necessitated the search for more effective and less toxic combinations. The research into new forms of therapy has taken into consideration drug resistance issues, along with matters pertaining to clonal evolution of disease. Currently, novel and less lethal emerging therapies are being tested as a part of several clinical trials, with or without the addition of stem cell transplantation. Moving forward, this review explains several chemotherapy options available currently for RRMM.
## Chemotherapy: monotherapy and combination therapy novel pis
Proteolysis of cellular proteins is a highly regulated complex process and is important in maintaining cellular function and homeostasis. This process is streamlined by proteasomes, which selectively destroy proteins that are covalently labeled with ubiquitins through the ubiquitin proteasome pathway (UPP). [bib_ref] Protein degradation by the ubiquitin-proteasome pathway in normal and disease states, Lecker [/bib_ref] Defects within this pathway could result in numerous diseases, including various malignancies. The UPP pathway consists of two essential steps, namely ubiquitin tagging and proteolytic degradation by the 26S proteasome, which is composed of 20S core and 19S regulator. The first step, ubiquitin tagging, includes the binding of an E1 ubiquitin-activating enzyme with ubiquitin, and the subsequent transfer of ubiquitin to protein by an E3 ubiquitin ligase. In the second step, the ubiquitin molecules are released and the unfolded protein is released into the inner chamber of the 26S proteasome. In this review, chymotrypsin-like, trypsin-like, and caspase-like enzymes catalyze the proteins to yield small peptides [fig_ref] Figure 1: Mechanism of action of Pis, immunomodulators, and HDACi [/fig_ref]. [bib_ref] Protein degradation by the ubiquitin-proteasome pathway in normal and disease states, Lecker [/bib_ref] Initially, PIs were probes facilitating research on the catalytic activity of proteasomes. Thereafter, their critical role in cell function heralded their role as potential therapeutic agents. The first mechanism of action identified in PIs was the inhibition of the inflammation-associated protein NFκB, a transcription factor that plays a critical role in carcinogenesis by activating angiogenesis, proliferation, migration, and suppression of apoptosis. In the cytoplasm, NFκB is bound to its inhibitor IκB and is activated only when IκB undergoes proteolysis. [bib_ref] A proteasome inhibitor prevents activation of NF-kappa B and stabilizes a newly..., Traenckner [/bib_ref] The inhibition of proteasome activity prevents the proteolytic cleavage of IκB and subsequently affects the activation and translocation of NFκB to the nucleus, thereby preventing the activation of downstream pathways. In addition, noteworthy is proteasome inhibition that leads to accumulation and aggregation of misfolded proteins in endoplasmic reticulum, which results in stress permitting the signaling to switch from pro-survival to pro-apoptotic in the cells. [bib_ref] Targeting the integrated networks of aggresome formation, proteasome, and autophagy potentiates ER..., Moriya [/bib_ref] PIs also affect UPP-regulated DNA repair mechanisms such as nucleotide excision repair, postreplication repair, and homologous recombination, by depleting the available nuclear ubiquitin. This results in an accumulation of nondegraded polyubiquitinated proteins, and consequently less free ubiquitin in the cell. [bib_ref] The ubiquitin-proteasome pathway and proteasome inhibitors, Myung [/bib_ref] Concurrently, the reduced levels of free ubiquitin in the cell lead to a loss of monoubiquitinated histones within the nucleus, thereby affecting DNA repair. PIs sensitize tumor cells to various therapies. Proteasome inhibition is also associated with angiogenesis inhibition by affecting the secretion of vascular endothelial growth factor (VEGF) and at the same time causes upregulation of pro-apoptotic molecules such as p53, Bcl-2-associated X protein, and NOXA along with the reduction in the levels of anti-apoptotic proteins such as Bcl-2 inhibitor (inhibitors of apoptosis protein). [bib_ref] The proteasome inhibitor PS-341 inhibits growth, induces apoptosis, and overcomes drug resistance..., Hideshima [/bib_ref] BTZ is a reversible dipeptidyl-boronic-acid-based specific PI that targets the chymotrypsin-and caspase-like active sites. The efficacy and safety of BTZ were established through a number of important clinical trials [fig_ref] Table 4: Landmark trials of BTZ [/fig_ref]. It was initially approved by the Food and Drug Administration (FDA) in 2003 for refractory MM and subsequently expanded for use in combination therapy in the first-line setting and its role is
## 4847
Salvage therapies in relapsed and/or refractory myeloma currently being explored in the setting of post-ASCT maintenance therapy as well as induction therapy. Preclinical studies not only revealed the synergy between BTZ and lenalidomide but also potentiated the action of dexamethasone on tumor cells. This was tested in a Phase II trial in patients with relapsed MM or RRMM. When challenged with BTZ combined to lenalidomide and dexamethasone (VRd), these patients showed a remarkable progression-free survival (PFS) and overall survival (OS) duration lasting up to 9.5 months and 30 months, respectively [fig_ref] Table 4: Landmark trials of BTZ [/fig_ref]. This triplet combination showed increased, but manageable, toxicities including sensory neuropathy, fatigue, and neutropenia. [bib_ref] A phase 2 trial of lenalidomide, bortezomib and dexamethasone in patients with..., Richardson [/bib_ref] This review laid the foundation for other ongoing combination studies with drugs that possess synergistic activity.
## Resistance to btz
BTZ has been an effective therapy for MM; however, prolonged therapy could result in toxicity, peripheral neuropathy (PN), and drug resistance mediated via the overexpression of the β5-subunit, mutation of active drug-binding sites, or downstream upregulation of survival pathways. Three factors appear to be strongly associated with resistance toward PIs. First, there is mounting evidence on the role of deubiquitinating enzymes, USP14 and UCHL5, in survival and resistance to BTZ. These deubiquitinating enzymes could be deubiquitinating misfolded/unfolded proteins in MM, thereby reducing the stress levels. Second, autophagy might be a protective mechanism in MM cells. Induction of autophagy indirectly results in proteasome inhibition. Third, heat shock proteins (HSPs) appear to be a contributing factor in the resistance toward PIs. In vitro results have suggested that the inhibition of HSP70, an important component of the chaperone-mediated autophagy cytosolic chaperone complex, leads to the apoptosis of MM cells. [bib_ref] The resistance mechanisms of proteasome inhibitor bortezomib, Lü [/bib_ref] These factors stimulated the development of other analogs of proteasome inhibitors with increased efficacy and decreased toxicity and at the same time overcoming bortezomib resistance.
## Second-generation pis carfilzomib
Carfilzomib (CFZ) is a second-generation irreversible PI that binds on the proteasome to a site different from BTZ. It is a peptidyl epoxyketone and primarily inhibits the chymotrypsin-like activity at the βS-subunit of the core 20S proteasome. [bib_ref] A phase 2 study of single-agent carfilzomib (PX-171-003-A1) in patients with relapsed..., Siegel [/bib_ref] Apart from triggering cell cycle arrest, inducing apoptosis, and activating stress response pathways in human tumor cell lines, CFZ has been shown to be active against BTZ -resistant cell lines. [bib_ref] Building on bortezomib: second-generation proteasome inhibitors as anti-cancer therapy, Dick [/bib_ref] Two Phase II trials investigated the efficacy of singleagent activity of CFZ in patients refractory to BTZ and lenalidomide. The first Phase II trial (PX-171-003-A0) studied 46 patients on CFZ at a dose of 20 mg/m 2 given intravenously for 2 days consecutively for 3 weeks, followed by 2 weeks of intermission. This constituted one cycle, and the study was designed to complete 12 cycles. Very favorable results (overall response rate [ORR] of 13% and median duration of response up to 7.3 months) led to the alteration of the study to incorporate more pretreated patients and extend dosing regimen. [bib_ref] An open-label single-arm pilot phase II study (PX-171-003-A0) of low-dose, single-agent carfilzomib..., Jagannath [/bib_ref] This extended single-arm, multi-center Phase II study (PX-171-003-A1) included 266 patients who had received a median of five prior anti-myeloma regimens. CFZ was administered at a starting dose of 20 mg/m 2 in cycle 1, followed by the escalated target dose of 27 mg/m 2 in the cycles 2-12. [bib_ref] A phase 2 study of single-agent carfilzomib (PX-171-003-A1) in patients with relapsed..., Siegel [/bib_ref] ORR was ∼23% and observed in 36% of treated patients. These results prompted the fast-track approval of CFZ for the treatment of RRMM in patients who had received at least two prior treatments, including BTZ and immunomodulatory drug in July 2012. Dosing regimens of CFZ in BTZ naive patients were also investigated in another Phase II trial. [bib_ref] An open-label, single-arm, phase 2 (PX-171-004) study of single-agent carfilzomib in bortezomib-naive..., Vij [/bib_ref] In this review, patients who met the aforementioned criteria either received a starting dose of 20 mg/m 2 for 12 cycles, each comprising of 28 cycle days (cohort 1), or 20 mg/m 2 in cycle 1, followed by the escalated target dose of 27 mg/m 2 in the cycles 2-12 (cohort 2). The analysis of their results revealed an ORR of 42% and 52% in cohorts 1 and 2, respectively. According to the authors,
## 4848
Thumallapally et al BTZ -naive patients had a better response to CFZ delivered in an escalated mode. The most common adverse effects noted in the safety analysis were fatigue (49%), anemia (46%), nausea (45%), and thrombocytopenia (39%). The results also demonstrated a clear superiority of this treatment option relative to BTZ in terms of inducing or worsening PN. More specifically, patients with preexisting neuropathy did not experience worsening of their symptoms following the CFZ therapy. CFZ is currently being tested in combination chemotherapy and in frontline settings, as listed in [fig_ref] Table 5: Key trials of second-generation Pis and immunomodulatorsAbbreviations [/fig_ref].
Data from Phase III ASPIRE study (aimed at comparing treatment consisting of the combination of CFZ and lenalidomide-dexamethasone with the effects of lenalidomide-dexamethasone alone in the relapsed setting) showed encouraging PFS rates (26.3 months vs 17.6 months; P,0.001) in the CFZ group, further spiraling its way up as the most promising PI. [bib_ref] ASPIRE Investigators. Carfilzomib, lenalidomide, and dexamethasone for relapsed multiple myeloma, Stewart [/bib_ref] In the ENDEAVOR trial, the aim was to assess the efficacy of CFZ and dexamethasone in comparison with the effects of a combination of BTZ and dexamethasone in patients with relapsed MM, revealing statistically significant and high PFS in the CFZ group (PFS 18.7 months vs 9.4 months). Although hematologic adverse effects are similar in both groups, hypertension, dyspnea, and cardiac failure were more frequently seen with CFZ in this trial.As these findings do not concur with those reported in other previous trials, they are of notable concern and thus require further monitoring and evaluation in the future studies.
In addition to the trials discussed earlier, two early-phase studies are also noteworthy, as their goal is to investigate the efficacy and safety of CFZ in frontline MM in combination with either thalidomide 34 or lenalidomide [bib_ref] Phase 2 dose-expansion study (PX-171-006) of carfilzomib, lenalidomide, and low-dose dexamethasone in..., Wang [/bib_ref] and dexamethasone. These trials are important for gaining a better understanding of CFZ, and their findings may facilitate an expanded indication for this agent, as well as regulatory approval in the future.
ENDURANCE trial, which is a Phase III trial, is also of interest for the current discussion, as it involves 756 newly diagnosed patients with myeloma, with the aim of comparing CFZ triplet with BTZ triplet.Once its findings are reported, they are likely to help elucidate which PI is most appropriate for use in the frontline settings.
## Marizomib
Findings of existing research studies suggest that marizomib, a novel PI, can be an effective replacement for BTZ. Results reported in 2012, pertaining to two parallel, Phase I, doseescalation studies conducted in Australia and the US, in which 34 patients diagnosed with RRMM took part, were highly encouraging. To note is the fact that ∼70% patients in this trial were BTZ refractory and still a partial response (PR) was observed in 20% of patients, when treated with marizomib. [bib_ref] Phase I clinical evaluation of twice weekly marizomib (NPI-0052), a novel proteasome..., Richardson [/bib_ref] Most of the patients reported fatigue, dizziness,
## 4849
Salvage therapies in relapsed and/or refractory myeloma and headache as the prevalent drug-related adverse effects. No evidence of PN or thrombocytopenia was noted. Preliminary data from these trials imply that the adverse effects of marizomib are quite different when compared with other PIs. Moreover, there is strong evidence that it is active in BTZrefractory patients. Following this research, the effectiveness of a twice-weekly regimen of marizomib 0.5 mg/m 2 in combination with Pomalidomide low-dose dexamethasone is presently being investigated. [bib_ref] Synergistic anti-myeloma activity of the proteasome inhibitor marizomib and the IMiD((R)) immunomodulatory..., Das [/bib_ref] ixazomib
The preliminary data suggest that ixazomib produces durable responses in patients treated with other agents. The FDA approval of ixazomib (MLN9708) was based on the data yielded by the ongoing Phase III TOURMALINE-MM trial that included 722 patients across 26 countries. All patients included in the trial had received at least one prior therapy.
In addition, 50% of the cohort had undergone a stem cell transplant and their condition was progressing at the time of their recruitment to the study. The reported results indicate that a triplet combination of ixazomib, lenalidomide, and dexamethasone resulted in a significant improvement in PFS which substantially exceeded the benefits of the therapy based on the combination of lenalidomide and dexamethasone (with the median PFS of 20.6 months vs 14.7 months; [fig_ref] Table 6: Key trials on emerging novel therapies [/fig_ref]. 39
## Oprozomib
Oprozomib is an irreversible PI, which is currently being evaluated in Phase Ib/II studies as a single agent and in combination with dexamethasone for patients with RRMM and other hematologic malignancies. 40 CEP18770 (delanzomib) is an oral/intravenous, reversible boronate peptide agent which has shown promising antitumor activity in mouse models, with some treated specimens experiencing complete regression and significant survival benefits. Subsequently, Phase I trial, in which 38 patients took part, yielded encouraging findings, as none exhibited neurotoxicity commonly associated with BTZ [fig_ref] Table 6: Key trials on emerging novel therapies [/fig_ref]. [bib_ref] CEP-18770: a novel, orally active proteasome inhibitor with a tumor-selective pharmacologic profile..., Piva [/bib_ref] [bib_ref] Clinical and marketed proteasome inhibitors for cancer treatment, Zhang [/bib_ref] Novel immunomodulators Immunomodulators (IMiDs) exhibit antitoxic effects by binding to cereblon, a key protein of the E3 ubiquitin ligase complex, resulting in rapid ubiquitination and subsequent degradation of transcription factors, IKZF3 and IKZF1, as shown in [fig_ref] Figure 1: Mechanism of action of Pis, immunomodulators, and HDACi [/fig_ref]. [bib_ref] Mechanism of action of immunomodulatory agents in multiple myeloma, Reske [/bib_ref] Immunomodulators also regulate cell adhesion and bone marrow angiogenesis by inhibiting VEGF, tumor necrosis factor-α, interleukin-6, and interferon gamma, which in turn cause natural killer (NK) cell activation and NK cell-dependent cytotoxicity. [bib_ref] Thalidomide and its analogs overcome drug resistance of human multiple myeloma cells..., Hideshima [/bib_ref] Thalidomide was the first IMiD to emerge as a treatment option for MM. However, it is highly aggressive and
## 4850
Thumallapally et al is associated with considerable toxicity, particularly in older patients. This motivated researchers to explore potential analogs exhibiting greater clinical efficacy and better toxicity profile. Among the IMiDs that emerged from this effort, the most notable are lenalidomide (Revlimid) and pomalidomide (Pomalyst). In the past decade, these two agents were subjected to numerous trials investigating their effectiveness in the treatment of MM, which established their safety and efficacy profiles. Thus, this review focuses on key trials that have provided a solid conceptual foundation for the use of newer analogs of thalidomide in RRMM. Thalidomide was first approved by the FDA in 2006. The decision was based on the results yielded by Phase III trial involving 207 newly diagnosed patients with MM who were randomly assigned to the treatment or control group, thus receiving a combination of thalidomide and dexamethasone vs dexamethasone alone. The results revealed that the patient response rates (based on serum or urine paraprotein measurements) were significantly higher in the combination arm than in the control group receiving dexamethasone alone (51.5% vs 35.6%, respectively, P=0.025). Since the FDA approval, several trials using thalidomide as a single agent, or in combination with dexamethasone or melphalan and prednisone, have been conducted, and their findings confirm the benefits of this therapy mode. Thus, thalidomide is currently recommended for the treatment of newly diagnosed patients with MM in both the transplant-eligible and non-transplant eligible populations, as well as in patients with relapsed disease. Based on the available evidence, PN and venous thromboembolism are the main side effects stemming from the extended use of thalidomide. [bib_ref] International Myeloma Working Group. Prevention of thalidomide-and lenalidomide-associated thrombosis in myeloma, Palumbo [/bib_ref] In 2006, the FDA also approved the more potent lenalidomide in combination with dexamethasone for the treatment of patients with MM who had received at least one prior therapy. The approval was based on the results yielded by the pooled analysis of two randomized, placebo-controlled Phase III trials, denoted as MM009 and MM010; the findings of which revealed longer PFS and OS in patients with RRMM. [bib_ref] Multiple Myeloma (009) study investigators. Lenalidomide plus dexamthasone for relapsed multiple myeloma..., Weber [/bib_ref] The incidence of thromboembolic events in these trials was reported at 16%, where in prophylactic anticoagulation was not mandatory. In additional trials, the role of maintenance lenalidomide following the ASCT was also evaluated, and the results indicate the improvement in event-free survival. [bib_ref] Lenalidomide maintenance after stem-cell transplantation for multiple myeloma, Attal [/bib_ref] Pomalidomide is the most recent contribution to the development of IMiDs and is presently available in the market for the treatment of patients with RRMM. It is derived by adding an amino group to the fourth carbon of the phthaloyl ring of thalidomide, resulting in increased potency of both anti-inflammatory and antiangiogenic properties, accompanied by reduced toxicities. [bib_ref] Pomalidomide for the treatment of multiple myeloma, Clark [/bib_ref] Once its efficacy was ascertained in Phase I studies, pomalidomide was further evaluated in combination with dexamethasone. One such trial is a Phase II study involving patients with double-refractory myeloma; the findings of which indicate superior PFS in the pomalidomide/dexamethasone group, as shown in [fig_ref] Table 5: Key trials of second-generation Pis and immunomodulatorsAbbreviations [/fig_ref]. In this trial, treatment was discontinued primarily due to myelosuppression, in particular, grade .3 neutropenia. [bib_ref] Pomalidomide alone or in combination with low-dose dexametha-sone in relapsed and refractory..., Richardson [/bib_ref] Findings of the vital Phase III MM-003 trial also indicated longer PFS and OS in patients treated by pomalidomide/ low-dose dexamethasone compared with those who received high-dose dexamethasone therapy, which led to the accelerated approval by the FDA in 2013. Presently, pomalidomide is approved for the treatment of patients with MM who have received at least two prior therapies, including lenalidomide and BTZ, and have demonstrated disease progression on or within 60 days of completion of the last therapy. [bib_ref] Pomalidomide plus low-dose dexamethasone versus high-dose dexamethasone alone for patients with relapsed..., San Miguel [/bib_ref] As a condition implicit to this accelerated approval, the FDA will require submission of the results of the clinical trial CC-4047-MM-007, a randomized Phase III trial examining the effects of treatment with 1) pomalidomide added to BTZ and low-dose dexamethasone compared with 2) a combination of BTZ and a low-dose dexamethasone in patients with previously treated MM, as shown in [fig_ref] Table 6: Key trials on emerging novel therapies [/fig_ref]. Available evidence indicates that pomalidomide produces tolerable grade 3-4 side effects, with neutropenia, thrombocytopenia, and anemia being the most common. Pomalidomide is also currently being evaluated as a combination therapy in various trials aiming to elucidate its role in RRMM in the near future [fig_ref] Table 5: Key trials of second-generation Pis and immunomodulatorsAbbreviations [/fig_ref].
## Novel histone deacetylase inhibitors
Epigenetic modification is defined as an alteration of gene expression without the alteration of DNA sequences. It is further characterized into two major types: DNA methylation and histone modification. [bib_ref] Epigenetic modulating agents as a new therapeutic approach in multiple myeloma, Maes [/bib_ref] Although aberrancy in DNA methylation is evident in MM pathogenesis, uncertainty prevails in its mechanism. On the other hand, histone acetylation is comparatively well defined. The lysine residues within the N-terminal of histone tails can undergo various chemical modifications, including acetylation, methylation, ubiquitination, phosphorylation, and sumoylation. [bib_ref] Lysine acetylation targets protein complexes and co-regulates major cellular functions, Choudhary [/bib_ref] The acetylation status of histones is controlled by the opposing actions of two classes of enzymes: histone acetyl transferases, which transfer acetyl groups to lysine residues in histones, and histone deacetylases (HDACs), which remove the acetyl groups [fig_ref] Figure 1: Mechanism of action of Pis, immunomodulators, and HDACi [/fig_ref]. This acetylation status of histones further OncoTargets and Therapy 2016:9 submit your manuscript | www.dovepress.com
## Dovepress
## Dovepress
## 4851
Salvage therapies in relapsed and/or refractory myeloma influences chromatin conformation, which in turn regulates the expression of tumor suppressors, oncogenic proteins, and transcription factors. [bib_ref] Lysine acetylation targets protein complexes and co-regulates major cellular functions, Choudhary [/bib_ref] Three standard HDAC classes (I, II, and IV) containing eleven HDACs have been identified thus far and are classified according to their homology to yeast proteins, subcellular location, and enzymatic activities. [bib_ref] Panobinostat: a novel pandeacetylase inhibitor for the treatment of relapsed or relapsed..., Richardson [/bib_ref] Recent studies have shown that the overexpression of specific HDAC isoenzymes, such as HDAC1-4, HDAC6, and HDAC11, is associated with decreased PFS in patients with MM. In addition, epigenetic silencing of tumor suppressor genes, such as GPX3, RBP1, SPARC, and TGRBI, is associated with decreased survival in MM.
Of all HDAC enzymes, HDAC6 needs a special mention. Apart from streamlining epigenetic modification, it plays a key role in the aggresome protein degradation pathway. This pathway is another alternative proteolytic process that allows myeloma cells to develop resistance to PIs such as BTZ. [bib_ref] The aggresome pathway as a target for therapy in hematologic malignancies, Simms-Waldrip [/bib_ref] HDAC6 identifies proteins tagged with polyubiquitinated tails and targets them toward the dynein motor complex which in turn directs to the autophagosome for destruction. [bib_ref] HDAC6 control autophagosome maturation essential for ubiquitin-selective quality control autophagy, Lee [/bib_ref] Consequently, in a therapy based on a combination of HDAC and BTZ, both agents work synergistically to increase the cytotoxicity in MM cells.
At present, several HDAC inhibitors (HDACi) are being evaluated as potential candidates for MM therapy. In clinical trials, panobinostat and vorinostat in particular have shown promising results. Thus, these agents are reviewed in the following sections in more detail, to elucidate their potential therapeutic effects and usage in the treatment of RRMM.
## Panobinostat
Panobinostat is among one of the most potent pan-HDAC inhibitors in clinical development. It is believed to act primarily through epigenetic modulation of gene expression and inhibition of protein metabolism in the nanomolar range. In terms of the effects on all HDACs, potency of panobinostat is established to be at least tenfold greater than that observed in vorinostat, another nonselective HDACi. [bib_ref] Panobinostat: a novel pandeacetylase inhibitor for the treatment of relapsed or relapsed..., Richardson [/bib_ref] It is active against classes I, II, and IV HDAC enzymes.
On February 23, 2015, the FDA approved panobinostat for use in combination with BTZ and dexamethasone in the treatment of patients with MM who have previously received at least two drug regimens. [bib_ref] Panobinostat: a novel pandeacetylase inhibitor for the treatment of relapsed or relapsed..., Richardson [/bib_ref] The approval was based on the results yielded by the Phase III PANORAMA 1 trial that examined effectiveness of panobinostat, when administered in combination with BTZ and dexamethasone.The study sample was composed of 768 adult patients with RRMM who had received one to three prior treatments. The results demonstrated a clinically significant improvement in median PFS among the patients assigned to the panobinostat arm, which was statistically significantly greater relative to that observed in the placebo arm. Among the reported side effects, the most prominent were primarily gastrointestinal (diarrhea, nausea, and vomiting) and hematologic (thrombocytopenia) issues. However, the researchers cautioned that prolonged usage of panobinostat could result in cardiac toxicity, which could be a barrier to a wider application of this medication. Available evidence indicates that QTc prolongation and arrhythmias are the most likely cardiac issues, which are observed predominantly in elderly patients with MM.Following from this investigation, PANORAMA 2 was initiated. This single-arm Phase II trial aims at evaluating the effects of treatment comprised of adding panobinostat to BTZ and dexamethasone as a means of establishing whether BTZ-refractory patients respond to BTZ when HDACi is added. Among the 55 patients enrolled in this trial, ORR of ∼34.5% was noted, with a median response duration of ∼6 months [fig_ref] Table 6: Key trials on emerging novel therapies [/fig_ref]. [bib_ref] PANORAMA 2: panobinostat in combination with bortezomib and dexamethasone in patients with..., Richardson [/bib_ref] In addition to the aforementioned studies, several further trials are presently being conducted, with the goal of evaluating effectiveness of panobinostat in combination with other novel agents, such as next-generation protease inhibitors (CFZ or ixazomib), an IMiD (lenalidomide), and BTZ plus an IMiD (thalidomide or lenalidomide) with dexamethasone in relapsed/refractory MM. [bib_ref] Panobinostat: a novel pandeacetylase inhibitor for the treatment of relapsed or relapsed..., Richardson [/bib_ref] vorinostat Vorinostat is another nonselective HDACi that inhibits the enzymatic activity of HDACs, HDAC1, HDAC2, and HDAC3 (class I), and HDAC6 (class II) at nanomolar concentrations. [bib_ref] HDAC6 control autophagosome maturation essential for ubiquitin-selective quality control autophagy, Lee [/bib_ref] This drug was approved for use in the US in 2006 and has since then been offered to patients with cutaneous T-cell lymphoma. Vorinostat has also been tested in patients with MM and has been shown to be a potent inducer of apoptosis in MM cells. When combined with BTZ, it has been shown to enhance inhibition of protein breakdown. Moreover, in clinical trials, patients have exhibited increased response. [bib_ref] Vorinostat in combination with lenalidomide and dexamethasone in patients with relapsed or..., Siegel [/bib_ref] [bib_ref] Vorinostat or placebo in combination with bortezomib in patients with multiple myeloma..., Dimopoulos [/bib_ref] Of particular interest for the current investigation are VANTAGE 099 and VANTAGE 095 multicenter clinical trials that assessed the efficacy and safety of treatment based on a combination of vorinostat and BTZ in patients with relapsed or refractory MM.
In the VANTAGE 095 trial, an open-label, single-arm Phase IIb trial, 143 patients with heavily pretreated RRMM who had previously undergone a median of two treatment submit your manuscript | www.dovepress.com
## Dovepress
## Dovepress
## 4852
Thumallapally et al regimens were evaluated for efficacy of vorinostat in combination with BTZ. Study patients had disease that was refractory to BTZ and IMiD. Patients were treated until they started exhibiting signs of disease progression, showed evidence of unacceptable toxicities, or withdrew from the study. The primary end point was ORR ($PR). This combination demonstrated an ORR of 17% with a median response duration of 6.3 months. The PFS and OS were 3.1 months and 11.2 months, respectively. [bib_ref] Vantage 095: vorinostat in combination with bortezomib in salvage multiple myeloma patients:..., Siegel [/bib_ref] This study was followed by VANTAGE 088 trial.
In the VANTAGE 088 trial, which was a Phase III randomized, double-blinded, placebo-controlled study, 637 patients, with a median of two previous treatments, were randomized to receive BTZ together with vorinostat or placebo (BTZ alone). As in VANTAGE 095 trial, patients were treated until disease progression, unacceptable toxicities, or withdrawal from the study. The primary end point for this trial was PFS, whereas secondary and exploratory end points included ORR ($PR), clinical benefit response (ORR + minimal response), OS, time to progression, and safety/tolerability. The study yielded findings indicating marked improvement in PFS in patients receiving the combination of vorinostat and BTZ compared with those who were treated with BTZ alone. Median PFS was 7.63 months in the vorinostat group, which was significantly higher than that in the placebo group (6.83 months). The ORR was also greater in the vorinostat group than the placebo group (56.2% vs 40.6%). [bib_ref] Vorinostat or placebo in combination with bortezomib in patients with multiple myeloma..., Dimopoulos [/bib_ref] At present, however, the vorinostat pharmaceutical manufacturer is not pursuing FDA approval for the treatment of MM. Nonetheless, its efficacy in combination with other myeloma agents is still being evaluated in clinical studies. Further research is clearly needed to define the role of vorinostat in RRMM as well as to determine its maximum-tolerated dose. Its side effects should also be explored, given the evidence of thrombocytopenia, diarrhea, nausea, fatigue, and anemia.
Because HDACs regulate a plethora of cellular functions, their use as multi-targeted therapeutic agents is appealing. Consequently, they have emerged as relevant clinical targets in MM treatment. In trials based on these drugs, especially when administered in combination with a PI, patients have demonstrated impressive outcomes, in particular, those with RRMM.
## Immunotherapy
In the spectrum of new agents that are being developed and tested as potential candidates for the treatment of MM, monoclonal antibodies (mAbs) targeting highly expressed tumor antigens have emerged as a promising strategy. Targeted immunotherapy with mAbs is seen as a revolutionary approach to the successful treatment of many forms of cancer. mAbs designed to act against cell surface proteins such as CD20 (rituximab), HER2 (trastuzumab), and cytokines such as VEGF (bevacizumab) have been critical in the therapeutic strategies for both hematologic malignancies and solid tumors.
The benefit of mAbs stems from their ability to function as anticancer agents through a number of different mechanisms. According to the available evidence, they target cancer cells for destruction by the immune system by engaging immune effector cells via antibody-dependent cellular cytotoxicity or complement-dependent cytotoxicity. Existing research also shows that mAbs are also able to transmit death signals by binding to and cross-linking surface receptors on the target cancer cell or by blocking an activation signal that is necessary for continued cancer growth or viability, thereby inducing apoptosis, as shown in [fig_ref] Figure 2: emerging therapies in MM [/fig_ref]. [bib_ref] Monoclonal antibodies in myeloma, Sondergeld [/bib_ref] [bib_ref] The development of potential antibody-based therapies for myeloma, Sherbenou [/bib_ref] In considering mAbs-based therapy in myeloma, a number of targets have been identified, including components of the bone marrow microenvironment and molecules at the myeloma cell surface. mAbs are able to target the tumor cells directly or indirectly, by interfering with the interaction between the myeloma and the bone marrow stromal cells. 64
## Antibodies against myeloma cells anti-slamf7 (anti-cd2 subset i) mab
Elotuzumab is a humanized recombinant monoclonal IgG1 antibody that targets signaling lymphocyte activation molecule (SLAMF7), also known as CSI. SLAMF7 is a cell surface glycoprotein that is highly expressed on MM plasma cells, and to some extent, on lymphocytes, such as NK cells. Elotuzumab is believed to work through several modes of action, including through targeting the antigen SLAMF7 on MM cells, mediating damage through complementdependent cytotoxicity and antibody-dependent cellular cytotoxicity, and activating SLAMF7-expressing NK cells, which increases tumor destruction. [bib_ref] Monoclonal antibodies in myeloma, Sondergeld [/bib_ref] When administered as a single agent, elotuzumab does not exhibit significant clinical activity. However, its use in combination with other drugs produced promising results in a recent Phase III study, ELOQUENT-2. [bib_ref] ELOQUENT-2 Investigators. Elotuzumab therapy for relapsed or refractory multiple myeloma, Lonial [/bib_ref] In this trial, the researchers compared the efficacy and safety of a combination of elotuzumab, lenalidomide, and dexamethasone with treatment with lenalidomide and dexamethasone in patients with relapsed disease. The trial included 646 patients with relapsed or refractory MM who had previously received a median of two therapies, including BTZ (70%), thalidomide (48%), and lenalidomide (6%). Moreover, based on their OncoTargets and Therapy 2016:9 submit your manuscript | www.dovepress.com
## Dovepress
## Dovepress
## 4853
Salvage therapies in relapsed and/or refractory myeloma medical history, 35% of patients were resistant to their most recent therapy. The median PFS was 19.4 months for patients who received the elotuzumab regimen vs 14.9 months for those who received lenalidomide and dexamethasone alone. However, side effects were more pronounced in the elotuzumab group and included infusion reaction, coughing, fever, diarrhea, fatigue, constipation, decreased appetite, headache, and weight loss. [bib_ref] ELOQUENT-2 Investigators. Elotuzumab therapy for relapsed or refractory multiple myeloma, Lonial [/bib_ref] In December 2015, the FDA approved elotuzumab for use in combination with lenalidomide and dexamethasone in the treatment of patients with MM who have received one to three prior therapies [fig_ref] Table 6: Key trials on emerging novel therapies [/fig_ref].
## Anti-cd38 mabs
Daratumumab is a human IgG1κ monoclonal antibody that targets CD38, an antigen highly expressed on MM cells. [bib_ref] Daratumumab, a novel therapeutic human CD38 monoclonal antibody, induces killing of multiple..., De Weers [/bib_ref] CD38 is a transmembrane glycoprotein found on lymphoid and myeloid cells, which is involved with calcium flux and signal transduction. In several preclinical studies, daratumumab was identified as having unique potent activity against MM. [bib_ref] Monoclonal antibodies in myeloma, Sondergeld [/bib_ref] [bib_ref] The development of potential antibody-based therapies for myeloma, Sherbenou [/bib_ref] [bib_ref] Daratumumab, a novel therapeutic human CD38 monoclonal antibody, induces killing of multiple..., De Weers [/bib_ref] The drug was approved by the FDA in November 2015 for the treatment of MM. This decision was based on the results yielded by the Phase II SIRIUS study involving 106 patients whose disease did not respond to three or more prior therapies, including a PI and an immunomodulatory agent. The overall response to daratumumab was 29%, with three patients experiencing a complete remission during the trial period. The median PFS was 3.7 months, and 65% of patients survived for at least 1 year. The most common side effects of daratumumab were infusion-related reactions, fatigue, nausea, back pain, fever, and cough. These findings
## 4854
Thumallapally et al established daratumumab as the first monoclonal antibody to have single-agent activity and led to its designation as a bespoke agent.SAR650984 (SAR, isatuximab) is a humanized IgG1 monoclonal antibody also targeting CD38. In a first-inhuman Phase I dose-escalation trial, 40 patients with heavily pretreated relapsed/refractory MM received SAR at varying doses. According to the preliminary findings, 33% of patients obtained a clinical benefit and the ORR was 27% in the entire cohort (based on the results pertaining to all dose groups). The encouraging activity seen in this heavily pretreated population prompted further trials in which the efficacy and safety of SAR650984 alone or in combinations with CFZ and pomalidomide are assessed [fig_ref] Table 6: Key trials on emerging novel therapies [/fig_ref].
## Conjugated mabs
Indatuximab ravtansine (BT-062) is another promising mAb. It is an antibody-drug conjugate, comprising the anti-CD138 chimerized mAb (nBT062) and the maytansinoid DM4 as a cytotoxic agent. It is designed to bind to CD138 on cancer cells, releasing DM4 upon internalization to cause cell death. Numerous clinical trials assessing its safety and efficacy are presently being conducted. Preliminary data from the Phase I dose-escalation portion of a Phase I/IIa study of BT-062 in combination with lenalidomide and dexamethasone were reported at American society of hematology annual meeting. [bib_ref] Indatuximab Ravtansine (BT062) In Combination With Lenalidomide and Low-Dose Dexamethasone In Patients..., Kelly [/bib_ref] According to these findings, the ORR among 36 patients evaluated was 78% [fig_ref] Table 6: Key trials on emerging novel therapies [/fig_ref].
## Emerging therapies
Several preclinical studies suggested that, for rapid replication of tumor cells, a synchronized microtubule assembly and disassembly is required. This led to developing agents that disrupt the mitosis spindle causing cell death. Researchers have identified kinesin spindle protein (KSP) as a key target in mitotic process and were successful in developing a novel drug ARRY-520 [fig_ref] Figure 2: emerging therapies in MM [/fig_ref]. It is a potent inhibitor of KSP, and its efficacy as a single agent has been validated in Phase I studies in patients with RRMM. [bib_ref] ARRY-520 Shows Durable Responses in Patients with Relapsed/Refractory Multiple Myeloma in a..., Shah [/bib_ref] [bib_ref] Combination of the KSP inhibitor ARRY-420 with bortezomib or revlimid causes sustained..., Woessner [/bib_ref] With these encouraging results, it was pursued in a Phase II study where patients with RRMM were divided into two cohorts. The first cohort received ARRY-520 only and the second cohort received a combination of ARRY-520 and dexamethasone. ORR was 16% and 22% in cohorts 1 and 2, respectively. [bib_ref] The novel KSP inhibitor ARRY-520 is active both with and without low..., Shah [/bib_ref] Hematologic adverse effects were commonly seen but, most importantly, no PN has been reported. KSP inhibitors hold a promising future for the patients in whom previous therapies have failed to yield benefits, those with RRMM in particular.
## Chimeric antigen receptor t-cell therapy
Recently, chimeric antigen receptor T-cell (CAR T-cell) therapy showed encouraging results in the treatment of MM. CARs are single-chain variable fragments from a monoclonal antibody consisting of intracellular T-cell receptor complex, CD3z chain, or FcR receptor and extracellular domain that recognizes cellsurface-specific antigen. [bib_ref] From humble beginnings to success in the clinic: Chimeric antigen receptor-modified T-cells..., Firor [/bib_ref] The second-and third-generation CAR T-cells also include co-stimulatory domains such as CD28. T-cells are genetically transduced through oncoretroviral and lentiviral vectors to express CARs on their surface. These T-cells are then activated to kill target cells either directly or through the engagement of other components of the immune system. [bib_ref] From humble beginnings to success in the clinic: Chimeric antigen receptor-modified T-cells..., Firor [/bib_ref] A variety of cancers can be targeted by simply substituting various antigen-binding domains on the CARs, encoded by single-chain variable fragments [fig_ref] Figure 2: emerging therapies in MM [/fig_ref].
In MM, CAR T-cells are directed toward various antigens. In 2013, Carpenter et al conducted an experimental design assessment of B-cell maturation antigen expression in normal human tissues and myeloma cells and demonstrated that adoptive transfer of anti-B-cell maturation antigen-CARexpressing T-cells is a promising strategy in the treatment of MM. [bib_ref] The basic principles of chimeric antigen receptor design, Sadelain [/bib_ref] This was followed by several other preclinical and in vivo studies eliciting CAR T-cell therapy as potential therapy for myeloma. [bib_ref] B-cell maturation antigen is promising target for adoptive T-cell therapy of multiple..., Carpenter [/bib_ref] [bib_ref] Gene modified T cells as immunotherapy for multiple myeloma and acute myeloid..., Peinert [/bib_ref] [bib_ref] CS1-specific chimeric antigen receptor(CAR)-engineered natural killer cells enhance in vitro and in..., Chu [/bib_ref] Although CAR T-cell therapy appears as a very reassuring strategy in tackling MM, there are certain challenges. These include nonspecific toxicity, as many of the surface molecules are expressed in normal tissues. For example, CD38 targeted in MM is also present in hematopoietic stem cells, NK cells, dendritic cells, pancreatic islet cells, and prostrate; CD138 is present in bronchial epithelia; CD56 in central nervous system neurons. Immunogenicity of CARs, cytokine release syndrome, and immune escape phenomena where tumor cells escape despite the presentation of antigen, are other drawbacks. The technicality and cost further add to these issues. Finally, factors that limit the usage of CAR T-cell therapy include uncertainty of the cell dose, timing of cell infusion and withdrawal, type of immune cell to be used for CAR T-cell delivery, and the durability of CAR T-cell-induced remissions.
## Vaccine therapy
Vaccine therapy is the newest in the category of the emerging therapies in the treatment of MM. Different types of vaccines are utilized, including cell-based vaccines, protein vaccines, idiotype vaccines, and ASCT and donor vaccines. The underlying mechanism of action is based on the fact that vaccines stimulate immune response against tumor-specific OncoTargets and Therapy 2016:9 submit your manuscript | www.dovepress.com
## Dovepress
## Dovepress
## 4855
Salvage therapies in relapsed and/or refractory myeloma antigens expressed by MM clone such as WT-1, CS-1, and idiotype protein, thereby attacking MM cells. [bib_ref] Effector memory and central memory NY-ESO-1-specific re-directed T cells for treatment of..., Schuberth [/bib_ref] Benefit of vaccine therapy is demonstrated in multiple Phase II trials. Lacy et al 81 compared 27 patients with myeloma after auto-Hematopoietic stem cell transplant (HCT) who were vaccinated with antigen-presenting cells pulsed with idiotype protein with 124 patients who were not vaccinated after auto-HCT. OS in the vaccine group was 5.3 years vs 3.4 years (P=0.02). Another Phase II trial showed that 24% patients who showed PR to ASCT showed complete/near complete response following vaccine administration. [bib_ref] Dendritic cell cancer vaccines: from the bench to the bedside, Katz [/bib_ref] A prior study has demonstrated an augmenting effect of IMiDs such as lenalidomide in patients treated with vaccines. [bib_ref] Lenalidomide-Induced Immunomodulation in Multiple Myeloma: Impact on Vaccines and Antitumor Responses, Noonan [/bib_ref] Despite these significant advances made in this arena, it is still in a budding stage and will unfold its tale in the near future with the help of current ongoing research.
## Factors associated with treatment selection
Treatment horizons in MM have expanded significantly over the past decade. However, even with extensive research, prior and current, there is often apprehension in choosing the most appropriate salvage regimen as there are no preset guidelines. [bib_ref] Treatment option for relapsed and refractory multiple myeloma, Nooka [/bib_ref] summarizes the key aspects when physicians arrive at cross-roads in making decisions on therapies. 84
# Conclusion
MM is heterogeneous, associated with complex gene abnormalities, and multiple signaling anomalies. Thus, it remains a serious malady, whereby most patients eventually succumb to disease. Several new drugs with novel mechanism of action and less toxic profile have been developed in the past decade, with the potential for use as single agents or in synergy with other treatment modes in MM therapy. Although these therapies produced better responses compared with traditional chemotherapy in controlled trials, the efficacy and safety of these novel drugs have not been sufficiently investigated, making it difficult to determine whether one agent is superior to another. Consequently, selecting the most appropriate initial management remains a challenge.
In particular, when treating relapsing patients, one of the most important goals is enhancing their quality of life (QoL). However, this kind of evaluation is often lacking in many studies investigating these novel therapies. In addition, as QoL is a subjective measure, developing tools for accurate assessment remains a challenge. Moreover, the discussions presented earlier revealed that most newly developed drugs target specific mechanisms of neoplastic cell growth. Thus, when considering their efficacy, it is important to highlight that myeloma is a highly heterogeneous disease, whereby neoplastic plasma cells can use several metabolic pathways in order to maximize growth potential. In addition, available evidence also indicates that different neoplastic clones may emerge in different phases of disease and it is possible that each clone has a different profile of drug sensitivity. Hence, it is conceivable that each of the drugs that have emerged as potential MM treatment candidates is effective in a subgroup of patients only and exhibits benefits only during a specific phase of disease. Given these limitations, the challenge for future investigations is identifying the specific range and timing of activity of each new drug.
Finally, it is essential for clinicians to recognize that the mechanism of action of these new drugs is typically Factors to consider in choosing salvage regimen Patient-related factors
- Assess for comorbidities such as diabetes and heart failure especially in elderly population as they are more vulnerable to drug toxicity - Assess for renal impairment as many novel drugs need adjustment. BTZ and CFZ do not need adjustment while LeN should be used with caution. Interestingly, POM can be used in full doses without any dose reduction in renal insufficiency - Assess for hepatic impairment. Cautious use of Pi and POM is recommended Treatment-related factors - Assess prior received therapies and toxicity associated with them. Use combination drugs that exhibit sensitization and synergistic activity such as HDACi along with Pi, and Pi with LeN - Avoid the same drugs that caused toxicity. if PN is an issue, choose CFZ instead of BTZ in salvage regimens - Duration of prior response is always critical when deciding on salvage regimens. intensive therapy is used if relapse occurs within last 12 months of prior therapy with triple regimen (PI, IMiDs with DEX) Disease-related factors
- Clinicians should know about high-risk cytogenetics. Most importantly, t(4;14) and del(17p). BTZ has better prognosis in both t(4;14) and del(17p). POM showed longer survival in del(17p) patients than in t . Several secondary variations in genetics have been identified (MYC dysregulation, del(18p)). However, studies are underway in elucidating the benefit of novel drugs in patients harboring high-risk genetics Availability of clinical trial
## 4856
Thumallapally et al different from that exhibited by chemotherapeutic drugs. Thus, when determining the optimal dose and identifying patients who would benefit the most from these novel therapies, it is unwise to draw parallels with traditional treatment modalities. In addition, when deciding whether to use these emerging agents, the proven efficacy of "old" chemotherapy against myeloma should not be overlooked. In many cases, it is likely that the patient would benefit the most from adding new treatments to the chemotherapy. Thus, the goal should not necessarily be to find completely new treatment options, but rather work on establishing the most advantageous combination of traditional and novel therapies.
As a general rule, physicians are advised to choose therapy based on patient's age, comorbidities, cytogenetic abnormalities, and locally available drugs, while aiming to enhance QoL. However, owing to rapid medical advances, they should also keep abreast with the latest developments and ongoing clinical trials, as these may also be potential avenues to pursue with their patients, if eligible for inclusion. Based on the discussions presented in this work, it can be concluded that the key to MM treatment may be combining various medications with different mechanisms of action to target not only the MM cells but also the tumor microenvironment.
# Disclosure
The authors report no conflicts of interest in this work.
## Oncotargets and therapy
Publish your work in this journal Submit your manuscript here: http://www.dovepress.com/oncotargets-and-therapy-journal OncoTargets and Therapy is an international, peer-reviewed, open access journal focusing on the pathological basis of all cancers, potential targets for therapy and treatment protocols employed to improve the management of cancer patients. The journal also focuses on the impact of management programs and new therapeutic agents and protocols on patient perspectives such as quality of life, adherence and satisfaction. The manuscript management system is completely online and includes a very quick and fair peer-review system, which is all easy to use. Visit http://www.dovepress.com/testimonials.php to read real quotes from published authors.
[fig] Figure 1: Mechanism of action of Pis, immunomodulators, and HDACi. Notes: Proteasomes are intracellular structures which catabolize proteins that are marked with ubiquitin. Pi block this mechanism of action causing accumulation of proteins triggering cell death. New molecular target identified for IMiD is cereblon, which is a key component of E3 ubiquitin ligase. This binding catalyzes ubiquitination of transcription factors iKZF1/iKZF3 eventually degrading them. iMiDs also modulate and inhibit angiogenesis and activate NK cells. HDAC enzymes regulate acetylation of the N-terminals of histones and other transcription factors. This is blocked by HDACi causing downregulation of protein expression. HDAC is also known to activate aggresome-proteasome pathway, which is blocked by HDACi.Reprinted by permission from Macmillan Publisher Ltd: Bone Marrow Transplant. Cornell RF, Kassim AA. evolving paradigms in the treatment of relapsed/refractory multiple myeloma: increased options and increased complexity. 2016;51(4):479-491. 95 Copyright 2016. Abbreviations: DAC, deacetylase; DACi, deacetylase inhibitor; HDAC, histone deacetylase; HDACi, histone deacetylase inhibitor; iMiD, immunomodulatory drug; MM, multiple myeloma; MOA, mechanism of action; NK, natural killer; Pi, proteasome inhibitor. OncoTargets and Therapy 2016:9 submit your manuscript | www.dovepress.com Dovepress Dovepress [/fig]
[fig] Figure 2: emerging therapies in MM. Notes: Current agents under development: CAR T-cells recognize target tumor cells and induce cell death. mAbs with the help of antibody-dependent cell-mediated toxicity induce apoptosis causing cell lethality. Oncolytic virotherapy acts by inducing direct virus-mediated cytotoxicity along with indirect enhancement of host immune responses. KSP inhibitors serve as antimitotic agents in rapidly dividing cells. Reprinted by permission from Macmillan Publisher Ltd: Bone Marrow Transplant. Cornell RF, Kassim AA. Evolving paradigms in the treatment of relapsed/refractory multiple myeloma: increased options and increased complexity. 2016;51(4):479-491. 95 Copyright 2016. Abbreviations: ADCC, antibody-mediated cell toxicity; CAR, chimeric antigen receptor; KSP, kinesin spindle protein; mAb, monoclonal antibody; MM, multiple myeloma; MOA, mechanism of action; NK, natural killer; TCR, T cell receptor. submit your manuscript | www.dovepress.com Dovepress Dovepress [/fig]
[fig] •: Always consider enrolling patient in clinical trial if they are eligible Note: Data from Nooka et al. 84 Abbreviations: BTZ, bortezomib; CFZ, carfilzomib; DEX, dexamethasone; HDACi, histone deacytalase inhibitors; IMiDs, immunomodulatory drugs; LEN, lenalidomide; Pi, proteasome inhibitor; POM, pomalidomide; PN, peripheral neuropathy. submit your manuscript | www.dovepress.com Dovepress Dovepress [/fig]
[table] Table 1: Snapshot of current and upcoming therapies Abbreviations: BTZ, bortezomib; CFZ, carfilzomib; CAR, chimeric antigen receptor; HDACi, histone deacetylase inhibitor; KSP, kinesin spindle protein; Pi, proteasome inhibitor. [/table]
[table] Table 2: Revised diagnostic criteria for MM [/table]
[table] Table 3: Defining myeloma [/table]
[table] Table 4: Landmark trials of BTZ [/table]
[table] Table 5: Key trials of second-generation Pis and immunomodulatorsAbbreviations: BTZ, bortezomib; CFZ, carfilzomib; DEX, dexamethasone; IMiD, immunomodulatory drug; LEN, lenalidomide; NA, not available; ORR, overall response rate; OS, overall survival; PFS, progression-free survival; Pi, proteasome inhibitor; POM, pomalidomide; THAL, thalidomide. submit your manuscript | www.dovepress.com [/table]
[table] Table 6: Key trials on emerging novel therapies [/table]
|
Substance abuse among high school learners in a rural education district in the Free State province, South Africa
Background: In South Africa, many studies conducted on substance abuse among in-school youth focus on urban areas. However, anecdotal evidence suggests that rural areas are experiencing an increase in substance abuse, though there is dearth of studies in these areas.Methods:This study used a quantitative design to collect data from 629 high school learners who were in Grades 10 and 11 in public schools in rural Free State Province, to determine the prevalence of, and factors associated with substance use.Results:The sample consisted of 46% males and 54% females. Their ages ranged from 14 to 20 years, with a mean of 16.9 years. The prevalence of substance abuse was 47% (n = 295) with alcohol consumption, cigarette and dagga smoking being the most common substances used. Socio-demographically, age and gender were significantly associated with substance abuse. While behavioural variables of physical fights, serious problems with parents and friends, poor academic performance, trouble with police, having sex without condom, and having sex and regretted the next day were significantly associated with substance abuse (p = 0.05).Conclusion:The prevalence of substance abuse is very high for this rural school community, which highlights the need to pay attention to rural schools regarding substance abuse challenges.
# Introduction
South Africa has a high rate of substance abuse among young people, which includes both in and out of school youth. Despite reported stigma and associated reluctance to seek treatment for substance abuse,an increase in young people aged 20 years seeking treatment for substance abuse, 2 is an indication of the gravity of the problem. Variations in the trends of substance use have been reported between urban and rural youth,with substance use among learners in urban area being more than those in rural areas.However, rural areas are increasingly experiencing problems of adolescent substance use.Substance abuse among learners is associated with a range of criminal acts,which includes violence and bullying.It has also been associated with a range of mental disorders; while socially, it is associated with social disorganisation, deviant behaviour, and social interaction with deviant groups,depending on which substances are used. Risky sexual behaviours are also likely to manifest into sexually transmitted infections.
In the school environment and academic context, substance abuse has been associated with challenges in school discipline,appetitive aggression 7 and other classroom management challenges.These frustrate the achievement of intended education outcomes, and result in poor academic performance, including possible dropping out of school.
The social environment is often a significant determinant of substance abuse,with adolescents being often influenced by their peers.Other determinants of drug use among young people include: curiosity, sense of growing up, amount or lack of parental discipline and monitoring, and family cohesion.Availability and easy access of illicit drugs within the community or the household, 17 economic hardship, high unemployment, lack of adequate social support networks, pressure to meet daily family needs, family conflicts,were all associated with use of illicit drugs.
In the context of schooling, substance abuse has been significantly associated with poor academic performance,which often results in dropping out of school.The perceptions that rural areas experience less substance abuse among learners often result in inadequate attention being paid to such areas. With the precise intention of expanding substance abuse studies to areas other than cities, the purpose of this study was to determine the prevalence of substance abuse, as well as explore the associations between substance abuse and a range of demographic variables among learners in a rural school district of Free State Province, South Africa.
# Methodology
## Study design
A cross-sectional quantitative survey, using a self-developed questionnaire, was conducted among learners attending local high schools in the rural school district of Free State Province, South Africa.
## Study setting
The study was conducted at high schools at Setsoto local municipality, Thabo Mofutsanyane District of the Eastern Free State, which is largely rural. The municipality is comprised of four rural towns namely Clocolan/ Hlohlolwane, Marquard/Moemaneng, Senekal/Matwabeng, and Ficksburg/Meqheleng. According to the Census 2011, the municipality has a population of 1 10 335, and of those aged 20 years and above approximately 8.7% have no formal schooling, 22.6% have completed matric, and 6.9% have some form of tertiary education. Agriculture is the main economic activity in the municipality, and the unemployment rate of those aged between 15 and 34 is high at 46%.
## Study population and sample
The study population was high school learners in public schools of Setsoto municipality. There are eight public schools in the sub-district, and using the hat method, four schools from each rural town were randomly selected. An additional school was used for the pilot study. The estimated population of Grades 10 and 11 in the eight schools is 2100.
## Sample and sampling technique
The sample consisted of learners who were in Grades 10 and 11 at the time of data collection. From estimated population size of 2100 from eight schools, the Raosoft sample calculator was used to determine a minimum sample for the study. Using a 5% margin of error, a confidence level of 95% and a distribution of 50%, a minimum sample size of 323 was calculated. Because a survey was used, in which all learners willing to participate in the study were invited, 800 learners in 42 classrooms participated, but 629 were analysed, with the rest having missing information of more than 10%.
## Recruitment
Recruitment was done at the identified school, with the researcher addressing the Grades 10 and 11 learners by telling them about the study and requesting them to participate. Those who agreed to participate were given letters for their parents to provide informed consent.
## Data collection tool
An English self-administered questionnaire, which was modified from a risk behaviour survey, was used to collect data. The tool was pilot tested among 20 learners at another school before data collection. The tool collected learnerrelated demographic data like age, gender, grade, and whether they have ever repeated a class, as well as the socioeconomic data of the family, such as employment status of parents, highest education attained by the parents and who the participants live with. Substance use related data collected included the substances of current use, age at which they first experimented with substances, use of substances in their social environment and ease of access of substances. Behaviour related data included whether they were involved in physical fights, were in trouble with police, were engaged in risky sexual behaviour, (such as having sex without using a condom or having sex and regretted it the next day), and problems with parents and friends.
## Data collection
On the day of data collection, learners whose parents had provided the informed consent were assembled in the school hall or classroom and an explanation about the study was repeated. The learners were given an opportunity to ask questions or seek clarification. Informed consent was administered to learners who were over the age of 18, while learners under the age of 18, whose parents had provided consent, were requested to provide assent by signing the appropriate forms. The data collection tool was then distributed to all the learners. Adequate time was given to complete the questionnaires, and the learners left the venue after all had completed the process.
# Data analysis
The data were captured into Microsoft Excel and transported to STATA version 13 for analysis. Descriptive statistics were used to analyse socio-demographic data prevalence of substance use, and these were reflected in the form of frequencies and percentages. Chi-square test was used to explore associations between a range of demographical variables and substance abuse among the sample. Statistical significance was set at ≤ 0.05.
## Validity, reliable and bias
At sub-district district level, selection bias was minimised by random selection of schools. The use of the survey at school minimised selection bias as all learners in the selected grades who were prepared to participate were included in the study.
https://www.safpj.co.za Open Access
The questionnaire was pilot tested to identify any challenges before the actual data collection commenced.
# Ethical considerations
Ethical approval for the study was obtained from the Sefako Makgatho Health Sciences University Research and Ethics Committee (number: SMUREC/H/95/2016). Permissions to conduct the study were obtained from the offices of the Provincial Department of Education, the Thabo Mofutsanyane District of Education, the Setsoto sub-district and the management of each participating school. Informed consent was obtained from parents for participants who were younger than 18 years of age, and these minors provided assent to participate in the study. Informed consent was obtained from participants who were 18 years and above.
# Results
## Characteristics of the sample
Eight hundred (800) learners participated in the survey, and of these, 629 were analysed, with the others being excluded because of missing data of 10% or more. Of the 629 students whose data were analysed, more than half (55%) were in the age group 16-17 years followed by those aged 18 years and above (33%). Their ages ranged from 14 to 20 years, with a mean of 16.9.
## Prevalence and type of substance use among learners
The prevalence of substance use among the study participants was 47%, and of those using substances, the highest proportion consume alcohol (87%) followed by cigarette (45%) and (24%) dagga smoking.
## Factors associated with substance use
The association between substance use and selected demographic characteristics are shown in. Substance use significantly increased with age (p < 0.05), while the prevalence was significantly higher among males than females (53% vs. 42%, p < 0.05). A significant higher proportion of the participants were introduced to substance use by friends (p < 0.05). There was no statistical association between substance use and grades, employment status of the parent, and whether they stay with their parents or not.illustrates behaviours that were significantly associated with substance use.
# Discussion
The purpose of the study is to determine the prevalence of substance abuse, as well as explore the association between substance abuse and a range of demographic variables. The prevalence of substance abuse among this sample is high at 47%, which is close to the 47.9% reported in a similar sample in Ethiopia,but higher than the 6% reported in another study conducted in a rural setting in South Africa.The finding that many learners started using substances at a young age of 15 years confirms findings of a previous study conducted in South Africa.Although cigarettes and alcohol are legal, they are still illegal for minors such as most of the sample. Of greater concern is the use of illicit drugs such as dagga, nyaope, ecstasy and cocaine, which indicates Note: The total exceeds 100% because some participants reported more than one substance.
https://www.safpj.co.za Open Access criminality as these are prohibited substances. Dagga is easily cultivated and commonly used in South Africa, and nyaope is a cocktail drug that has destroyed many lives among Black communities in South Africa because of its high addictive characteristics.Adolescents who use drugs have been reported to have significantly lower levels of psychological well-being and life satisfaction, 31,32 which implies mental and social risks for the sample. Although peer pressure may influence young people to use drugs, they still feel guilty and stigmatised by family and community, which increases the chances of social illhealth,which increases the shame associated with the behaviour of using substance.The challenge of substance abuse should therefore be understood comprehensively as a problem of adolescent social ill-health.
Males had a significant higher prevalence of substance use than females (p = 0.007), which is similar to previous studies which reported that males were up to 10 times more likely to use substances than females.The finding that older learners are more likely to use substances than younger ones is similar to a study conducted among learners in the Western Cape, which reported the odds ratio of 1.6 among older learners.Also similar is the finding that substance use by other members of the household and friends, increases the risk of use among learners two-fold.These findings highlight the need for comprehensive interventions to influence the comprehensive well-being of young people, especially among young learners. Such interventions also need a community component,which is likely to improve the effectiveness of substance use prevention amongst learners, and thus improve the overall well-being of these young people.
Substance use has been shown to be associated with poor academic performance,a serious barrier to reaching the goals of the education system. The finding of a significant association between substance use and risky sexual behaviourconfirms the negative impact of substance abuse on overall youth health.
The significant association of substance abuse with a range of anti-social behaviours of physical fights, serious problems with parents and friends, poor academic performance, trouble with police, having sex without condom and having sex and regretted the next day, all with p-values of 0.001, are similar to the findings reported in previous studies, which reported statistically significant associations ranging from p-values of 0.001-0.05.These findings put substance abuse at the centre of various problems experienced at South African schools and communities. These associations also identify the need to target substance abuse as a barrier to overall social development because the outcomes, be they academic, physical health, social and/or mental have long-term implications for the affected learners.
Alcohol, cannabis and cigarettes were found to be the most commonly used substances, which is similar to another study conducted in Durban, South Africa.The ease of access for these substances increases the levels of challenges as this cannot be addressed without the involvement of other sectors, including the law enforcement and behavioural scientists. The complexity of the situation also indicates the urgency of stakeholders to work together to develop interventions that are focused on both prevention and management. Currently, there are no such interventions accessible by the general learner in any South African publicschool setting.
Despite the challenges of substance abuse in schools, it does not seem that the Department of Education has any specific intervention to address the problem, other than relying on the Life Orientation (LO) learning area, which aims to address a wide range of learner developmental areas. which include personal, psychological, neuro-cognitive, motor, physical, moral, spiritual, cultural and socio-economic areas.However, the delivery of LO has been reported to have serious challenges because of constraints at the individual, interpersonal, school, and community levels.Another shortcoming of LO as a resource for substance abuse is that it is general, and does not address personal experiences,and therefore is limited in assisting learners with substance abuse challenges. Of serious concern is that the Department of Basic Education's policy on management of substance abuse in schools is neither known nor implemented,which implies that there is not much at school level that addresses the serious problem of substance abuse among learners.
## Limitations of the study
As with other survey studies, 21% of the questionnaires could not be analysed because of missing data. However, the high response rate of 78% and a relatively large sample size of 629 counteracted the non-usable questionnaires. A limitation which applies to other survey studies is the response bias, in which the sample may under-or overestimate the population parameter. However, this potential bias was minimised by sampling from various schools and more than one class in a school. Societal lack of approval of substance abuse by learners may have contributed to bias in their responses, but this was minimised by the privacy afforded to the participants, which meant that others would not have known about their responses.
# Conclusion
The result of the study contributes to highlight the need for interventions to address the challenge of substance abuse in schools, which will improve the academic outcomes with long-term social and career impacts. As substance abuse is more of a societal rather than just a school's challenge, the required interventions need not be limited to schools, but extend to other young people in communities, including rural areas.
## Recommendations
It is recommended that the substance abuse problem be outsourced to public health and/or behavioural health specialist and not be left to the Department of Education, as this is not their focus areas. This will enable consistent application, monitoring and evaluation of such interventions, and enhance the implementation of necessary modifications. |
Presence of Li Clusters in Molten LiCl-Li
## Supporting information
The Raman spectra of LiCl-Li 2 O-Li were recorded with varying incident laser power between 0 and 10mW to quantify the dependence of the signal on the power of the laser. 1 shows the spectra recorded at various laser powers, along with a plot of the intensity of the Raman feature at 302 cm -1 as a function of laser power. A linear response was observed with an R-squared value of 0.979. To investigate the stability of the LiCl-Li melt as a function of time, Raman spectra of LiCl-Li were recorded 5, 15, 45, and 90 minutes after adding 1-wt% Li to molten LiCl. The spectra recorded at these times are shown in 2. Melts of LiCl-Li were investigated spectroscopically in Mo and Ta crucibles to demonstrate that the observed Raman spectrum of the melt was independent of the crucible material. Raman spectra of LiCl-1-wt% Li recorded in Ta and Mo crucibles are shown in 3. The Raman spectrum shown inwas restricted to a narrow spectral band for clarity. [fig_ref] Figure SI 4: Raman spectrum of LiCl-Li 2 O-Li at 923 K shown inFigure 1across... [/fig_ref] shows the spectrum inacross a larger spectral range, and it includes both the Raman features attributed to the presence of Li 8 as well as the Na fluoresce lines exhibited in [fig_ref] Figure SI 2: Raman spectra of LiCl-Li at 923 K recorded 5, 15, 45, and... [/fig_ref].
[fig] Figure SI 1: (Top) Raman spectra of LiCl-Li at 923 K recorded with an incident laser power increasing from 0 to 10 mW. (Bottom) Plot of the intensity of the 302 cm -1 Raman shift as a function of incident laser intensity, showing the linear dependence of signal intensity on the laser power. [/fig]
[fig] Figure SI 2: Raman spectra of LiCl-Li at 923 K recorded 5, 15, 45, and 90 minutes after adding 1-wt% Li to LiCl. The minimal variation in signal intensity over 90 minutes demonstrates the quasi-stability of the LiCl-Li mixture. [/fig]
[fig] Figure SI 3: Raman spectra of LiCl-Li melt at 923 K contained in Mo and Ta crucibles. The results demonstrate that the observed spectrum is independent of the crucible material. [/fig]
[fig] Figure SI 4: Raman spectrum of LiCl-Li 2 O-Li at 923 K shown inFigure 1across a larger spectral range. The spectrum exhibits the Raman modes attributed to the presence of Li 8 as well as Na fluoresce. [/fig]
|
Transferable Coarse-Grained Potential for De Novo Protein Folding and Design
Protein folding and design are major biophysical problems, the solution of which would lead to important applications especially in medicine. Here we provide evidence of how a novel parametrization of the Caterpillar model may be used for both quantitative protein design and folding. With computer simulations it is shown that, for a large set of real protein structures, the model produces designed sequences with similar physical properties to the corresponding natural occurring sequences. The designed sequences require further experimental testing. For an independent set of proteins, previously used as benchmark, the correct folded structure of both the designed and the natural sequences is also demonstrated. The equilibrium folding properties are characterized by free energy calculations. The resulting free energy profiles not only are consistent among natural and designed proteins, but also show a remarkable precision when the folded structures are compared to the experimentally determined ones. Ultimately, the updated Caterpillar model is unique in the combination of its fundamental three features: its simplicity, its ability to produce natural foldable designed sequences, and its structure prediction precision. It is also remarkable that low frustration sequences can be obtained with such a simple and universal design procedure, and that the folding of natural proteins shows funnelled free energy landscapes without the need of any potentials based on the native structure. OPEN ACCESS Citation: Coluzza I (2014) Transferable Coarse-Grained Potential for De Novo Protein Folding and Design. PLoS ONE 9(12): e112852.Recently we have presented many results that point to the existence of a ''maximum valence principle'' (MVP)[41][42][43], according to which for a heteropolymer to be designable and foldable it is sufficient that chain is decorated with directional (low valence) interactions that shape the configurational space. In De Novo Protein Folding and Design PLOS ONE |
# Introduction
Computer simulations of the protein folding process have in the last ten years reached amazing level of description and accuracy [bib_ref] Alpha-Helix Formation: Discontinuous Molecular Dynamics on an Intermediate-Resolution Protein Model, Smith A [/bib_ref] [bib_ref] Atomistic protein folding simulations on the submillisecond time scale using worldwide distributed..., Pande [/bib_ref] [bib_ref] Accelerated molecular dynamics: a promising and efficient simulation method for biomolecules, Hamelberg [/bib_ref] [bib_ref] Reproducible polypeptide folding and structure prediction using molecular dynamics simulations, Seibert [/bib_ref] [bib_ref] Coarse-grained models for proteins, Tozzini [/bib_ref] [bib_ref] Dynamic personalities of proteins, Henzler-Wildman [/bib_ref] [bib_ref] Effects of hydrophobic and dipole-dipole interactions on the conformational transitions of a..., Mu [/bib_ref] [bib_ref] Convergence of folding free energy landscapes via application of enhanced sampling methods..., Huang [/bib_ref] [bib_ref] Metadynamics: a method to simulate rare events and reconstruct the free energy..., Laio [/bib_ref] [bib_ref] Generic coarse-grained model for protein folding and aggregation, Bereau [/bib_ref] [bib_ref] Atomic-level characterization of the structural dynamics of proteins, Shaw [/bib_ref] [bib_ref] Enhanced sampling and applications in protein folding in explicit solvent, Zhang [/bib_ref] [bib_ref] Ab initio simulation of a 57-residue protein in explicit solvent reproduces the..., Ikebe [/bib_ref] [bib_ref] How Fast-Folding Proteins Fold, Lindorff-Larsen [/bib_ref] [bib_ref] Folding helical proteins in explicit solvent using dihedral-biased tempering, Zhang [/bib_ref] [bib_ref] Folding and stability of helical bundle proteins from coarse-grained models, Kapoor [/bib_ref]. The power of the computers and the understanding of the physics that governs folding allows now for a large screening of the experimental data for instance collected in the Protein Data Bank [bib_ref] The Protein Data Bank, Berman [/bib_ref]. From a theoretical point of view a successful approach is the''minimal frustration principle'' (MFP) [bib_ref] Spin glasses and the statistical mechanics of protein folding, Bryngelson [/bib_ref] [bib_ref] Engineering of stable and fast-folding sequences of model proteins, Shakhnovich [/bib_ref] [bib_ref] From the Cover: Chemical physics of protein folding, Wolynes [/bib_ref] in which protein folding is described as a downhill sliding process in a low frustration energy landscape (''funnelled'' shaped) towards the native state. While MFP has been proven for lattice heteropolymers [bib_ref] Engineering of stable and fast-folding sequences of model proteins, Shakhnovich [/bib_ref] [bib_ref] Ground-state of random copolymers and the discrete Random Energy-model, Gutin [/bib_ref] [bib_ref] Proteins with selected sequences fold into unique native conformation, Shakhnovich [/bib_ref] [bib_ref] Implications of thermodynamics of protein folding for evolution of primary sequences, Shakhnovich [/bib_ref] [bib_ref] Designing refoldable model molecules, Coluzza [/bib_ref] [bib_ref] Designing specificity of protein-substrate interactions, Coluzza [/bib_ref] [bib_ref] Monte Carlo study of substrate-induced folding and refolding of lattice proteins, Coluzza [/bib_ref] [bib_ref] Disordered flanks prevent peptide aggregation, Abeln [/bib_ref] , in more realistic protein representations a residual frustration which prevents the systematic prediction of the native structure of natural sequences is often observed. Off-lattice instead MFP is used as a main justification for the use of structure-based potentials such as the GO [bib_ref] Theoretical studies of protein folding, Go [/bib_ref] and elastic models [bib_ref] Anisotropy of fluctuation dynamics of proteins with an elastic network model, Atilgan Ar [/bib_ref]. In fact, there is still space for development of transferable models that are capable of systematic associating the experimentally determined native structure to natural sequence. Surprisingly, with the exception of few notable examples [bib_ref] Probing the role of packing specificity in protein design, Dahiyat [/bib_ref] [bib_ref] De novo design of the hydrophobic cores of proteins, Desjarlais [/bib_ref] [bib_ref] Side-chain and backbone flexibility in protein core design, Desjarlais [/bib_ref] [bib_ref] Construction of new ligand binding sites in proteins of known structure, Hellinga [/bib_ref] [bib_ref] Collective many-body van der Waals interactions in molecular systems, Distasio [/bib_ref] [bib_ref] Kemp elimination catalysts by computational enzyme design, Rothlisberger [/bib_ref] [bib_ref] De Novo Protein Design: Fully Automated Sequence Selection, Dahiyat [/bib_ref] [bib_ref] Design of a novel globular protein fold with atomic-level accuracy, Kuhlman [/bib_ref] [bib_ref] Evaluation and optimization of discrete state models of protein folding, Kellogg [/bib_ref] , it has also been extremely difficult to artificially construct sequences capable of folding into given target protein structures. The group of David Baker [bib_ref] Evaluation and optimization of discrete state models of protein folding, Kellogg [/bib_ref] introduced a novel procedure to select sequences with low frustration capable of correctly refolding in vitro to their target structure with a success rate between 8% and up to 40% of the total trials. In their work the authors have introduced a set of rules for the design of the local amino acids interactions to disfavour non-native states. After many iterations, a refolding calculation filters out about 90% of the initial sequences that are found not to have a funnelled energy landscape. The complexity of Baker's procedure demonstrates that is not easy to produce sequences with low frustration.
Here we present a novel protein model where low frustration folding is observed both for natural and designed sequences, the latter obtained without the need of negative design. The novel model is obtained from the optimization of the residue-residue and residue-solvent interaction energy terms under the condition that a large number of sequences designed for 125 test proteins are equal to the corresponding natural sequences. As a result, designed sequences with our model are for several properties comparable to natural ones and fold with a low frustration free energy landscape. We additionally demonstrated that for 15 additional randomly selected proteins, notoriously difficult to fold [bib_ref] An improved protein decoy set for testing energy functions for protein structure..., Tsai [/bib_ref] [bib_ref] CASP9 target classification, Kinch [/bib_ref] , the natural sequences correctly refolded to their corresponding native structures with a remarkable precision between 2.5 and 5 Å . In other words both quantitative protein design and folding are possible simultaneously. We anticipate that our methodology will have direct application for protein design and structure prediction, but also we expect that it will become a reference point for the development of alternative protein models. For instance, a more or less accurate description can be obtained by adding or removing details from our model, under the condition that the maximum valence principle remains satisfied. the case of proteins we have shown (Caterpillar model [bib_ref] A coarse-grained approach to protein design: learning from design to understand folding, Coluzza [/bib_ref] that the minimum set of directional interactions translates into the combination of just the backbone molecular geometry and the backbone hydrogen bond interactions (see [fig_ref] Figure 1: Real-space representation of the backbone of the Caterpillar model [/fig_ref].
In what follows we will show that by optimizing the interactions under the condition that natural and designed sequences are the same at constant amino acid composition for a large set of proteins, we will also quantitatively predict the folded structures of natural and designed sequences with similar accuracy. This is possible because our model includes the correct set of interactions that satisfy the MVP and, accordingly, the design procedure [bib_ref] A coarse-grained approach to protein design: learning from design to understand folding, Coluzza [/bib_ref] alone is capable of predicting if a sequence, either natural or artificial, will fold to the target structure. Since we cannot model the particular evolutionary pressure that determined the natural amino acid composition, we chose to keep it constant. Such pressure could be due to many factors such as the particular function of the protein or the difficulty of synthesizing each amino acid type. The ansatz of this work is that folding and design can occur also outside such conditions and that is possible to design a foldable artificial protein from an infinite bath of amino acids. Hence, the above evolutionary pressure is taken into account by fixing the composition to the natural one. The optimization scheme that we used is the maximum entropy principle (MEP) already tested for proteins by Seno et al. [bib_ref] Maximum Entropy Approach for Deducing Amino Acid Interactions in Proteins, Seno [/bib_ref]. MEP states that the more information is used to model a system the lower the associated entropy will be [bib_ref] A Mathematical Theory of Communication, Shannon [/bib_ref]. Hence, in order to find the optimal parameters that require the least amount of information, all is needed is to maximize the entropy associated with the probability of observing a given protein P(S i ,C j ), where S i indicates to the sequence and C j the three dimensional structure, under the sequence similarity and normalization constraints defined in work of Seno et al. [bib_ref] Maximum Entropy Approach for Deducing Amino Acid Interactions in Proteins, Seno [/bib_ref]. The derivation follows closely the one used in the work of Seno et al. [bib_ref] Maximum Entropy Approach for Deducing Amino Acid Interactions in Proteins, Seno [/bib_ref] (the full derivation is in the Supplemental Material together with the details about the model and simulations techniques) and we determined that the entropy maximum corresponds to the values of the model parameters (E, E HOH , and V in Eq. 1) at which the amino acid hydrophobic/hydrophilic (HP) profileand the interaction energy of each residue with all other are simultaneously equal to the natural ones:
[formula] F score~X N Prot j X N j k X N Seq i P(S i ,C j )E ik Sol {E Real jk Sol 0 @ 1 A 2 z X N Prot j X N j k X N Seq i P(S i ,C j )c i k {c Real j k 0 @ 1 A 2 zE Shannon X H(E) log H(E)ð1Þ [/formula]
where the index i runs over the N Seq designed sequences for each protein j of length N j , the E Sol is the hydrophobicity scale of each residue (see Eq.S3 in the SM), while the c i k 's are the contribution to the total energy of each residue calculated within the Caterpillar model. The last term instead guarantees that the Shannon entropy associated to the matrix elements E kl (H are the histograms) is maximized to avoid an uniform matrix. We phenomenologically determined E Shannon~8 :0 for the scaling term to be a good value. Note that here and in the following, energies are given in units of k B T Ref , where T Ref is a reference temperature that sets the scale of the interactions, hence all simulation temperatures are given in units of T Ref . It is important to stress that T Ref is not necessarily the folding temperature or the environment temperature, but all the energies can be rescaled to have T Ref matching the physical temperature. In fact, in what follows we will show that all proteins studied fold approximately at the same temperature, one could think to rescale the energies to set the folding temperature to the one observed in nature. A schematic representation of the algorithm is reported in [fig_ref] Figure 2: Schematic representation of the MEP algorithm [/fig_ref].
To the best of our knowledge our work is the first of his kind to optimize the model parameters by reducing the differences between natural and designed sequences and, thanks to the MVP, is the simplest (in terms of the number of parameters needed) to successfully and quantitatively reproduce both sequences and structures of natural proteins to high precision.
# Results
## Parameters optimization
We began by selecting a protein training set from the Protein Data Bank (PDB) [bib_ref] The Protein Data Bank, Berman [/bib_ref] , which includes all the proteins that obey the following conditions: a X-Ray structural resolution below 1.5 Å , are made of single chains of length ranging from 20 to 200 residues, and do not contain any DNA or RNA. According to the stated conditions we selected 125 proteins (see Tab. S2 of the SM for the complete list of the PDB id's). It is important to stress that we did not select for specific experimental conditions, in particular pH and temperature during the measurements fluctuate significantly among the proteins in the set. In [fig_ref] Figure 3: Comparison between the total residue energy ,E Tot /k B T Ref [/fig_ref] and [fig_ref] Figure 3: Comparison between the total residue energy ,E Tot /k B T Ref [/fig_ref] we plot the comparison of the natural to the designed sequences, the latter obtained with the MEP optimized interaction parameters. The plot shows strong correlation (.0.9) between the total energy of the designed (abscissa) and natural (ordinate) sequences, and between the profiles of the residue the HP profiles and the energy contribution ( For a trial set of the E, E HOH , and V parameters and for each protein in the training set a large number (10 5 ) of sequences with composition fixed to the natural one are generated following the design scheme in the SM. The scoring function F score (Eq.S16 in the SM) is then evaluated and the trial parameters are accepted or reject according to a Metropolis like scheme. New parameter sets are generated at each iteration, and the sequences of the proteins in the training set are redesigned by 10 5 simple pair residue swapping moves, which are accepted or rejected according to a standard Metropolis algorithm with the energy defined in Eq.S4 (see SM). During each design iteration, the HP and energy profiles (Eq.S16 in the SM) are averaged over the observed sequences weighted by their Boltzmann weight. The averaging guarantees that the profiles are calculated over the most probable sequences that, as we showed previously [bib_ref] A coarse-grained approach to protein design: learning from design to understand folding, Coluzza [/bib_ref] , are robust against mutations and are more thermally stable. After ,10 8 iterations the interaction parameters converged to their final values: V~21:0+0:5 and E HOH~0 :015+0:001, and the residue-residue E interaction parameters which are listed in Tab. S1 of the SM. conclude that, for all 125 proteins in the training set, the designed proteins and natural proteins are equally compatible sequences to their respective target structures, strongly suggesting that our procedure may now be used to design realistic protein sequences. We applied the MEP derived parameters to design 15 randomly selected from independent training sets [bib_ref] An improved protein decoy set for testing energy functions for protein structure..., Tsai [/bib_ref] [bib_ref] CASP9 target classification, Kinch [/bib_ref] , and characterized by different secondary and tertiary motives. The top five resulting sequences for each target structures are listed in Tab. S4 of the SM. It is important to stress for this design we relaxed the constraint on the amino acid composition used during the optimization. Hence, the folding of the designed sequences does not depend on the previous knowledge of the natural amino acid composition, nevertheless the amino acids composition of the artificial sequences is similar to the natural one (see Tab. S3). It has to be said that the artificial sequences appear unusual with repeats of the same amino acid (e.g. for 1gab WDDMIIRRRRFVVYYLWGSMTAEVEAEKGTNGFYYHHHD-FGTKKKAQQQSNNL). Such repeats could be due to the approximations of the model, however it is important to remember that we did not include in the design any information about the function of the protein. In fact there is no reason to expect that natural sequences are the only one capable of folding, and we want to stress again that additional constraints applied during the design procedure would dramatically reduce the volume of the sequence space [bib_ref] Analytic markovian rates for generalized protein structure evolution, Coluzza [/bib_ref] reducing the probability of repeats. We believe the latter to be the main cause of the repeats and we tested this hypothesis forcing into the design procedure and additional constraint expressly rejecting mutations that would result into a repetition of the same amino acid for 3 residues forward and backwards. The resulting sequences are listed in Tab. S6 of the SM. We were surprised to find that sequences with energy comparable to the unconstrained ones had a lower number of permutations (log(N P )*107 instead of 108). Nevertheless, the now more reasonable looking sequences folded in a similar fashion with respect to their unconstrained alternatives (see [fig_ref] Figure 4: Folding free energy landscape F [/fig_ref]. Finally, we will show below the model is capable of refolding also several natural sequences, demonstrating that in the model the presence of repeats is not necessary to stabilize natural protein structures. It is important to stress that during the last step of the MEP optimization the fewer sequences generated with fixed composition do not present the repeating patterns (see . However, this is an interesting problem and deserves a dedicated study that is beyond the objective of this work. Objective of ongoing research is also to experimentally test whether such sequences are capable of folding to the predicted target structures.
## Protein design
In order to apply the model to the folding of both designed and natural sequences, we need to balance the residue energy term with the backbone hydrogen bond term (parameter a in Eq.S4 in the SM). The energies can be rescaled by choosing the value of a for which designed sequences fold best to their target structures [bib_ref] A coarse-grained approach to protein design: learning from design to understand folding, Coluzza [/bib_ref]. Hence, we selected four designed sequences from Tab. S4 (PDB ids 2l09, 3mx7,chain A of 3obh, and 1qyp), and for each sequence we [fig_ref] Figure 1: Real-space representation of the backbone of the Caterpillar model [/fig_ref] in the SM for details). The plot shows for each protein a funnelled profile with a global minimum very close to the respective target structure (DRMSD [ ½1:5{2:0 Å ). So at least below the folding temperature the proteins seems to follow a downhill process. This observation would need a verification with a study of the folding dynamics. The refolding free energy profiles shown in [fig_ref] Figure 4: Folding free energy landscape F [/fig_ref] prove that realistic protein sequences with low frustration folding free energy landscapes can now be designed with a straightforward positive design scheme.
## Refolding of natural sequences
The next logical step is to asses the behaviour of the model when refolding natural sequences and prove that folding as well can be performed to a quantitative level with the model. For this we randomly selected 15 proteins known to be difficult to fold (from Tsai et al. [bib_ref] An improved protein decoy set for testing energy functions for protein structure..., Tsai [/bib_ref] and from the 9th edition of the well known Critical Assessment of Techniques for Protein Structure Prediction [bib_ref] CASP9 target classification, Kinch [/bib_ref] and we performed folding simulations of their natural sequences. The results are plotted in [fig_ref] Figure 5: Folding free energy landscape F [/fig_ref] , where we have superimposed all the computed free energy profiles. Although, the details of each profile might not be clearly visible, a first fundamental feature is apparent, namely the concentration of the free energy minima in the region between 1.5 and 2 Å DRMSD which remarkably is also the same regions observed for the design proteins. A second important result is the funnel shape common to all free energy profiles providing definite proof of the capability of the model of capturing the low frustration folding of natural proteins with a rather high precision. In fact when the predicted conformations of the folded states are compared to the experimentally determined structures, the two overlapped with a precision between 2.4 and 4.1 Å RMSD (see top inset of [fig_ref] Figure 5: Folding free energy landscape F [/fig_ref] which is surprisingly accurate especially considering the simplicity of the model. An alternative comparison of the refolded structures to their native targets is reported in Tab. S7 produced with the ''MaxCluster'' program from Alex Herbert (http://www.sbg.bio.ic.ac.uk/maxcluster/index.html) and the TM-scoring function introduced by Zhang et al. [bib_ref] Scoring function for automated assessment of protein structure template quality, Zhang [/bib_ref] [bib_ref] How significant is a protein structure similarity with TM-score 50, Xu [/bib_ref]. It is important to note that the configurations with the lowest energy are not necessarily equal to the ones corresponding to the minimum of the free energy, however, in most cases, they are very similar. This is due to the strong directional nature of the hydrogen bonds which makes them very sensitive to thermal fluctuations. As a consequence, there are isolated structures that might have a lower energy but are not very stable at finite temperature. We would like to stress that, since the native sequences fold in a similar fashion compared to the designed ones and we did not observe the native sequences themselves as an outcome of the design process (see Tab. S4-S6), we could speculate that, at least within the Caterpillar model, the space of folding sequences is much wider than the one comprising only of natural sequences.
# Discussion
To the best of our knowledge our coarse-grained protein model is the simplest, in terms of the number of parameters needed, with a transferable energy function capable of achieving such precision for the prediction of the native folded structures. Also it is one of the very few models that allows for both quantitative proteins design and folding, the latter demonstrated by free energy calculations. It is remarkable that low frustration sequences can be obtained with such a simple and universal design procedure, and that the folding of natural proteins shows funnelled free energy landscapes without the need of any potentials based on the native structure [bib_ref] Respective roles of short-range and long-range interactions in protein folding, Go [/bib_ref].
Although, the artificial sequences present some unnatural features like repetitions of some amino acids, the sequences designed with a natural amino acid composition share many features with the natural occurring ones, and the native structures of the latter are correctly predicted by our model. Hence, we expect that our designed proteins (see Tab. S4), once synthesized, may fold to the structures used as design targets, which may also represent the ultimate and most important test of our methodology. We hope that our methodology will become an useful tool in experiments requiring alterations of natural proteins, or the total redesign of target protein structures. Of course, constraints on the composition can always be applied to the design procedure with no major changes in the procedure. Moreover, the prediction power of the model gives us high confidence that our design methodology may be directly used to tackle important open problems of drug design, or used in a multi-scale approach where the results from our model could be refined with a more accurate but also a computationally more expensive protein model.
Finally, this work not only extends our previous results obtained with the Caterpillar model, but also strengthens the connection among all our work on lattice heteropolymers and protein unrelated systems such as patchy polymers [bib_ref] Design and folding of colloidal patchy polymers, Coluzza [/bib_ref] [bib_ref] Sequence Controlled Self-Knotting Colloidal Patchy Polymers, Coluzza [/bib_ref]. The success of the same design strategy for all these systems demonstrates that the maximum valence principle is a sufficient condition to satisfy for the generalized design of low frustration sequences and the prediction of their proper native state. The profiles have a common funnel shape and show a clustering of the free energy minima in the region 1.5 and 2 Å DRMSD consistent with the results obtained for designed sequences. In b) we plot the free energies for proteins with the worst (2ptl) and the best (3nmd-E) distance of the folded structure from the native one. For the latter the free energy profile shows a minimum remarkably close to the native state probably due to the highly simplified structure of protein 3nmd-E. The minimum of 2ptl, on the other hand, is located further away from the low DRMSD values than the other proteins. This apparent discrepancy is due to the definition of the DRMSD which includes the contribution from the C a atoms located in the long unstructured tail from the residue 1 to 18. Since the probability of observing that particular conformation in solution is very low, it follows that the particular realization of the native structure has a large entropy penalty. However if we measure the overlap ignoring the contribution from the tail we see that the predicted structure of the protein core is again reasonably close to the experimentally determined one (<5.2 Å RMSD). In the insets we compare the experimental structures (in yellow) superimposed to the equilibrium configurations (in red), and we show that the proteins refolded with a precision between 2.4 and 4.1 Å RMSD. . Comparison of the average composition of the designed sequences and the natural sequences used in the parameter optimization. It is important that since we do not model Cys-Cys bond and the Proline rigid bond we have excluded them from the design alphabet. This is why the frequency associated to those amino acids is zero in the designed sequences. We are currently working on implementing such special cases in the Caterpillar model. We have highlighted in bold the amino acids types with the largest discrepancies namely: Histidine, Methionine, Tryptophan, Tyrosine. Such amino acids are know to be the one with the lowest appearance frequency in nature. Since we did not impose any restriction on the design procedure over the relative abundance of amino acids in nature it is not surprising to find the largest discrepancies in the composition for such amino acids. doi:10.1371/journal.pone.0112852.S007 (PDF) . Sequences obtained during the last step of the matrix optimization procedure. The amino acid composition is identical for all sequences and the first is the natural sequences taken from the pdb file. doi:10.1371/journal.pone.0112852.S009 (PDF) . Designed sequences under the additional constraint that local repetition of up to 5 residues are forbidden. doi:10.1371/journal.pone.0112852.S010 (PDF . Summary of the refolded structures with the natural sequences. The DRMSD value is taken form the minimum of the folding free energy (see [fig_ref] Figure 5: Folding free energy landscape F [/fig_ref] , while the Overlap, the gRMSD and the TM-score are calculated using the Max Cluster program from Alex Herbert (http://www.sbg.bio.ic.ac.uk/maxcluster/ index.html). The Overlap is a measure of percentage of matched structural elements between the native and the refolded structures. The gRMSD and TMscore are calculated over the overlapping structural elements. TM-score was defined by Zhang et al. [bib_ref] Scoring function for automated assessment of protein structure template quality, Zhang [/bib_ref] [bib_ref] How significant is a protein structure similarity with TM-score 50, Xu [/bib_ref]. The low overlapping value for 2ptl and 1vif are due to the unstructured sections of the proteins. doi:10.1371/journal.pone.0112852.S011 (PDF) Text S1. Supplemental Material containing the details about the model and simulations techniques together with the derivation of the scoring function with the Maximum Entropy Principle. doi:10.1371/journal.pone.0112852.S012 (PDF)
[fig] Figure 1: Real-space representation of the backbone of the Caterpillar model. The large blue sphere represent the self-avoidance volume R HC~2 :0A of the C a atoms, while the interaction radius of each residue is represented by the large dashed circle or radius 6 Å (see Eq. S2 in Methods). The H and O atoms interact through a 10-12 Lennard-Jones potential tuned with a quadratic orientation term that selects for alignment of the C, H, O, and N atoms involved in a bond (see top right inset and Eq. S1 in Methods). The backbone fluctuates only around the torsional angles w 1 and w 2 . [/fig]
[fig] Figure 2: Schematic representation of the MEP algorithm. [/fig]
[fig] Figure 3: Comparison between the total residue energy ,E Tot /k B T Ref . (Eq.S4 in SM) averaged over all the 10 5 designed sequences per target(abscissa) and the same energy calculated over the native sequence of same target (ordinate). Each point corresponds to one protein in the data set and shows a strong linear trend verified by the fit (red line) with a correlation coefficient of ,0.995 and a slope of ,1.000 indicating that two energies are perfectly correlated. In the insets we show the comparison of the HP profiles (top left) and interaction energy E=k B T Ref of each residue with all other (bottom right), this time each point corresponds to a single residue of each test protein. In both cases the data follow a remarkable linear trend (fits in green and blue lines respectively), and a positive correlation close to unity. For the HP profiles the correlation coefficient (,0.98) indicates that when in natural proteins we find an hydrophobic residue also the design procedure will put one and vice versa. While the correlation coefficient (,0.90) of E=k B T Ref demonstrates that each natural residue has a very similar contribution to the total energy compared to the designed ones. A perfect match cannot be expected since natural sequences might have experience a selection pressure influenced by interactions not represented in the model, different environmental conditions or simply unknown functional requirements. Nevertheless the accordance is remarkable. doi:10.1371/journal.pone.0112852.g003We now have obtained the optimized parameters for our model: a~0:10+0:01k B T Ref , V~21:0+0:5, E HOH~0 :015+0:001 and for the E see Tab. S1 in the SM. [/fig]
[fig] Figure 4: Folding free energy landscape F(DRMSD)/k B T Ref as a function of DRMSD of the four designed proteins (PDB ids 2l09, 3mx7, chain A of 3obh, and 1qyp). All profiles have a global minimum around 1.5 and 2 Å DRMSD with a smooth funnelled shape. Due the approximations present in the model and to thermal fluctuations is shifted with respect to DRMSD~0 (note that to the value DRMSD~0 of each profile will correspond a different native structure). Because of the definition of DRMSD, the smaller the value the fewer are the possible structures that can have this value of DRMSD. Ultimately, DRMSD~0 is possible only for the target structure itself. The funnelled profiles with single minimum implies that both an ensemble of arrested structures and a single alternative fold are less stable compared to the desired configuration. In the bottom right inset we plot the folding free energy landscape F(DRMSD,Q H )=k B T Ref for 3mx7 as a function of both the DRMSD and the number of hydrogen bonds Q H , to give a visual example of the funnel nature of the folding landscapes. On the left we compare the experimentally determined structures (in yellow) with a typical folded conformation selected as the sampled configurations with the lowest energy at the free energy minimum (in red). The RMSD value is indicated in the middle. The structures were aligned using the RMSD calculator tool in VMD[52], while the secondary structure elments where identified with STRIDE[53].doi:10.1371/journal.pone.0112852.g004 [/fig]
[fig] Figure 5: Folding free energy landscape F(DRMSD)/k B T Ref of the 15 proteins set selected to test the accuracy of the MEP optimized parameters. [/fig]
[fig] Figure S1, Figure S2, Figure S3, Figure S4: Folding free energy landscape F(DRMSD)=k B T Ref as a function of DRMSD of the designed protein PDB ids 1CTF close to the folding temperature. doi:10.1371/journal.pone.0112852.S001 (TIFF) On the left: correlation plot between the DRMSD and the RMSD collective variable. The estimated correlation coefficient from a linear regression fitting (in red) is <0.8 which increases to <0.98, if we exclude the configurations for values of DRMSD ,1.5 Å which is below the model resolution, indicating that the free energy profile should be qualitatively similar if the states are projected over RMSD instead of DRMSD. On the right: Free Energy folding profile of the protein 3NMD-E projected over the collective variables DRMSD and RMSD. The profiles are not identical because the RMSD is more sensitive to local distortions of the protein with respect to the DRMSD. This is also demonstrated by the wider free energy minimum which reflects the thermal fluctuations. However, overall the qualitative shape of the profiles is very similar with between each other in particular since both have a clear global free energy minimum. doi:10.1371/journal.pone.0112852.S002 (TIFF) Correaltion between designed and real E sol profiles. The correlation coefficient has been estimated from a linear regression fitting (in red) to be be very high <0.98. doi:10.1371/journal.pone.0112852.S003 (TIFF) Folding free energy landscape F(DRMSD)=k B T Ref as a function of DRMSD of the four designed proteins (PDB ids 2l09, 3mx7, chain A of 3obh, and 1qyp). All profiles have a global minimum around 1.5 and 2 Å DRMSD with a smooth funnelled shape. Due the approximations present in the model and to thermal fluctuations is shifted with respect to DRMSD~0 (note that to the value DRMSD~0 of each profile will correspond a different native structure). Because of the definition of DRMSD, the smaller the value the fewer are the possible structures that can have this value of DRMSD. Ultimately, DRMSD~0 is possible only for the target structure itself. The funnelled profiles with single minimum implies that both an ensemble of arrested structures and a single alternative fold are less stable compared to the desired configuration. In the bottom right inset we plot the folding free energy landscape F(DRMSD,Q H )=k B T Ref for 3mx7 as a function of both the DRMSD and the number of hydrogen bonds Q H , to give a visual example of the funnel nature of the folding landscapes. On the left we compare the experimentally determined structures (in yellow) with a typical folded conformation selected as the sampled configurations with the lowest energy at the free energy minimum (in red). The RMSD value is indicated in the middle. doi:10.1371/journal.pone.0112852.S004 (TIFF) [/fig]
[table] Table S1: Optimized values of the residue-solvent E Sol and residue-residue E(S k )(S l ) interaction parameters. The uncertainty on the values is <¡0.01. doi:10.1371/journal.pone.0112852.S005 (PDF) [/table]
[table] Table S2: List of PDB id's used as training set for the maximum entropy parameters optimization. doi:10.1371/journal.pone.0112852.S006 (PDF [/table]
[table] Table S4: Designed sequences. doi:10.1371/journal.pone.0112852.S008 (PDF) [/table]
|
Effect of Front-of-Package Information, Fruit Imagery, and High–Added Sugar Warning Labels on Parent Beverage Choices for Children
IMPORTANCE Fruit drinks are widely consumed by young children, and many parents mistakenly believe that these drinks are healthy, potentially due to front-of-package claims and imagery.Research is needed on the influence of this marketing and how labeling regulations could change behavior.OBJECTIVE To assess the effects of a front-of-package 100% vitamin C claim, fruit imagery, percentage juice and teaspoons of added sugar disclosures, and high-added sugar warnings on parents' choices, knowledge, and perceptions of beverages.DESIGN, SETTING, AND PARTICIPANTSThis randomized clinical trial was conducted May to July 2021 as a single-exposure (no follow-up) online survey of primary caregivers of children ages 0 to 5 years throughout the US.INTERVENTIONS Participants were shown no-, low-, and high-added sugar beverages and asked to choose 1 for their child. Participants were randomized to see high-added sugar beverages with 1 of 7 front-of-package conditions: (1) claim and imagery (control); (2) no claim; (3) no imagery; (4) no claim or imagery; (5) claim, imagery, and percentage juice disclosure; (6) claim, imagery, and warning; or (7) claim, imagery, warning, and teaspoons of added sugar disclosure.MAIN OUTCOMES AND MEASURESPrimary outcomes were type of beverage chosen (eg, highadded sugar beverage) and resulting calories and added sugar (in grams). Secondary outcomes were fruit drink knowledge (added sugar and percent juice) and perceptions.RESULTSThere were 5005 participants included in the final analysis (mean [SD] age, 31.5 [8.3] years; 3587 female participants [71.7%]), including 714 participants in group 1, 717 participants in group 2, 710 participants in group 3, 717 participants in group 4, 708 participants in group 5, 729 participants in group 6, and 710 participants in group 7. Compared with participants in the control group, who had a mean (standard error [SE]) of 9.4 (0.5) g of added sugar and 81.9 (1.6) kcal in chosen beverages, only participants who saw warnings with teaspoons of added sugar disclosures had significantly reduced added sugar (−1.3 g; 95% CI, −2.6 to −0.1 g [−14.2%; 95% CI, −26.7% to −1.8%]; P = .04) and calories (−5.3 kcal; 95% CI, −9.8 to −0.9 kcal [−6.5%; 95% CI, −11.8% to −1.3%]; P = .02) in selected beverages. In warning conditions (ie, 6 and 7) compared with the control group (mean [SE] 41.0% [1.8%]), the proportion of participants choosing high-added sugar beverages was significantly reduced, by 5.5 percentage points (95% CI, 0.5 to 10.5 percentage points [13.4%; 95% CI, 1.2% to 25.6%]; P = .03) and 6.4 percentage points (95% CI, 1.4 to 11.4 percentage points [15.6%; 95% CI, 3.3% to 27.8%]; P = .01), respectively. The no claim or imagery condition (4) significantly reduced the proportion of parents choosing high-added sugar beverages (−7.6 percentage points; (continued) Key Points Question What are the effects of a front-of-package 100% vitamin C claim, fruit imagery, percentage juice and teaspoons of added sugar disclosures, and high-added sugar warnings on parents' beverage choices for their children? Findings In this randomized clinical trial of 5005 parents of children aged 0 to 5 years, 15.6% fewer parents chose a high-added sugar beverage when packages displayed warnings with teaspoons of added sugar disclosures, reducing added sugar and calories in chosen beverages by 14.2% and 6.5%, respectively; 18.4% fewer chose a highadded sugar beverage when shown packages without a 100% vitamin C claim or fruit imagery.Meaning These findings suggest that adding warnings or removing claims and imagery from high-added sugar beverages may reduce parents' purchases of these beverages for their children. Abstract (continued) 95% CI, −12.6 to −2.6 percentage points [−18.4%; 95% CI, −30.6% to −6.3%]; P = .003). Percentage juice disclosures did not affect beverage choice.CONCLUSIONS AND RELEVANCEThese findings suggest that added sugar warnings and prohibitions of front-of-package claims and imagery may reduce parents' purchases of high-added sugar beverages for their young children but that percentage juice disclosures may not change behavior.TRIAL REGISTRATION ClinicalTrials.gov Identifier: NCT04811690JAMA Network Open. 2022;5(10):e2236384.
# Introduction
Most young children in the US consume sugar-sweetened beverages (SSBs), [bib_ref] Trends and patterns in sugar-sweetened beverage consumption among children and adults by..., Dai [/bib_ref] [bib_ref] Consumption patterns of sugar-sweetened beverages in the United States, Han [/bib_ref] [bib_ref] Added sugars intake among US infants and toddlers, Herrick [/bib_ref] predisposing these children to tooth decay, obesity, type 2 diabetes, and other chronic diseases. [bib_ref] Sugar-sweetened beverages and cardiometabolic health: an update of the evidence, Malik [/bib_ref] [bib_ref] The association of sugar-sweetened beverage intake during infancy with sugarsweetened beverage intake..., Park [/bib_ref] [bib_ref] The negative impact of sugar-sweetened beverages on children's health: an update of..., Bleich [/bib_ref] The most frequently consumed SSBs among children aged 1 to 8 years are fruit drinks (ie, fruit-flavored drinks containing <100% juice), which account for 20% of added sugars consumed by children aged 1 year 3 and 11% of added sugars consumed by children aged 2 to 8 years. [bib_ref] Beverage consumption patterns among infants and young children (0-47.9 months): data from..., Kay [/bib_ref] Reducing fruit drink consumption among young children could thus help prevent chronic diseases later in life.
One potential driver of fruit drink consumption is the use of misleading front-of-package (FOP) claims and imagery. Nutrient content claims (eg, 100% vitamin C) and fruit imagery are prevalent on fruit drink packages 9-11 but are not reliable indicators of product healthfulness or juice content. [bib_ref] Nutrition claims on fruit drinks are inconsistent indicators of nutritional profile: a..., Duffy [/bib_ref] [bib_ref] Front-of-package claims & imagery on fruit-flavored drinks and exposure by household demographics, Musicus [/bib_ref] [bib_ref] Children's fruit "juice" drinks and FDA regulations: opportunities to increase transparency and..., Pomeranz [/bib_ref] Research has shown that this marketing misleads consumers. [bib_ref] Ignorance is bliss: how parents of preschool children make sense of front-ofpackage..., Abrams [/bib_ref] [bib_ref] Effects of nutrient content claims, sports celebrity endorsements and premium offers on..., Dixon [/bib_ref] [bib_ref] A systematic review, and meta-analyses, of the impact of health-related claims on..., Kaur [/bib_ref] [bib_ref] Mothers' perceptions of toddler beverages, Rigo [/bib_ref] [bib_ref] Marketing of sugar-sweetened children's drinks and parents' misperceptions about benefits for young..., Fleming-Milici [/bib_ref] [bib_ref] Nutrition-related claims lead parents to choose less healthy drinks for young children:..., Hall [/bib_ref] In a 2022 study, [bib_ref] Nutrition-related claims lead parents to choose less healthy drinks for young children:..., Hall [/bib_ref] parents who viewed fruit drinks with claims were more likely to incorrectly believe that these products did not contain added sugar or were 100% juice. Correcting these misperceptions could enhance parents' knowledge and encourage healthier beverage purchases.
There are multiple regulatory options to address misleading FOP marketing, but effects of such options alone and in combination have not been well-studied. Some options include restricting nutrient content claims or fruit imagery on products with high levels of added sugar, requiring FOP percentage juice content disclosures [bib_ref] Misperceptions about added sugar, non-nutritive sweeteners and juice in popular children's drinks:..., Harris [/bib_ref] (in addition to currently mandated back-of-pack disclosures 20 ), or requiring added-sugar warnings if these added sugars exceed a predetermined level. Evidence suggests that added-sugar warnings may reduce SSB purchases, [bib_ref] The effect of graphic warnings on sugary-drink purchasing, Donnelly [/bib_ref] and the US Congress has introduced a bill to mandate such warnings. Combining warnings with quantitative added-sugar disclosures (eg, teaspoons of added sugar/serving) may further discourage SSB consumption. Prior consumer research on FOP marketing has largely been conducted with unbranded or mock beverages, so there are limited data on the impact of FOP changes in actual products in a realistic retail setting, in which FOP claims, imagery, disclosures, warnings, and branding interact to affect behavior. [bib_ref] Ignorance is bliss: how parents of preschool children make sense of front-ofpackage..., Abrams [/bib_ref] The goal of this study was to compare independent and combined effects of common FOP marketing tactics (ie, 100% vitamin C claims and fruit imagery), nutrition disclosures (ie, teaspoons of added sugar/serving and percentage juice content), and high-added sugar warnings on parents' beverage selections for their young children and on parents' knowledge and perceptions of fruit drinks with varied levels of added sugar.
Institutional Review Board approved this study, and all participants gave electronic informed consent. The trial protocol appears in Supplement 1.
## Participants
Participants were recruited for an online randomized clinical trial (single exposure, no follow-up) from May to July 2021 through Qualtrics, a survey-sampling and administration company that recruits participants from a network of research panels via online ads, promotions, social networks, and online and mobile games. Participants were sampled to reflect the US educational distribution based on 2010 Census data, with oversampling for Black and African American and Hispanic participants because Black and Hispanic children have the highest national fruit drink consumption rates. 31,32 Participants self-reported race and ethnicity using investigator-defined options. Race and ethnicity options were Asian; Black or African American; Latinx/o/a or Hispanic; Native American; Native Hawaiian, Pacific Islander, or Alaska Native; White, and other. Other represents anyone who self-identified as other or as more than 1 category (excluding Hispanic). The Hispanic category included anyone who indicated they were Hispanic or Hispanic and 1 or more other races. For example, if someone said they were White and Hispanic, that was counted toward White and Hispanic separately, but if they said they were White, Asian, and Hispanic, that was counted toward Hispanic and other separately. Participants had to be living in the US, aged 18 years or older, and primary caregivers of at least 1 child aged 0 to 5 years. We excluded participants for completing the survey in less than 5 minutes (one-third of the median completion time of 15 minutes), failing an attention check (multiple choice question asking, "What month is it?"), or refusing to share their data.
There were no missing data. The final sample had 5005 participants based on a priori sample size calculations informed by previous literature 24,33,34 [fig_ref] Table 1: Participant Characteristics [/fig_ref] and .
## Survey procedure
After providing written informed consent, participants completed an online survey (all questions and image information are in eAppendix 1 and 2 in Supplement 2). In a choice task, participants viewed 12 commonly purchased 11 multipacks of real beverages in a random order and were instructed to "click on the drink you would like to purchase for your oldest child aged 0-5." To incentivize realistic shopping behavior, participants were told that at the end of the survey a computer would randomly select whether they would have their selected beverage shipped to them or be given $5. Beverages included 6 high-added sugar beverages (>20% daily value added sugar/serving 35 : 4 fruit drinks, 1 cola, and 1 fruit-flavored soda) and 6 low-or no-added sugar beverages (1 low-added sugar fruit drink, 1 no-added sugar fruit drink, two 100% juices, 1 milk, and 1 water). After participants selected a beverage, an enlarged image appeared and participants were asked to confirm their selection or go back and choose a different beverage.
Participants then answered knowledge and perception questions about 4 fruit drinks (2 high, 1 low, and 1 no added sugar), which they viewed 1 at a time in random order. These drinks had not appeared in the previous choice task but featured the same brands. Participants then reported whether they noticed different package elements and provided demographic information, including age, gender, race and ethnicity, height, weight, chronic disease history, annual household income, household size, number of children aged 0 to 5 years, and Supplemental Nutrition Assistance Program (SNAP) and Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) participation. They also reported the frequency of any fruit drink consumption for their oldest child in that age range.
## Label randomization
Participants were randomized (simple randomization) to view high-added sugar beverages throughout the survey with 1 of 7 FOP label designs [fig_ref] Figure 2: Study Conditions [/fig_ref] ; eAppendix 2 in Supplement 2).
Although images of real branded products were used in this study, product images in [fig_ref] Figure 2: Study Conditions [/fig_ref] vector images created by macrovector, designed by Freepik. In the study, a graphic designer modified real beverage package images to make each condition look as realistic as possible. The control condition included high-added sugar beverages with 100% vitamin C claims and fruit imagery because most sweetened drinks for children contain claims and imagery and 100% vitamin C claims are among the most common. [bib_ref] Nutrition claims on fruit drinks are inconsistent indicators of nutritional profile: a..., Duffy [/bib_ref] [bib_ref] Front-of-package claims & imagery on fruit-flavored drinks and exposure by household demographics, Musicus [/bib_ref] [bib_ref] Children's fruit "juice" drinks and FDA regulations: opportunities to increase transparency and..., Pomeranz [/bib_ref] High-added sugar beverages in conditions 2 to 4 lacked 1 or both of these elements (2: no claim; 3: no imagery; 4: no claim or imagery). Conditions 5 to 7 had additional labeling elements (5: claim, imagery, and percentage juice disclosure; 6: claim, imagery, and high-added sugar warning; 7: claim, imagery, high-added sugar warning, and teaspoons of added sugar disclosure). Throughout the survey, 100% vitamin C claims and fruit imagery were shown on all low-and no-added sugar fruit drinks and 100% juices, but these elements were not shown on milk, water, or soda in any condition because those products do not typically display those claims or imagery. In condition 5, the FOP percentage juice disclosure was shown on all fruit drinks, 100%
juices, and fruit-flavored sodas. All beverages additionally displayed calorie labels in all conditions because that is the current industry standard.
## Outcomes
Primary outcomes were the selection of a high-added sugar beverage in the choice task and selected beverages' calories and added sugar (in grams) content. We also examined selected beverages' total sugar (in grams) and the proportion of participants who chose each beverage category: fruit drinks (high, low, and no added sugar), 100% juice, soda, milk, and water.
Secondary outcomes included fruit drink knowledge and perceptions, which were assessed for each drink individually and the mean found for the 2 high-added sugar fruit drinks. Knowledge outcomes included participant estimates of drink juice content (0%-100%; 22 options in 5% intervals) and teaspoons of added sugar/serving (open-text response restricted to 0-100). Estimated juice content values were assigned as the middle value of participant-chosen intervals. We also reported the percentage of participants in each condition who believed the drink was 100% juice.
For perception outcomes, participants rated how much added sugar/serving each drink contained (none, a little, some, or a lot), as well as how likely they were to buy each drink for their child in the next month and how appealing the drink would be to their child based on the packaging (both on a 7-point Likert scale, not at all to extremely). Participants also rated how healthy they and increase their child's risk of dental cavities (reverse coded) (1 = strongly disagree to 7 = strongly agree). These 4 health-related questions were summed to create a health perceptions index (4 = least healthy to 28 = healthiest).
We also assessed perceived message effectiveness (PME), 38 a scale shown to be sensitive to differences between warnings in online studies and associated with behavior change. We adapted the scale's language to assess the entire package instead of a single label on the package. We created a mean PME score as the mean of responses to the following 3 items (1 = strongly disagree to Although images of real branded products were used in this study, product images in [fig_ref] Figure 2: Study Conditions [/fig_ref] are brandless mock-ups for the purposes of publication. The grapefruit images in the figure mock-ups are vector images created by macrovector, designed by Freepik.
In the study, a graphic designer modified real product images for each condition to ensure that all packages looked as realistic as possible. For more information about images used in the study, please see eAppendix 2 in Supplement 2.
added sugar seem unpleasant to me"; and "This package discourages me from wanting to let my child drink beverages with added sugar." We also measured package element recall by assessing whether participants noticed images of fruit, added-sugar warnings, and information about vitamin C, percent juice, or teaspoons of added sugar when selecting a drink for their child.
# Statistical analysis
Analyses were preregistered with clinicaltrials.gov (NCT04811690) and AsPredicted (trial protocol in Supplement 1; eAppendix 3 in Supplement 2). We tested for differences in participant characteristics across conditions using 2-sided analyses of variance for continuous variables and χ 2 tests for categorical variables. Linear and logistic regression were used to compare continuous and categorical outcomes, respectively, across conditions. For beverage knowledge and perception outcomes, we prespecified that models would control for the frequency of purchasing each beverage (participants reported how often they had given each brand of drink to their oldest child aged 0-5 years in the last month).
Exploratory analyses examined interactions between condition and potential moderators on selected beverage calories and grams of added sugar using linear regression models with indicators for condition, levels of the potential moderator, and interaction terms between condition and levels of the moderator. Potential moderators tested were educational attainment (Նcollege vs <college), SNAP status, WIC status, annual household income (Ն$75 000 vs <$75 000), frequency of child fruit drink consumption (Նonce/wk vs <once/wk), race (Black or other vs White), and Hispanic ethnicity.
These race and ethnicity groups were created because we were interested in how outcomes may differ among Black and Hispanic parents compared with White parents. An exploratory logistic regression analysis was also conducted to assess differences in package choice (bottles vs pouches) among participants who chose high-added sugar fruit drinks. All analyses were conducted using Stata/MP statistical software version 17.0 (StataCorp). Statistical significance was set at P < .05, and all tests were 2-tailed.
# Results
Among 5005 for their children [fig_ref] Figure 3: Beverage Choice by Category, Calorie Content, and Total and Added [/fig_ref]. In exploratory analyses, 2 factors modified the effect: WIC participation and Hispanic ethnicity (eFigures 1 and 2 in Supplement 2). Participants who had participated in WIC selected beverages with fewer calories (β = −11.9; P for interaction = .02) and
added sugar (in grams; β = −3.9; P for interaction = .007) than participants who had not participated in WIC when exposed to no claim or imagery compared with participants in the control group.
Hispanic participants selected beverages with fewer calories than non-Hispanic participants in the no claim (β = −14.1; P for interaction = .009), no imagery (β = −11.4; P for interaction = .03), and warning (β = −13.1; P for interaction = .02) conditions compared with participants in the control group.
## Jama network open | pediatrics
## High-added sugar fruit drink knowledge
Compared with participants in the control group (who estimated a mean [SE] of 46.1% [1.0%] juice), participants in the percentage juice disclosure and both warning conditions had lower (ie, more
## Jama network open | pediatrics
## High-added sugar fruit drink perceptions
Compared with participants in the control group, participants in the percentage juice disclosure and both warning conditions perceived high-added sugar fruit drinks to be significantly higher in added sugar and less appealing and healthy [fig_ref] Table 2: Effect of Package Modifications on Beverage Choice, Knowledge, Perceptions, and Package Element... [/fig_ref]. Participants in the no claim or imagery and both warning conditions reported a lower likelihood of buying high-added sugar fruit drinks for their children. Participants who viewed high-added sugar fruit drinks without fruit imagery (no imagery or no claim or imagery) perceived those drinks to be significantly less appealing to their children compared with participants in the control group.
## Perceived message effectiveness and package element recall
Participants in the percentage juice disclosure and both warning conditions rated high-added sugar fruit drink packages to have significantly higher PME compared with participants in the control group. Most participants reported noticing the package elements they were randomized to see [fig_ref] Table 2: Effect of Package Modifications on Beverage Choice, Knowledge, Perceptions, and Package Element... [/fig_ref].
# Discussion
This randomized clinical trial tested the effects of high-added sugar warnings, percentage juice and teaspoons of added sugar disclosures, and removal of a 100% vitamin C claim or fruit imagery from high-added sugar beverages on parents' choices for their young children. The warning with teaspoons of added sugar disclosure was the only condition that led parents to select beverages with significantly fewer calories and added sugars. This builds on previous research showing that teaspoon disclosures can reduce sugar in selected beverages. We also found that warnings alone and with teaspoons of added sugar disclosures led to 13.4% and 15.6% reductions, respectively, in the proportion of parents selecting high-added sugar beverages for their children. These results are consistent with in-person and online studies of high-added sugar beverage warnings. 22,25 Adding a percentage juice disclosure, removing a 100% vitamin C claim, or removing fruit imagery had no behavioral effects independently, but the removal of both the claim and imagery resulted in 18.4% fewer parents selecting high-added sugar beverages for their children.
When behavioral effects were observed, parents substituted a no-added sugar fruit drink (not 100% juice, a low-added sugar fruit drink, milk, water, or soda) for high-added sugar fruit drinks. This is a promising finding given that the no-added sugar fruit drink contained less total sugar than 100% juice, no added sugar, and no non-nutritive sweeteners (NNSs). NNSs are not recommended for children 40 but are often added to fruit drinks, [bib_ref] Front-of-package claims & imagery on fruit-flavored drinks and exposure by household demographics, Musicus [/bib_ref] especially in response to policies requiring high- perceptions of fruit drinks' healthfulness and appeal, and increased PME.
## Jama network open | pediatrics
Although removing both a 100% vitamin C claim and fruit imagery from high-added sugar fruit drinks reduced parents' selections of high-added sugar drinks, it had no effect on calories. This was because many participants shifted to 100% juice, which contained more calories than some highadded sugar fruit drinks. Claim and imagery removal did not significantly reduce added sugar, which may be due to sample size limitations. Claim and imagery removal had no knowledge or health perception effects combined or independently. While previous findings on whether fruit imagery impacts health perceptions are mixed, 13,44 our findings differ from previous studies that found that nutrient content claims, [bib_ref] A systematic review, and meta-analyses, of the impact of health-related claims on..., Kaur [/bib_ref] 100% vitamin C claims in particular, [bib_ref] Nutrition-related claims lead parents to choose less healthy drinks for young children:..., Hall [/bib_ref] increased perceptions that drinks were healthy and selection of high-added sugar beverages. These differences are likely due to product type (branded vs mock or unbranded products) and choices offered (many different beverages vs 1 other beverage). Our findings suggest that regulations restricting both fruit imagery and nutrient content claims may reduce purchases of high-added sugar beverages but that restricting claims or imagery alone may be insufficient.
FOP percentage juice content disclosures did not change behavior in this study, but consistent with prior work, [bib_ref] Misperceptions about added sugar, non-nutritive sweeteners and juice in popular children's drinks:..., Harris [/bib_ref] they resulted in the most accurate estimations of fruit drink juice content and reduced misperceptions that fruit drinks were 100% juice. These disclosures also increased PME and perceived added-sugar content and reduced perceived healthfulness but to a lesser extent than warnings.
Results from our exploratory analyses suggested that warnings and claim or fruit imagery removal may be particularly effective for Hispanic parents, but the driving mechanism remains unknown. While it is possible that individuals with low English language use may rely more on icons and images when making decisions, 45 Hispanic ethnicity is not necessarily an appropriate proxy for language use. We also found that the removal of both claims and imagery was particularly effective for individuals who participated in WIC, who may associate fruit juice with health because it is provided in WIC food packages. More research is needed among specific populations to ensure that regulations have equitable impacts.
Strengths of this study include its randomized design in a large, racially and ethnically diverse sample. This study is the first, to our knowledge, to test the effects of FOP percentage juice disclosures and warnings with teaspoons of added sugar disclosures on simulated purchasing behavior. It also is the first, to our knowledge, to compare the effects of adding FOP warnings and disclosures and removing FOP claims and imagery and the first to use real products with branding, claims, and imagery.
# Limitations
This study has several limitations. First, hypothetical purchases were measured, but realistic behavior was incentivized by informing participants that they might receive their selected beverages.
However, we did not measure consumption; some parents may not have served an entire resealable fruit drink bottle to their young child at 1 time. Second, social desirability bias may have encouraged participants to select healthier options than they otherwise would have, but this may be similar across randomized groups. Third, we may have been underpowered to detect significant reductions in added sugar between conditions. Fourth, we examined a 1-time exposure to FOP modifications.
Future research should test how in-person, repeated exposure to these modifications influences
## Jama network open | pediatrics
# Conclusions
This randomized clinical trial's findings suggest that added sugar warnings, especially those that disclose added sugar content in teaspoons, may reduce parents' purchases of high-added sugar beverages for their young children, improve their understanding of high-added sugar fruit drinks added sugar and percentage juice content, and correct health misperceptions of these drinks. Our findings suggest that prohibiting FOP nutrient claims and fruit imagery on high-added sugar beverages may also reduce parents' purchases but is unlikely to modify knowledge or perceptions of high-added sugar fruit drinks. Our results also suggest that FOP percentage juice disclosures may improve understanding of juice and added-sugar content but are unlikely to change consumer behavior.
## Article information
[fig] 5: = strongly agree): "This package makes me concerned about the health effects of my child drinking beverages with added sugar"; "This package makes the idea of my child drinking beverages with [/fig]
[fig] Figure 2: Study Conditions [/fig]
[fig] Figure 3: Beverage Choice by Category, Calorie Content, and Total and Added [/fig]
[table] Table 1: Participant Characteristics [/table]
[table] Table 2: Effect of Package Modifications on Beverage Choice, Knowledge, Perceptions, and Package Element Recall [/table]
|
Fractional laser therapy – the next step in alleviating the symptoms of skin aging (own observations)
Skin aging is a natural process of the skin, which accelerates in menopause and is additionally intensified by accumulating effects of repeated exposure to solar UV radiation and other external factors. Anti-aging skin treatment and constant improvement of its methods have become an important area of current research.The need to apply effective skin anti-aging methods that minimize traumatization resulted in the development of fractional laser technology delivering a laser beam to microscopic column skin zones in order to achieve skin photo-remodeling.
The process of natural skin aging tends to significantly accelerate in menopause and is additionally intensified by accumulating effects of repeated exposure to solar UV radiation and other external factors. The skin turns wrinkled, dry, thin, inelastic and contains hyperpigmentation spots. Decrease in the thickness of the epidermis and the number of eccrine glands, sebaceous gland hypertrophy, decreased number of fibroblasts and accumulation of abnormal elastin known as elastosis are visible in the histopathological image. Dysplastic lesions in the epidermal cells can also occur [bib_ref] Wpływ hormonów na procesy starzenia się skóry, Trznadel-Budźko [/bib_ref] [bib_ref] Dermatologiczne aspekty menopauzy, Trznadel-Budźko [/bib_ref].
Skin rejuvenation procedures and constant improvement of methods connected with this subject have become considerable fields of scientific research. The need for effective methods of skin photorejuvenation with minimized traumatization resulted in the development of fractional laser technology. The concept of fractional laser technology involves exposure to a microscopic laser beam 100-300 μm in diameter: Micro Epidermal Necrotic Debris (MEND) in the case of ablative lasers or Micro Thermal Zones (MTZ) for nonablative lasers [bib_ref] Fractional Photothermolysis: A New Concept for Cutaneous Remodeling Using Microscopic Patterns of..., Manstein [/bib_ref] [bib_ref] Fractional photothermolysis: a novel aesthetic laser surgery modality, Hantash [/bib_ref] [fig_ref] Figure 1: Columns of fractional laser beamsFig [/fig_ref].
A laser is a device emitting an intensive, coherent and monochromatic light in a constant or pulsed way. An impulse of proper wavelength that lasts shorter than the period of thermal relaxation of the target tissue makes it possible to act in a precisely defined area without damaging the surrounding tissues [bib_ref] Selective photothermolysis: precise microsurgery by selective absorption of pulsed radiation, Anderson [/bib_ref].
In photorejuvenation procedures the water contained in the skin and collagen are the skin chromo-phores that absorb most of the emitted energy. In general, laser fractional skin rejuvenation can be divided into ablative, with vaporization of the epidermis and possibly the upper layers of the dermis (MEND), where the depletion is filled by collagen fibers from the surrounding undamaged areas, and non-ablative, with controlled overheating of tissues (MTZ). The temperature of the overheated area reaches 55-56ºC, which results in reduced tension caused by the constriction of the cruciate ligaments between spirally twisted polypeptide chains creating collagen fibers. New collagen is produced at the same time as the ongoing skin regeneration takes place in areas subjected to laser, which was confirmed in histopathological examinations. The epidermis desquamates superficially after a few days of erythema and edema of the tissue. Skin tension improves, skin surface turns smoother and color becomes more uniform. The facial skin is most frequently subjected to the procedure, but the neck, neckline and dorsal surface of hands are also common localizations. While preparing for the procedure the skin is covered with Prilocaine cream (EMLA). The first dermatological fractional laser, which was introduced in 2004, was non-ablative Fraxel SR 750, 1540 nm, currently replaced by the higher power Fraxel SR 1500 that is capable of achieving greater penetration. Currently, erbium glass lasers 1540-1550 nm (Fraxel, Mosaic, etc.), which can create columns of controlled overheating at the depth of about 1 mm without vaporization or breaking the skin surface integrity Fractional laser therapy -the next step in alleviating the symptoms of skin aging (own observations) are used in fractional technology for non-ablative skin rejuvenation.
Ablative Erb-YAG 2940 nm and CO 2 10 600 nm lasers are characterized by forming column skin vaporization areas of -depending on the impulse energy set -up to approximately 90 μm and 2.5 mm in depth in the case of Erb-YAG and CO 2 lasers, respectively, which makes it possible to reduce large wrinkles. A new trend aimed at intensification of therapeutic effects in as short a time as possible is to use two lasers during one session, e.g., a deep penetrating CO 2 (10 600 nm) laser and Erb-YAG (2940 nm) laser operating on the surface or a non-ablative (1550 nm) laser and Erb-YAG (2940 nm) ablative laser [bib_ref] Fractional Photothermolysis: A New Concept for Cutaneous Remodeling Using Microscopic Patterns of..., Manstein [/bib_ref] [bib_ref] Consensus recommendations on the use of an erbium-doped 1,550-nm fractionated laser and..., Sherling [/bib_ref] [bib_ref] Novel use of erbium:YAG (2,940-nm) laser for fractional ablative photothermolysis in the..., Lapidoth [/bib_ref].
Qualifying a patient for a given type of procedure includes considering the degree of skin aging symptoms, the patient's expectations and therapeutic possibilities, which depend on the equipment available. Information on the possible side effects of laser therapy with the absolute need of SPF 50 photo protection seems extremely important. Pregnancy, active infectious lesions (e.g. herpes, mycosis), active neoplastic disease, psoriasis and lichen planus (active phase of the disease), proneness to scarring, use of phototoxic substances, treatment with retinoid drugs (less than 6 months following the end of treatment), tanning (approximately 8 weeks following the exposure to UV radiation), epilepsy (in the case of procedures using visible spectrum waves, i.e. 400-780 nm), mental diseases that limit the possibility of giving informed consent to the procedure and unreal expectations are contraindications to the procedure. A patient should be informed about possible results of the procedures and should be aware of the fact that no laser treatment or any other methods can stop time.
The Glogau Classification of Photoaging is usually used to assess the degree of skin aging.
Grade 1 -minimal wrinkles, particularly around the eyes and lips. The wrinkles appear during movements of the facial muscles and the skin. Such lesions are characteristic of the skin of thirty-year-old people.
Grade 2 -mild pigment changes appear, telangiectasias occur, wrinkles are more visible and can be noticed independently of facial muscle movements -wrinkles at rest.
Grade 3 -characteristic features also include gravitational changes; loss of skin firmness and proper tension are visible, the skin gets more flabby compared to the previous grade.
Grade 4 -numerous wrinkles visible. Gravitational changes and photoaging occur; the skin is flabby, thin, dry, devoid of firmness; spots of skin discolorations are present.
We present a 47-year-old patient treated twice a year with 532 nm KTP laser for three years due to numerous facial telangiectasias. She was qualified for treatment with fractional CO 2 laser due to skin discolorations and visible wrinkles occurring mostly around the eyes. Three procedures were performed in 4-6 week intervals. The energy of 35 mJ and density of 100 ablation columns of 300 μm × 1 cm² were applied. The area of the procedure was covered with EMLA anes-thetic cream 30-40 minutes before the treatment. The patient tolerated the procedures well, with minimal discomfort, and the healing was without complications. Self-limiting scattered bleeding occurred on the second day after the procedure, about the possibility of which the patient had been informed. During the period after the treatment she was administered preventive external therapy for 7 days, creams accelerating skin healing subsequently and general antibiotic therapy for 5 days. A prophylactic cream with UV SPF 50 was used between the procedures and 30 days after the last procedure and it was recommended to prolong the treatment for the period of increased exposure to sun in order to prevent skin discoloration. Laser treatment resulted in harmonization of the skin color, increased flexibility and a significant reduction of wrinkles, which reduced the effects of skin aging from degree 3 to 2 in the Glogau classification (Figs. [bib_ref] Consensus recommendations on the use of an erbium-doped 1,550-nm fractionated laser and..., Sherling [/bib_ref] [bib_ref] Novel use of erbium:YAG (2,940-nm) laser for fractional ablative photothermolysis in the..., Lapidoth [/bib_ref] [bib_ref] Fractional photothermolysis: a novel aesthetic laser surgery modality, Hantash [/bib_ref].
Due to the idea of fractionality a part of skin in columns and their direct surrounding is only remodeled with each procedure. Several procedures, usually 3-5 in 4-week intervals, must be performed to optimize the effect and rebuild the whole skin. Skin smoothing is often visible directly after the treatment, which is connected partially with tissue edema. Remodeling of the collagen and improvement in the skin smoothness take place during several months after the procedure and the long-term results can be evaluated afterwards. The period of healing and the risk of undesired effects such as scarring, discoloration, depigmentation or infections have been significantly reduced by introducing the fractional technology. Local treatment with antibiotic creams or silver sulfadiazine are applied until the skin is fully healed, especially after ablative procedures to minimize the risk of possible complications. After nonablative procedures treatment with creams accelerating skin healing is usually sufficient. Photo protection with SPF 50 should absolutely be applied for 4 weeks following the procedure, regardless of its type. Preven- . A 47-year old patient before the first procedure using fractional CO 2 laser . The same patient directly after CO 2 fractional laser treatment tive antibiotic therapy and treatment for Herpes S virus are introduced after extensive ablative procedures. After non-ablative procedures, the post-treatment period is usually uncomplicated and healing faster (so-called lunch time treatment). Therefore, it is more frequently chosen by patients who cannot be excluded from their professional life even for several days. The ablative procedure involves a longer healing period due to scattered vaporization of the epidermis (Erb-YAG laser) and creation of necrotic columns (CO 2 laser). Scattered bleeding can occur and the period of re-epithelization coming from the undamaged areas takes 1-2 days, compared to 7-10 days after non-fractional non-ablative procedures. Subsequently, the skin becomes remodeled, with the activated fibroblasts forming new collagen and filling the microscopic column defects [bib_ref] Consensus recommendations on the use of an erbium-doped 1,550-nm fractionated laser and..., Sherling [/bib_ref] [bib_ref] Novel use of erbium:YAG (2,940-nm) laser for fractional ablative photothermolysis in the..., Lapidoth [/bib_ref] [bib_ref] Fractional photothermolysis: a novel aesthetic laser surgery modality, Hantash [/bib_ref] [bib_ref] Recent advances In laser dermatology, Butani [/bib_ref].
Aging is a progressive phenomenon, which still can be alleviated and delayed by systemic or local use of estrogens in women. These improve the nourishment of the skin, which contains a large number of estrogen receptors [bib_ref] Wpływ hormonów na procesy starzenia się skóry, Trznadel-Budźko [/bib_ref] [bib_ref] Dermatologiczne aspekty menopauzy, Trznadel-Budźko [/bib_ref].
Apart from hormone therapy, which is applied in the menopausal period and is usually recommended for other reasons than skin aging, fractional laser therapy can be conducted, which at present makes it possible to induce natural processes of tissue regeneration leading to remodeling the aging skin with minimum traumatization. As a result, skin tension improves, the skin sur-face becomes smoother and the color uniform, which on the whole creates the image of so-called photorejuvenation.
# Disclosure
The authors report no conflict of interest. . The same patient before the second CO 2 fractional laser treatment . The same patient one month after three CO 2 fractional laser procedures (skin smoothing, wrinkles reversed)
[fig] Figure 1: Columns of fractional laser beamsFig. 2. The picture of the skin. Necrotic columns in case of ablative fractional laser useFig. 3. 10 600 nm CO 2 fractional ablative laser Fig. 4. 1550 nm erbium glass non-ablative fractional laser (Mosaic) [/fig]
|
In vitroassembly of the bacterial actin protein MamK from ‘CandidatusMagnetobacterium casensis’ in the phylumNitrospirae
# Supplemental methods
Cloning, expression and purification. The mamK gene was amplified from Mcas genomic DNA using the primers WB643 (AAGGCATATGACAAAAACAAAGATACTTAACA) and WB567
(TTCAAGCTTTTATCGCTTGCTTACTTCGTCCCA) and then digested with NdeI and HindIII. The digested products were purified and ligated with the pET28a vector (Invitrogen) and digested with the same restriction enzymes to create the plasmid pWYE369 for expressing N-terminal his-tagged MamK. To construct C-terminal his-tagged MamK expressing plasmid pWYE403, the mamK gene was amplified using primers WB643 and WB860 (GTGAAGCTTTCGCTTGCTTACTTCG) and then cloned into the pET22b vector (Invitrogen) at restriction sites NdeI/HindIII.
The plasmids pWYE369 and pWYE403 were transformed into Escherichia coli C43 (DE3) cells and verified by DNA sequencing. The cells were grown in LB medium at 37°C and induced with 0.5 mM IPTG after reaching an OD 600 of 0.8. After 4 hours of induction, the cells were harvested and disrupted by sonication in buffer A containing 150 mM NaCl, 10% glycerol and 20 mM Tris-HCl (pH 8.0). After centrifugation at 15,000 ×g for 30 min, the his-tagged MamK protein in the supernatant was purified by Ni 2+ affinity chromatography (GE healthcare, USA) and gel filtration (Superdex 75 10/300 GL high-performance column, GE healthcare, USA). To obtain protein without his tag, purified N-terminal his-tagged MamK was subsequently incubated with immobilizing bovine thrombin (Sigma, USA) at 4°C to 3 remove the poly-histidine tags. The resulting MamK protein only with three residues at the N-terminus was eluted from the thrombin-agarose, and passed through a Ni 2+ affinity column again to remove the remove uncleaved protein and the cleaved His tags. The purified protein was centrifuged at 500,000 ×g for 30 min at 4°C to remove the aggregates and stored in aliquots at -80°C for further study. The protein concentration was determined using the BCA protein assay kit (Pierce, Rockford) (1) where n is the dilution factor and 25,120 cm -1 M -1 is the molar extinction coefficient of the MamK protein at 280 nm (estimated using the ProtParam tool, http://web.expasy.org/protparam/). The degree of labeling (D, moles dye per mole protein) was calculated as equation 2:
[formula] 494 An D= 71000 M (2) 280 494 A (A 0.11) n M 25120 4 [/formula]
where n is the dilution factor and 71,000 cm -1 M -1 is the molar extinction coefficient of the Alexa Fluor 488 dye at 494 nm.
Characterization of MamK ATPase features. Enzyme activity was determined using a modified malachite-green assay as described above, with slight further modifications. Purified protein at a final concentration of 2.0 μM in the nucleotide hydrolysis reaction was used to determine ATPase activity under various conditions. The optimum temperature was determined after incubating the reaction mixtures at different temperatures (30-70°C). The effect of pH on enzyme activity was measured at different pH values (5-11) at 37°C. The relative enzyme activity was calculated as a percentage of the maximal activity. For the thermal stability assay, MamK was incubated at various temperatures (30-70°C) for 30 min, and the residual enzyme activity was subsequently determined. The residual enzyme activity was calculated as a percentage of the starting activity.
The effects of various metal ions on the enzyme activity were determined after incubating the reaction mixtures with MgCl 2 , CaCl 2 , FeSO 4 , CuCl 2 and NiSO 4 at concentrations ranging from 0 to 2.0 mM. To avoid the oxidation of iron (II), FeSO 4 was dissolved in the reaction buffer containing a reducing agent (2 mM ascorbic acid), which had been determined to have no effects on the MamK ATPase activity.
The enzyme activity assayed under the same condition, but in the absence of metal ions, was set at 100%.
## 5
To evaluate the effect of salt, the enzyme activity was measured in the presence of 0-500 mM KCl and NaCl. The enzyme activity assayed under the same conditions, but in the absence of salts, was set at 100%.
The kinetic parameters were determined through Lineweaver-Burk curves using varying concentrations of ATP from 0.01 to 0.50 mM. Substrates at different concentrations were incubated with MamK using a standard ATPase activity assay as described above. The velocity (μM min -1 Pi /μM protein, y axis) was defined as the change in catalytic rate of the concentration of released phosphate per minute and per μM protein, which was plotted against the concentration of the substrate (μM, x axis). Subsequently, the curves were fitted using the Lineweaver-Burk curve method, from which the Michaelis constants (Vmax and Km) for each enzymatic reaction were derived. The initial velocities were used to determine Vmax and Km.
The catalytic constant (Kcat) was determined after dividing the Vmax by the concentration of the enzyme. The catalytic efficiency was defined as Kcat/Km (min -1 μM -1 ). Approximately 4.5 μM of MamK was doped with 15% Alexa 488-labelled monomer in polymerization buffer and loaded onto a cover slip, followed by polymerization using saturating amounts of ATP at 37°C. The fluorescently labelled filaments were observed at 0 to 25 μm from the bottom of the slide using a Leica SP8 confocal microscopy.
## Determination of the physicochemical conditions for polymerization. various
Determination of the polymerization by pelleting assay. Because the ATPase 7 activity and assembly of MamK from Mcas was inhibited at 4°C , polymerization was assessed using a pelleting assay. To simultaneously determinate the ATP hydrolysis and the polymerization of MamK protein, the polymerizing sample at the same time point was separated in two parts. One was mixed with 4 volumes of chromogenic reagent and incubated at 37°C for 30 min, while another was centrifuged at 500,000 ×g at 4°C for 30 min. Subsequently, they were subjected to determine the amount of Pi (630 nm) and protein monomers concentration (595 nm) by spectrophotometer and the pellet by SDS-PAGE.
Statistical analyses. Significant differences were analyzed using SPSS software (version 13.0; http://www-01.ibm.com/software/analytics/spss/). The data conducted under various conditions were compared by one-way analysis of variance (ANOVA) and Duncan's multiple range tests. All measurements were performed at least in triplicate, and each value represented the mean ± standard deviation (SD). P-values less than 0.05 were considered statistically significant. |
Genome-scale target identification in Escherichia coli for high-titer production of free fatty acids
To construct a superior microbial cell factory for chemical synthesis, a major challenge is to fully exploit cellular potential by identifying and engineering beneficial gene targets in sophisticated metabolic networks. Here, we take advantage of CRISPR interference (CRISPRi) and omics analyses to systematically identify beneficial genes that can be engineered to promote free fatty acids (FFAs) production in Escherichia coli. CRISPRi-mediated genetic perturbation enables the identification of 30 beneficial genes from 108 targets related to FFA metabolism. Then, omics analyses of the FFAs-overproducing strains and a control strain enable the identification of another 26 beneficial genes that are seemingly irrelevant to FFA metabolism. Combinatorial perturbation of four beneficial genes involving cellular stress responses results in a recombinant strain ihfA L− -aidB + -ryfA M− -gadA H− , producing 30.0 g L −1 FFAs in fed-batch fermentation, the maximum titer in E. coli reported to date. Our findings are of help in rewiring cellular metabolism and interwoven intracellular processes to facilitate high-titer production of biochemicals.
M icrobial biosynthesis of a desired product is a sophisticated process that usually involves the coordination of multiple metabolic intermediates and reactions that are subject to complex interactions [bib_ref] Engineering cellular metabolism, Nielsen [/bib_ref] [bib_ref] DCEO biotechnology: tools to design, construct, evaluate, and optimize the metabolic pathway..., Chen [/bib_ref]. On the one hand, each of these intermediates flows to other metabolic pathways for the biosynthesis of cellular components or metabolites, acting as one of the multibranched nodes in the topology of metabolic landscapes [bib_ref] It is all about metabolic fluxes, Nielsen [/bib_ref] [bib_ref] Engineering synthetic recursive pathways to generate non-natural small molecules, Felnagle [/bib_ref]. On the other hand, each reaction in the biosynthetic pathway relies greatly on many cellular processes, such as gene transcription and translation to enable enzyme and metabolic functions [bib_ref] Quantifying absolute protein synthesis rates reveals principles underlying allocation of cellular resources, Li [/bib_ref] [bib_ref] Evaluation of 244,000 synthetic sequences reveals design principles to optimize translation in..., Cambray [/bib_ref] , biomass accumulation, and DNA replication to ensure cell survival and proliferation [bib_ref] Growth-limiting intracellular metabolites in yeast growing under diverse nutrient limitations, Boer [/bib_ref] [bib_ref] Metabolic regulation of cell growth and proliferation, Zhu [/bib_ref]. Additionally, biosynthesis may place cells under nutrition or redox stresses, and metabolic rewiring must occur to address such challenges to maintain cellular viability and functions [bib_ref] Engineering cellular metabolism, Nielsen [/bib_ref] [bib_ref] Metabolic regulation of cell growth and proliferation, Zhu [/bib_ref] [bib_ref] Strategies and challenges for metabolic rewiring, Yu [/bib_ref]. As such, manipulation of other metabolic pathways or cellular processes might lead to an unexpected improvement in the biosynthesis of the desired product due to distant effects of genetic modulations or unknown regulatory interactions [bib_ref] Efficient protein production by yeast requires global tuning of metabolism, Huang [/bib_ref] [bib_ref] Lipid production in Nannochloropsis gaditana is doubled by decreasing expression of a..., Ajjawi [/bib_ref] [bib_ref] Morphology engineering of bacteria for bioproduction, Jiang [/bib_ref]. Thus, to construct a superior microbial cell factory for enhanced product biosynthesis, it is highly desirable to perform genome-wide identification of beneficial gene targets in complex intracellular interaction networks that can be reengineered for this purpose.
Free fatty acids (FFAs) are feedstocks for the manufacture of detergents, lubricants, cosmetics, and pharmaceutical ingredients [bib_ref] Microbial production of fatty acid-derived fuels and chemicals, Lennen [/bib_ref]. By engineering metabolic pathways, FFAs production was enhanced in various microorganisms, such as the model microorganisms Escherichia coli [bib_ref] Modular optimization of multi-gene pathways for fatty acids production in E. coli, Xu [/bib_ref] and Saccharomyces cerevisiae [bib_ref] Reprogramming yeast metabolism from alcoholic fermentation to lipogenesis, Yu [/bib_ref] and the oleaginous microorganisms Yarrowia lipolytica [bib_ref] Combining metabolic engineering and process optimization to improve production and secretion of..., Ledesma-Amaro [/bib_ref] and Rhodococcus opacus [bib_ref] Engineering of an oleaginous bacterium for the production of fatty acids and..., Kim [/bib_ref]. Taking E. coli as an example, overexpression of key enzymes in the FFA biosynthetic pathway, such as acetyl-coenzyme A (CoA) carboxylase [bib_ref] A process for microbial hydrocarbon synthesis: overproduction of fatty acids in Escherichia..., Lennen [/bib_ref] [bib_ref] Overproduction of free fatty acids in E. coli: implications for biodiesel production, Lu [/bib_ref] and fatty acyl-ACP thioesterase [bib_ref] A process for microbial hydrocarbon synthesis: overproduction of fatty acids in Escherichia..., Lennen [/bib_ref] [bib_ref] Overproduction of free fatty acids in E. coli: implications for biodiesel production, Lu [/bib_ref] [bib_ref] Microbial production of fatty-acid-derived fuels and chemicals from plant biomass, Steen [/bib_ref] [bib_ref] Microbial production of short-chain alkanes, Choi [/bib_ref] , or deletion of key genes in the beta-oxidation pathway, such as the fadD [bib_ref] A process for microbial hydrocarbon synthesis: overproduction of fatty acids in Escherichia..., Lennen [/bib_ref] [bib_ref] Overproduction of free fatty acids in E. coli: implications for biodiesel production, Lu [/bib_ref] [bib_ref] Microbial production of fatty-acid-derived fuels and chemicals from plant biomass, Steen [/bib_ref] [bib_ref] Microbial production of short-chain alkanes, Choi [/bib_ref] or fadE 20,21 gene, was implemented to improve FFAs production. In particular, optimal combinatorial expression of 15 crucial genes in the FFA biosynthetic pathway with fadD deletion in E. coli BL21 resulted in a titer of 8.6 g L −1 FFAs in fed-batch cultivation with glucose as the carbon source [bib_ref] Modular optimization of multi-gene pathways for fatty acids production in E. coli, Xu [/bib_ref]. Despite these efforts, further improvement of FFAs production is hampered by the limited understanding of cellular rewiring mechanism and linkage between FFA biosynthesis and other cellular processes [bib_ref] Strategies and challenges for metabolic rewiring, Yu [/bib_ref] [bib_ref] Engineering microbial fatty acid metabolism for biofuels and biochemicals, Marella [/bib_ref]. In addition, it remains difficult to systematically identify beneficial gene targets using laborious genome engineering methods, such as homologous recombination-based gene knockout [bib_ref] One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products, Datsenko [/bib_ref]. These hurdles severely restrict the scope and speed of engineering microbial hosts to obtain a superior overproducer.
CRISPR interference (CRISPRi) enables sequence-specific repression of gene transcription by pairing dCas9 with a synthetic sgRNA [bib_ref] Repurposing CRISPR as an RNA-guided platform for sequencespecific control of gene expression, Qi [/bib_ref] [bib_ref] CRISPR interference (CRISPRi) for sequence-specific control of gene expression, Larson [/bib_ref]. CRISPRi can be easily designed and readily implemented, enabling rapid screening of beneficial knockdown gene targets. As a promising tool for gene modulation, CRISPRi has been generally used for improved biosynthesis of desired products in E. coli [bib_ref] Enhancing flavonoid production by systematically tuning the central metabolic pathways based on..., Wu [/bib_ref] [bib_ref] Redirecting metabolic flux via combinatorial multiplex CRISPRi-mediated repression for isopentenol production in..., Tian [/bib_ref] [bib_ref] CRISPR interference-guided multiplex repression of endogenous competing pathway genes for redirecting metabolic..., Kim [/bib_ref] [bib_ref] A systematic optimization of medium chain fatty acid biosynthesis via the reverse..., Wu [/bib_ref] and many other microbial hosts, such as Corynebacterium glutamicum [bib_ref] Corynebacterium glutamicum metabolic engineering with CRISPR interference (CRISPRi), Cleto [/bib_ref] and cyanobacteria of the genus Synechocystis 31 . However, in most of these cases, the candidate gene targets were restricted to direct competitive pathways [bib_ref] Redirecting metabolic flux via combinatorial multiplex CRISPRi-mediated repression for isopentenol production in..., Tian [/bib_ref] [bib_ref] CRISPR interference-guided multiplex repression of endogenous competing pathway genes for redirecting metabolic..., Kim [/bib_ref] [bib_ref] A systematic optimization of medium chain fatty acid biosynthesis via the reverse..., Wu [/bib_ref] [bib_ref] Corynebacterium glutamicum metabolic engineering with CRISPR interference (CRISPRi), Cleto [/bib_ref] [bib_ref] Multiple gene repression in cyanobacteria using CRISPRi, Yao [/bib_ref] , which cannot fully unleash the cellular potential for biochemical overproduction [bib_ref] Strategies and challenges for metabolic rewiring, Yu [/bib_ref] [bib_ref] Engineering microbial fatty acid metabolism for biofuels and biochemicals, Marella [/bib_ref] [bib_ref] Systems metabolic engineering strategies: integrating systems and synthetic biology with metabolic engineering, Choi [/bib_ref].
As an important supplementary method, omics analysis of differentially performing strains allows the acquisition of comprehensive data related to the mechanisms of cellular metabolism [bib_ref] Efficient protein production by yeast requires global tuning of metabolism, Huang [/bib_ref] [bib_ref] Systems metabolic engineering strategies: integrating systems and synthetic biology with metabolic engineering, Choi [/bib_ref] , providing additional available clues regarding candidate gene targets in complex intracellular interaction networks for enhanced biosynthesis of the desired product [bib_ref] Lipid production in Nannochloropsis gaditana is doubled by decreasing expression of a..., Ajjawi [/bib_ref] [bib_ref] Dynamic control of the mevalonate pathway expression for improved zeaxanthin production in..., Shen [/bib_ref].
Here, we take advantage of CRISPRi to readily and rapidly downregulate gene expression and omics analyses to identify potential gene targets in cellular interaction networks, enabling rapid, systematic, and effective identification of chromosomal gene targets that can be engineered for enhanced production of FFAs in E. coli. Upon CRISPRi-mediated gene repression, we first identify beneficial genes that are directly related to the FFA metabolism. Next, transcriptomic and proteomic analyses are performed for three recombinant FFAs-overproducing strains and a control strain, and beneficial genes in cellular networks are further identified. Upon combinatorial modulation of the identified genes, the engineered strain ihfA L− -aidB + -ryfA M− -gadA H− produces 30.0 g L −1 FFAs (0.689 g L −1 h −1 productivity) in fedbatch fermentation. We speculate that the high titer of FFAs obtained in the ihfA L− -aidB + -ryfA M− -gadA H− strain benefits from the enhanced cellular stress responses, which protect cells from cytotoxic byproducts during the course of microbial fermentation.
# Results
Identification of beneficial gene targets in the pathways related to FFA metabolism using the CRISPRi system. We employed glycerol as the carbon source for the biosynthesis of FFAs. In E. coli, glycerol is catalyzed to the key intermediates glyceraldehyde-3P, pyruvate, acetyl-CoA, and malonyl-ACP, in that order, and then converted to acyl-ACP through the fatty acid biosynthetic pathway . However, these intermediates are also involved in other metabolic pathways, resulting in the diversion of carbon flux away from FFAs. Expression of the truncated fatty acyl-ACP thioesterase TesA′ (with the leader sequence deleted) could release FFA intermediates from ACP in the cytosol [bib_ref] Defective export of a periplasmic enzyme disrupts regulation of fatty acid synthesis, Cho [/bib_ref]. Nevertheless, synthetic FFAs are degraded naturally by the betaoxidation process. To channel carbon flux toward the production of FFAs and repress the degradation of FFAs, we chose 108 chromosomal genes in a hypothesis-driven manner for modulation by CRISPRi. The 108 targets comprise 26 genes in the module of upstream carbon flux diversion (those in the competitive pathways consuming glycerone-P, glyceraldehyde-3P, or pyruvate); 15 genes in the module of downstream carbon flux diversion (those in the competitive pathways consuming acetyl-CoA or malonyl-ACP); 24 genes in the module of amino acid metabolism that consume glycerate-3P, pyruvate, or acetyl-CoA; 6 genes in the module of beta-oxidation that degrade FFAs; and 37 genes in the module of transcription factors that regulate multiple pathways or cellular physiology and Supplementary Data 1).
To facilitate target identification, we constructed a library of 108 synthetic sgRNAs that represses the expression of the 108 chromosomal genes in the above five modules with high efficiency [bib_ref] Repurposing CRISPR as an RNA-guided platform for sequencespecific control of gene expression, Qi [/bib_ref]. The P Trc promoter was selected to control the expression of catalytically dead Cas9 (dCas9) in CRISPRi, which could achieve the expected gene repression without affecting cell growth (see details in the Supplementary Note 1). Plasmid pCF expressing TesA′ and dCas9 was transformed into the BL21(DE3) strain to generate the starting strain CF. We introduced plasmids harboring each of the 108 synthetic sgRNAs individually into the CF strain to implement gene perturbation . Plasmid Sg-0 expressing sgRNA0 without the target site complementary sequence was introduced into the CF strain to generate the Control strain, which did not modulate any chromosomal genes and produced 631 mg L −1 FFAs. Any CRISPRi-engineered strain that could increase the FFAs titer by over 20% compared with the Control strain was sorted out. As a result, 30 beneficial targets were identified, and the fadE H− and fadR H− strains achieved FFAs titers of 1232 and 1193 mg L −1 , which were 95% and 89% higher than that of the Control strain, respectively [fig_ref] Figure 2: CRISPRi-based identification of beneficial genes for FFAs production [/fig_ref].
The identified beneficial genes were mostly distributed in the modules of upstream carbon flux diversion [fig_ref] Figure 2: CRISPRi-based identification of beneficial genes for FFAs production [/fig_ref] , downstream carbon flux diversion [fig_ref] Figure 2: CRISPRi-based identification of beneficial genes for FFAs production [/fig_ref] , beta-oxidation [fig_ref] Figure 2: CRISPRi-based identification of beneficial genes for FFAs production [/fig_ref] , and transcription factors [fig_ref] Figure 2: CRISPRi-based identification of beneficial genes for FFAs production [/fig_ref]. This result suggested that reducing the carbon flux to the byproducts phosphatidylglycerol (gpsA, plsY, pgpB, and pgpC), D-lactate (dld), acetate (pta, yccX, and eutE), and biotin (bioC, bioH, and bioD), or blocking fatty acid degradation (fadE, fadB, atoB, and yqeF) contributed to the enhanced production of FFAs. Reducing the expression of a few transcription factors (namely, fabR, fadR, asnC, fnr, cra, gadW, glcC, glpR, ihfA, ihfB, and torR) also enhanced the production of FFAs [fig_ref] Figure 2: CRISPRi-based identification of beneficial genes for FFAs production [/fig_ref] , which might benefit from the global regulation of gene expression. However, inhibiting amino acid biosynthetic pathways seemed unfavorable for FFAs production, as most of the engineered strains exhibited reduced titers of FFAs [fig_ref] Figure 2: CRISPRi-based identification of beneficial genes for FFAs production [/fig_ref] , and some of these strains were even unable to grow (Supplementary [fig_ref] Figure 2: CRISPRi-based identification of beneficial genes for FFAs production [/fig_ref]. A possible reason is that while these amino acids were necessary for cell viability, the reduced amount of amino acids obtained upon CRISPRi perturbation was insufficient to maintain normal cell physiology.
Among the 30 beneficial gene targets, 20 (gpsA, plsY, pgpB, pgpC, dld, yccX, bioC, bioH, bioD, gabD, gadA, asnC, fnr, cra, gadW, glcC, glpR, ihfA, ihfB, torR) have not been engineered for FFAs production in previous studies. We found three favorable metabolic pathways that have not been previously modulated for FFAs production, namely, the biotin, phosphatidylglycerol, and glutamate biosynthetic pathways [fig_ref] Figure 2: CRISPRi-based identification of beneficial genes for FFAs production [/fig_ref]. Even in the previously modulated pathways, we identified beneficial gene targets that have not been engineered for FFAs production in previous studies, such as dld in the lactate biosynthetic pathway and yccX and eutE in the acetate formation pathway [fig_ref] Figure 2: CRISPRi-based identification of beneficial genes for FFAs production [/fig_ref].
Additionally, we found that the high repression of pta in the acetate formation pathway (strain pta H− ) increased FFAs production by 86% [fig_ref] Figure 2: CRISPRi-based identification of beneficial genes for FFAs production [/fig_ref] , while deletion of this gene in a previous study had a negative effect on FFAs production 35 . Collectively, these results demonstrate that CRISPRi is useful for the systematic identification of beneficial targets that can be engineered for improved biosynthesis of the desired product.
Fine tuning of gene repression is enormously important for improving the biosynthesis of desired products. Regarding the genes in the competitive pathways, the gene repression level needs to be fine-tuned so that the target genes can still be expressed to a certain degree to maintain cellular viability while resources are maximally redirected toward the target metabolite [bib_ref] Enhancing flavonoid production by systematically tuning the central metabolic pathways based on..., Wu [/bib_ref] [bib_ref] Redirecting metabolic flux via combinatorial multiplex CRISPRi-mediated repression for isopentenol production in..., Tian [/bib_ref] [bib_ref] A systematic optimization of medium chain fatty acid biosynthesis via the reverse..., Wu [/bib_ref]. A few approaches have been developed to tune the efficiency of CRISPRi-mediated gene repression [bib_ref] Repurposing CRISPR as an RNA-guided platform for sequencespecific control of gene expression, Qi [/bib_ref]. Truncation of the basepairing region of sgRNA from the 5′ end could severely decrease the repression efficiency; however, it is difficult to achieve a medium repression level [bib_ref] Repurposing CRISPR as an RNA-guided platform for sequencespecific control of gene expression, Qi [/bib_ref]. A single mismatch in the seed region or the other 8-nt sequence of the base-pairing region could achieve low or medium repression efficiency, respectively, but the mismatch in the seed region might confer off-target effects [bib_ref] Repurposing CRISPR as an RNA-guided platform for sequencespecific control of gene expression, Qi [/bib_ref] [bib_ref] Guide-target mismatch effects on dCas9-sgRNA binding activity in living bacterial cells, Feng [/bib_ref] , thus failing to repress the desired target. Many studies have demonstrated that the repression level in gene expression and the target distance from the transcription start site generally had an inverse relationship [bib_ref] Repurposing CRISPR as an RNA-guided platform for sequencespecific control of gene expression, Qi [/bib_ref] [bib_ref] Programmable repression and activation of bacterial gene expression using an engineered CRISPR-Cas..., Bikard [/bib_ref] [bib_ref] Engineering Escherichia coli for malate production by integrating modular pathway characterization with..., Gao [/bib_ref]. Most recently, bases surrounding the protospacer adjacent motif (PAM) on the target sequence were also shown to affect the CRISPRi efficiency [bib_ref] On-target activity predictions enable improved CRISPR-dCas9 screens in bacteria, Calvo-Villamanan [/bib_ref]. We here tuned the repression of each beneficial gene with medium or low efficiency by applying the sgRNA gene M− or gene L− , binding the nontemplate DNA strand at the middle or terminal region, respectively [fig_ref] Figure 3: Tuning of the repression of beneficial genes for the improvement of FFAs... [/fig_ref]. Plasmids harboring each of the synthetic Gene targets related to FFA metabolism that were selected for modulation by the CRISPRi system. a Schematic of the metabolic or regulatory pathways related to FFA biosynthesis in E. coli BL21(DE3), which were classified into five modules, namely, upstream carbon flux diversion, downstream carbon flux diversion, amino acid metabolism, beta-oxidation, and transcription factor, represented in blue, orange, purple, green, and golden backgrounds, respectively. The 108 candidate genes in the above five modules are marked with blue, orange, purple, green, and golden dots, respectively (see Supplementary Data 1). 2854 represents the gene numbered ECD_02854 in BL21(DE3). The fatty acid biosynthetic pathway is shown as a red line, from glycerol to acyl-ACP and conversion to FFAs by TesA′ (encoded by the tesA′gene marked with a red dot). The key metabolites glycerone-P, glyceraldehyde-3P, glycerate-3P, pyruvate, acetyl-CoA, and malonyl-ACP are represented as numbers 1 to 6, respectively. b Plasmid constructs for the expression of tesA′, dCas9 and the sgRNAs gene H/M/L− . Cm R , chloramphenicol-resistance gene; Amp R , ampicillin-resistance gene; P15A and ColE1, replication origin. sgRNAs were thus individually introduced into the starting strain CF, and FFAs production of the corresponding recombinant strains was determined. For example, in strain gpsA L− , the sgRNA gpsA L− was expressed, and the gpsA gene was repressed with low efficiency. The results showed that the gpsA L− , pfkA L− , yccX M− , gabD L− , gadA M− , atoB M− , atoB L− , glpR L− , ihfA M− , and ihfA L− strains produced higher FFAs titers than strains with the corresponding gene repressed with high efficiency [fig_ref] Figure 3: Tuning of the repression of beneficial genes for the improvement of FFAs... [/fig_ref] , and the gpsA L− , yccX M− , atoB M− , and ihfA L− strains achieved the highest FFAs titers, which were > 100% higher than that of the Control strain [fig_ref] Figure 3: Tuning of the repression of beneficial genes for the improvement of FFAs... [/fig_ref]. These results suggested the importance of fine-tuning gene expression in the production of target metabolites.
Identification of potential targets in the cellular network via proteomic and transcriptomic analyses. To further explore potential gene targets that could facilitate FFAs production, we focused on the strains with high production of FFAs. In the above studies on CRISPRi-mediated genetic perturbation, the six engineered strains (fadE H− , fadR H− , yccX M− , atoB M− , gpsA L− , and ihfA L− ) achieved the highest FFAs titers [fig_ref] Figure 2: CRISPRi-based identification of beneficial genes for FFAs production [/fig_ref] and [fig_ref] Figure 3: Tuning of the repression of beneficial genes for the improvement of FFAs... [/fig_ref] , which were > 89% higher than that of the Control strain. Among these gene targets, fadR encodes the transcription factor FadR, which regulates the processes of fatty acid biosynthesis, degradation, and transport 40 ; ihfA encodes the alpha subunit of integration host factor, which affects many cell functions (including DNA replication, recombination, and transcription) and influences gene expression on a global scale 41 ; and the other four genes are involved in the competitive pathways of FFA biosynthesis (yccX and gpsA) and the beta-oxidation pathway (fadE and atoB). Considering the role of transcription factors in the extensive regulation of metabolic processes, we selected the fadR H− and ihfA L− strains for further analysis. Since overexpression of FadR could significantly enhance FFAs production 40 , we constructed a FadR-overexpressing strain (fadR + ) that could produce a high titer of FFAs (1462 mg L −1 ) [fig_ref] Figure 4: ihfA Lvs [/fig_ref]. Although these three FFAsoverproducing strains have different genotypes, they might have convergent regulatory responses that facilitate FFA biosynthesis. Thus, these three engineered strains (ihfA L− , fadR H− and fadR + ) and the Control strain were applied for proteomic and transcriptomic analyses to investigate the potential target genes and their underlying mechanisms in FFAs overproduction.
Cells were sampled at the early stationary phase (24 h after IPTG induction) [fig_ref] Figure 4: ihfA Lvs [/fig_ref]. Hundreds of proteins and transcripts were differentially abundant in ihfA L− vs. Control, fadR H− vs. Control, and fadR + vs. Control (Supplementary Figs. 5 and 6). The results indicated that FFAs overproduction might be affected not only by the metabolic
[formula] Control fabR H- fadR H- adiY H- asnC H- crp H- csiR H- cytR H- deoR H- citB H- fis H- flhD H- fnr H- cra H- fur H- gadE H- gadW H- gadX H- glcC H- glpR H- ihfA H- ihfB H- ilvY H- lexA H- lrp H- marA H- metJ H- nac H- narL H- nsrR H- pdhR H- phoP H- purR H- rbsB H- rob H- rutR H- soxS H- torR H- 0 800 1600 Transcription factor FFAs (mg L -1 ) e * * * * * * * * * * * Control fadD H- fadE H- fadB H- fadA H- atoB H- yqeF H- 0 800 1600 Beta-oxidation FFAs (mg L -1 ) d * * * * Control serA H- serC H- ilvL H- ilvI H- ilvH H- ilvN H- ilvB H- ilvG H- ilvM H- ilvC H- ilvD H- ilvE H- gltA H- sucC H- gabD H- sad H- puuE H- gabT H- gadB H- gadA H- fumA H- fumC H- argH H- argG H- 0 800 1600 [/formula]
Amino acid metabolism processes in which the modulated gene is directly involved, but also by secondary cellular regulatory responses to genetic perturbation. To identify common rewiring in these FFAsoverproducing strains, we plotted the differentially expressed genes of the three strain pairs in a Venn diagram [fig_ref] Figure 4: ihfA Lvs [/fig_ref].
[formula] FFAs (mg L -1 ) * c * Control glpK H- gpsA H- plsX H- plsY H- plsB H- plsC H- cdsA H- ynbB H- pgsA H- pgpA H- pgpB H- pgpC H- glpX H- fbp H- pfkA H- lpxC H- tktA H- gloA H- gloB H- pykF H- poxB H- ybiW H- pflB H- tdcE H- ldhA H- dld H- 0 800 1600 Upstream carbon flux diversion FFAs (mg L -1 ) * * * * * * * a Control pta H- eutD H- ackA H- yccX H- mhpF H- adhE H- eutE H- aldB H- bioC H- bioH H- bioF H- 2854 H- bioA H- bioD H- bioB H- [/formula]
Considering the poor correlation between the proteome and the transcriptome, the candidate gene targets were selected based on their significantly differential abundance at the protein level (a and c) or the transcript level (b and d) [fig_ref] Figure 4: ihfA Lvs [/fig_ref] (see details in the Supplementary Note 2). As a result, 41 targets were selected, and their differential abundances at the protein and transcript levels are summarized in [fig_ref] Figure 4: ihfA Lvs [/fig_ref]. We found that these 41 genes included none of the 108 genes that we first selected from the genes related to FFA metabolism. Although most of the 108 genes were expressed with decreased abundance (Supplementary Data 2), none met the selection standards for significantly differential abundance (fold change < 0.67 at the protein level) in three FFAsoverproducing strains. Thus, these 41 gene targets were newly identified and were not directly related to FFA biosynthesis.
To test the function of these 41 candidate targets in FFAs production, the corresponding modulation of each gene was conducted based on the differential abundance at the protein level (the a and c sets in [fig_ref] Figure 4: ihfA Lvs [/fig_ref] or the transcript level (the b and d sets in [fig_ref] Figure 4: ihfA Lvs [/fig_ref]. We first applied CRISPRi to repress gene expression with high efficiency or utilized the P BAD promoter to upregulate gene expression [fig_ref] Figure 5: Reverse engineering of the candidate targets derived from omics analyses for identification... [/fig_ref]. Considering that the selected genes were all differentially expressed in the ihfA L− vs. Control strain pair [fig_ref] Figure 4: ihfA Lvs [/fig_ref] , these targets were also manipulated in the ihfA L− strain to test their effects on FFAs production [fig_ref] Figure 5: Reverse engineering of the candidate targets derived from omics analyses for identification... [/fig_ref]. The plasmids carrying the cassettes for each candidate gene were transformed into the starting strain CF or the ihfA L− strain [fig_ref] Figure 5: Reverse engineering of the candidate targets derived from omics analyses for identification... [/fig_ref]. An FFA assay of the resulting strains showed that modulation of 26 of the identified genes enhanced FFAs production by over 20% in comparison to that in the Control or ihfA L− strain [fig_ref] Figure 5: Reverse engineering of the candidate targets derived from omics analyses for identification... [/fig_ref]. In particular, the highest FFAs production was obtained in the ihfA L− -aidB + strain, in which the ihfA gene was repressed with low efficiency and the aidB gene was overexpressed under the P BAD promoter. The titer reached up to 2052 mg L −1 , which was 57% and 225% higher than the titers in the ihfA L− and Control strains, respectively [fig_ref] Figure 5: Reverse engineering of the candidate targets derived from omics analyses for identification... [/fig_ref]. In summary, 26 (norR, nrdE, npr, yihU, glpG, ftsQ, creC, sdhB, pal, rplW, tyrU, crsC, rnpB, ssrS, csrB, ryfA, sdsR, folk, aidB, yphF, deoD, nnr, rihC, hofP, ECD_01466, and ulaE) newly identified genes were determined to be beneficial targets that enhanced FFAs production.
Considering the significance of the gene expression level for improving product biosynthesis, we further tuned the expression level of each beneficial gene. We replaced the P BAD promoter with the stronger promoter P trc or P T7 (P T7 > P trc > P BAD ) [bib_ref] Biosynthesis of odd-chain fatty alcohols in Escherichia coli, Cao [/bib_ref] to control the overexpression of the beneficial genes. The results showed that only three strains exhibited increased FFAs production compared with strains with the corresponding gene controlled by the P BAD promoter [fig_ref] Figure 5: Reverse engineering of the candidate targets derived from omics analyses for identification... [/fig_ref]. Additionally, we utilized sgRNAs that bind the nontemplate strand at the middle or terminal region to repress the beneficial gene with medium or low efficiency [fig_ref] Figure 3: Tuning of the repression of beneficial genes for the improvement of FFAs... [/fig_ref]. We found that more than half of the engineered strains with the tuned gene repression level obtained further improved FFAs production compared with the strains with the corresponding gene repressed with high efficiency [fig_ref] Figure 5: Reverse engineering of the candidate targets derived from omics analyses for identification... [/fig_ref].
From the candidate gene targets provided by omics analyses, 26 beneficial genes that could improve FFAs production were verified [fig_ref] Figure 5: Reverse engineering of the candidate targets derived from omics analyses for identification... [/fig_ref]. However, these newly identified beneficial gene targets were not present in the metabolic pathways directly related to FFA metabolism. To mine the related regulatory mechanisms for FFAs production, the functions of the 26 beneficial genes were analyzed by using the NCBI, KEGG, UniProt, and Gene Ontology database, and summarized into eight sets [fig_ref] Figure 6: Functional network plot of the beneficial genes derived from omics analyses [/fig_ref]. Surprisingly, processes such as cell division, signal transduction, protein metabolism, and nucleic acid metabolism are seemingly irrelevant to FFA metabolism and thus generally inaccessible to target in the hypothesis-driven methods. Then, we further dissected the underlying connections between these enriched cellular processes and FFAs production. High repression of ftsQ or pal increased the intracellular FFAs concentration by 88% or 113%, respectively, compared with that in the Control strain , which might be associated with the increased cell volume for storing FFAs, since inactivation of either of the two genes could result in a deficiency in cell division and lead to the formation of long multiseptated cell chains [bib_ref] Structural and mutational analysis of the cell division protein FtsQ, Van Den Ent [/bib_ref] [bib_ref] The transenvelope Tol-Pal complex is part of the cell division machinery and..., Gerding [/bib_ref]. Beneficial effects of the downregulation of membrane proteins (such as those encoded by the genes glpG, ftsQ, creC, and pal) might be associated with microbial tolerance to FFAs [bib_ref] Proteomic analysis of the response of Escherichia coli to short-chain fatty acids, Rodriguez-Moya [/bib_ref] response to phosphate starvation [bib_ref] Transcript and protein level analyses of the interactions among PhoB, PhoR, PhoU..., Baek [/bib_ref] , was favorable for enhancing FFAs production, which is consistent with the promotive effects of cultivation under phosphate limitation [bib_ref] Free fatty acid production in Escherichia coli under phosphate-limited conditions, Youngquist [/bib_ref]. The effect of npr downregulation and yphF upregulation may be associated with the reduction in nitrogen assimilation [bib_ref] Regulatory roles of the bacterial nitrogen-related phosphotransferase system, Pfluger-Grau [/bib_ref] and the elevation of carbohydrate uptake [bib_ref] ABC transporters catalyzing carbohydrate uptake, Schneider [/bib_ref] , indicating a regulatory role of nitrogen and carbon utilization in enhancing FFAs production. Repression of rplW (encoding 50 S ribosomal subunit protein L23), tyrU (encoding tRNA-Tyr), and rnpB (encoding M1 RNA, the catalytic component of the tRNA-processing enzyme RNase P) was associated with the inhibition of protein synthesis, which increased the partitioning of total carbon from protein to lipids [bib_ref] Lipid production in Nannochloropsis gaditana is doubled by decreasing expression of a..., Ajjawi [/bib_ref]. ihfA . High repression was achieved by applying sgRNA gene H− , and overexpression was enabled by the P BAD promoter. TesA′, dCas9, and sgRNAs were controlled by the P T7 , P Trc , and P R promoters, respectively. Effect of perturbation of the target genes on FFAs production in the CF strain (c) or in the ihfA L− strain (d).
[formula] norR (ECD_02559) aroM (ECD_00337) waaF (ECD_03477) nrdE (ECD_02531) npr (ECD_03071) yihU (ECD_03767) lpp (ECD_01646) glpG (ECD_03275) cirA (ECD_02084) ftsQ (ECD_00094) creC (ECD_04275) yijO (ECD_03839) sdhB (ECD_00684) pal (ECD_00701) rplW (ECD_03169) recN (ECD_02504) tyrU (ECD_t00075) csrC (ECD_n04383) rnpB (ECD_n04368) ssrS (ECD_n04358) csrB (ECD_n04352) ssrA (ECD_n04348) ryfA (ECD_n04343) sdsR (ECD_n04325) folK (ECD_00141) aidB (ECD_04054) tam (ECD_01476) yphF (ECD_02440) deoD (ECD_04260) ybhC (ECD_00725) nnr (ECD_04034) hpaG (ECD_04230) rihC (ECD_00034) yajD (ECD_00358) hofP (ECD_03244) nikB (ECD_03326) 1466 (ECD_01466) ulaE (ECD_04064) torD (ECD_01001) cheR (ECD_01855) wcaF(a dcas9 tesA' T T gene H-T P BAD gene T dcas9 tesA' T T b gene H- ihfA L-T T dcas9 tesA' T T dcas9 tesA' T T ihfA L-T T PL--gene L- ihfA L--gene M- h sgRNA0 norR H- aroM H- waaF H- nrdE H- npr H- yihU H- lpp H- glpG H- cirA H- ftsQ H- creC H- yijO H- sdhB H- pal H- rplW H- recN H- tyrU H- csrC H- rnpB H- ssrS H- csrB H- ssrA H- ryfA H- sdsR H- folK + aidB + tam + yphF + deoD + ybhC + nnr + hpaG + rihC + yajD + hofP + nikB + 1466 + ulaE + torD + cheR + wcaF [/formula]
Asterisk represents the strain in which the FFAs titer increased by over 20% compared with that in the reference strain in each bar chart, and the targeted genes are referred to as beneficial genes. Effect of tuning the expression of beneficial genes on FFAs production in the CF strain (e, g) or in the ihfA L− strain (f, h). The overexpression was tuned by utilizing the P Trc or P T7 promoter, and the repression was tuned by applying the sgRNA gene M− or gene L− . Red bar, overexpression; green bar, repression. The titers were obtained in batch cultivation of 30 g L −1 glycerol. Data are presented as mean ± SD (n = 3 biological replicates). Source data underlying (c-h) are provided as a Source Data file. Although seemingly irrelevant to FFA metabolism, modulating these genes could markedly boost FFAs production due to the complex interactions of cellular networks. These findings provide insights into linkages between cell functions and product biosynthesis, providing avenues for the construction of a superior microbial cell factory for biochemical overproduction.
Enhancing FFAs production by combinatorial genetic perturbation. In the above studies on the modulation of candidate targets, we identified 56 beneficial genes (Figs. 2, 5c, d) and verified the corresponding beneficial sgRNAs that could be expressed to improve FFAs production [fig_ref] Figure 2: CRISPRi-based identification of beneficial genes for FFAs production [/fig_ref] , d, and Supplementary . All these beneficial sgRNAs are listed in Supplementary Data 4. The resulting strain ihfA L− -aidB + achieved the highest FFAs titer of 2,052 mg L −1 [fig_ref] Figure 5: Reverse engineering of the candidate targets derived from omics analyses for identification... [/fig_ref]. To further enhance FFAs production, we performed combinatorial multiplex perturbation of these beneficial genes. First, FFAs production of the ihfA L − -aidB + strain in combination with each of the 88 beneficial sgRNAs was investigated . The final resulting strain ihfA L− -aidB + -ryfA M− -gadA H− (ihfA L− -aidB + -ryfA M− in combination with gadA H− , overexpressing aidB via the P BAD promoter and repressing ihfA, ryfA, and gadA with low, medium, and high efficiency, respectively) produced 2,901 mg L −1 FFAs , the highest titer among all our engineered strains, 41% and 360% higher than that of the ihfA L− -aidB + and Control strains, respectively. These results strongly demonstrate the significance of these identified beneficial targets for FFAs overproduction.
To elucidate the mechanism of FFAs overproduction in the ihfA L − -aidB + -ryfA M− -gadA H− strain, we analyzed the effect of modulating the target genes. ihfA encodes the alpha subunit of the integration host factor, which affects DNA replication, recombination, and transcription [bib_ref] Genomic analysis of DNA binding and gene regulation by homologous nucleoid-associated proteins..., Prieto [/bib_ref]. Low repression of ihfA induces comprehensive regulation of gene expression (including aidB and ryfA) that promotes FFAs production (Supplementary Figs. 5, 6 and [fig_ref] Figure 5: Reverse engineering of the candidate targets derived from omics analyses for identification... [/fig_ref]. aidB encodes a DNA repairing protein involved in adaptive response of cells. Overexpression of aidB protects cells from cytotoxic effects of acidified cytoplasm or alkylating agents (released as byproducts of cellular metabolism) [bib_ref] Induction of the alkylation-inducible aidB gene of Escherichia coli by cytoplasmic acidification..., Smirnova [/bib_ref] [bib_ref] Preferential DNA damage prevention by the E. coli AidB gene: a new..., Rippa [/bib_ref]. ryfA encodes a small RNA (sRNA) with unknown functions. Nevertheless, bacterial sRNAs have been characterized as native mediators of stress responses [bib_ref] coli 6S RNA: a universal transcriptional regulator within the centre of growth..., Geissen [/bib_ref] [bib_ref] Integration of bacterial small RNAs in regulatory networks, Nitzan [/bib_ref] [bib_ref] Regulatory non-coding sRNAs in bacterial metabolic pathway engineering, Leistra [/bib_ref]. Repression of ryfA decreases membrane damage under stress conditions. gadA encodes glutamate decarboxylase A, which catalyzes the decarboxylation of glutamate and produces carbon dioxide. sgRNA targeting gadA would also inhibit the expression of the downstream gene gadX (encoding an activator of the glutamate decarboxylase system) in the same operon [bib_ref] A CRISPRi screen in E. coli reveals sequence-specific toxicity of dCas9, Cui [/bib_ref] , resulting in coordinated repression of gadA. Repression of gadA could decrease the production of carbon dioxide, a toxic byproduct inhibitory to cellular metabolism [bib_ref] The inhibition by CO 2 of the growth and metabolism of micro-organisms, Dixon [/bib_ref]. Overall, we speculate that genes modulation in the ihfA L− -aidB + -ryfA M− -gadA H− strain improves the cellular stress responses to protect cells from toxic effects and support the cellular phenotype of FFAs overproduction. In pursuit of a high titer of the desired product, microbial cells often encounter harsh and complex stressors, such as toxic byproducts and polytropic environments 1 . Thus, enhancing stress responses to reinforce microbial stress tolerance is a promising method to improve biochemicals production.
FFAs production by fed-batch fermentation. To further evaluate the potential of our optimized strain ihfA L− -aidB + -ryfA M− -gadA H− in a scaled-up process, fed-batch fermentations were performed in 5 L fermenters using the dO 2 -stat feeding strategy (details described in the "Methods"). The ihfA L− -aidB + -ryfA M− -gadA H− strain produced an exceptionally high titer of FFAs (30.0 g L −1 titer and 0.689 g L −1 h −1 productivity) in comparison to both the Control strain (8.6 g L −1 titer and 0.231 g L −1 h −1 productivity) and previously published values of FFAs fermentation performance in E. coli [bib_ref] Exploiting nongenetic cell-to-cell variation for enhanced biosynthesis, Xiao [/bib_ref]. Duplicate fermentation of the ihfA L− -aidB + -ryfA M− -gadA H− strain resulted in a similar titer and productivity . Extracellularly secreted FFAs were clearly visible at the top of the culture medium after centrifugation . Moreover, floating dead cells or fatty acid particles 58,59 were precipitated and stuck on the inner wall of the fermenter and on the sensors over the course of the fermentation process (Supplementary , exhibiting an additional amount of FFAs that is difficult to accurately quantify and not included in the FFAs titers.
# Discussion
Genome-scale identification of beneficial gene targets is especially necessary for improving the production of desired products since microbial biosynthesis is subject to sophisticated interactions. In our work, we took advantage of the CRISPRi system and omics , exhibiting a similar titer and productivity. b Solid FFAs layer after centrifugation. After fermentation, 1 mL of sample was picked up and centrifuged at 5000 × g for 15 min. Source data underlying (a) are provided as a Source Data file. analyses to identify beneficial genes for FFAs overproduction from candidate targets that are directly related or seemingly irrelevant to FFA metabolism (Supplementary and Supplementary Data 1). First, a set of 108 candidate targets was selected in a hypothesis-driven manner based on the competitive or regulatory roles of the genes in FFA biosynthesis . We identified 30 beneficial genes from these targets via CRISPRi-mediated repression with high efficiency [fig_ref] Figure 2: CRISPRi-based identification of beneficial genes for FFAs production [/fig_ref]. We further evaluated each of these 30 genes at two repression levels and obtained strains with further enhanced FFAs production [fig_ref] Figure 3: Tuning of the repression of beneficial genes for the improvement of FFAs... [/fig_ref]. Considering the high titer of FFAs and the extensive regulation of genetic networks, three FFAs-overproducing strains (ihfA L− , fadR H− , and fadR + ) that modulated transcription factors and the Control strain were applied for comparative transcriptomic and proteomic analyses, leading to the discovery of a set of 41 candidate targets that were significantly differentially expressed [fig_ref] Figure 4: ihfA Lvs [/fig_ref]. From these gene targets, we identified 26 beneficial genes [fig_ref] Figure 5: Reverse engineering of the candidate targets derived from omics analyses for identification... [/fig_ref] , which were associated with cell division, ncRNAs, signal transduction, transport, protein metabolism, and nucleic acid metabolism, etc. [fig_ref] Figure 6: Functional network plot of the beneficial genes derived from omics analyses [/fig_ref]. We further evaluated each of these 26 genes at two expression levels and obtained strains with further enhanced FFAs production [fig_ref] Figure 5: Reverse engineering of the candidate targets derived from omics analyses for identification... [/fig_ref]. Overall, we identified 56 beneficial genes that could significantly improve FFAs production, 30 of which were directly related to FFA metabolism and 26 of which were seemingly irrelevant to FFAs production. By combinatorial perturbation of the identified genes (repressing ihfA, ryfA, and gadA and overexpressing aidB), we obtained the optimized strain ihfA L− -aidB + -ryfA M− -gadA H− , producing 30.0 g L −1 FFAs (0.689 g L −1 h −1 productivity) in fed-batch fermentation . These titer and productivity of the microbial FFAs production are high in comparison to previous studies (Supplementary . Our findings shed light on the connection between target metabolite biosynthesis and complex interaction networks, aiding in exploiting the full cellular potential for further improvement in the production of target chemicals. This combined method for identifying beneficial targets and unleashing the cellular potential may be extended to other microbes to optimize the production of other desired chemicals.
In this work, we utilized CRISPRi to readily downregulate gene expression by blocking transcription elongation [bib_ref] Repurposing CRISPR as an RNA-guided platform for sequencespecific control of gene expression, Qi [/bib_ref] [bib_ref] CRISPR interference (CRISPRi) for sequence-specific control of gene expression, Larson [/bib_ref]. Several other strategies could also be utilized to achieve gene repression [bib_ref] Optimization of a yeast RNA interference system for controlling gene expression and..., Crook [/bib_ref] [bib_ref] Metabolic engineering of Escherichia coli using synthetic small regulatory RNAs, Na [/bib_ref]. RNA interference (RNAi) can downregulate gene expression by translational inhibition or transcript degradation; however, RNAi is limited to eukaryotes that have the proper host machinery [bib_ref] Optimization of a yeast RNA interference system for controlling gene expression and..., Crook [/bib_ref]. In bacteria, gene repression can be achieved by knocking down the translation of selected target mRNAs by using synthetic sRNAs [bib_ref] Metabolic engineering of Escherichia coli using synthetic small regulatory RNAs, Na [/bib_ref]. Both CRISPRi and sRNAs can be easily designed and readily implemented with varied efficiency and multiplex combination, enabling rapid screening of beneficial knockdown targets, including genes involved in essential metabolism, for enhanced production of biochemicals. An advantage of CRISPRi is that diverse regions, such as trans-acting small ncRNAs and cis-encoded untranslated regions, can be targeted, enabling the acquisition of additional regulatory knowledge. As a result, several beneficial ncRNAs, such as those encoded by ryfA and ssrS, were identified in our work, and the corresponding regulation greatly improved FFAs production [fig_ref] Figure 5: Reverse engineering of the candidate targets derived from omics analyses for identification... [/fig_ref].
In the process of target identification, we observed an intriguing phenomenon of FFAs overproduction upon modulation of the fadR gene. Overexpression of fadR could increase the output of FFAs, as shown in a previous study [bib_ref] Enhancing fatty acid production by the expression of the regulatory transcription factor..., Zhang [/bib_ref] , which was also verified in our fadR + strain in this study [fig_ref] Figure 4: ihfA Lvs [/fig_ref]. Conversely, we found that high repression of fadR (fadR H− ) could also enhance FFAs production [fig_ref] Figure 4: ihfA Lvs [/fig_ref]. To reveal the mechanisms underlying FFAs overproduction in the fadR + and fadR H− strains, we analyzed the data acquired from proteomic and transcriptomic analyses of these two strains compared with the Control strain (Supplementary . FadR is a transcription factor that controls the expression of genes involved in FFA metabolism [bib_ref] Enhancing fatty acid production by the expression of the regulatory transcription factor..., Zhang [/bib_ref]. We found that the abundance of AccD, FabB, and FabI in the FFA biosynthetic pathway was significantly increased (fold change of 1.96, 2.14, and 1.63, respectively) in fadR + vs. Control (Supplementary . Thus, overexpression of fadR strengthened the FFA biosynthetic pathway, resulting in FFAs overproduction in the fadR + strain . Regarding the fadR H− strain, we speculated that high repression of fadR could enhance cellular responses to reconcile intracellular and environmental stresses, maintaining a functional cellular state to facilitate FFAs overproduction . On the one hand, we observed that the abundance of several sRNAs, such as those encoded by ssrS and ryfA, was significantly decreased (fold change < 0.5) in fadR H− vs. Control . Modulation of these sRNAs, native mediators of stress responses, enables the cell to respond to stresses to support the overproduction phenotype [bib_ref] coli 6S RNA: a universal transcriptional regulator within the centre of growth..., Geissen [/bib_ref] [bib_ref] Integration of bacterial small RNAs in regulatory networks, Nitzan [/bib_ref] [bib_ref] Regulatory non-coding sRNAs in bacterial metabolic pathway engineering, Leistra [/bib_ref]. On the other hand, we also found that the protein and transcript abundances of aidB were increased (fold change of 2.02 and 1.47, respectively) in fadR H− vs. Control . Increased expression of AidB serves to protect cells from the cytotoxic effects of an acidified cytoplasm or alkylating agents [bib_ref] Induction of the alkylation-inducible aidB gene of Escherichia coli by cytoplasmic acidification..., Smirnova [/bib_ref] [bib_ref] Preferential DNA damage prevention by the E. coli AidB gene: a new..., Rippa [/bib_ref]. In fact, a similar phenomenon wherein both up-and downregulating the expression of a gene could facilitate product biosynthesis has been previously observed, for example, the improved production of lycopene upon overexpressing or deleting the POX2 gene (encoding acyl-CoA oxidase) [bib_ref] Constructing yeast chimeric pathways to boost lipophilic Terpene synthesis, Liu [/bib_ref] [bib_ref] Production of lycopene in the non-carotenoid-producing yeast Yarrowia lipolytica, Matthaus [/bib_ref].
Our study demonstrates that the identification of crucial gene targets is of enormous importance in engineering microbes for enhanced production of desired biochemicals. Further effort should be made for multilevel and large-scale identification of beneficial gene targets in the chromosomes of microbes. For example, the CRISPRa (CRISPR activation) method could be utilized to identify genes that would enhance target chemical production by upregulation of specific genes. CRISPRa enables the activation of endogenous targets by expressing the activator and the dCas9-sgRNA complex [bib_ref] Programmable repression and activation of bacterial gene expression using an engineered CRISPR-Cas..., Bikard [/bib_ref] [bib_ref] Engineering complex synthetic transcriptional programs with CRISPR RNA scaffolds, Zalatan [/bib_ref] [bib_ref] Synthetic CRISPR-Cas gene activators for transcriptional reprogramming in bacteria, Dong [/bib_ref] [bib_ref] Engineered CRISPRa enables programmable eukaryote-like gene activation in bacteria, Liu [/bib_ref] [bib_ref] Effective CRISPRa-mediated control of gene expression in bacteria must overcome strict target..., Fontana [/bib_ref] [bib_ref] Programmable CRISPR-Cas transcriptional activation in bacteria, Ho [/bib_ref]. In bacteria, however, CRISPRa currently has stringent requirements for effective target sites and lacks general functional activators 37,65,66 , which constrain its wide applications. Additionally, interconnected with high-throughput screening of a desired phenotype, it would be better to apply a genome-scale sgRNA library for identifying beneficial targets from the complete list of directly related and seemingly irrelevant genes [bib_ref] Pooled CRISPRi screening of the cyanobacterium Synechocystis sp PCC 6803 for enhanced..., Yao [/bib_ref] [bib_ref] Pooled CRISPR interference screening enables genome-scale functional genomics study in bacteria with..., Wang [/bib_ref]. Future advances in genome-scale technologies and high-throughput screening methods will enable systematic and comprehensive identification of potential chromosomal gene targets, unleashing the full cellular potential for microbial biosynthesis.
# Methods
Experimental materials. All strains and plasmids used in this study are listed in Supplementary Data 5 and Supplementary Data 6, respectively. All genes targeted in this study are listed in Supplementary Data 1. In consideration of the polar effect of CRISPRi 56 , the operon context of each gene is also summarized in Supplementary Data 1. E. coli Trans1-T1 was used for cloning. E. coli BL21(DE3) and its derived strains were used for fermentation. Plasmids pACYCDuet-1, YX210, YX212, and YX213 were used as the vectors for the expression of dCas9, GFP, or TesA′. Plasmid Sg-S (with two BsaI restriction sites) was constructed and used as a platform vector for the expression of various sgRNAs. Plasmids Sg-BAD, Sg-Trc, and Sg-T7 were used as the vectors for the expression of chromosome genes. Plasmids, excluding that expressing sgRNAs, were constructed by standard enzyme digestion and ligation. Target site complementary sequences of sgRNAs were designed according to a previous study [bib_ref] CRISPR interference (CRISPRi) for sequence-specific control of gene expression, Larson [/bib_ref]. Custom-designed spacers were inserted into plasmid Sg-S by one-step Golden Gate assembly [bib_ref] Programmable repression and activation of bacterial gene expression using an engineered CRISPR-Cas..., Bikard [/bib_ref] [bib_ref] A one pot, one step, precision cloning method with high throughput capability, Engler [/bib_ref] , allowing rapid construction of plasmids for the expression of sgRNAs targeting any genomic locus of interest. Briefly, the designed forward and reverse primers were annealed to obtain a double-stranded inserted fragment, which could be cleaved by BsaI and ligated into plasmid Sg-S by T4 DNA ligase, resulting in the desired sgRNA expression plasmid Sg-gene H/M/L− . Primers annealed to custom-designed spacers and primers used to amplify genes from E. coli genomic DNA are listed in Supplementary Data 7 and Supplementary Data 8, respectively. DNA sequences used in this study are listed in Supplementary Data 9. The construction of plasmids is presented in [fig_ref] Figure 2: CRISPRi-based identification of beneficial genes for FFAs production [/fig_ref].
Culturing conditions and media. To assess FFAs production by different recombinant E. coli strains, fermentation was performed in 34 mL glass tubes (180 mm in length and 18 mm in caliber) containing 5 mL of modified M9 medium with oblique and orbital shaking at 220 rpm and 30°C. Each strain was inoculated from a freshly transformed single colony in LB agar plate to 2 mL of LB medium as seed culture. When cell accumulation reached stationary phase, 1% (V/V) of seed culture was re-inoculated to 5 mL of modified M9 medium in a glass tube. The tube cultures were induced with 1 mM IPTG at an OD 600 of about 1.0 and allowed to grow for an additional 40 h. When the P BAD promoter was applied, 10 mM arabinose was utilized to induce gene expression at an OD 600 of 0.5-0.6. This sequential addition of arabinose and IPTG was to avoid the inhibition of the arabinose-inducible promoter P BAD expression system by IPTG [bib_ref] Directed evolution of AraC for improved compatibility of arabinose-and lactose-inducible promoters, Lee [/bib_ref] , which was also applied in many previous studies [bib_ref] Overproduction of free fatty acids in E. coli: implications for biodiesel production, Lu [/bib_ref] [bib_ref] Quantitative analysis and engineering of fatty acid biosynthesis in E. coli, Liu [/bib_ref]. Tube cultivations were performed in three biological replicates. For proteomic and transcriptomic analyses, fermentation was implemented in 250 mL unbaffled Erlenmeyer flasks containing 50 mL of modified M9 medium with the identical culturing procedures of tube fermentation. Flask cultivations were performed in two biological replicates. Glycerol, a byproduct of soap manufacturing, is utilized here as a common feedstock for FFAs production in E. coli (Supplementary . Modified M9 medium used for tube and flask fermentation 74 was as described: 17.1 g L −1 Na 2 HPO 4 ·12H 2 O, 3 g L −1 KH 2 PO 4 , 0.5 g L −1 NaCl, 2 g L −1 NH 4 Cl, 2 g L −1 yeast extract, 30 g L −1 glycerol, 0.25 g L −1 MgSO 4 ·7H 2 O, 11.1 mg L −1 CaCl 2 , 10 mg L −1 thiamine, and 0.1% (v/v) Triton-X100. 1 mL L −1 metal trace stock solution was also supplemented, which contained 27 g L −1 FeCl 3 ·6H 2 O, 2 g L −1 ZnCl 2 , 2 g L −1 Na 2 MoO 4 ·2H 2 O, 1.9 g L −1 CuSO 4 ·5H 2 O, and 0.5 g L −1 H 3 BO 3 . pH was adjusted to about 7.2 by Tris. If necessary, 34 μg mL −1 chloramphenicol or 100 μg mL −1 carbenicillin was supplemented.
For fed-batch fermentation, an overnight LB culture (1 mL) of the freshly transformed single colony in 34 mL glass tubes was re-inoculated into 100 mL of modified mineral medium (6 g L −1 NH 4 Cl, 8.5 g L −1 KH 2 PO 4 , 0.5 g L −1 citrate, 5 g L −1 yeast extract, 15 g L −1 glycerol, 1 g L −1 MgSO 4 ·7H 2 O, 0.07 g L −1 CaCl 2 ·2H 2 O, 4 mL L −1 metal trace stock solution, and 100 mg L −1 thiamine) 74 with 34 μg mL −1 chloramphenicol and 100 μg mL −1 carbenicillin in 500 mL unbaffled flasks. When OD 600 reached 3-4, 200 mL of culture was re-inoculated into a 5 L bioreactor (T&J, China) with 1.8 L of modified mineral medium, 34 μg mL −1 chloramphenicol, and 100 μg mL −1 carbenicillin. Fermentation temperature was set at 30°C and pH was controlled at 7 by feeding 6 N ammonium hydroxide via an auto pump. Air flow rate was maintained at around 2 L min −1 . The dissolved oxygen (dO 2 ) concentration was controlled above 30% by agitation cascade (300-800 rpm). When cell density (OD 600 ) reached about 6 and 17, 10 mM arabinose (when the P BAD promoter was applied) and 1 mM IPTG were added into the fermentation cell culture, respectively. Feeding medium (2.47 g L −1 MgSO 4 ·7H 2 O, 500 g L −1 glycerol, and 100 g L −1 yeast extract) was fed to the fermentation culture when the initial glycerol was almost depleted and the dissolved oxygen showed a sharp increase. The carbon source restriction strategy was carried out by dO 2 concentration, given that the dO 2 level immediately spiked once the carbon source was run out. Once the dO 2 > 60%, the feeding would automatically start at 5 mL min −1 until dO 2 < 60%. Antifoam (Sigma) was added automatically as needed. Broth samples (about 5 mL) were collected at a series of time points to measure cell density and stored at −20°C for further measurements of glycerol and FFAs.
GFP fluorescence assay. The fluorescence signal of GFP was used to characterize the repression efficiency of the CRISPRi system. Culturing procedures and induction process of the GFP reporting strains were the same as tube fermentation of the FFAs-producing strains. For each tube, 200 μL of the sample was diluted into the linear range of the detector with phosphate-buffered saline (PBS) at 20 h after IPTG induction. Fluorescence intensity (excitation at 485 nm and emission at 520 nm) and cell density (OD 600 ) were detected using 96-well polystyrene plates (black plate with a clear bottom) (Corning Incorporated 3603, USA) and a microplate reader (SpectraMax M2, Molecular Devices, USA). The relative fluorescence intensity was first normalized using OD 600 and then subtracted that of blank E. coli BL21(DE3). Experiments were performed in three biological replicates.
Metabolite extraction and analysis. FFAs titers in whole-cell culture (only FFAs were measured in this study) were quantified following previously published methods [bib_ref] Design of a dynamic sensorregulator system for production of chemicals and fuels..., Zhang [/bib_ref]. Specifically, 0.5 mL of cell culture (or an appropriate volume of cell culture diluted to 0.5 mL) was acidified with 50 μL of concentrated HCl, spiked with 120 μg of heptadecanoic acid as internal standard. The cell culture was extracted twice with 0.5 mL of ethyl acetate. The extracted FFAs were then determined using a Thermo Scientific TRACE 1300 gas chromatograph (GC) equipped with a TG-WaxMS A column (30 m × 0.32 mm × 0.25 µm; Thermo Scientific) and a Flame Ionization Detector (FID) operating under constant flow rate of the carrier gas (nitrogen) at 1 mL min −1 . The following temperature program was used: hold at 50°C for 1 min, then heat to 245°C at 30°C min −1 and hold at this temperature for 22.5 min. Individual fatty acid species were qualified by authentic homologous standards and quantified by comparing the peak areas with that of the internal standard using the Chromeleon 7.1 software. Total concentrations of FFAs were calculated as the sum of C 12 to C 18 (saturated and monounsaturated).
Glycerol concentration was determined by high-performance liquid chromatography (HPLC) following Waters standard protocols. Briefly, filtered culture supernatants were analyzed by a Waters HPLC system including a Waters e2695 separation module, a Waters 2414 refractive index detector (RID) and an Aminex HPX-87H column (Bio-Rad). The separation was performed through elution with 5 mM H 2 SO 4 at a flow rate of 0.6 mL min −1 at 65°C for 30 min.
Proteomic analysis. Cells were harvested by centrifugation at 3000 ×g for 5 min at 24 h after IPTG induction and flash frozen in liquid nitrogen. For protein preparation, cells were resuspended in 600 μL of lysis buffer (0.05 M Tris, pH 8, 1% SDS, and 8 M Urea) and disrupted by ultrasonication for 5 min (cycles with 2 s work and 3 s pause). After centrifugation at 18,000 × g and 4°C for 20 min, the supernatant was reduced with 10 mM dithiothreitol for 1 h at 56°C, and subsequently alkylated with sufficient iodoacetamide for 1 h at room temperature in the dark. Samples were subjected to protein precipitation by being vortexed with precooled acetone. After centrifugation, the precipitate was collected, dried, and redissolved in 600 μL of Urea buffer (0.05 M Tris, pH 8, and 8 M Urea). The protein samples were quantified by the Bradford method. A total of 100 μg of each sample was trypsin-digested overnight. After desalination, the tryptic peptides were lyophilized, re-dissolved in 0.5 M triethylammonium bicarbonate (TEAB) buffer, and then labeled using 8-plex iTRAQ reagents (AB SCIEX) as manual described. The tryptic peptides from the Control, ihfA L− (LihfA), fadR H− (HfadR), and fadR + (OfadR) strains were labeled with iTRAQ tags 113, 114, 115, and 116, respectively. After labeling, the samples were fractioned and then analyzed using a Q Exactive HF mass spectrometer (Thermo Scientific). The raw data obtained from mass spectrometry detection were deposited as 'raw.7z'. Peptide identification and quantification were then conducted by Beijing Novogene Bioinformatics Technology Co., Ltd (China) using the Proteome Discoverer 2.2 software (Thermo Scientific), and the results were deposited as 'search.7z'. The raw MS data 'raw.7z'and the peptide identification and quantification data 'search.7z' are associated with the ProteomeXchange Consortium accession PXD017890.
Transcriptomic analysis. Cells were harvested after 24 h of IPTG induction by quick centrifugation at 10,000 × g for 1 min and immediately frozen in liquid nitrogen. Total RNA was extracted using the RNAprep pure Cell/Bacteria Kit (Tiangen) following lysozyme treatment. RNA degradation and contamination were monitored on 1% agarose gels. RNA concentration was measured by Qubit ® RNA Assay Kit in Qubit ® 2.0 Flurometer (Life Technologies, CA, USA) and RNA integrity was assessed on a 2100 Bioanalyzer (Agilent Technologies, CA, USA). rRNA was removed using a Ribo-zero kit that left the mRNA. A total of 3 μg RNA was used as an input per sample. Sequencing libraries were generated using a NEBNext ® UltraTM RNA Library Prep Kit for Illumina ® (NEB, USA) following the manufacturer's instructions, and index codes were added to attribute sequences to each sample. Clustering of the index-coded samples was performed on a cBot Cluster Generation System using the TruSeq PE Cluster Kit v3-cBot-HS (Illumina) according to the manufacturer's instructions. After cluster generation, the library preparations were sequenced on an Illumina Hiseq 4000 platform and paired-end reads were generated. The data were analyzed by Beijing Novogene Bioinformatics Technology Co., Ltd (China). Control, LihfA, HfadR, and OfadR represent the Control, ihfA L− , fadR H− and fadR + strains, respectively.
Reporting summary. Further information on research design is available in the Nature Research Reporting Summary linked to this article.
## Data availability
Data supporting the findings of this work are available within the paper and its Supplementary Information files. A reporting summary for this article is available as a Supplementary Information file. The mass spectrometry-based proteome data generated in this study have been deposited in ProteomeXchange Consortium via the iProX partner repository under accession code PXD017890. The transcriptome data generated in this study have been deposited in Gene Expression Omnibus under accession code GSE146162. Nucleotide sequences of pCF, Sg-S, sgRNA0-biobrick, T7-T1, Trc-T1, and BAD-T1 have been deposited in NCBI Genbank under accession codes MZ567118, MZ567119, MZ567120, MZ567121, MZ567122, and MZ567123 respectively. GraphPad Prism 8 was used to plot data. Source data are provided with this paper.
[fig] Figure 2: CRISPRi-based identification of beneficial genes for FFAs production. All 108 genes in the upstream carbon flux diversion module (a), downstream carbon flux diversion module (b), amino acid metabolism module (c), beta-oxidation module (d), and transcription factor module (e) were repressed with high efficiency. Asterisk represents the strain in which the FFAs titer increased by over 20% compared with that in the Control strain (631 mg L −1 ), and the targeted genes are referred to as beneficial genes. The titers were obtained in batch cultivation of 30 g L −1 glycerol. Data are presented as mean ± SD (n = 3 biological replicates). Source data are provided as a Source Data file. [/fig]
[fig] Figure 3: Tuning of the repression of beneficial genes for the improvement of FFAs production. a Design of sgRNAs with high (gene H− ), medium (gene M− ), or low (gene L− ) repression efficiency toward the target gene by binding the nontemplate DNA strand at the initial, middle, or terminal region, respectively. b Improvement of the FFAs titer upon repressing genes with medium or low efficiency compared with the titer obtained upon repressing genes with high efficiency. The gene targets were the genes marked with an asterisk inFig. 2. The titers were obtained in batch cultivation of 30 g L −1 glycerol. Data are presented as mean ± SD (n = 3 biological replicates). Source data underlying (b) are provided as a Source Data file. [/fig]
[fig] Figure 4: ihfA Lvs. Control 2: fadR Hvs. Control 3: fadR + vs. Control X: Proteomic and transcriptomic analyses for exploration of available targets for FFAs production. a FFAs titer of the overproducing strains ihfA L− , fadR H− , and fadR + and the Control strain. The titers were obtained in batch cultivation of 30 g L −1 glycerol. These strains were applied to the omics analyses. Data are presented as mean ± SD (n = 3 biological replicates). b Venn diagram of differentially expressed genes. Blue numbers, genes expressed with decreased abundance; red number, genes expressed with increased abundance. Genes in sets a, b, c, and d had decreased abundance at the protein level, decreased abundance at the transcript level, increased abundance at the protein level, and increased abundance at the transcript level, respectively, in at least two FFAs-overproducing strains compared with the Control strain. c Abundance of the selected candidate genes at the protein and transcript levels. 1466 represents the gene numbered ECD_01466 in BL21(DE3). Genes in sets a, b, c, and d were selected from sets a, b, c, and d in (b). Source data are provided as a Source Data file. [/fig]
[fig] Figure 5: Reverse engineering of the candidate targets derived from omics analyses for identification of beneficial genes and tuning of gene expression for enhanced FFAs production. Schematic of cassettes used to repress or overexpress the target genes in the CF strain (a), or in the ihfA L− strain (b) [/fig]
[fig] Figure 6: Functional network plot of the beneficial genes derived from omics analyses. The size of the node represents the number of genes within the function set. The node color represents the W value of the function set. The thickness of the line indicates the number of genes overlapping between two function sets. The W value was calculated by averaging the maximum FFAs ratio of the engineered strains corresponding to each gene in the function set. Genes and the FFAs ratio are shown in Supplementary Data 3. Source data are provided as a Source Data file. [/fig]
[fig] Figure 7, Figure 8: gadA H-glcC H-fabR L-ssrS H-ssrS MCombinatorial genetic perturbation for improved FFAs production. a FFAs production of ihfA L− -aidB + in combination with each beneficial sgRNA. The gray line represents the FFAs titer of the ihfA L− -aidB + strain. Numbers 1 to 88 represent the sgRNAgpsA H− , plsY H− , pgpB H− , pgpC H− , pfkA H− , pykF H− , dld H− , pta H− , yccX H− , eutE H− , bioC H− , bioH H− , bioD H− , gabD H− , gadA H− , fadE H− , fadB H− , atoB H− , yqeF H− , fabR H− , fadR H− , asnC H− , fnr H− , cra H− , gadW H− , glcC H− , glpR H− , ihfB H− , torR H− , gpsA M− , yccX M− , bioC M− , gadA M− , atoB M− , cra M− , glpR M− , gpsA L− , pfkA L− , pta L− , bioH L− , bioD L− , gabD L− , gadA L− , atoB L− , fabR L− , asnC L− , cra L− , glpR L− , norR H− , nrdE H− , npr H− , yihU H− , glpG H− , ftsQ H− , creC H− , sdhB H− , pal H− , rplW H− , tyrU H− , csrC H− , rnpB H− , ssrS H− , csrB H− , ryfA H− , sdsR H− , norR M− , nrdE M− ,glpG M− , creC M− , pal M− , rplW M− , csrC M− , rnpB M− , ssrS M− , csrB M− , ryfA M− , sdsR M− , npr L− , yihU L− , glpG L − , creC L− , sdhB L− , rplW L− , tyrU L− , csrC L− , rnpB L− , ssrS L− , and csrB L− , respectively. Blue number, FFAs titer in the corresponding strain was increased compared with that in the ihfA L− -aidB + strain; red number, the corresponding strain obtained the highest FFAs titer, which was applied for the next round of combination. b FFAs production of ihfA L− -aidB + -ryfA M− in combination with gadA H− , glcC H− , fabR L− , ssrS H− , or ssrS M− . Asterisk represents the strain that produced the highest FFAs titer, and this strain was applied to the next fed-batch fermentation. The titers were obtained in batch cultivation of 30 g L −1 glycerol. Data are presented as mean ± SD (n = 3 biological replicates). Source data are provided as a Source Data file. Fed-batch fermentation of the ihfA L− -aidB + -ryfA M− -gadA H− strain for FFAs production in a 5 L bioreactor. a Time courses of cell growth, glycerol consumption, and FFAs production of the ihfA L− -aidB + -ryfA M− -gadA H− strain during fed-batch fermentation (Batch 1). The result of a duplicate fermentation (Batch 2) is shown in [/fig]
|
Opportunities for Refinement of Non-Human Primate Vaccine Studies
Citation: Prescott, M.J.; Clark, C.; Dowling, W.E.; Shurtleff, A.C. Opportunities for Refinement of Non-Human Primate Vaccine Studies. Vaccines 2021, 9, 284. https://
# Introduction
Animal models remain an integral part of the immunogenicity, efficacy, and safety assessment of new vaccines and drugs. They also provide possibilities for research on host-pathogen interactions and the interplay with the host immune system. In some cases, non-human primates (NHPs) provide the best, or even only, models to study these aspects in infectious disease research. Over the last decades, NHP studies have been instrumental in gaining an understanding of the pathogenesis of various infectious diseases and have provided relevant models to develop new therapies (e.g., vaccines against polio, yellow fever, Hepatitis B, and Ebola; identification of the causative agents of infectious diseases such as SARS, typhoid fever, and mumps; and more in-depth understanding of infections such as HIV. With several emerging viral infections becoming epidemic, NHPs continue to be important animals for investigating human viral diseases. NHPs are highly favored for modeling Nipah, Middle Eastern Respiratory Syndrome (MERS), and Rift Valley fever (RVF), and have been species of choice in the last year for SARS-CoV-2 research and the discovery of vaccines and countermeasures to combat the coronavirus disease 2019 (COVID-19) pandemic. Indeed, the global research community has turned to NHP species so heavily for SARS-CoV-2 research that the availability of these animals is scarce.
The Coalition for Epidemic Preparedness Innovations (CEPI) is a global partnership between public, private, philanthropic, and civil society organizations. It funds the development of vaccines against deadly diseases such as Nipah, RVF, Lassa fever, and Chikungunya, for which no licensed vaccines are currently available. Of very current interest, two coronaviruses are also in the portfolio of pathogens of importance to CEPI: MERS-CoV, the etiologic agent of MERS, and more recently SARS-CoV-2. The COVID-19 pandemic has brought CEPI to the forefront of not only vaccine development for this novel pathogen, but also in playing a main role in the global initiative COVAX, which aims to guarantee fair and equitable access to novel licensed SARS-CoV-2 vaccines for every country in the world.
Before the pandemic, a clear mandate from CEPI has been the responsible use of NHPs as a valuable research resource, knowing that many of the CEPI priority pathogens must be evaluated in laboratory models of NHP infection because some of these viruses do not infect other laboratory animal species in a way that adequately models human disease. For some pathogens, NHP studies for drug and vaccine development may be considered the pivotal efficacy studies upon which product licensure decisions are made under the US Food and Drug Administration (FDA) "Animal Rule" regulatory approval pathway. Product development studies must be performed under high-quality research conditions with an emphasis on data integrity, and the responsible care and use of the NHP models is integral to performance of successful studies.
Depending on the pathogen's course of infection and severity of clinical signs, study type (e.g., drug candidate, prophylactic vaccine, therapeutic vaccine) and experimental endpoints chosen, product development studies can potentially result in severe suffering for the NHPs involved, especially untreated or unvaccinated control animals. Optimizing the welfare of NHPs used in such studies is important for CEPI and its funders. Experiments performed without the best welfare approaches risk the collection of poor-quality experimental data, potentially compromising model validity and the ability to detect positive effects of the experimental product under test. Sub-standard housing, husbandry, and care practices also result in unacceptable reputational risks to all organizations involved. Public opinion polls show high concern about the use of NHPs in research, and greater approval for in vivo research where steps are taken to reduce animal use and suffering, in line with the 3Rs principles, replacement, reduction, and refinement.
To help ensure appropriate attention to the 3Rs and high-quality experimental design in funded NHP studies, CEPI relies on the valuable peer review and advice service provided by the UK's National Centre for the Replacement, Refinement, and Reduction of Animals in Research (NC3Rs). NC3Rs is a Government-backed organization that works with scientists and institutions across the bioscience sector to discover and implement new technologies and approaches that replace, reduce, and refine the use of animals in research. Under the service, individual research proposals are reviewed by expert staff for opportunities to implement the 3Rs and for compliance with principles in the NC3Rs guidelines "Non-Human Primate Accommodation, Care and Use", adopted as a condition of funding by CEPI and other funders. The guidelines reflect contemporary good practice and are aligned with housing and husbandry standards provided for NHPs in the European Union (EU). NC3Rs' feedback is used during funding decisions and when drafting the terms and conditions of grant awards. The Centre is in the privileged position of being able to compare research facilities and practices internationally, and this process has identified opportunities to refine NHP vaccine development studies submitted for CEPI funding and to improve animal welfare without compromising the science.
The NC3Rs has itself funded several research projects aimed at advancing the 3Rs in vaccine testing with NHPs, including awards for the development of an immunologicallycompetent in vitro model of the human liver to replace NHPs in the assessment of yellow fever vaccine attenuation, and for the transfer between laboratories of an in vitro functional assay to refine efficacy testing of tuberculosis (TB) vaccine candidates by avoiding the need for in vivo challenge. The Centre has also worked with the international pharmaceutical and biotechnology industry to embed the 3Rs in the development of monoclonal antibodies (mAbs) as therapies for disease. Non-clinical testing of mAbs poses challenges because their high degree of target specificity can mean that there is either no relevant species to use or the NHP is the only option. By acting as an honest broker for cross-sector data sharing, and analyzing data on over 100 biologics from 15 companies, the Centre has identified opportunities to halve the number of non-human primates used in a typical mAb development program from 144 to 64, whilst supporting patient safety. This collaborative work has changed company practice and influenced the addendum to the ICH S6 guidelines on the nonclinical safety evaluation of biotechnology-derived pharmaceuticals. Also of relevance to this special issue, the NC3Rs has been tasked by the World Health Organization (WHO), and funded by the Bill and Melinda Gates Foundation, to carry out an independent and comprehensive review of WHO guidelines for biologics. The international expert working group is evaluating which animal tests are recommended for the batch release and quality control testing of biologics, including vaccines, and what opportunities exist for better implementation of 3Rs principles and alternative test methods, generating recommendations to WHO on how this could be best achieved.
CEPI, along with the Bill and Melinda Gates Foundation and US Government agencies such as Biomedical Advanced Research and Development Authority (BARDA) and National Institute of Allergy and Infectious Diseases (NIAID), are funding vaccine development programs for SARS-CoV-2 and other pathogens, yet also increasingly leading the world in discussions of how vaccine development impacts availability of research resources such as NHPs, and the responsibility to use these resources ethically and conservatively. As a way for CEPI-funded laboratories and vaccine developers to speak directly about how to refine their NHP vaccine development studies, CEPI and the NC3Rs organized a joint, non-public workshop in June 2019, hosted by the University of Texas Medical Branch (UTMB), Galveston, TX, USA. The objectives of the workshop were to:
- explore opportunities for refinement of CEPI-funded NHP vaccine development studies, in order to optimize animal welfare and scientific outcomes; - better align CEPI-funded studies with the NC3Rs NHP guidelines and deliver on public commitments to the 3Rs; - share relevant data and experience from international laboratories; - and provide a platform for follow-up work on important concepts such as refinement and standardization of humane endpoints.
Here we share key findings and best practice, and summarize the full recommendations from the workshop to help support other funders, regulators, researchers and laboratory staff to refine NHP vaccine studies and facilitate better animal wellbeing, research quality, data integrity, and public support.
## Refinement opportunities
## Social housing and socialization
It is well established that social housing is crucial for the welfare of NHPs, including macaques, marmosets, and other species used in bioscience research. There is a large body of literature on the negative impact of individual housing on the health and psychological well-being of these highly social animals. Accordingly, regulations and guidelines mandate or encourage social housing, and this is the default housing configuration at most NHP facilities. A recent international survey of behavioral management practices found that the proportion of facilities housing all of their NHPs socially (i.e., kept with one or more compatible conspecifics in the same cage or enclosure) was 83% in the UK (n = 6 facilities), 46% in the EU (n = 11), and 32% in the US (n = 25).
While exemptions to social housing can be granted for justifiable reasons by the local oversight body (e.g., Institutional Animal Care and Use Committee-IACUC), there is variation in what is considered reasonable grounds for exemption in infectious disease studies, which can last several months. In the aforementioned survey, across all facility types, concerns about cross-infection between animals, monitoring of clinical signs, and anticipated rapid clinical decline of animals were reported as factors preventing social housing by 50%, 29%, and 40% of EU facilities and 75%, 65% and 53% of US facilities respectively, but were not considered constraints by UK facilities. Research facilities in the UK and EU, generally socially house their NHPs throughout natural history, pathogenesis, and therapeutic intervention studies, including during the phase when animals are challenged with an infectious agent (i.e., when under high containment), and this was reflected by UK and French laboratories at the workshop. Perceived problems such as aggression during clinical disease manifestations, or healthy animals injuring sick companions, have not been observed. Similarly, the National Microbiology Laboratory of the Public Health Agency of Canada (PHAC), which at the encouragement of its IACUC has pairhoused macaques in BSL-4 for over seven years, also reported that with familiarization and acclimation processes, the pair-housed animals are not aggressive towards each other. When reporting study findings, researchers should give details of the social configurations used (e.g., pair housing), the timings of these if changed through different phases of the study, and the reasons for any individual housing.
In the USA, research with non-human primates is regulated by the Animal Welfare Act and its regulations, which are administered by the Animal and Plant Health Inspection Service (APHIS) of the US Department of Agriculture (USDA). Separately, regulations for handling pathogens are set by the Division of Select Agents and Toxins (DSAT) of the Centers for Disease Control and Prevention (CDC). Most facilities tend to separate NHPs before being moved into rooms for infectious disease studies or before the challenge phase. Practicality, staff safety, and scientific concerns (around animal cross-contaminations and accurately tracking and managing animal biosamples that potentially contain infectious material/select agents) are also identified as barriers to social housing. However, it is unclear why this should be the case in some world regions and not in others. It is notable that in the UK, vaccine efficacy studies conducted with pair-or group-housed NHPs have been accepted by the FDA, and the facilities have been inspected by CDC personnel to select agent criteria, with no issues raised regarding social housing.
The consensus is that for reasons of both good animal welfare and good science, NHPs should be socially-housed during vaccine studies, and with appropriate infrastructure and expertise this can be achieved in most cases, especially during the phase of the study where animals are held vaccinated but prior to the challenge phase. Long-standing experience shared at the workshop by PHAC, Public Health England (PHE), the Defence Science and Technology Laboratory (DSTL), and the French National Institute of Health and Medical Research (INSERM) demonstrates that social housing is also possible during the challenge phase. In the opinion of these research teams, pair-and group-housed animals are more content, calmer, and may recover more quickly from procedures than singly housed animals on studies involving high containment pathogens. Social buffering studies (including in BSL-2) have shown the powerful role of social housing in mitigating reactions of NHPs to stressful events, enhancing immune responses, and optimizing the ability to cope with disease, potentially leading to models that are more representative of the human condition. It is conceivable that working with calmer, less stressed subjects could also result in less variable data and more reproducible findings, which may enable a reduction in NHP use, though to our knowledge, this has not yet been proven in specifically designed studies. It is incumbent upon organizations like CEPI and the NC3Rs to fund such studies to validate the scientific benefits of social housing.
There is a large literature on how to set up and maintain compatible pairs/groups of NHPs. Breeding on-site, or else a strong relationship with external suppliers who can provide biographical information and behavioral compatibility data, enables the identification of compatible animals for experimental use, minimizing the likelihood of aggression occurring later. These practices also facilitate training of the animals in preparation for studies. Such animals may come at higher prices, but obtaining pre-established pairs or groups takes the burden off researchers and care staff at the experimental facility, reducing the investment of time and effort. In addition, cage groups can be organized such that animals are familiar with their nearest neighbors before being moved into the high containment area.
Laboratories that group house up to six animals per group report no issues in terms of results being confounded by cross-infection provided there is careful allocation of treatments (i.e., all animals housed in each group receive the same treatment, such as the infectious dose and/or vaccination regimen) and refinement of challenge doses. Blinded studies sometimes require mixing of experimental treatment groups, or at least coded cage-labeling, which is operationally more difficult, but possible. One solution could be personnel-based: it may be possible to have two technical teams, one which is aware of animal group assignments, for the purpose of administering vaccines or therapeutics, and a second blinded team unaware of group assignments to perform the daily clinical health assessments, especially if observations which can be subjective are a primary endpoint for the study. Obviously, socially housed individuals need to be well-identified (e.g., via tattoo). Where surgical procedures are necessary, compatible pairs/group can be temporarily separated for post-surgical recovery and wound healing. Some experiments will be designed to observe potential disease recrudescence; for example, after completion of an experimental treatment, animals may be held for an extra month to determine if latent infection might reactivate. Pair housing might confound this situation because the second animal may be re-exposed to the infection, and could raise questions about the efficacy of the experimental vaccine or immune status post-challenge. This situation would require careful observation, immune assessments, and evaluation of both pair-housed animals before the study were brought to a close.
Monitoring several animals in a social group simultaneously is time efficient for the researcher and care staff. In addition, behavioral signs of illness (e.g., changes in locomotion, balance, coordination, and peer interaction) are often easier to detect in socially housed NHPs because they express a wider range of natural behaviors than do singlyhoused animals. Any deterioration in well-being shows a marked contrast against both the behaviors of cage mates and the previous behaviors of the affected animals. Measurement of individual food intake and excreta is a challenge in social groups, but these parameters may not be primary endpoints in vaccine or therapeutic efficacy studies, but rather toxicology studies not involving an infectious agent. If food and water intake are important endpoints, then planning for paired housing and quantitative intake assessment methods should be carefully considered in the study design. Intake can be observed on a pair/cage basis instead and used along with other indicators of well-being as part of a holistic approach, such as body weight measurement, noting of feeding habits during feeding times, behavioral observations via closed-circuit television (CCTV) by experienced staff (i.e., familiar with the individual animals and their "normal" behavior), and clinical chemistry and virology following blood sampling.
## Enclosures and enviromental enrichment
For good health and psychological wellbeing, captive NHPs require a complex and stimulating environment that provides them with opportunity for social interactions, exercise, and the ability to express a wide range of behaviors appropriate to the species. Unfortunately, this is not always provided to NHPs housed under high biocontainment. Where this is the case, pre-challenge work, such as the vaccination phase of a study, should be done in the normal colony environment or high-quality standard accommodation area, because this approach can act as a refinement by minimizing the amount of time spent in relatively small, unenriched cage units of the high containment environment.
In the experience of EU NHP facilities, high containment studies can be run in the group-housed condition in relatively large and well-enriched pen-style enclosures (i.e., 3.6 m 3 /127 ft 3 for two macaques >3 years of age, which is the legal minimum in the EU under Directive 2010/63/EUyet still be compliant with safety and good laboratory practice (GLP) standards. Structural enrichment (e.g., ropes/rope ladders), toys/manipulanda (e.g., balls, wooden dumbbells, mirrors), and foraging devices (e.g., puzzle feeders, cardboard forage boxes, forage trays with floor substrate such as Enviro-dri ® or SizzleNest ® ) are commonly provided, all of which are easily decontaminated or disposed of at the end of studies. A high level of environmental enrichment increases the scope for scoring relevant clinical signs because NHPs often change the level of interaction with enrichment items when they experience adverse effects, and this may occur before other detectable clinical signs are present.
Cages used for studies involving infectious agents need to be able to withstand robust decontamination processes, such as autoclaving and fumigation. Whilst stainless steel is the obvious choice as a material for such cages, other materials can be used to provide a softer and less noisy environment, such as Trespa ® for cage partitions and polypropylene for platforms, shelves, and perches.
PHE has developed a state-of-the-art containment system which balances the need for staff safety at all stages of the experimental process, whilst addressing the requirements of macaques for space, environmental enrichment, and social interaction. The system has been used to study a variety of infectious agents, including SARS-COV-2. It allows group housing whilst offering quantifiable operator protection during infection, experimental sampling, and husbandry. The containment system utilizes the principle of directional airflow away from the operator towards the rear of the cage in order to protect staff (staff still wear full personal protective equipment [PPE]). With correct operation, this flow is maintained at a minimum of 0.7 M/second and the air extracted by a total-loss room air handling system via high efficiency particulate air (HEPA) filtration. A series of interconnected cages are placed inside modular booths that have a clear rigid plastic screen at the front of the system. This screen allows good observation of the caged macaques whilst providing protection from physical contamination by urination, defecation, cough, throwing of objects, or by reaching through from the cage. Each screen is fitted with flap valves that control the velocity of the air into the system and each has a number of small access doors placed in strategic alignment with the cage front to allow animal care staff access to replace food and water, change bedding, and sedate the occupants for subsequent procedures. Further operator protection is provided by using a transfer box for sedated animals and a downdraught table for more risky procedures, such as blood sampling, bronchioalveolar lavage, X-rays, close clinical assessment, and necropsy. Additional details are given under "Housing" on the NC3Rs Macaque website. DSTL has also developed bespoke caging for its BSL-3/4 studies with vasectomized male-female pairs of marmosets. Cages are constructed of stainless steel, with a polypropylene section at the rear to ensure full telemetry coverage, and housed in a rigid half-suit isolator maintained at negative pressure. The animals are provided with a good amount of cage furniture (e.g., removable nest box, veranda, and wooden/plastic perches), treats, and a varied diet in high containment. A hammock positioned at the front of the cage facilitates observations of the animals, even by CCTV and at night. A cassette of nylon netting can be inserted at the top of the cage and wound down to bring the animal to the base of the enclosure quietly and gently, which enables safe restraint and handling, avoiding the need for hand capture or use of a heavy cage squeeze-back mechanism.
Where facility architecture (e.g., size of rooms, dimensions of doors, size and configuration of autoclaves) prevents the purchase of more spacious pen-style enclosures, existing caging systems can be customized to meet the animals' needs for companionship and provide greater space without decreasing housing capacity or incurring major financial cost; for example, side panels can be removed to link and combine standard oneover-one caging units. Pair housing need not decrease animal capacity since the pair can occupy the combined space previously given to the two animals separately. The squeezeback mechanism can be retained in one of the units, so that animals can still be restrained, enabling greater cage enrichment in the other/adjacent unit. Animals can be readily trained before the study to shift on command into the unit with the squeeze-back and will reliably do so. In this way, there is no greater risk to staff safety and no need for new caging or larger squeeze-backs, which would be unsafe and unwieldy in biocontainment.
## Animal training, sedation, and selection
When working with dangerous pathogens, physically restraining NHPs to administer substances is not a safe option, and where significant volumes of blood are required it would be difficult to sample safely from an unsedated individual. Chemical restraint is therefore widely used, but researchers need to be aware of the impact of sedation on physiological processes and translational relevance to clinical settings. For instance, ketamine sedation was recently found to blunt cytokine levels in rhesus and cynomolgus macaques compared to awake cooperating animals. It is associated with a prolonged reduction in daily food intake in NHPs, which can lead to significant body weight loss when used repeatedly. Where sedation is used, fully and quickly reversible anesthetic agents are preferable since these enable rapid recovery (e.g., eating and drinking) within minutes and swift return to the social group.
Regardless of whether sedation is used, NHPs should be trained to cooperate with study procedures and adapt to the high containment environment. Effort put into training and acclimation will be recouped later in facilitating the performance of the study and animal monitoring and welfare, leading to higher quality, less variable data. NHP training can be conducted at the breeding unit, in the pre-study phase, and/or before transfer to containment housing or high level containment, and should be performed consistently amongst researchers and technicians. Training approaches could be considered important methods that impact the results of the study; therefore, training methods should be standardized through the use of standard operating procedures (SOPs), staff training sessions, and CCTV monitoring to deliver consistency in technique. There are some excellent resources on humane training methods available in the literature.
Good practice identified at the workshop includes:
- Training animals using positive reinforcement (e.g., food rewards) to station (approach a specific location) or target (touch a specific object) on command, including when in pairs/groups, so that there is reliable voluntary control of animal movement when housed in the containment environment and all individuals can be dosed without interference.
## -
Training to come forward and present themselves for intramuscular injection of a sedative, so that procedures such as X-ray, weighing and temperature measurement, which are vital to clinical assessment, can be conducted safely and with minimal stress caused to the animals. Marmosets have also been trained within 7-10 days to enter a removable capture box which can then be connected to a gaseous anesthesia unit in a biosafety cabinet for sedation without handling the animal.
## -
Training to sit on a scale or in a weighing bucket (for small monkeys, such as marmosets), so that staff members do not need to handle the animals for weighing once they are on-study (resulting in less stress for animals and staff).
## -
Training to take liquid (e.g., fruit juice, milkshake) from a syringe, which can later be used to administer potentially bitter tasting therapy or medication as required, rather than requiring anesthesia to administer dose. - Assisting adaptation to the "new" containment environment by transfer of familiar items (e.g., scent-marked nest box of marmosets).
## -
Interacting with the animals pre-study to familiarize them with the personnel involved, the appearance of staff in full PPE, and human behavior. This also enables staff to get to know individual animals and how each behaves, meaning they can then more easily identify changes from normal, facilitating humane endpoints based on changes in behavior.
## -
Integrating animal training into the experimental protocol.
PHE has cynomolgus macaques (two origins) and rhesus macaque (Indian) breeding colonies on site, which means they can select the best species/model for the disease. They have published on species/strain differences in model development, and it is worth considering what implications colony origin and disease differences may have on model refinement and translation. Use of animals from different populations or vendors may result in wide variation in endpoints; an example of this was given at the workshop. In general, it is desirable to use animals from one source on a given study or series of studies, to support collection of consistent data.
## Humane endpoints
A key component of refining NHP vaccine studies is the identification and implementation of humane endpoints. These can be defined as the earliest reliable indicator in an animal of pain, distress, suffering, or impending death on the basis of which the animal is euthanized, treated, or removed from the study. By substituting these planned triggers for intervention in place of more severe experimental outcomes, such as advanced pathology or death, humane endpoints prevent or alleviate unnecessary pain and distress, whilst still meeting the desired experimental objectives.
Participants at the workshop agreed that it is essential to identify accurate humane endpoints prior to study initiation, and to monitor animals at an appropriate time and frequency, to enable the earliest possible euthanasia decision and avoid moribundity and spontaneous animal death in severe studies. The humane endpoint guidelines of the US Association of Primate Veterinarians, p.7) state: "A moribund condition indicates an animal is in a severely debilitated state and in terminal distress. Moribund condition and death should be avoided as study endpoints (unless there are no alternatives) and must be scientifically justified and approved by the IACUC. Unless scientifically justified and approved by the IACUC, all moribund NHPs should be immediately evaluated by a veterinarian and euthanized."
Aside from the animal welfare benefits, identification of early endpoints can improve scientific results by helping to ensure that biologic samples are minimally degraded and of limited variability due to adverse effects of severe illness (e.g., inability to obtain food or water, dehydration, hypothermia). Using early endpoints also minimizes the influence of adverse effects unrelated to the infection on disease development and animal death, which would otherwise complicate determination of an organism's pathogenicity.
A comprehensive review of humane endpoints used in NHP vaccines studies is beyond the scope of this article. Typical endpoints for disease models of infection with CEPI-prioritized pathogens include, but are not limited to, body temperature, inappetence, weight loss, slowness or lack of responsiveness, and/or lying prostrate. Different viral models, and even different strains of the same virus, may require different endpoints due to varying aspects of unique pathogenesis caused by those viruses/strains. The goal of the study/vaccine, inoculation route, and pathogen dose level may also influence the choice of endpoints. For example, if the goal of a vaccine is to prevent all signs of disease, then selection of early endpoints, such as a change in body temperature, could suffice for drawing the study conclusions; whereas late onset therapeutic intervention would necessitate significant and more advanced clinical signs.
Humane endpoints should be continually refined over time as more data become available from studies using the specific model. Compiling and reviewing historical datasets can identify trends, enabling further or earlier endpoints to be developed. Participants at the workshop felt the field needs more natural history studies of CEPI pathogens for the specific purposes of defining the exact disease progression in the models and facilitating endpoint identification and validation. Greater data sharing and collaboration between research and veterinary staff could also assist the development and refinement of humane endpoints. It would help to standardize endpoint criteria across institutions so that similar study outcomes could be observed. For endpoints to be used consistently across laboratory sites, thorough training of staff members is required. All personnel performing endpoint criteria assessments need to be able to recognize the signs of disease for the infection models and apply consistent scores to clinical signs of illness and changes in behavior. Veterinarians need to also take a consistent approach to observations and treatment, when required. Structured welfare assessment/humane endpoint score sheets are valuable tools that can be used to monitor and document behavioral and physiologic parameters that are predictive of changes in clinical conditions. These should be available to all relevant staff and preferably be included in training on standard operating procedures (SOPs). One group that worked to harmonize endpoint criteria across multiple institutions was the Filovirus Animal Non-clinical Group (FANG), a US interagency group focused on facilitating the development of filovirus medical countermeasures. Full adoption of a standardized scoring system was not achieved but there was greater harmonization of endpoints across laboratories. Discussions regarding changes in established endpoints need to include not only the researchers but also veterinary staff and IACUC members.
Clinical scores are sometimes accompanied by secondary endpoint measurements, if available, such as hematology and serum chemistry parameters measured on laboratory instruments. At this time, reliable secondary blood measurement endpoints are not uniformly described for Lassa, Nipah, MERS-CoV, or other CEPI pathogens. Better datasets, such as those that could be collected through natural history studies, would help inform changes in blood chemistry parameters that may be predictive of lethal disease and/or serve as early biomarkers for euthanasia decisions. Such an approach was employed for assessment of the Ebola virus rhesus macaque model, which led to several institutions adopting secondary endpoints of temperature change and specific clinical chemistry values.
When endpoints are measured during a study, baseline (pre-pathogen exposure) measurements should be taken for each animal and changes relative to that animal's baseline should inform decisions about clinical scoring. Animals should be assessed at times and frequencies that will best help to identify the early onset of harmful effects. Similarly, experiments should be scheduled so that an adequate number of trained research personnel are available to assess animals when they are most severely affected and require intensive monitoring. For marmoset infectious disease studies at DSTL, welfare monitoring is performed at 8-h intervals, including CCTV and physical observation, and there is 24-h coverage by experienced scientists and technicians. Once animals begin to show clinical signs, observation is increased to every 4 h and then moves up to continuous monitoring until the predetermined humane endpoint criterion is reached.
Telemetry is an innovation that is not as widely used as it could be for implementing humane endpoints in NHP infectious disease studies, perhaps due to the cost for instal-lation in high containment laboratories. Telemetry offers a large, accurate, and objective dataset that can be used to refine endpoints such as body temperature, blood pressure, ECG and heart rate, respiratory rate, activity levels, and even glucose in some implant models. Pre-implanted animals can be ordered if implantation surgery is not possible at the receiving BSL-4 laboratories. If more facilities used telemetry, then model development could perhaps be better standardized across facilities through use of the same endpoint criteria (e.g., temperature, hypotension).
At DSTL, telemetry devices are implanted 4 weeks prior to challenge to enable monitoring of the core body temperature of individual marmosets in real time. The facility can record from 16 pair-housed animals at once. Marmosets exposed to Burkholderia pseudomallei show clinical signs 12-24 h after temperature increase. Once there is observation of a sharp temperature decline following the febrile response, the animals are euthanized before more severe disease signs are observed. A decline to 39 - C has been defined as the humane endpoint criterion (in combination with clinical signs), as the disease model demonstrates little variation between animals. Marmosets at this facility also wear an activity tracking device (Actiwatch) on a collar, from which data can be downloaded at the end of the study. The animals become less active as disease progresses, so this change in behavior may be used as an endpoint in some models, together with other data. The CCTV camera at the top of the cage, which enables monitoring with minimum disturbance, is particularly useful at the end stage of disease, when external stimuli such as entrance of a caretaker into the room may cause the animal to react and hide its illness/vulnerability. The laboratory is currently investigating a Data Sciences International (DSI) telemetry implant that enables measurement of core body temperature, EEG/ECG, HR, BP, and activity, and can synchronize with camera data.
Telemetry has also been used extensively in the establishment of a model which has been submitted under the FDA Animal Model Qualification program. The cynomolgus macaque model of inhalational tularemia for assessment of therapeutics was developed at multiple institutions under a NIAID program. As with the marmosets described above, telemetry devices were implanted in the animals to allow monitoring of core body temperature. In this model, a fever early in the disease course was used as a trigger for antibiotic treatment. Late in the disease course, hypothermia was consistently observed and was incorporated into euthanasia criteria.
Medical imaging (such as CT and PET scanning) is another useful innovation, but it is less widely available than telemetry, especially in biocontainment facilities. For its aerosol challenge studies with macaques (e.g., TB), PHE hires portable clinical scanners to come on site, and the animals are transferred into a negative pressure imaging pod. This allows them to monitor disease progression and look for disease effects (lesions and nodal enlargement), instead of leaving animals to suffer severe disease, euthanizing at an anticipated timepoint, and doing histopathology. For tuberculosis studies, CT scanning also allows evaluation of the challenge dose received by counting primary tuberculin lesions, and this has in turn enabled the use of much lower, clinically realistic challenge doses that can be used to evaluate efficacy whilst improving welfare through reduced disease burden. PET-CT methods represent a longer process for imaging capture, so the laboratory uses high frequency jet ventilation (HFJV) to stop imaging motion artefacts caused by breathing. It has published a methodology for disease evaluation from CT images that aligns with clinical scores seen at necropsy, allowing much earlier detection of treatment effects and hence shorter time courses for studies in high containment. CT and PET/CT is also being used by multiple facilities for analysis of COVID-19 disease in macaques which allows assessment of disease onset and timing of treatment interventions.
## Supportive care
Supportive care, such as fluid replacement and pain relief, is another means of refinement. Participants at the workshop considered that provision of supportive care is important for humane reasons and expected by ethics committees (e.g., IACUCs). Vac-cinated animals need to be healthy upon entry into the challenge phase of the study, so animals displaying idiopathic dehydration or diarrhea during the vaccine phase are generally placed under watch of a veterinarian, and given fluids and common medications for relief of symptoms. Some veterinary treatments, such as immunosuppressants, may be contraindicated in the challenge phase, though palliative care should always be possible and does not seem to affect disease outcomes in the experimental model (e.g., offering soft or moistened food; additional, hydrating fresh fruit/vegetables; ice cubes; oral rehydration solution; liquid via a syringe; or moving the water bottle closer to an animal showing muscle weakness).
The type of supportive care offered should be selected carefully since treatments may interfere with the humane endpoint criterion, correlate of protection or biomarker that is being measured. So, for example, where the febrile response is the primary indicator of disease and essential for the humane endpoint, fever reducing medications or non-steroidal anti-inflammatory drugs (NSAIDs) may not be appropriate. The plan for supportive care should be detailed prospectively in the study design and agreed upon among the stakeholders (e.g., product development sponsors, regulatory agencies) at the outset. Considerations such as whether each animal will receive care when meeting a trigger point, whether every animal will receive the same supportive care uniformly, and whether staff are suitably trained and competent to detect key signs of illness, should be discussed during protocol design. Some respondents at the workshop commented that administration of supportive care might increase the risk to staff (e.g., if it involves handling of sharps, or extra hours spent in high containment), so the biosafety considerations need to be considered alongside animal welfare.
Where the effect on correlates of infection or protection is unknown, it may be necessary to test each intervention to determine its effect on the model. Participants had experience with administering subcutaneous or intravenous fluid therapy to models of Lassa or Ebola virus infection in macaques. In one specific case, supportive care did not offer a statistically significant survival benefit or extend mean time to death, which means that adopting these practices during experimental studies of a particular therapeutic's efficacy should not confound the study results and may importantly afford the animal some alleviation of pain or suffering. This group also tested whether Tylenol (Acetaminophen), as a pain reliever that is processed by the liver, had any effects on liver enzyme levels during Ebola virus infection of rhesus macaques, which it did not (even though Ebola virus is known to cause virus-induced liver cytopathicity, which might potentiate further damage after Tylenol administration). Other drugs may potentiate liver damage in virally infected animals, so a plan for testing is important.
The use of suitable analgesics should be considered when it is expected and predicted that the challenge process could be painful at any stage. If analgesia is used, the type of analgesic, dose, and treatment regime need to be empirically examined to assess not only whether it is effective and beneficial, but also to confirm that it does not materially interfere with the course of the disease or identification of humane endpoints.
# Conclusions
There is considerable scope to further refine NHP vaccine studies to minimize harm to the animals involved and to maximize data quality, in line with societal, regulatory, and funder expectations. Researchers, veterinarians, technical and biosafety staff, regulators, and funders should therefore work together to improve existing practices and challenge dogma. With sufficient will, knowledge, training, and resources, NHP vaccine studies can be conducted to GLP, or other systems that assure data quality and integrity, while meeting genuine high standards of animal welfare. For example, experience has shown pair-or group-housing of NHPs does not have a negative effect on the science or the ability to do experiments in BSL-4 conditions, and biosafety departments can adequately quantify the methods used for keeping staff safe. Safety is, of course, a prime consideration at high biocontainment levels and training of staff to ensure consistent use of de-risked practices is essential. Institutional SOPs should be written and reviewed for safety and risk management around all practices, including social housing.
Funders of NHP research are placing increasing emphasis on the 3Rs and are important drivers for improving standards, helping to ensure such studies are legitimate, ethical, and high-quality. Discussion at the workshop led CEPI in August of 2019 to release a call for proposals "High Containment Studies to Support Product Development" with the goal of setting up contracts with preferred laboratories that meet standards laid down by the NC3Rs and high regulatory quality standards, such as GLP. As part of this, CEPI agreed to make investments in infrastructure and training to help laboratories achieve the required standards. Ultimately contracts were signed with several facilities in the EU for animal studies at BSL-3/4, where novel animal telemetry, plethysmography, and activity-monitoring equipment were purchased, to enrich the animals' housing as well as expand the data endpoints collected at these laboratories. In the USA, a CEPI contract to the University of Pittsburgh facilitated an upgrade to its NHP caging, which enabled social housing in an expanded-size enclosure that approaches EU caging standards. These investments, made at the early stages of the pandemic, have already contributed to achievements in performance of 3Rs compliant studies supporting COVID-19 vaccine development. In parallel, the NC3Rs has established a working group with the US National Toxicology Program Interagency Center for the Evaluation of Alternative Toxicological Methods (NICEATM) and other agencies to coordinate global efforts to use microphysiological systems (organon-a-chip and other complex multicellular in vitro models) to reduce reliance on in vivo studies of COVID-19 and future infectious diseases.
Given their importance for both animal welfare and study quality, refinement measures employed in NHP vaccine studies should be properly reported in published manuscripts, in accordance with the ARRIVE guidelines. For further guidance and resources on improving NHP care and use, including dedicated websites, an annual symposium, and data sharing working groups, visit the NC3Rs website (www.nc3rs.org.uk/nhpwelfare; accessed on 18 March 2021). Funding: M.J.P. was supported by the NC3Rs. A.C.S., C.C. and W.E.D. were supported by CEPI. |
Epstein-Barr Virus-Encoded RNAs: Key Molecules in Viral Pathogenesis
The Epstein-Barr virus (EBV) is known as an oncogenic herpesvirus that has been implicated in the pathogenesis of various malignancies. EBV-encoded RNAs (EBERs) are non-coding RNAs expressed abundantly in latently EBV-infected cells. Herein, I summarize the current understanding of the functions of EBERs, including the interactions with cellular factors through which EBERs contribute to EBV-mediated pathogenesis. Previous studies have demonstrated that EBERs are responsible for malignant phenotypes in lymphoid cells, and can induce several cytokines that can promote the growth of various EBV-infected cancer cells. EBERs were also found to bind retinoic acid-inducible gene I (RIG-I) and thus activate its downstream signaling. Furthermore, EBERs induce interleukin-10, an autocrine growth factor for Burkitt's lymphoma cells, by activating RIG-I/interferon regulatory factor 3 pathway, suggesting that EBER-mediated innate immune signaling modulation contributes to EBV-mediated oncogenesis. Recently, EBV-infected cells were reported to secret EBERs, which were then recognized by toll-like receptor 3 (TLR3), leading to the induction of type I interferon and inflammatory cytokines, and subsequent immune activation. Furthermore, EBER1 was detected in the sera of patients with active EBV-infectious diseases, suggesting that EBER1-meidated TLR3 signaling activation could account for the pathogenesis of active EBV-infectious diseases.
# Introduction
Epstein-Barr virus (EBV), a ubiquitous human herpes virus that has infected more than 90% of the global populations, is associated with a variety of malignancies, including endemic Burkitt's lymphoma (BL), Hodgkin's lymphoma, nasopharyngeal carcinoma (NPC), gastric carcinoma (GC), and opportunistic lymphoma. These "EBV-associated" cancers are characterized by the proliferation of monoclonal EBV-infected cells with restricted latent viral gene expression [bib_ref] Epstein-Barr virus, Rickinson [/bib_ref].
EBV encodes small nonpolyadenylated, non-coding (nc) RNAs called EBV-encoded RNA (EBER) [bib_ref] Epstein-Barr virus, Rickinson [/bib_ref] [bib_ref] Two small RNAs encoded by Epstein-Barr virus and complexed with protein are..., Lerner [/bib_ref] ; these are the most abundant viral transcripts in latently EBV-infected cells [bib_ref] Identification of transcribed regions of Epstein-Barr virus DNA in Burkitt lymphoma-derived cells, Rymo [/bib_ref]. EBER1 and EBER2, which are respectively 167 and 172 nucleotides long, are transcribed by RNA polymerase III (pol III) [bib_ref] Striking similarities are exhibited by two small Epstein-Barr virus-encoded ribonucleic acids and..., Rosa [/bib_ref]. EBERs can be used as target molecules for the in situ hybridization (ISH) detection of EBV-infected cells in tissues [bib_ref] Description of an in situ hybridization methodology for detection of Epstein-Barr virus..., Chang [/bib_ref] and are considered a reliable marker of the presence of EBV, as they are abundantly expressed in latently EBV-infected cells. Previous studies have demonstrated the roles of EBERs in EBV oncogenicity. Expression of EBER in B lymphocytes induces colony formation in soft agar and tumorigenesis in nude mice [bib_ref] Oncogenic role of Epstein-Barr virus-encoded RNAs in Burkitt's lymphoma cell line Akata, Komano [/bib_ref] [bib_ref] Epstein-Barr virus small RNAs potentiate tumorigenicity of Burkitt lymphoma cells independently of..., Ruf [/bib_ref]. In addition, EBERs confer resistance to RNA-dependent protein kinase (PKR)-induced apoptosis upon BL cells [bib_ref] Epstein-Barr virus RNA confers resistance to interferon-alpha-induced apoptosis in Burkitt's lymphoma, Nanbo [/bib_ref]. Previous studies have also demonstrated that EBERs induce the transcription of various cytokines depending on cell type, such as interleukin-10 (IL-10) in BL cells, insulin-like growth factor-1 (IGF1) in epithelial cells and IL-9 in T cells; these cytokines subsequently act as autocrine growth factors for the EBV-infected cancer cells [bib_ref] Epstein-Barr virus-encoded poly(A)(-) RNA supports Burkitt's lymphoma growth through interleukin-10 induction, Kitagawa [/bib_ref] [bib_ref] Autocrine growth of Epstein-Barr virus-positive gastric carcinoma cells mediated by an Epstein-Barr..., Iwakiri [/bib_ref] [bib_ref] Epstein-Barr virus-encoded small RNA induces insulin-like growth factor 1 and supports growth..., Iwakiri [/bib_ref] [bib_ref] Epstein-Barr virus (EBV)-encoded RNA promotes growth of EBV-infected T cells through interleukin-9..., Yang [/bib_ref]. More recent studies have shown that EBER1 is sufficient to elicit these phenotypes [bib_ref] Growth-promoting properties of Epstein-Barr virus EBER-1 RNA correlate with ribosomal protein L22..., Houmani [/bib_ref] [bib_ref] Lymphoid hyperplasia and lymphoma in transgenic mice expressing the small non-coding RNA,..., Repellin [/bib_ref]. Regarding the molecular action mechanisms of EBREs, these ncRNAs were reported to contribute to the EBV infection pathogenesis by modulating of innate immune signals [bib_ref] EB virus-encoded RNAs are recognized by RIG-I and activate signaling to induce..., Samanta [/bib_ref] [bib_ref] Epstein-Barr virus-encoded small RNA induces IL-10 through RIG-I-mediated IRF-3 signaling, Samanta [/bib_ref] [bib_ref] Epstein-Barr virus (EBV)-encoded small RNA is released from EBV-infected cells and activates..., Iwakiri [/bib_ref].
## Structure of ebers
EBER1 and EBER2 are 166 and 172 nucleotides long, respectively [bib_ref] Striking similarities are exhibited by two small Epstein-Barr virus-encoded ribonucleic acids and..., Rosa [/bib_ref] , and only 54% sequence homology exists between these two RNAs. The EBER genes are separated by 161 base pairs and are transcribed from left to right on the EBV map. Both EBER genes carry intragenic transcription control regions for RNA polymerase (pol) III, and can be transcribed by this molecule. Despite the low (54%) primary sequence similarity, both EBER1 and EBER2 exhibit striking similarities in their secondary structures and feature extensively base-paired structures with several short stem loops [fig_ref] Figure 1: Secondary structures of EBERs [/fig_ref]. The secondary structure of EBERs is similar to that of adenovirus-associated RNAs (VAs), which are also small non-polyadenylated RNAs. In addition, the function of VA1 can be partially replaced by EBERs [bib_ref] Striking similarities are exhibited by two small Epstein-Barr virus-encoded ribonucleic acids and..., Rosa [/bib_ref]. It has also been reported the similarity between the primary RNA structure of EBER1 and herpesvirus papio type 1 [bib_ref] Isolation and characterization of the genes for two small RNAs of herpesvirus..., Howe [/bib_ref]. The differences in the secondary structures suggest that EBER1 and EBER2 have potentially distinct functions, although secondary structure conservation seems to be crucial for the functions of EBERs. The replacement of guanine residues with inosine in EBER1 was reported to disrupt the secondary structure of EBER1, leading to lost ability to form complexes with PKR [bib_ref] Functional expression and RNA binding analysis of the interferon-induced, double-stranded RNA-activated, 68,000-Mr..., Katze [/bib_ref].
The primary sequences of EBERs are highly conserved among a number of EBV strains, with only two reported changes within EBER2 and none in EBER1. A total of eight changes have been observed outside of the coding regions [bib_ref] Functional expression and RNA binding analysis of the interferon-induced, double-stranded RNA-activated, 68,000-Mr..., Katze [/bib_ref]. Family 1 and family 2 parallel type A and type B EBVs (also called type I and type II EBVs) are defined by variations in EBV-nuclear antigen 2 (EBNA2) and EBNA3; however, some isolates appear to be intertypic recombinants, which are often observed in isolates from HIV patients [bib_ref] Two families of sequences in the small RNA-encoding region of Epstein-Barr virus..., Arrand [/bib_ref]. The high level of sequence conservation among EBERs suggests the importance of EBERs in the virus life cycle. [bib_ref] Striking similarities are exhibited by two small Epstein-Barr virus-encoded ribonucleic acids and..., Rosa [/bib_ref]. EBER, Epstein-Barr virus-encoded RNA.
## Synthesis and expression of ebers
EBERs are primarily transcribed by RNA pol III [bib_ref] Identification of transcribed regions of Epstein-Barr virus DNA in Burkitt lymphoma-derived cells, Rymo [/bib_ref]. Class III promoters are characterized by their intragenic location and, in the case of EBERs, have sequences that are nearly identical to the consensus sequences derived from boxes A and B. In addition, EBERs contain three upstream elements that together stimulated in vivo expression 50-fold and contain a TATA box and ATF-and Sp1-like promoter elements, which resemble sites associated with typical class II promoters [bib_ref] Isolation of intertypic recombinants of Epstein-Barr virus from T-cell-immunocompromised individuals, Yao [/bib_ref] , thus indicating the possibility that EBERs can also be transcribed by RNA pol II [bib_ref] Isolation of intertypic recombinants of Epstein-Barr virus from T-cell-immunocompromised individuals, Yao [/bib_ref] [bib_ref] Epstein-Barr virus small RNA (EBER) genes: Unique transcription units that combine RNA..., Howe [/bib_ref].
The EBER copy number appears to be related to the DNA molecule copy number in each cell type [bib_ref] Characterization of the major Epstein-Barr virus-specific RNA in Burkitt lymphoma-derived cells, Arraud [/bib_ref]. Following the EBV infection of primary B lymphocytes, EBV-determined nuclear antigen 2 (EBNA2) appears first at 6 h post infection, followed by other viral genes, latent membrane proteins (LMPs) and EBERs. On the other hand, the non-transforming P3HR-1 strain, from which the EBNA2 gene is deleted, expresses only the EBNA-leader protein (EBNA-LP) and trace amounts of EBER1 in primary B lymphocytes, whereas the same virus can express EBNA1, EBNA3, EBNA-LP and EBERs in EBV-genome-negative BL cell lines [bib_ref] Influence of Burkitt's lymphoma and primary B cells on latent gene expression..., Rooney [/bib_ref]. These findings suggest that EBER expression depends on the host cell, likely through products specific to the cell cycle or the state of B-cell differentiation.
EBER expression appears to be affected by the state of viral life cycle status. A previous study demonstrated downregulated EBER transcription during the switch from latent infection to lytic viral replication [bib_ref] Epstein-Barr virus small RNA (EBER) genes: Differential regulation during lytic viral replication, Greifenegger [/bib_ref]. In contrast, the amounts of EBERs remain unaltered within 72 h after the induction of lytic replication. Although both EBERs are transcribed at approximately equal rates, EBER1 has been reported to be present at 10-fold higher levels relative to EBER2, presumably because of the longer half-life of EBER1. In the presence of actinomycin D, the half-lives of EBER1 and EBER2 are 8 to 9 h and 45 min, respectively [bib_ref] Expression of genes for the Epstein-Barr virus small RNAs EBER-1 and EBER-2..., Clarke [/bib_ref].
EBERs are not expressed in the tissues of oral hairy leukoplakia, Sjögren's syndrome, salivary gland lymphoma or oral papilloma, in which EBV actively replicates [bib_ref] Epstein-Barr virus small nuclear RNAs are not expressed in permissively infected cells..., Gilligan [/bib_ref] [bib_ref] Epstein-Barr virus (EBV) infection in salivary gland tumors: Lytic EBV infection in..., Wen [/bib_ref] [bib_ref] Association of Epstein-Barr virus (EBV) with Sjögren's syndrome: Differential EBV expression between..., Wen [/bib_ref] [bib_ref] Detection of Epstein-Barr virus in oral papilloma, Mizugaki [/bib_ref] expressed in latently infected cells. Therefore, an ISH targeted EBER1 has been extensively used as a reliable marker of EBV infection in tissue specimens [bib_ref] Description of an in situ hybridization methodology for detection of Epstein-Barr virus..., Chang [/bib_ref]. A previous study reported the heterogeneous EBER expression in individual NPC tumor cells; specifically, some tumor cells expressed high levels of EBER whereas adjacent tumor cells expressed very little or no EBER [bib_ref] Heterogeneity of HLA and EBER expression in Epstein-Barr virus-associated nasopharyngeal carcinoma, Yao [/bib_ref] , although it was unclear whether EBERnegative cells were positive for the EBV genome.
## Localization of ebers
Although EBERs are located primarily in the nucleus, as was demonstrated via intense nuclear EBER ISH staining [bib_ref] Description of an in situ hybridization methodology for detection of Epstein-Barr virus..., Chang [/bib_ref] [bib_ref] Localization of Epstein-Barr virus-encoded small RNAs by in situ hybridization, Howe [/bib_ref] , high-resolution ISH using confocal laser scanning microscopy has demonstrated that EBERs are found in both the cytoplasm and nuclei of interphase cells [bib_ref] Localization of Epstein-Barr virus-encoded RNAs EBER-1 and EBER-2 in interphase and mitotic..., Schwemmle [/bib_ref] , showing the close association of EBER with the perinuclear region corresponding to the location of the rough endoplasmic reticulum (ER) and Golgi apparatus [bib_ref] Localization of Epstein-Barr virus-encoded RNAs EBER-1 and EBER-2 in interphase and mitotic..., Schwemmle [/bib_ref]. More recently, EBERs were shown to be confined to the nucleus [bib_ref] Epstein-Barr virus noncoding RNAs are confined to the nucleus, whereas their partner,..., Fok [/bib_ref]. On the other hand, Iwakiri et al. [bib_ref] Epstein-Barr virus (EBV)-encoded small RNA is released from EBV-infected cells and activates..., Iwakiri [/bib_ref] reported a new finding that suggested that EBERs are positively secreted in a complex with the lupus antigen (La) protein. Given that EBERs can bind various proteins that are not only restricted to the nucleus but are also located in the cytoplasm, EBERs would be able to localize in the cytoplasm [bib_ref] EB virus-encoded RNAs are recognized by RIG-I and activate signaling to induce..., Samanta [/bib_ref] [bib_ref] Binding of Epstein-Barr virus small RNA EBER-1 to the double-stranded RNA-activated protein..., Clarke [/bib_ref] [bib_ref] AUF1/hnRNP D is a novel protein partner of the EBER1 noncoding RNA..., Lee [/bib_ref] [bib_ref] The La antigen shuttles between the nucleus and the cytoplasm in CV-1..., Bachmann [/bib_ref].
## Interactions with cellular proteins
## La protein
EBERs exist in nuclear ribonucleoprotein (RNP) complexes that can be precipitated using anti-La antibodies, which are associated with systemic lupus erythematosus (SLE) [bib_ref] Two small RNAs encoded by Epstein-Barr virus and complexed with protein are..., Lerner [/bib_ref]. In these complexes, La binds the oligouridylate stretch at the 3'-termini of all mammalian RNA pol III transcripts; this binding is transient for most RNAs but stable in the case of the EBERs [bib_ref] Isolation and characterization of the genes for two small RNAs of herpesvirus..., Howe [/bib_ref]. Although the significance of this interaction is unknown, it is expected to affect the interaction between La and RNA pol III in EBV-infected cells. La is primarily localized in the nucleus; however, it can localize in the cytoplasm under certain condition [bib_ref] The La antigen shuttles between the nucleus and the cytoplasm in CV-1..., Bachmann [/bib_ref] [bib_ref] La autoantigen translocates to cytoplasm after cleavage during granzyme B-mediated cytotoxicity, Huang [/bib_ref] [bib_ref] La autoantigen enhances and corrects aberrant translation of poliovirus RNA in reticulocyte..., Meerovitch [/bib_ref]. The poliovirus-mediated cleavage of the La nuclear localization signals was reported to result in the shuttling of La from the nucleus to the cytoplasm [bib_ref] La autoantigen enhances and corrects aberrant translation of poliovirus RNA in reticulocyte..., Meerovitch [/bib_ref]. A recent study suggested that the EBER-La interaction is significant in terms of the secretion of EBER from EBV-infected cells, as since EBER was found to be mainly released in complex with La [bib_ref] Epstein-Barr virus (EBV)-encoded small RNA is released from EBV-infected cells and activates..., Iwakiri [/bib_ref].
## Pkr
The adenovirus-associated RNAs, VA1 and VA2, are small RNAs that are transcribed by RNA pol III [bib_ref] Intracellular redistribution of truncated La protein produced by poliovirus 3Cpro-mediated cleavage, Shiroki [/bib_ref]. Although no striking nucleotide sequence homology exists between EBERs and VAs, similarities exist with respect to size, degree of secondary structure and genomic organization [bib_ref] Structure of genes for virus-associated RNAI and RNAII of adenovirus type 2, Akusjärvi [/bib_ref]. Similar to VA RNAs, EBERs bind double-stranded RNA-dependent protein kinase (PKR), an interferon (IFN)-inducible serine/threonine kinase that acts as a key mediator of the antiviral effects of IFN [bib_ref] Epstein-Barr virus noncoding RNAs are confined to the nucleus, whereas their partner,..., Fok [/bib_ref] [bib_ref] Two small RNAs encoded by Epstein-Barr virus can functionally substitute for the..., Bhat [/bib_ref] [bib_ref] Molecular cloning and characterization of the human double-stranded RNA-activated protein kinase induced..., Meurs [/bib_ref] [bib_ref] Comparative analysis of the regulation of the interferon-inducible protein kinase PKR by..., Sharp [/bib_ref]. PKR has been reported to bind to the stem-loop IV of EBER1 [bib_ref] The binding site of the RNA-dependent protein kinase (PKR) on EBER1 RNA..., Vuyisich [/bib_ref]. IFNs activates PKR, thus leading to phosphorylation of the α-subunit of the protein synthesis initiation factor eIF2 and causing translational inhibition at the level of initiation. In vitro assays have demonstrated that EBERs can inhibit PKR activation and block the phosphorylation of eIF2α, thus blocking the eiF2α-mediated inhibition of protein synthesis [bib_ref] Functional expression and RNA binding analysis of the interferon-induced, double-stranded RNA-activated, 68,000-Mr..., Katze [/bib_ref] [bib_ref] Comparative analysis of the regulation of the interferon-inducible protein kinase PKR by..., Sharp [/bib_ref] [bib_ref] Translational control by the Epstein-Barr virus small RNA EBER-1. Reversal of the..., Clarke [/bib_ref]. In BL cells, EBERs were also reported to confer resistance to IFN α induced apoptosis by directly binding to PKR and inhibiting its phosphorylation [bib_ref] Epstein-Barr virus RNA confers resistance to interferon-alpha-induced apoptosis in Burkitt's lymphoma, Nanbo [/bib_ref]. McKenna et al. [bib_ref] Viral dsRNA inhibitors prevent self-association and autophosphorylation of PKR, Mckenna [/bib_ref] reported that EBERs/VAs bind preferentially to the latent dephosphorylated form of PKR with a similar affinity to that of dsRNA activators. However, EBERs/VAs prevent PKR dimerization, which is required for efficient PKR trans-autophosphorylation. Consequently, PKR substrate phosphorylation is blocked, allowing protein synthesis to proceed.
## L22
A second highly abundant protein designated EBER-associated protein (EAP), was identified in La-containing RNP complexes [bib_ref] EAP, a highly conserved cellular protein associated with Epstein-Barr virus small RNAs..., Toczyski [/bib_ref]. EBER1 was reported to primarily bind to EAP [bib_ref] The cellular RNA-binding protein EAP recognizes a conserved stem-loop in the Epstein-Barr..., Toczyski [/bib_ref] , and EAP was subsequently shown to be the ribosomal protein L22 [bib_ref] The Epstein-Barr virus (EBV) small RNA EBER1 binds and relocalizes ribosomal protein..., Toczyski [/bib_ref]. Although its functions are not well understood, L22 was identified as the target of chromosomal translocation in certain leukemia-associated proteins [bib_ref] Fusion between transcription factor CBF beta/PEBP2 beta and a myosin heavy chain..., Liu [/bib_ref] [bib_ref] The 3;21 translocation in myelodysplasia results in a fusion transcript between the..., Nucifora [/bib_ref] , suggesting that L22 levels might be a determinant in cell transformation. A previous study suggested that the cellular functions of L22 might involve its association with the human telomerase [bib_ref] Identification of two RNA-binding proteins associated with human telomerase RNA, Le [/bib_ref] , and L22 has also been shown to interact with a number of small viral RNAs other than EBERs [bib_ref] Association of herpes simplex virus regulatory protein ICP22 with transcrip-tional complexes containing..., Leopardi [/bib_ref] [bib_ref] Hepatitis C virus 3'X region interacts with human ribosomal proteins, Wood [/bib_ref].
L22 localizes to the nucleoli and cytoplasm in uninfected human B lymphocytes, however, in EBV-infected cells, in which approximately 30%-50% of the L22 has been found to associate with EBER1, L22 relocalizes to the nucleoplasm [bib_ref] The Epstein-Barr virus (EBV) small RNA EBER1 binds and relocalizes ribosomal protein..., Toczyski [/bib_ref] , suggesting that the cellular distribution of L22 in EBV-infected cells results from the L22-EBER1 interaction. A recent study reported that the distribution of L22 was predominately cytoplasmic in EBV-immortalized lymphoblastoid cell lines (LCLs), but independent of EBER1 [bib_ref] Cellular gene expression that correlates with EBER expression in Epstein-Barr virus-infected lymphoblastoid..., Gregorovic [/bib_ref]. Previous studies have shown that EBER1 contains multiple L22-binding domains, including stem-loop III [bib_ref] The cellular RNA-binding protein EAP recognizes a conserved stem-loop in the Epstein-Barr..., Toczyski [/bib_ref] , stem-loop IV [bib_ref] In vitro selection of RNA ligands for the ribosomal L22 protein associated..., Dobbelstein [/bib_ref] , and stem-loop I [bib_ref] Multiple domains of EBER 1, an Epstein-Barr virus noncoding RNA, recruit human..., Fok [/bib_ref]. These multiple L22 binding domains suggest the possibility that most EBERs form complexes with L22 in vivo and thereby EBERs might modulate protein translation [bib_ref] Multiple domains of EBER 1, an Epstein-Barr virus noncoding RNA, recruit human..., Fok [/bib_ref]. Elia et al. [bib_ref] Ribosomal protein L22 inhibits regulation of cellular activities by the Epstein-Barr virus..., Elia [/bib_ref] reported that L22 and PKR compete for a common binding site on EBER1. L22 hampers the ability of EBERs to inhibit PKR activation by dsRNA through this competition. Transient EBER1 expression in murine embryonic fibroblasts results in reporter gene upregulation and partially blocks the inhibitory effects of PKR. However, EBER1 is also stimulatory when transfected into PKR-null cells, suggesting a PKR-independent manner of EBER function. L22 expression prevents both the PKR-dependent and -independent effects of EBER1 in vivo. These findings suggest that the L22-EBER1 interaction can reduce the biological effect of viral ncRNA, including PKR inhibition and another mechanism by which EBER1 induces gene expression.
## Other interacting proteins
Recent studies have reported the other cellular proteins that interact with EBERs. The host evokes innate immune responses that eliminate invading pathogens by detecting the presence of infection. Cells express a limited number of germ line-encoded pattern-recognition receptors (PRR) that specifically recognize pathogen-associated molecular patterns within microbes. The retinoic acid-inducible gene I (RIG-I) like receptor (RLR) family, which includes RIG-I [bib_ref] The RNA helicase RIG-I has an essential function in double-stranded RNA-induced innate..., Yoneyama [/bib_ref] , melanoma differentiation-associated gene (Mda)-5 [bib_ref] mda-5: An interferon-inducible putative RNA helicase with double-stranded RNA-dependent ATPase activity and..., Kang [/bib_ref] , and LGP2 [bib_ref] Function of RIG-I-like receptors in antiviral innate immunity, Yoneyama [/bib_ref] , comprises cytoplasmic proteins that recognize viral RNA. RLRs are known to play as a key role in IFN-inducible antiviral effects [bib_ref] Toll-like receptors and RNA helicases: Two parallel ways to trigger antiviral responses, Meylan [/bib_ref]. When RIG-I is activated via an interaction with viral dsRNA, it initiates signaling pathways that lead to the induction of protective cellular genes, including type I IFNs and inflammatory cytokines. RIG-I contains a C-terminal DExD/H-box RNA helicase domain and an N-terminal caspase recruitment domain (CARD). The helicase domain is responsible for dsRNA recognition, and the CARD domain activates downstream signaling cascades via the mitochondrial adaptor IFN-β promoter stimulator (IPS)-1, leading to activating the transcription factors, nuclear factor (NF)-κB and interferon regulatory factor 3 (IRF3) [bib_ref] mda-5: An interferon-inducible putative RNA helicase with double-stranded RNA-dependent ATPase activity and..., Kang [/bib_ref] [bib_ref] IPS-1, an adaptor triggering RIG-I-and Mda5-mediated type I interferon induction, Kawai [/bib_ref]. RIG-I is known to recognize 5'-triphosphate RNAs [bib_ref] 5'-Triphosphate RNA is the ligand for RIG-I, Hornung [/bib_ref] , and therefore the EBERs as 5'-triphosphate RNA molecules could also interact with RIG-I [bib_ref] EB virus-encoded RNAs are recognized by RIG-I and activate signaling to induce..., Samanta [/bib_ref]. A recent study suggested that EBER-mediated RIG-I activation contributes to EBV oncogenesis [bib_ref] EB virus-encoded RNAs are recognized by RIG-I and activate signaling to induce..., Samanta [/bib_ref] [bib_ref] Epstein-Barr virus-encoded small RNA induces IL-10 through RIG-I-mediated IRF-3 signaling, Samanta [/bib_ref] , see below for further details.
The AU-rich element binding factor 1 (AUF1) has the ability to bind to AU-rich elements present in the 3'-untranslated regions of precursor RNA [bib_ref] Endotoxic shock in AUF1 knockout mice mediated by failure to degrade proinflammatory..., Lu [/bib_ref]. The association between AUF1 and pre-mRNAs in the nucleus was reported to influence pre-mRNA processing, metabolism, and transport [bib_ref] The role of AUF1 in regulated mRNA decay, Gratacós [/bib_ref] , whereas AUF1 alone might play a role in stabilizing certain transcripts [bib_ref] Concurrent versus individual binding of HuR and AUF1 to common labile target..., Lal [/bib_ref]. A recent study reported that AUF1 is a novel EBER1 binding protein. The EBER1/AUF1 interaction prevents AUF1 from binding to short-lived mRNAs [bib_ref] AUF1/hnRNP D is a novel protein partner of the EBER1 noncoding RNA..., Lee [/bib_ref]. How this interaction might affect the EBV-mediated regulation of various genes remains to be clarified.
## Oncogenic roles of ebers
EBERs have been linked to the malignant phenotypes of BL cells. The transfection of EBER genes into EBV-negative BL-derived Akata cells restored the capacity for cell growth in soft agar, tumor formation in severe combined immunodeficiency mice, resistance to apoptotic inducers, and the upregulation of bcl-2 that could protect Akata cells from c-Myc induced apoptosis [bib_ref] Oncogenic role of Epstein-Barr virus-encoded RNAs in Burkitt's lymphoma cell line Akata, Komano [/bib_ref] [bib_ref] Epstein-Barr virus small RNAs potentiate tumorigenicity of Burkitt lymphoma cells independently of..., Ruf [/bib_ref] [bib_ref] Isolation of Epstein-Barr virus (EBV)-negative cell clones from the EBV-positive Burkitt's lymphoma..., Shimizu [/bib_ref] [bib_ref] Malignant transformation of B lymphoma cell line BJAB by Epstein-Barr virus-encoded small..., Yamamoto [/bib_ref].
A recent study of transgenic mice that expressed EBERs in the lymphoid compartment demonstrated the development of lymphoid hyperplasia and, in some cases, lymphoma [bib_ref] Lymphoid hyperplasia and lymphoma in transgenic mice expressing the small non-coding RNA,..., Repellin [/bib_ref].
Previous studies have demonstrated that EBERs can induce the expression of various cytokines. EBERs induce human IL-10 expression in BL cells [bib_ref] Epstein-Barr virus-encoded poly(A)(-) RNA supports Burkitt's lymphoma growth through interleukin-10 induction, Kitagawa [/bib_ref] ; moreover, EBV-positive BL biopsies consistently expressed IL-10, whereas EBV-negative BL biopsies did not. Further analysis revealed that IL-10 acts as an autocrine growth factor for BL cells, suggesting that EBERs contribute to BL development via IL-10 induction [bib_ref] Epstein-Barr virus-encoded poly(A)(-) RNA supports Burkitt's lymphoma growth through interleukin-10 induction, Kitagawa [/bib_ref]. Additionally, EBERs were reported to be responsible for the induction of IL-9, which acts as an autocrine growth factor for T cell proliferation, suggesting that EBERs affect the development of EBV-associated T cell lymphoma [bib_ref] Epstein-Barr virus (EBV)-encoded RNA promotes growth of EBV-infected T cells through interleukin-9..., Yang [/bib_ref]. On the other hand, EBERs have been shown to promote the growth and proliferation of epithelial cell lines derived from NPC and GC [bib_ref] Autocrine growth of Epstein-Barr virus-positive gastric carcinoma cells mediated by an Epstein-Barr..., Iwakiri [/bib_ref] [bib_ref] Epstein-Barr virus-encoded small RNA induces insulin-like growth factor 1 and supports growth..., Iwakiri [/bib_ref] [bib_ref] Stable expression of EBERs in immortalized nasopharyngeal epithelial cells confers resistance to..., Wong [/bib_ref] [bib_ref] Oncogenic role of Epstein-Barr virus-encoded small RNAs (EBERs) in nasopharyngeal carcinoma, Yoshizaki [/bib_ref] [bib_ref] Epstein-Barr virus infection in immortalized nasopharyngeal epithelial cells: Regulation of infection and..., Tsang [/bib_ref]. Iwakiri et al. [bib_ref] Autocrine growth of Epstein-Barr virus-positive gastric carcinoma cells mediated by an Epstein-Barr..., Iwakiri [/bib_ref] [bib_ref] Epstein-Barr virus-encoded small RNA induces insulin-like growth factor 1 and supports growth..., Iwakiri [/bib_ref] demonstrated that EBERs induce IGF1, which acts as an autocrine growth factor for NPC and GC cells. Further, high levels of IGF1 expression were observed in EBV-positive but not in EBV -negative NPC or GC biopsies, suggesting that EBERs contribute to epithelial carcinogenesis via the induction of IGF1 expression [bib_ref] Autocrine growth of Epstein-Barr virus-positive gastric carcinoma cells mediated by an Epstein-Barr..., Iwakiri [/bib_ref] [bib_ref] Epstein-Barr virus-encoded small RNA induces insulin-like growth factor 1 and supports growth..., Iwakiri [/bib_ref]. A more recent study also demonstrated EBER-mediated IGF1 induction [bib_ref] Cellular gene expression that correlates with EBER expression in Epstein-Barr virus-infected lymphoblastoid..., Gregorovic [/bib_ref].
Regarding the role of EBERs in the process of EBV-induced B cell transformation, an early study reported that EBERs were not essential for primary infection, viral replication, or the B lymphocytes transformation [bib_ref] Recombinant Epstein-Barr virus with small RNA (EBER) genes deleted transforms lymphocytes and..., Swaminathan [/bib_ref]. Other study reported that the 50% transforming dose of an EBER-deleted virus was approximately 100-fold less than that of the EBER-positive EBV; this difference was due to the significantly reduced growth potential of LCLs established with an EBER-deleted virus [bib_ref] Critical role of Epstein-Barr Virus (EBV)-encoded RNA in efficient EBV-induced B-lymphocyte growth..., Yajima [/bib_ref]. Subsequently, EBER2 was found to contribute to the efficient growth transformation of B lymphocytes via the induction of IL-6, which is required for LCL growth [bib_ref] Epstein-Barr virus (EBV)-encoded RNA 2 (EBER2) but not EBER1 plays a critical..., Wu [/bib_ref]. More recently, EBER deletion was reported to have no effect on LCL transformation efficiency or growth [bib_ref] Cellular gene expression that correlates with EBER expression in Epstein-Barr virus-infected lymphoblastoid..., Gregorovic [/bib_ref]. This discrepancy might be due to differences in the EBV strain backgrounds used in these two experiments.
## Ebers-mediated pathogenesis via modulation of innate immune signals
Interactions between EBERs and host dsRNA sensors have been reported to play a significant role in EBV-mediated pathogenesis . Samanta et al. [bib_ref] EB virus-encoded RNAs are recognized by RIG-I and activate signaling to induce..., Samanta [/bib_ref] reported that EBER, which is expected to form dsRNA structures, activates RIG-mediated signaling. The results suggest that in BL cells, RIG-I is constitutively activated by EBERs, leading to the activation of NF-κB and IRF-3, and subsequent induction of type-I IFN. Although IFN induction appears to be disadvantageous for the virus, EBV can maintain a latent infection state because of resistance to IFN, such as that provided by EBER-mediated PKR inhibition [bib_ref] Epstein-Barr virus RNA confers resistance to interferon-alpha-induced apoptosis in Burkitt's lymphoma, Nanbo [/bib_ref]. Furthermore, a subsequent study demonstrated the significant role EBER-induced RIG-I activation in oncogenesis. EBER promotes the BL cell growth by inducing expression of anti-inflammatory and growth-promoting cytokine IL-10, in a manner dependent on RIG-I-mediated IRF3 signaling but independent of NF-κB [bib_ref] Epstein-Barr virus-encoded small RNA induces IL-10 through RIG-I-mediated IRF-3 signaling, Samanta [/bib_ref]. . EBER-mediated modulation of innate immune signaling contributes to EBV-mediated pathogenesis. Reproduced from Iwakiri et al. [bib_ref] Role of EBERs in the pathogenesis of EBV infection, Iwakiri [/bib_ref]. Left, in BL cells, EBERs are recognized by RIG-I via the RNA helicase domain of RIG-I and following recognition, RIG-I associates with the adaptor IPS-1 via its CARD. IPS-1 initiates signaling leading to the activation of IRF3 and NF-κB to induce type I IFNs and inflammatory cytokine expression. EBERs induce the expression of the growth-promoting cytokine IL-10 via RIG-I-mediated IRF3 activation and might support BL development. EBERs also bind to IFN-inducible PKR and block its activity, which is required for the IFN-mediated antiviral effect; therefore, EBV might maintain a latent infection state. Right, activation of innate immunity via TLR3 signaling in response to secreted EBER. During an active EBV-infection, EBER1 is released from EBV-infected lymphocytes primarily in a complex with La. Circulating EBER induces DC maturation via TLR3 signaling and induces type I IFN and inflammatory cytokine production by activating IRF3 and NF-κB. DC activation leads to T cell activation and systemic cytokine release. Furthermore, TLR3-expressing T and NK cells including EBV-infected T or NK cells could be activated by EBER1 through TLR3, thus leading to inflammatory cytokine production. Therefore, immunopathologic diseases caused by active EBV infections including T or NK cell activaiton and hypercytokinemia, could be attributed to EBER1-induced TLR3-mediated T cell activation and cytokinemia. EBER, Epstein-Barr virus encoded RNA; EBV, Epstein-Barr virus; BL, Burkitt's lymphoma; IPS-1, interferon-β promoter stimulator-1; CARD, caspase recruitment domain; RIG-I, retinoic acid-inducible gene I; DC, dendritic cell; IFN, interferon; NK cell, natural killer cell; IL, interleukin; TLR, Toll-like receptor; PKR, RNA-dependent protein kinase; IRF 3, interferon regulatory factor 3.
Toll-like receptors (TLRs) comprise a distinct family of PRRs that sense virus-derived nucleic acids and trigger antiviral innate immune responses by activating signaling cascades via Toll/IL-1 receptor (TIR) domain-containing adaptors [bib_ref] Toll-like receptor signaling, Akira [/bib_ref]. The role of TLR3 in the dsRNA recognition was demonstrated in a study of TLR3-deficient mice [bib_ref] Recognition of double-stranded RNA and activation of NF-kappaB by Toll-like receptor 3, Alexopoulou [/bib_ref]. dsRNA-induced signal transduction via TLR3 leads to the recruitment of TIR domain-containing adaptor inducing IFN-β (TRIF) and the subsequent phosphorylation of downstream molecules such as IRF3 and NF-κB [bib_ref] Toll-like receptors and RNA helicases: Two parallel ways to trigger antiviral responses, Meylan [/bib_ref]. Iwakiri et al. [bib_ref] Epstein-Barr virus (EBV)-encoded small RNA is released from EBV-infected cells and activates..., Iwakiri [/bib_ref] demonstrated that EBERs are released extracellular environment and are recognized by TLR3, leading to the induction of type I IFN and inflammatory cytokines. The majority of the released EBER1 exists as a complex with La, suggesting that EBER1 is released from the cells via the active secretion of La. EBV has been known to cause active infectious diseases such as infectious mononucleosis (IM), chronic active EBV infection (CAEBV) and EBV-associated hemophagocytic lymphohistiocytosis (EBV-HLH). IM is characterized by the expansion of reactive T-cells and is most likely an immunopathologic disease, in which the general symptoms are caused by inflammatory cytokines [bib_ref] Epstein-Barr virus, Rickinson [/bib_ref]. CAEBV and EBV-HLH are also active EBV infections with persistent or recurrent IM-like symptoms. EBV-HLH is characterized by an EBV infection in CD4-positive T cells or naural killer (NK) cells and the systemic release of inflammatory cytokines, leading to blood cell hemophagocytosis via the activation of macrophages [bib_ref] Epstein-Barr virus, Rickinson [/bib_ref] [bib_ref] Fatal Epstein-Barr virus-associated hemophagocytic syndrome, Kikuta [/bib_ref] [bib_ref] Differential cellular targets of Epstein-Barr virus (EBV) infection between acute EBV-associated hemophagocytic..., Kasahara [/bib_ref]. On the other hand, in CAEBV, CD8 + -T cells are primary EBV infection targets [bib_ref] Differential cellular targets of Epstein-Barr virus (EBV) infection between acute EBV-associated hemophagocytic..., Kasahara [/bib_ref]. Iwakiri et al. [bib_ref] Epstein-Barr virus (EBV)-encoded small RNA is released from EBV-infected cells and activates..., Iwakiri [/bib_ref] demonstrated that sera from patients with IM, CAEBV and EBV-HLH contained EBER1. A further analysis revealed that serum EBER1 activates TLR3 signaling in immune cells, including dendritic cells, suggesting that EBER1, which is released from EBV-infected cells is responsible for EBV-mediated immune activation, and induction of type I IFN and inflammatory cytokines [bib_ref] Epstein-Barr virus (EBV)-encoded small RNA is released from EBV-infected cells and activates..., Iwakiri [/bib_ref]. Because CD8 + -T cells and NK cells express TLR3 and can be activated by TLR3 signaling [bib_ref] Human effector CD8 + T lymphocytes express TLR3 as a functional coreceptor, Tabiasco [/bib_ref] [bib_ref] APC-independent activation of NK cells by the Toll-like receptor 3 agonist double-stranded..., Schmidt [/bib_ref] , TLR3-expressing T and NK cells could potentially be activated by EBER1 through TLR3 to produce inflammatory cytokines. Therefore, EBER1-induced activation of innate immunity would account for the immunopathologic diseases caused by active EBV infection. A more recent study of a humanized mice model of fetal EBV-infectious disease observed the EBER1 in the serum, thus suggesting that EBER1 contribute to the disease pathology [bib_ref] A novel animal model of Epstein-Barr virus-associated hemophagocytic lymphohistiocytosis in humanized mice, Sato [/bib_ref]. In summary, EBERs contribute to EBV infection-related pathogenesis, including cancer and active infectious diseases, through interactions with RIG-I and TLR3 [bib_ref] Role of EBERs in the pathogenesis of EBV infection, Iwakiri [/bib_ref].
## Conclusions and future direction
Recently, the bioactive extracellular small membrane vesicles also known as exosomes were reported to contribute to cancer development. Exosomes has been shown to contain miRNA, DNA, proteins, and even virus particles [bib_ref] Extracellular vesicles: Exosomes, microvesicles, and friends, Raposo [/bib_ref] [bib_ref] The multi-faceted exosome: Biogenesis, role in normal and aberrant cellular function, and..., Pant [/bib_ref]. Additionally, EBV-infected cells were also reported to produce exosomes that contained viral proteins and miRNA [bib_ref] Functional delivery of viral miRNAs via exosomes, Pegtel [/bib_ref]. Previous studies have demonstrated that the La protein can be excreted in exosomes, thus supporting the possibility that EBER is also secreted and transferred to neighboring cells via exosomes [bib_ref] Salivary gland epithelial cell exosomes: A source of autoantigenic ribonucleoproteins, Kapsogeorgou [/bib_ref]. The interactions between EBERs and cellular factors are involved in EBV-mediated pathogenesis, including cancer. A further study of the detailed molecular mechanism is required in order to develop a new therapeutic approach that targets these ncRNAs in EBV-associated cancers.
[fig] Figure 1: Secondary structures of EBERs. Reproduced from Rosa et al. [/fig]
|
Sequence Inversion to Facilitate Concurrent Radiotherapy and Systemic Therapy. A Proof of Principle Study in the Setting of a Phase II Randomized Trial in Prostate Cancer
Background: Concomitant chemo-radiation for pelvic cancers remains challenging to be delivered at full doses. We hypothesized that fewer delays in chemotherapy would occur if the sequence of radiotherapy would be reversed, starting with the boost volume followed by the elective nodal volume. We report the result of a Phase II randomized study for high risk prostate cancer.Patients and Method: The study was a double-blinded phase II randomized trial. Patients were eligible if they had non-metastatic high-risk prostate cancer. All patients received 2.5 years of hormonal therapy and 46.5 Gy in 25 fractions to the pelvic lymph nodes. Patients received a radiation boost to the prostate, either before or after whole pelvic irradiation. Concurrent (20 mg/m 2 ) Docetaxel was given on the first day of radiotherapy and weekly thereafter for a total of eight treatments until predefined toxicity stopping rules.Results: Ninety patients were included and randomized. Four were ineligible for the analysis. In total, 42 patients were randomized to the standard sequence, 44 patients to the experimental sequence. There were statistically fewer GI or GU toxicities leading to a docetaxel dose reduction or omission in the experimental sequence compared to the standard sequence, 5 vs. 15 events (p = 0.027). There was no difference in overall survival, cause-specific survival, or biochemical-relapse free survival between the two sequences.Conclusions:This is the first study to test sequence inversion for pelvic radio-chemotherapy in a randomized double-blind trial. Less chemotherapy interruptions or dose reductions occurred by inverting the radiation sequence of the large field and the boost.The trial was registered with Clinicaltrials.gov: NCT00452556
# Background
Chemo-radiation is a standard treatment for several pelvic malignancies, including those arising from the cervix, rectum, and anal canal, and has also been tested for high risk prostate cancer (1-6). When regional nodes are included in the radiotherapy field, and systemic therapy itself has bowel or urinary toxicity, often there are delays or dose-reductions, in one or both modalities, and either treatment can be interrupted or even discontinued. However, by tradition, the radiotherapy treatment consists of sequentially large fields treated at low doses, followed with a boost on the gross target volume. It is unknown if the order of treatment between boost and large field may have an impact on treatment tolerance and could enable fewer delays or dose-reductions of chemotherapy, by postponing the time when patients will present with significant bowel side effects.
In developed countries, prostate cancer is the most frequent non-cutaneous cancer in men. Twenty percent are classified as high risk, meaning they have a 30-50% risk of nodal involvementand a 20-30% chance of microscopic distant metastases. For those high-risk patients, treatment options include radiotherapy, usually combined with hormonal therapy, and less frequently surgery in some highly selected patients. Chemotherapy and concomitant chemo-radiation, which could target both the loco-regional disease and the distant micrometastasis remains experimental. Among various systemic therapies used in prostate cancer, Docetaxel is a radio-sensitizer with activity against prostate cancer. However, Docetaxel is also known to cause gastrointestinal toxicity, so there is concern that when delivered concomitantly to radiotherapy this would lead to excessive, dose-limiting toxicity. In a study on 22 patients, Kumar et al.showed that when delivered concomitantly, the full chemotherapy regimen could only be given in 50% of patients largely because of an excess of gastrointestinal toxicities.
We hypothesized that there would be fewer dose reductions or delays in docetaxel chemotherapy if the sequence of radiotherapy would be reversed, starting with the boost volume followed by the large elective nodal volume. This is a proof of principle study of sequence inversion, using prostate cancer as an example. In this manuscript, we report the result of a phase II randomized study of concomitant chemoradiation in high risk prostate cancer patients comparing the standard sequence treating the large volume followed by the boost volume, with an experimental sequence delivering the boost dose before the loco-regional treatment. Outcomes include the number of dose reductions and/or delays in docetaxel, as a result of gastrointestinal or genitourinary toxicity, survival, and health-related quality of life measured by the Expanded Prostate Cancer Index Composite (EPIC).
# Materials and methods
## Eligible and ineligible patients
Patients were eligible if they had a high risk prostate cancer, defined in the current trial either as untreated patients with a 2003 TNM clinical stage T2c, T3a, or T3b, or a Gleason score 8 to 10, or a PSA ≥ 20 µg/L but less than 50 µg/L (16), a life expectancy of at least 5 years, and an ECOG (Eastern cooperative oncology group) performance status of 0 or 1. Also, patients who had radical prostatectomy (RP) were offered to receive regional radiation as part of the study if they had more than a 50% chance of biochemical recurrence following the Kattan et al.Nomogram. Post radical prostatectomy patients had to have a post-operative PSA of < 1.0 µg/L and be able to start the study protocol within 6 months from surgery. In all cases, patients must have had no evidence of metastatic disease after screening bone scan, chest X-ray, and CT scan of the abdomen and pelvis, and adequate end organ function in terms of bone marrow, liver, and kidneys. Patients were excluded if they had a PSA > 50 µg/L, prior pelvic radiotherapy, grade ≥ 2 peripheral neuropathy, prior malignancy, or known hepatitis B or C.
## Study design
The study was a single center double-blinded phase II randomized design. Patients were approached for the study during the initial consultation by the study co-ordinator, and after informed consent was obtained, they were referred for medical oncology consultation. When deemed eligible for combined chemo-radiation treatment, patients were randomized in a 1:1 ratio to treatment sequence at the time of registration into the study. Randomization was performed by a computer algorithm, using SAS software version 9.4 (Cary, NC), using the permuted block design, using block sizes of four and six patients. To ensure blinded assessment of toxicities, the attending radiation oncologist completed the delineation of the target volumes and organs at risk and approved the final plans for each phase. All subsequent quality assurance, including verification dosimetry and daily image guidance, was reviewed by an independent radiation oncologist involved in the study, but not in the patient's treatment nor the assessment of toxicity. The attending physician was responsible for assessing and scoring toxicity at the time of weekly review within the hospital and was blinded to patient treatment sequence. Prior to study initiation, approval was obtained from the institutional ethics review board, and the trial was registered with ClinicalTrials.gov.
## Treatment protocol: chemotherapy and radiation prescriptions
Patients underwent radiation simulation following institutional guidelines. In brief, patients were immobilized supine in a Vak-Loc (CIVCO, Coralville, Iowa) and CT simulation. A planning MRI was performed in the treatment position using a 1.5 T magnet without an endorectal coil. All patients, including those having had radical prostatectomy, had insertion of gold fiducial markers in the prostate or prostate bed. The clinical target volume (CTV) for the larger volume which included the pelvic nodes included the external and internal iliac vessels plus 7 mm except where vessels were in direct abutment to bone or muscle, where the CTV included only the vessel. The CTV for the larger volume included the nodes and either the prostate plus 3 mm, except at the prostate-rectal interface, or the prostate bed, from the bottom of the anastomosis, superiorly, to the inferior aspect of the proximal vas deferens. The CTV for the smaller boost volume included only the prostate bed, or the prostate plus any extraprostatic extension, if present. The planning target volume (PTV) was equal to the CTV plus 7 mm. The dose fractionation was intended to be biologically equivalent to 70 Gy in 35 fractions, with the elective nodal dose equivalent to 46 Gy in 23 fractions. All patients received 46.5 Gy in 25 fractions to the pelvic lymph nodes. Patients who had not had previous radical prostatectomy received a boost, either before, or after whole pelvic irradiation, to the prostate and to the seminal vesicles for T3b tumors to a dose of 26.78 Gy in 13 fractions. The boost dose was reduced to 20.6 Gy in 10 fractions, if patients had previous radical prostatectomy. There were no specified dose constraints to organs at risk. Plans were approved if 95% of the dose covered 98% of the PTV volume. All patients were treated with static port intensity-modulated radiotherapy (IMRT), using either five or seven ports, with six MV photons. All radiation plans were verified by Medical Physics prior to treatment as per institutional policy. Daily image guidance was performed with on-board kV imaging, matching to the fiducial markers.
All patients received hormonal therapy with leuprolide acetate 45 mg subcutaneously every 6 months. This was given 4 months prior to starting concurrent chemoradiation and continued for two years post-treatment. In addition, patients received 4 weeks of daily bicalutamide 50 mg at the time of the first leuprolide acetate administration. Concurrent Docetaxel was given on the first day of IMRT (week 16 of protocol therapy) and weekly thereafter for a total of eight treatments at a dose of 20 mg/m 2 over 30 min. Chemotherapy was withheld for grade 3 or greater: diarrhea, thrombocytopenia, absolute neutrophil count (ANC) < 500× 10 9 /L, febrile neutropenia with ANC < 1.0 × 10 9 /L, nausea, and vomiting, stomatitis, grade 2 peripheral neuropathy, or abnormal liver function tests. When acceptable toxicity was reached, docetaxel was restarted at 16 mg/m 2 . If grade 3 toxicity recurred, or if docetaxel was delayed more than 2 weeks, docetaxel was discontinued.
## Outcomes
The primary endpoint of the study was the comparison between sequences of the patients' proportion experiencing Docetaxel dose reductions or omissions due to gastrointestinal or genitourinary toxicity. The number of dose reductions and/or delays was compared using Poisson regression. Secondary endpoints included the time to selected grade 2 and 3 NCI CTCAE version 3.0 toxicity, the difference in incidence of grade 2 and 3 toxicity, as well as the difference in overall bowel domain score of the EPIC at weeks 16, 20, and 24. Differences in rates of toxicity were compared using Poisson regression, when the modeling fit the data, and when the assumptions of Poisson regression were not met, the proportions of toxicity between sequences were calculated using the Chi-squared test. Differences in Overall survival and time to selected grades of toxicity were calculated using the Log Rank test, and differences in biochemical relapse-free survival and prostate cancer related mortality were calculated using competing risk proportional hazards modeling. The date of biochemical failure in the patients with no prior surgery occurred at the time the PSA reached a value of the PSA nadir +2, and the date of biochemical failure in the patients who had had a prior radical prostatectomy occurred at a PSA of 0.2. Statistical analyses were conducted using SAS software version 9.4 (Cary, NC).
## Sample size
Based on the Kumar et al.study, we assumed 50% of patients receiving the standard sequence would require docetaxel dose reductions or delays and 20% of patients in the experimental sequence would require dose reductions or delays. With a type 1 error rate of 0.05 and a power of 80%, 39 patients per sequence would be required. Assuming a 10% rate of withdrawal or discontinuation, 43 patients per sequence were calculated. Early trial stopping rules dictated that if there was a greater than 60% grade 2 or more gastrointestinal or genitourinary toxicity in the first 10 patients in either sequence, the dose of docetaxel would be reduced to 16 mg/m 2 weekly in all subsequently treated patients, and if more than 30% of patients subsequently experienced grade 3 toxicity in the next 10 patients treated after the dose reduction, the study would have been discontinued.demonstrates that the two treatment sequences were well-balanced for baseline characteristics. In total 90 patients were registered and randomized. Four were deemed ineligible for the analysis, including three who withdrew consent, and one found to have an invasive bladder cancer on the
# Results
## Patients characteristics
## Chemo-radiation delivery and toxicity
All patients received radiotherapy, as per protocol, without breaks or dose reductions and the vast majority were able to receive the full dose of chemotherapy, There were significantly fewer GI or GU toxicities leading to a docetaxel dose reduction or omission in the experimental sequence compared to the standard sequence, 5 vs. 15 events (p = 0.027), using Poisson regression. Secondary study endpoints included the total amount of Docetaxel that can be delivered. In the standard sequence 78.6% of patients received 8 weeks of chemotherapy compared to 81.8% in the experimental sequence (p = 0.88), and 76.2% of patients did not require a docetaxel dose reduction compared to 77.2% in the experimental sequence (p = 0.88). Goodness of fit testing indicated that the Poisson regression fit the data well (p = 0.44) for the primary endpoint of the study. Secondary endpoints also included individual GI and GU toxicities, selected ones of which are listed in. There were significantly more cumulative GI grade 2 and 3 toxicities in the experimental sequence, 91%, compared to the standard sequence, 69% (p = 0.0109). The rates of Grade 2 or higher diarrhea corresponding to 3 or more bowel movements per day above baseline were similar with the standard sequence, 50%, compared to 52.2% in the experimental sequence (p = 0.83). Conversely, there were more combined grade 2 and 3 gastrointestinal toxicities in the experimental sequence, with a non-statistically significant trend to more proctitis in the experiment sequence, 22.7 vs. 11.9% (p = 0.186). Constipation was the only statistically significant individual item (p = 0.045). There was a statistically significant delay in time to grade 1 diarrhea in the experimental sequence (p = 0.04). There was also a non-statistically significant trend to more combined grade 2 and 3 urinary frequency corresponding to 2 times increase of the normal voiding frequency, 70.5 vs. 50% (p = 0.0525).
It is important to note that bone marrow toxicity was infrequent in both sequences, with no patients experiencing febrile neutropenia (see. One patient in the experimental sequence had a lower GI bleed, requiring transfusion, but this resolved without further intervention. There were no treatment related deaths.
## Survival
There was no difference in overall survival (Log-Rank p = 0.66), cause-specific survival (Gray's Test p = 0.47), or biochemical recurrence (Gray's Test p = 0.98; seebetween the two sequences. However, overall survival is encouraging with 90% of patients alive at 8 years, and only 22.5 and 21.0% of patients demonstrating biochemical recurrence at 8 years, in the experimental sequence and standard sequences, respectively.
## Heath-related quality of life
During the period of concurrent chemo-radiation (week 16-24 of protocol therapy) there was statistically significant declines in all domains of the EPIC, except on the Mental Component score of the SF12. For this domain there was a statistically decline in the standard sequence only using repeated measures analysis of variance (p = 0.01). For all the other domains, there was no difference between sequences in terms of score decline for the bowel, urinary, sexual, hormonal domains, nor in the AUA symptom scores, or Physical Component score.
# Discussion
This is the first proof of principle study, to test the inversion of the sequence of loco-regional radiation followed by the boost as the standard sequence or the reverse as the experimental sequence, in a randomized, double-blind trial. The study found that there were fewer dose-reductions or delays due to gastrointestinal or genitourinary toxicity when the radiotherapy started with the boost phase followed by the large loco-regional phase that includes the nodal volume. This could potentially enable more radio-sensitization of prostate cancer cells by docetaxel, and our study demonstrates that by inverting the order of the radiation sequences there was no detriment in cancer control to addressing only the grossly apparent tumor first. There was more grade 2 toxicity, mostly grade 2 urinary frequency, in the experimental sequence arm (70.4% vs. 50% combined grade 2 and 3 urinary frequency, p = 0.053, with 9.09% (four events in four patients) vs. 2.38% (one event in one patient) grade 3 urinary frequency, p = 0.18), but this excess was not sufficient to fit the pre-specified criteria to warrant chemotherapy dose reductions or delays, and resulted in 15 events resulting in dose reductions or delays in the standard sequence vs. five events in the experimental sequence, due to GU or GI toxicity. In essence, slightly more toxicity was seen in the experimental sequence, as a percentage of patient affected, but it didn't translate into more events that required reduction or omission of the systemic therapy, in the experiment arm. In fact, they had less events, and perhaps more bother, although this was not detected by bother as measured by the EPIC.
Clearly, however, docetaxel is out of favor to be combined with radiotherapy in a concurrent fashion, and is not recommended.
Similar sequence inversion of the boost and whole pelvic irradiation for concomitant chemo-radiation of prostate cancer with paclitaxel had been reported by Sanfilippo et al. (18), but this was not explored in a randomized study. Patients enrolled at the beginning of the study received the traditional sequence starting with whole pelvic lymphatics and prostate treated followed by a boost. Due to GI toxicity, the patients accrued later in the trial received inverted radiotherapy sequences starting with the prostate boost and treating the pelvic lymphatics last. The study reported a decrease in the incidence of grade 3 toxicity, however, a formal analysis of the toxicity rates before and after the sequencing switch was not provided. The patients in the Sanfilippo study also received biweekly paclitaxel and 9 months of androgen deprivation. They reported a 3 years biochemicalfree relapse rate of 74%, using the Phoenix definition, and an 18% rate of grade 3 diarrhea, which is very similar to the 20.4% rate of grade 3 diarrhea in our study experimental sequence. Conversely, to Sanfilippo study we were able to escalate the loco-regional radiation dose to 46.5 Gy in 25 fractions with a similar rate of toxicity utilizing IMRT.
The strength in the present study is its randomized nature, blinding of participants and investigators, which limits potential bias. One other important aspect is that the incidence of grade 2 or higher bone marrow toxicity was less than 10%, without febrile neutropenia, likely owing to the weekly docetaxel regimen compared to docetaxel given every 3 weeks, when marrow suppression is more pronounced. Ideally, the optimal systemic agent would be radio-sensitizing only to tumor cells, would have a high degree of independent anti-tumor effect, and itself would not cause treatment-related diarrhea, or urinary toxicity.
What is the meaning for other cancer sites? The same sequence inversion strategy could be tested for anal canal, cervix, or rectal cancer or used as option when patients present with significant co-morbidities presenting a challenge to protocol completion.
In conclusion, our study demonstrates the proof of principle that sequence inversion of the large and boost volumes results in fewer dose deductions or delays in systemic therapy when there is overlapping normal tissue toxicity between the two. There was no detriment in cancer control to addressing only the grossly apparent tumor first. While the study will not turns the heels of radiotherapy on its head, it does provide scientific proof that in special circumstances it may offer an approach of how to optimize combined modality therapy, with radiotherapy and a systemic agent, when there is overlapping toxicity.
# Data availability statement
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.
# Ethics statement
The studies involving human participants were reviewed and approved by Nova Scotia Health Authority, Research Ethics Board. The patients/participants provided their written informed consent to participate in this study.
# Author contributions
DW contributed to conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, resources, software, supervision, validation, visualization, writing-original draft, and writingreview and editing. LW contributed to conceptualization, data curation, funding acquisition, investigation, methodology, project administration, resources, supervision, validation, visualization, and writing-review and editing. SC contributed to data curation, investigation, project administration, and writing-review and editing. RR contributed to conceptualization, data curation, investigation, methodology, project administration, visualization, and writing-review and editing. HH contributed to conceptualization, data curation, investigation, methodology, project administration, and writing-review and editing. DB contributed to data curation, investigation, project administration, and writing-review and editing. NP contributed to data curation, investigation, project administration, and writing-review and editing. CA contributed to conceptualization, data curation, funding acquisition, investigation, methodology, project administration, resources, supervision, and writing-review and editing. J-PP contributed to formal analysis, investigation, methodology, project administration, resources, supervision, visualization, writing-original draft, and writing-review and editing. |
Trapping Conformational States Along Ligand-Binding Dynamics of Peptide Deformylase: The Impact of Induced Fit on Enzyme Catalysis
For several decades, molecular recognition has been considered one of the most fundamental processes in biochemistry. For enzymes, substrate binding is often coupled to conformational changes that alter the local environment of the active site to align the reactive groups for efficient catalysis and to reach the transition state. Adaptive substrate recognition is a well-known concept; however, it has been poorly characterized at a structural level because of its dynamic nature. Here, we provide a detailed mechanism for an induced-fit process at atomic resolution. We take advantage of a slow, tight binding inhibitor-enzyme system, actinonin-peptide deformylase. Crystal structures of the initial open state and final closed state were solved, as well as those of several intermediate mimics captured during the process. Ligand-induced reshaping of a hydrophobic pocket drives closure of the active site, which is finally ''zipped up'' by additional binding interactions. Together with biochemical analyses, these data allow a coherent reconstruction of the sequence of events leading from the encounter complex to the key-lock binding state of the enzyme. A ''movie'' that reconstructs this entire process can be further extrapolated to catalysis.
# Introduction
Flexibility of proteins around their active site is a central feature of molecular biochemistry [bib_ref] Enzyme catalysis: not different, just better, Knowles [/bib_ref] [bib_ref] Multiple conformational changes in enzyme catalysis, Hammes [/bib_ref] [bib_ref] A perspective on enzyme catalysis, Benkovic [/bib_ref] [bib_ref] Dynamic personalities of proteins, Henzler-Wildman [/bib_ref] [bib_ref] Functional aspects of protein flexibility, Teilum [/bib_ref]. Although this has been a central concept in biochemistry for half a century, the detailed mechanisms describing how the active enzyme conformation is achieved have remained largely elusive, as a consequence of their transient nature. Direct structural evidence and/or kinetic analyses have only recently emerged [bib_ref] Enzymes with lid-gated active sites must operate by an induced fit mechanism..., Sullivan [/bib_ref] [bib_ref] Selected-fit versus induced-fit protein binding: kinetic differences and mutational analysis, Weikl [/bib_ref] [bib_ref] Role of induced fit in enzyme specificity: a molecular forward/reverse switch, Johnson [/bib_ref] [bib_ref] Advances in kinetic protein crystallography, Bourgeois [/bib_ref] [bib_ref] Ramanassisted crystallography reveals end-on peroxide intermediates in a nonheme iron enzyme, Katona [/bib_ref]. Three classic ''textbook'' models are used to describe the formation of the ligand-enzyme complex: (i) the Fischer's ''lock-and key'' model, (ii) the Koshland's induced-fit model, and (iii) the selected-shift model or conformational selection mechanism [bib_ref] Enzymes with lid-gated active sites must operate by an induced fit mechanism..., Sullivan [/bib_ref] [bib_ref] Selected-fit versus induced-fit protein binding: kinetic differences and mutational analysis, Weikl [/bib_ref] [bib_ref] Role of induced fit in enzyme specificity: a molecular forward/reverse switch, Johnson [/bib_ref] [bib_ref] Application of a theory of enzyme specificity to protein synthesis, Koshland [/bib_ref] [bib_ref] Residence time of receptor-ligand complexes and its effect on biological function, Tummino [/bib_ref] [bib_ref] The role of dynamic conformational ensembles in biomolecular recognition, Boehr [/bib_ref]. In the Fischer's ''lock-and key'' model, the conformations of free and ligand-bound proteins are essentially the same. In the induced-fit model, ligand binding induces a conformational change in the protein, leading to the precise orientation of the catalytic groups and implying the existence of initial molecular matches that provide sufficient affinity prior to conformational adaptation [bib_ref] Molecular recognition by induced fit: how fit is the concept?, Bosshard [/bib_ref]. In contrast, the selected-fit model assumes an equilibrium between multiple conformational states, in which the ligand is able to select and stabilize a complementary protein conformation. In this case, the conformational change precedes ligand binding, in contrast to the induced-fit model in which binding occurs first. The conformational selection and/or induced-fit processes have been shown to be involved in a number of enzymes [bib_ref] Residence time of receptor-ligand complexes and its effect on biological function, Tummino [/bib_ref] [bib_ref] The role of dynamic conformational ensembles in biomolecular recognition, Boehr [/bib_ref] [bib_ref] Free-energy landscape of enzyme catalysis, Benkovic [/bib_ref] [bib_ref] Conformational selection or induced fit: a flux description of reaction mechanism, Hammes [/bib_ref]. For several of these studies, conformational selection is proposed because the experimental data support that, even in the absence of the ligand, the enzyme samples multiple conformational states, including the ligand-bound (active) state [bib_ref] Enzymes with lid-gated active sites must operate by an induced fit mechanism..., Sullivan [/bib_ref]. Although direct structural evidence and/or kinetic analyses have provided clues [bib_ref] Enzymes with lid-gated active sites must operate by an induced fit mechanism..., Sullivan [/bib_ref] [bib_ref] Selected-fit versus induced-fit protein binding: kinetic differences and mutational analysis, Weikl [/bib_ref] [bib_ref] Role of induced fit in enzyme specificity: a molecular forward/reverse switch, Johnson [/bib_ref] [bib_ref] Residence time of receptor-ligand complexes and its effect on biological function, Tummino [/bib_ref] [bib_ref] The role of dynamic conformational ensembles in biomolecular recognition, Boehr [/bib_ref] [bib_ref] Conformational selection or induced fit: a flux description of reaction mechanism, Hammes [/bib_ref] , how we can distinguish whether a protein binds its ligand in an induced-or selected-fit mechanism remains critical and often controversial.
The enzyme-inhibitor interaction is a form of molecular recognition that is more amenable to investigation than the enzyme-substrate interaction as there is no chemical transformation of the ligand during this process. In this context, slow, tightbinding inhibition is an interesting interaction process, as it closely mimics the substrate recognition process and has been shown to be commonly involved in adaptive conformational changes [bib_ref] Residence time of receptor-ligand complexes and its effect on biological function, Tummino [/bib_ref] [bib_ref] The behavior and significance of slow-binding enzyme inhibitors, Morrison [/bib_ref]. In slow, tight-binding inhibition, the degree of inhibition at a fixed concentration of compound varies over time, leading to a curvature of the reaction progress curve over time during which the uninhibited reaction progress curve is linear. Indeed, the slow, tight-binding inhibition is a two-step mechanism that depends on the rate and strength of inhibitor interactions with the enzyme. Binding of the inhibitor (I) to the enzyme (E) leads to the rapid formation of a non-covalent enzyme-inhibitor complex (E:I) followed by monomolecular slower step (k 5 ) in which the E:I is transformed into a more stable complex (E:I*) that relaxes and dissociates at a very slow rate, mainly inferred by the k 6 value when k 6 ,,k 5 ,,k 4 , [fig_ref] Figure 1: Slow, tight-binding inhibition of PDF by actinonin induces conformational change in the... [/fig_ref] ; see also footnote f in [fig_ref] Table 1: Comparison of the main kinetic and thermodynamic parameters describing the inhibition of... [/fig_ref].
Although only a few studies have investigated the mechanisms of slow, tight-binding inhibitors, such molecules are favored for use as therapeutics, as they usually exhibit unique inhibitory properties, including selective potency and long-lasting effects [bib_ref] Slow-tight binding inhibition of xylanase by an aspartic protease inhibitor: kinetic parameters..., Dash [/bib_ref] [bib_ref] The crystal structure of E.coli 1-deoxy-D-xylulose-5-phosphate reductoisomerase in a ternary complex with..., Sweeney [/bib_ref] [bib_ref] Pimelic diphenylamide 106 is a slow, tight-binding inhibitor of class I histone..., Chou [/bib_ref] [bib_ref] Structure of the deacetylase LpxC bound to the antibiotic CHIR-090: Time-dependent inhibition..., Barb [/bib_ref] [bib_ref] Structure-activity relationships among the nitrogen containing bisphosphonates in clinical use and other..., Dunford [/bib_ref] [bib_ref] Slow-onset inhibition of fumarylacetoacetate hydrolase by phosphinate mimics of the tetrahedral intermediate:..., Bateman [/bib_ref] [bib_ref] Drug-target residence time and its implications for lead optimization, Copeland [/bib_ref]. Here, we explore the precise structural inhibitory mechanism of actinonin [fig_ref] Figure 1: Slow, tight-binding inhibition of PDF by actinonin induces conformational change in the... [/fig_ref] ; [bib_ref] Actinonin: an antibiotic substance produced by an actinomycete, Gordon [/bib_ref] , which is a slow, tightbinding inhibitor of peptide deformylase (PDF), a metal cationdependent enzyme [bib_ref] Actinonin, a naturally occurring antibacterial agent, is a potent deformylase inhibitor, Chen [/bib_ref] [bib_ref] Mechanism of time-dependent inhibition of polypeptide deformylase by actinonin, Van Aller [/bib_ref]. The function of the active-site metal is to activate the reactive water molecule involved in peptide hydrolysis [bib_ref] Characterization of cobalt(II)-substituted peptide deformylase: function of the metal ion and the..., Rajagopalan [/bib_ref]. PDF is the first enzyme in the N-terminal methionine excision pathway, an essential and ubiquitous process that contributes to the diversity of N-terminal amino acids [bib_ref] Molecular recognition governing the initiation of translation in Escherichia coli. A review, Schmitt [/bib_ref] [bib_ref] Cotranslational processing mechanisms: towards a dynamic 3D model, Giglione [/bib_ref]. Actinonin is a natural product with antibiotic activity that inhibits PDF by mimicking the structure of its natural substrates (nascent peptide chains starting with Fo-Met-Aaa, where Fo is a formyl group and Aaa is any amino acid) in their transition state [fig_ref] Figure 1: Slow, tight-binding inhibition of PDF by actinonin induces conformational change in the... [/fig_ref]. The transition state inhibitor actinonin, as well as other structurally related inhibitors, has been shown to systematically exhibit a ''slow-binding'' inhibition behavior [fig_ref] Figure 1: Slow, tight-binding inhibition of PDF by actinonin induces conformational change in the... [/fig_ref] , regardless of the organism of origin of the PDF [bib_ref] Mechanism of time-dependent inhibition of polypeptide deformylase by actinonin, Van Aller [/bib_ref] [bib_ref] Slow-binding inhibition of peptide deformylase by cyclic peptidomimetics as revealed by a..., Nguyen [/bib_ref].
Using structural, biocomputing, and enzymatic analyses, we were able to (i) reveal that the free enzyme is in an open conformation and that actinonin induces transition of the enzyme into a closed conformation; (ii) show that there is no evidence for the occurrence of a closed conformation in the apostructure of the open enzyme, which, together with detailed kinetic analyses, makes the closed form fully compatible with an induced-fit model; and (iii) identify the sequence of molecular events leading to the final, bound, closed complex (E:I*). Moreover, using several rationally designed point mutants of the enzyme, ligand-induced intermediates, which mimic conformational states that normally would not be expected to accumulate with the wild-type (WT) enzyme, were trapped. These conformations recapitulate physical states that the WT enzyme must pass through during its overall transition from the apo-enzyme to the E:I* complex. ''Freezing'' of ligand-induced intermediate states provides direct evidence for an induced-fit mechanism and allows the reconstruction of a virtual ''movie'' that recapitulates this mechanism. Since PDF is one example of an enzyme remaining active in the crystalline state and because actinonin closely mimics the natural substrates bound to PDF in the transition state as shown previously with the Escherichia coli form (EcPDF; see [fig_ref] Figure 1: Slow, tight-binding inhibition of PDF by actinonin induces conformational change in the... [/fig_ref] [bib_ref] Iron center, substrate recognition and mechanism of peptide deformylase, Becker [/bib_ref] [bib_ref] The crystal structures of four peptide deformylases bound to the antibiotic actinonin..., Guilloteau [/bib_ref] , we propose a model suggesting that induced fit also contributes to efficient catalysis.
# Results
## Slow, tight binding of the transition-state analog actinonin to peptide deformylase
In the present study, at the atomic level we explored the precise inhibitory mechanism of actinonin on Arabidopsis thaliana PDF1B (AtPDF), a close eukaryotic homologue of EcPDF [fig_ref] Figure 1: Slow, tight-binding inhibition of PDF by actinonin induces conformational change in the... [/fig_ref] [bib_ref] Identification of eukaryotic peptide deformylases reveals universality of N-terminal protein processing mechanisms, Giglione [/bib_ref] [bib_ref] Distinctive features of the two classes of eukaryotic peptide deformylases, Serero [/bib_ref]. Measurements of the kinetic parameters of the second step of the binding mechanism (k 5 ) revealed a timescale in the 10-s range [fig_ref] Table 1: Comparison of the main kinetic and thermodynamic parameters describing the inhibition of... [/fig_ref] , which is consistent with the collective motion of a large domain [bib_ref] Dynamic personalities of proteins, Henzler-Wildman [/bib_ref] [bib_ref] Functional aspects of protein flexibility, Teilum [/bib_ref]. This finding is supported by NMR studies [bib_ref] Solution structure of nickel-peptide deformylase, Dardel [/bib_ref] [bib_ref] A new subclass of the zinc metalloproteases superfamily revealed by the solution..., Meinnel [/bib_ref] , which showed that actinonin binding induces drastic changes in the heteronuclear single quantum coherence (HSQC) spectrum of EcPDF, since most resonances undergo significant shifts that affect a large part of the structure [bib_ref] ) 1 H, 13 C and 15 N NMR assignments of the..., Larue [/bib_ref] [bib_ref] Ligandinduced changes in the structure and dynamics of Escherichia coli peptide deformylase, Amero [/bib_ref]. The existence of alternative conformational states of EcPDF is further supported by recent biophysical studies [bib_ref] Delineation of alternative conformational states in Escherichia coli peptide deformylase via thermodynamic..., Berg [/bib_ref]. Previously reported snapshots of a series of different conformations of the enlarged and mobile loop-the socalled CD loop-of the dimeric PDF from Leptospira interrogans PDF (LiPDF) in the presence or absence of inhibitor led to the hypothesis of the existence of an equilibrium between a closed and open form of the CD-loop of PDF enzymes, suggesting a selected-shift model to the authors [bib_ref] Novel conformational states of peptide deformylase from pathogenic bacterium Leptospira interrogans: implications..., Zhou [/bib_ref]. Taken together, these data suggest that the binding of actinonin to PDF is accompanied or preceded by conformational changes within the enzyme. Paradoxically, this proposal has not been currently supported by the available structural data. Indeed, free and complexed crystal structures have provided no evidence for any significant conformational change in PDF structure induced by the binding of ligand [bib_ref] The crystal structures of four peptide deformylases bound to the antibiotic actinonin..., Guilloteau [/bib_ref] [bib_ref] Novel conformational states of peptide deformylase from pathogenic bacterium Leptospira interrogans: implications..., Zhou [/bib_ref] [bib_ref] Antibiotic activity and characterization of BB-3497, a novel peptide deformylase inhibitor, Clements [/bib_ref] [bib_ref] Crystal structure of peptide deformylase from Staphylococcus aureus in complex with actinonin,..., Yoon [/bib_ref] [bib_ref] Structure analysis of peptide deformylase from Bacillus cereus, Moon [/bib_ref] [bib_ref] Peptidyl aldehydes as slow-binding inhibitors of dualspecificity phosphatases, Park [/bib_ref].
Tight inhibition in the closed state is associated with the K I * apparent equilibrium constant [fig_ref] Figure 1: Slow, tight-binding inhibition of PDF by actinonin induces conformational change in the... [/fig_ref]. A K I * value (see [fig_ref] Table 1: Comparison of the main kinetic and thermodynamic parameters describing the inhibition of... [/fig_ref] and Materials and Methods for the biochemical definition of K I *) of 0.9 nM for actinonin could be measured for AtPDF; that is, a value very similar to that obtained for bacterial PDFs, including EcPDF and Bacillus stearothermophilus PDF2 (BsPDF2, [fig_ref] Table 1: Comparison of the main kinetic and thermodynamic parameters describing the inhibition of... [/fig_ref]. Tightening of the initial encounter complex (E:I) resulted in a final complex (E:I*) in which the potency of actinonin (K I /K I* ) was enhanced by more than two orders of magnitude and exhibited a very slow off-rate (k 6 , [fig_ref] Table 1: Comparison of the main kinetic and thermodynamic parameters describing the inhibition of... [/fig_ref]. The dissociation constant value of AtPDF for actinonin was also assessed using isothermal titration calorimetry (ITC) experiments [fig_ref] Table 1: Comparison of the main kinetic and thermodynamic parameters describing the inhibition of... [/fig_ref] and [fig_ref] Figure 2: Four distinct conformational states of PDF enzymes [/fig_ref]. The corresponding ITC titration curves [fig_ref] Figure 2: Four distinct conformational states of PDF enzymes [/fig_ref] are consistent with a very strong affinity of the ligand for the enzyme [bib_ref] Isothermal titration calorimetry, Velazquez-Campoy [/bib_ref] , enabling us to determine an accurate K d . Moreover, these studies generated values similar to those measured by other means for AtPDF and EcPDF [bib_ref] Delineation of alternative conformational states in Escherichia coli peptide deformylase via thermodynamic..., Berg [/bib_ref] [bib_ref] Discovery and refinement of a new structural class of potent peptide deformylase..., Boularot [/bib_ref].
## Author summary
The notion of induced fit when a protein binds its ligandlike a glove adapting to the shape of a hand-is a central concept of structural biochemistry introduced over 50 years ago. A detailed molecular demonstration of this phenomenon has eluded biochemists, however, largely due to the difficulty of capturing the steps of this very transient process: the ''conformational change.'' In this study, we were able to see this process by using X-ray diffraction to determine more than 10 distinct structures adopted by a single enzyme when it binds a ligand. To do this, we took advantage of the ''slow, tight-binding'' of a potent inhibitor to its specific target enzyme to trap intermediates in the binding process, which allowed us to monitor the action of an enzyme in real-time at atomic resolution. We showed the kinetics of the conformational change from an initial open state, including the encounter complex, to the final closed state of the enzyme. From these data and other biochemical and biophysical analyses, we make a coherent causal reconstruction of the sequence of events leading to inhibition of the enzyme's activity. We also generated a movie that reconstructs the sequence of events during the encounter. Our data provide new insights into how enzymes achieve a catalytically competent conformation in which the reactive groups are brought into close proximity, resulting in catalysis.
## Ligand-induced conformational closure of atpdf in the crystalline state
Occurrence of a conformational change induced by drug binding was visualized via the resolution of several crystal structure forms of AtPDF, the free form and/or in a complex with actinonin . The data reveal a structural switch between the two forms that can account for both the thermodynamic and kinetic data. The enzyme was observed in two states, a novel open apo-form and a closed, induced, actinonin-bound complex [fig_ref] Figure 1: Slow, tight-binding inhibition of PDF by actinonin induces conformational change in the... [/fig_ref]. Binding of actinonin resulted in a tightening of the active site through the collective closure of the entire N-terminal portion of the protein (strands b1, b2, and b3; helix a1; and CD-loop, see Movies S1 and S2, [fig_ref] Figure 1: Slow, tight-binding inhibition of PDF by actinonin induces conformational change in the... [/fig_ref] , and [fig_ref] Figure 1: Slow, tight-binding inhibition of PDF by actinonin induces conformational change in the... [/fig_ref]. The amplitude of the structural change was maximal for Pro60 [fig_ref] Figure 1: Slow, tight-binding inhibition of PDF by actinonin induces conformational change in the... [/fig_ref] , the Ca of which was shifted 4 Å upon actinonin binding. This collective movement involved the formation of a ''super b-sheet'' as the result of the large rearrangement of b-strands 4 and 5 relative to the rest of the structure in which actinonin forms an additional strand bridging the two b-sheets (b1 andb2) on either side of the active site [fig_ref] Figure 1: Slow, tight-binding inhibition of PDF by actinonin induces conformational change in the... [/fig_ref] and [fig_ref] Figure 1: Slow, tight-binding inhibition of PDF by actinonin induces conformational change in the... [/fig_ref]. As actinonin is a peptide-like compound (see Introduction and [fig_ref] Figure 1: Slow, tight-binding inhibition of PDF by actinonin induces conformational change in the... [/fig_ref] , this behavior closely mimics what occurs in the natural protein substrates of PDF, which also form this strand-bridging interaction. This phenomenon also accounts for the strong stabilization of the protein by actinonin, which was also challenged by differential scanning calorimetry (DSC) experiments: the T m of AtPDF increased from 61uC to 81uC upon binding of the inhibitor [fig_ref] Figure 1: Slow, tight-binding inhibition of PDF by actinonin induces conformational change in the... [/fig_ref] , see also below).
Thus far, this closure of the enzyme induced by actinonin is part of the rare structural evidence for the slow, tight-binding mechanism at an atomic scale. The open state, which has never been observed, was captured not only in the two molecules of the asymmetric subunit but also in different crystals and under two distinct crystallization conditions (Table S2 and [fig_ref] Figure 2: Four distinct conformational states of PDF enzymes [/fig_ref]. All r.m.s.d. values were smaller than 0.25 Å . The closure is very unlikely to result from crystal packing constraints, as soaking the apo-AtPDF crystals in a solution containing actinonin induced the structural transition from the open to the closed state within the crystals without cracking them or altering their diffracting power. Thus, crystal packing is compatible with both states of the enzyme [fig_ref] Figure 3: Effect of actinonin binding on the conformation of key residues in PDF [/fig_ref]. Therefore, the open structure most likely corresponds to a stable state in solution.
The closed final conformation was identical to that previously reported for PDF complexes obtained either with actinonin or with a product of the reaction [bib_ref] Iron center, substrate recognition and mechanism of peptide deformylase, Becker [/bib_ref] [bib_ref] The crystal structures of four peptide deformylases bound to the antibiotic actinonin..., Guilloteau [/bib_ref] [bib_ref] Antibiotic activity and characterization of BB-3497, a novel peptide deformylase inhibitor, Clements [/bib_ref] [bib_ref] N-alkyl urea hydroxamic acids as a new class of peptide deformylase inhibitors..., Hackbarth [/bib_ref] , indicating that this structure is common for the ligands (compare [fig_ref] Figure 1: Slow, tight-binding inhibition of PDF by actinonin induces conformational change in the... [/fig_ref] , and [fig_ref] Figure 4: Evidence for an induced fit in crystalline and solution states of AtPDF [/fig_ref]. Hydrogen bonding was also conserved, especially the bond between the backbone nitrogen of Ile42 (corresponding to Ile44 in EcPDF, see [fig_ref] Figure 1: Slow, tight-binding inhibition of PDF by actinonin induces conformational change in the... [/fig_ref] and [fig_ref] Figure 5: Inhibition and enzymatic reactions progress through an induced fit pathway [/fig_ref] and the alkyl carbonyl chain of actinonin, which potently contributes to the formation of the super b-sheet (Movie S2 and [fig_ref] Figure 1: Slow, tight-binding inhibition of PDF by actinonin induces conformational change in the... [/fig_ref] , see also below). Between the open and closed states, the side chains of Ile42, Phe58, and Ile130 underwent significant structural changes [fig_ref] Figure 3: Effect of actinonin binding on the conformation of key residues in PDF [/fig_ref] and D and [fig_ref] Figure 6: Effect of 6b binding on the conformation of key residues of PDF [/fig_ref] , corresponding to a hydophobic pocket rearrangement, with Ile42 being the most affected [fig_ref] Figure 3: Effect of actinonin binding on the conformation of key residues in PDF [/fig_ref]. Interestingly, Ile42 is the second residue of the conserved active-site motif G 41 IGLAAXG (motif 1) that was previously shown to be essential for activity [bib_ref] Substrate recognition and selectivity of peptide deformylase. Similarities and differences with metzincins..., Ragusa [/bib_ref].
To assess and visualize the differences between the two states, two independent structural parameters were measured: the r.m.s.d. value with respect to the open form and the aperture angle (d ap ), which measures the angle made between the N-and C-domains through three fixed-points, corresponding to the Ca of three conserved residues, each sitting in one of the three conserved motifs [fig_ref] Figure 2: Four distinct conformational states of PDF enzymes [/fig_ref]. The bi-dimensional graph of these two parameters is a good representation of the closing motion snapshots [fig_ref] Figure 2: Four distinct conformational states of PDF enzymes [/fig_ref] shown in Movie S1. With this tool at this stage, two states could be defined: the closed (C) and open (O) states [fig_ref] Figure 2: Four distinct conformational states of PDF enzymes [/fig_ref].
Evidence for a Pure Induced-Fit Mechanism in the Binding of Actinonin to AtPDF Recent quantitative analyses of both conformational selection and induced fit have led to an integrated continuum-a so-called ''flux-description''-of these two limiting mechanisms [bib_ref] Conformational selection or induced fit: a flux description of reaction mechanism, Hammes [/bib_ref]. According to this model, conformation selection tends to be preferred at low ligand concentrations (mM range)-that is, using detailed kinetic studies-whereas induced fit dominates at high ligand and enzyme concentrations (mM range) obtained, for instance, in NMR or crystallographic approaches. Structural studies are most useful to reveal subpopulations of biological significance.
We investigated the existence of lowly populated, alternative conformations of apoPDF. To probe the occurrence of alternate conformers in the crystalline state of PDF, the new Ringer program is the most suitable investigation tool [bib_ref] Hidden alternative structures of proline isomerase essential for catalysis, Fraser [/bib_ref] [bib_ref] Automated electron-density sampling reveals widespread conformational polymorphism in proteins, Lang [/bib_ref]. Ringer searches for evidence of alternate rotamers by systematically sampling electron density maps-free of model bias-around the dihedral angles of protein side chains. Two independent WT open datasets of the apoenzyme, including a high-resolution set (1.3 Å ), were used in the analysis. Ringer analysis revealed the existence of only one rotamer of most side chains of either molecule in the asymmetric unit, including the three main residues primarily involved in conformation change-that is, Ile 42, Phe58, and Ile130 [fig_ref] Figure 4: Evidence for an induced fit in crystalline and solution states of AtPDF [/fig_ref]. Ringer analysis showed evidence for unmodeled alternate conformers for very few residues, including Ile121 and Phe87, or Phe119 to a much lesser extent . There is therefore no evidence for the occurrence of a closed conformation in the apostructure of AtPDF, supporting the hypothesis that the conformational change was essentially induced by the binding of actinonin rather than from conformational selection among multiple states occurring in the crystalline state.
To further investigate the mechanism involved, we followed a kinetic approach aimed at discriminating between induced fit and population shift at low ligand concentrations (sub-mM range) [bib_ref] Residence time of receptor-ligand complexes and its effect on biological function, Tummino [/bib_ref]. The experimentally observed pseudo-first-order rate constant for the approach to equilibrium between the free components and the binary AtPDF-actinonin complex (k obs ) was measured and plotted as a function of actinonin concentration. This plot yielded a hyperbolic saturation curve with a positive slope, as fully expected for a pure induced-fit mechanism [fig_ref] Figure 4: Evidence for an induced fit in crystalline and solution states of AtPDF [/fig_ref] and C). In contrast, if the enzyme sampled two or more conformational states, the curve would imply that the value of k obs decreases with increasing ligand concentration (see, for instance, curve C in [fig_ref] Figure 1: Slow, tight-binding inhibition of PDF by actinonin induces conformational change in the... [/fig_ref] in [bib_ref] Residence time of receptor-ligand complexes and its effect on biological function, Tummino [/bib_ref]. The same conclusion can be reached for EcPDF and BsPDF2 [fig_ref] Figure 4: Evidence for an induced fit in crystalline and solution states of AtPDF [/fig_ref] and C) and was already reported by others for S. aureus PDF [bib_ref] Mechanism of time-dependent inhibition of polypeptide deformylase by actinonin, Van Aller [/bib_ref].
Together, these data indicate that a pure induced-fit mechanism triggered by the binding of actinonin appears to direct the conformational change both in solution and in the crystalline state.
## Single variants at gly41 exhibit strongly reduced actinonin-binding potency and catalytic efficiency
When dealing with an induced-fit mechanism, knowledge of the initial O and final C state is crucial but does not provide direct information on the position of actinonin in the encounter complex or on the sequential mechanism of the transition process. We suspected that the conserved glycine-rich motif 1 (G 41 IGLAAXQ) could contribute to the flexibility required for the observed structural transition. Evidence for such flexibility comes from NMR analysis of EcPDF in which a few residues show exchange cross-peaks of an additional, alternative form [bib_ref] Solution structure of nickel-peptide deformylase, Dardel [/bib_ref]. The most strongly affected residues are Cys90, one of the metal ligands, its neighbor Leu91, and both of the alanines within the above conserved glycine-rich motif [fig_ref] Figure 1: Slow, tight-binding inhibition of PDF by actinonin induces conformational change in the... [/fig_ref] , suggesting that EcPDF undergoes conformational dynamics in a similar region.
To unravel the dynamics of the recognition process, we surmised that it should be possible to freeze the conformational
[formula] t 1/2 (min) f 2965 6 61 1 . 1 60.2 a [/formula]
The enzyme concentrations used in the assay were 100, 50, and 25 nM for AtPDF, EcPDF, and BsPDF2, respectively. b Data from [bib_ref] Discovery and refinement of a new structural class of potent peptide deformylase..., Boularot [/bib_ref]. c Prior to kinetic analysis for determination of the K I * value, actinonin was incubated at the final concentration in the presence of the studied enzyme set for 10 min at 37uC. The kinetic assay was initiated by the addition of a small volume of the substrate. d For determination of K I , k 5 , and k 6 values, actinonin was not preincubated with the enzyme. The kinetic assay was initiated by the addition of the enzyme. e k 4 corresponds to the kinetic constant of the dissociation of the primary enzyme-actinonin complex. It is assumed that the rate of complex association is diffusion-limited (see .3 in, that is, k 3 -the kinetic constant of the association of the primary enzyme-actinonin complex-is 10 9 M21 .s 21 . f t 1/2 is 0.693(k 4 +k 5 +k 6 )/k 4 k 6 (see case of induced fit and calculation in [fig_ref] Table 1: Comparison of the main kinetic and thermodynamic parameters describing the inhibition of... [/fig_ref] of [bib_ref] Residence time of receptor-ligand complexes and its effect on biological function, Tummino [/bib_ref]. In this case, t 1/2 ,0.693/k 6 because k 6 ,,k 5 ,,k 4 . doi:10.1371/journal.pbio.1001066.t001 A zoom is displayed on the right of the panel. Superimpositions were realized using ''module superpose'' in the CCP4i package and the ''secondary structure matching'' tool. The extent of aperture/closure of PDF enzymes was assessed primarily by the measurement of the aperture angle (d ap ), the angle made between the Ca of three strictly conserved residues (C, H, and I) of all PDFs, each characterizing a secondary crucial structure module of the active site crevice, namely b 4 , a 2 , and b1 (see [fig_ref] Figure 1: Slow, tight-binding inhibition of PDF by actinonin induces conformational change in the... [/fig_ref]. Each single residue belongs to one of the three conserved motifs (motifs 2, 3, and 1, respectively) and corresponds respectively to Cys91, His137, and Ile42 in AtPDF. change along the pathway by introducing selected, minor variations within the above-mentioned crucial residues involved in the collective motion. In this respect, site-directed mutagenesis of AtPDF was performed on Gly41, Ile42, and Ile130. Single substitutions were made at Gly41 (G41A/Q/M), Ile42 (I42A/F/ N/W), and Ile130 (I130A/F), and the variants were purified and characterized. These mutant proteins showed no change in overall stability, as evidenced by DSC experiments (unpublished data). However, two variants of G41, G41Q and G41M, showed dramatic effects; the k cat /K m values were reduced by three orders of magnitude due to large decreases in the k cat values compared to the WT enzyme [fig_ref] Figure 5: Inhibition and enzymatic reactions progress through an induced fit pathway [/fig_ref] and [fig_ref] Table 1: Comparison of the main kinetic and thermodynamic parameters describing the inhibition of... [/fig_ref]. The reduced k cat /K m values suggest an altered ability of these variants to attain the final enzyme-transition state complex and, as a result, to give rise to possible states different from the final E:I* complex. Substitutions at positions 42 and 130 only caused small reductions in the k cat values [fig_ref] Figure 5: Inhibition and enzymatic reactions progress through an induced fit pathway [/fig_ref] , [fig_ref] Figure 2: Four distinct conformational states of PDF enzymes [/fig_ref] , and [fig_ref] Table 1: Comparison of the main kinetic and thermodynamic parameters describing the inhibition of... [/fig_ref]. The actinoninbinding potency of both G41 variants was also greatly reduced (Table S1 and [fig_ref] Figure 2: Four distinct conformational states of PDF enzymes [/fig_ref]. The time-dependent inhibition by actinonin of the most active variants was then studied . Ringer plots of electron density (r) versus x 1 angle for representative residues of the 3D apo-structure of AtPDF. Data were obtained with the 3M6O, 3PN2, and 3PN3 datasets (2.0 and 1.3 Å resolution, respectively, see [fig_ref] Table 1: Comparison of the main kinetic and thermodynamic parameters describing the inhibition of... [/fig_ref]. The secondary peaks in the Ile residues are observed because Ile is a branched amino acid. To reveal an alternative conformation with Ile, three peaks should be observed. (B) k obs is a saturable function of actinonin with various PDFs, including AtPDF. Data obtained for k obs , the experimentally observed pseudo-first-order rate constant for the approach to equilibrium between the free components and the binary PDF-actinonin complex, were obtained at various concentrations of actinonin in the presence of EcPDF, AtPDF, and BsPDF2. A direct plot is shown. Inset, time-course measurement of deformylation as a function of varying actinonin concentrations. (C) Inverted plot of the data in panel B, which is expected to be a straight line if the k obs is ..k 6 in the case of induced fit. The correlation coefficient of each line is 1.00, 0.99, and 1.00 for AtPDF, BsPDF2, and EcPDF, respectively, indicative of the accuracy of the conclusion. doi:10.1371/journal.pbio.1001066.g004
The half-lives of the final complexes-as assessed by comparison of the 1 / k 6 values-were always significantly smaller , suggesting that the conformational change induced by actinonin binding still occurred, but the C state is destabilized relative to the O state in the mutants compared to the WT. Accordingly, actinonin strongly stabilized almost all of the variants; T m was increased by more than 20uC. This differs from the G41M and G41Q variants, which both showed increases in the T m of only 12uC, consistent with reduced binding potency [fig_ref] Table 1: Comparison of the main kinetic and thermodynamic parameters describing the inhibition of... [/fig_ref].
## Conformational changes of gly41 variants are affected on-pathway
The two most interesting variants, G41Q and G41M, could be crystallized under the same conditions as the WT protein. In the case of G41Q, the structure of the apo-protein did not show any modifications compared to the WT structure and remained in an O conformation [fig_ref] Figure 2: Four distinct conformational states of PDF enzymes [/fig_ref] ; ''O'' zone). In contrast, the 3D structure of the G41M variant showed that the asymmetric unit was composed of two molecules with distinct structures. One molecule (chain A) is in the O state and is similar to the structures of the WT and the G41Q variant [fig_ref] Figure 2: Four distinct conformational states of PDF enzymes [/fig_ref] ; zone ''O''). The second molecule (chain B) is in a C state, closer to that observed for the WT chain in the presence of actinonin (''C''), a so-called ''superclosed'' state [fig_ref] Figure 2: Four distinct conformational states of PDF enzymes [/fig_ref] ; zone ''S''), suggesting that the substitution modified the equilibrium between the two states in solution either (i) at the step of protein synthesis by providing two conformers, the inter-conversions of which are blocked due to steric hindrance brought by the new bulkier side-chain at position 41, or (ii) by dramatically unbalancing the free inter-conversion between the O and S conformers towards the S state. Ringer analysis indicates that in the free G41M variant, many residues show evidence for unmodeled alternate conformers-including positions 58, 42, and 130-in keeping with the second hypothesis.
For all variants of position G41, addition of actinonin to the crystal [fig_ref] Figure 3: Effect of actinonin binding on the conformation of key residues in PDF [/fig_ref] and [fig_ref] Figure 6: Effect of 6b binding on the conformation of key residues of PDF [/fig_ref] induced a closure of the protein within the crystal. Nevertheless, as expected from in silico graphic modeling followed by energy minimization, the occurrence of a bulky side chain at position 41 prevented the completion of the closure in the presence of the ligand and, hence, the formation of the hydrogen bond between the backbone nitrogen of Ile42 and actinonin. This finding is consistent with the strongly reduced T m of the complex of the variants with actinonin compared to WT as measured by DSC. Remarkably, both S and O forms of the G41M apo-structures in the asymmetric unit of the crystal yielded a unique intermediary structure (''I'' state) upon actinonin binding (r.m.s.d. between the molecules is ,0.25 Å ; see also [fig_ref] Figure 2: Four distinct conformational states of PDF enzymes [/fig_ref] , zone ''I''). In this case, it is likely that the induced-fit mechanism drives the equilibrium by capturing only the O population and closing it to an intermediary step, thus depleting the pool of O conformers that is shifted sequentially back from the remaining pool of S conformers and allows the complete binding of actinonin to the enzyme.
In line with the rational design of the PDF mutants, the extent of the structural differences suggests that the underlying motions are dependent on the length of the side chain . Together, these data account for the reduced catalytic rate, as the hydrogen bond is strictly required for the substrate to be efficiently cleaved by PDFs [bib_ref] Design and synthesis of substrate analogue inhibitors of peptide deformylase, Meinnel [/bib_ref]. Therefore, from both structural and kinetic analyses, each substitution most likely reproduces intermediates along the pathway that lead to the closure of PDF around its substrate [fig_ref] Figure 2: Four distinct conformational states of PDF enzymes [/fig_ref].
## Conformational changes of gly41 variants recapitulate closing intermediates
Analysis of the structures allows us to propose the following sequence of atomic events [fig_ref] Figure 2: Four distinct conformational states of PDF enzymes [/fig_ref] and [fig_ref] Figure 6: Effect of 6b binding on the conformation of key residues of PDF [/fig_ref]. To name the various sites of the ligand and subsites of PDF, we will use the usual nomenclature found in [bib_ref] On the size of the active site in proteases, Schechter [/bib_ref] , which defines the various binding pockets of a protease, where P1' is the first side chain at the C-terminal side of the cleavage site and its binding pocket is S1', also referred to as the hydrophobic pocket in the case of PDF. First, actinonin aligns along the S1' pocket to form the encounter complex, which shifts the Ile130 side chain to avoid steric hindrance in the S1' pocket, promotes rotation of the Ile42 side chain, and finally rearranges the phenyl group of Phe58. These events achieve an optimal hydrophobic S1' pocket conformation [fig_ref] Figure 3: Effect of actinonin binding on the conformation of key residues in PDF [/fig_ref] , and the concomitant closure leads to the formation of a hydrogen bond between the first carbonyl group of actinonin and the backbone nitrogen of Ile42. The initial N-O distance is reduced from 5 Å to 2.8 Å , which is an optimal value for hydrogen bonding (Movie S2 and . Thus, the primary driving force for the active site closure appears to be the P1':S1' hydrophobic interaction. The C state is ultimately locked by the super-b-sheet hydrogen bonds extending across the ligand, including those involving Ile42. The DDG binding value (2.2-2.4 kcal/mol, , as calculated from the K d values for actinonin binding to wild-type (WT) and G41M and G41Q, is consistent with the loss of a hydrogen bond that also contributes to the conformational stability of the protein [bib_ref] Hydrogen bonding and biological specificity analysed by protein engineering, Fersht [/bib_ref] [bib_ref] Contribution of hydrogen bonds to the conformational stability of human lysozyme: calorimetry..., Takano [/bib_ref]. Thus, this bond contributes to the major binding free energy difference between the two complexes (3.1 kcal/mol; , Tables S1 and S3, and [bib_ref] Mechanism of time-dependent inhibition of polypeptide deformylase by actinonin, Van Aller [/bib_ref]. Interestingly, the above DDG binding values also correlate with the DDG ES values derived from the k cat /K m and k cat measurements. This dataset strongly correlates with the . (B) Schematic model for actinonin binding to AtPDF in favor of an induced-fit pathway. PDF might exist in at least two conformational states, open (O) or closed (C). The relative abundance of each conformation would vary, depending on the enzyme type. With AtPDF, it is likely that the most abundant form is the O one, which is the only form leading to a productive complex. The superclosed form (S) is likely to show reduced affinity for the ligand because of steric occlusion of the active site. At the initial stage, the inhibitor (shown in red) binds to AtPDF (indicated in brown) in the O conformation. To reach the final key-lock state (productive closed conformation, C), two major and extreme pathways can be used. According to the conformational selection pathway, the inhibitor selects the C conformation. This pathway, which is represented by the dashed arrow, does not occur within the crystal. In contrast, the G41Q and G41M mutants, by providing the structure of the enzyme in intermediate conformations (I), prove the existence of the socalled encounter complex and confirm that the inhibitor binds to the enzyme when it is in the O conformation. The ligand-binding site is then reorganized to yield the C enzyme conformation, that is, the key-lock state. Indeed, the inhibitor binds to the enzyme through the induced-fit pathway. Each timescale was calculated using the data available in the text and corresponds to t 1/2 values deduced from the calculation of 0.693/ (kinetic constant of interest). The k cat value (k 2 ) was used to assess the timescale of catalysis in panel C, whereas, in (B), k 4 assesses the first step of inhibition, and k 6 is used in the case of the slow step. For the SO conversion (left, B), the lifetime of the minor form of EcPDF was used to assess the order of magnitude (see text and [bib_ref] Solution structure of nickel-peptide deformylase, Dardel [/bib_ref]. (C) Schematic model for the deformylation reaction catalyzed by PDF. Since actinonin is a pseudo-peptidic inhibitor, it is likely that a peptidic substrate can bind to the PDF enzyme through an induced-fit pathway, as described in (B). The key-lock state represents a transition state in which the N-formylated substrate is deformylated to yield the final reaction product. doi:10.1371/journal.pbio.1001066.g005 gyration and van der Waals radii of the side chain at position 41 as well as the N-O distance between the first carbonyl group of actinonin and the backbone nitrogen of Ile42 . These results suggest that the capacity of both G41M and G41Q variants to form the transition state is a consequence of their inability to reach the fully closed state.
Thus, our study of the designed Gly41 mutant enzymes reveals that, in addition to the initial and final states observed for the WT enzyme, the conformations of the Gly41 variants correspond indeed to on-pathway intermediates, thus providing snapshots along the trajectory from the O to the C state of the enzyme [fig_ref] Figure 2: Four distinct conformational states of PDF enzymes [/fig_ref]. The 3D structure of the variants in the absence of ligand is similar to that of WT, and a strict correlation exists between the completeness of the conformational change and both binding potency and catalytic efficiency. This suggests that both events require complete protein closure to generate a productive complex. The strong stabilization of AtPDF by actinonin [fig_ref] Figure 1: Slow, tight-binding inhibition of PDF by actinonin induces conformational change in the... [/fig_ref] closely mimics what occurs with its natural substrates when it reaches the transition state [bib_ref] Iron center, substrate recognition and mechanism of peptide deformylase, Becker [/bib_ref] [bib_ref] Deformylase as a novel antibacterial target, Yuan [/bib_ref]. Indeed, as expected, the enzyme facilitates the final C conformation by lowering its final energy [bib_ref] Enzymes with lid-gated active sites must operate by an induced fit mechanism..., Sullivan [/bib_ref]. Optimal arrangement of the S1' pocket [fig_ref] Figure 3: Effect of actinonin binding on the conformation of key residues in PDF [/fig_ref] proceeds along the reaction process towards the final C conformation, triggering the alignment of reactive groups in an optimal arrangement for ligand recognition. Upon binding, actinonin alters the thermodynamic landscape for the structural transition between the O and C states. This ligand is a potent inhibitor because it can trigger the above sequence of events similar to the substrate, but unlike the substrate, it is nonhydrolyzable. Thus, by mimicking the transition state and being non-hydrolyzable [fig_ref] Figure 1: Slow, tight-binding inhibition of PDF by actinonin induces conformational change in the... [/fig_ref] , the final C complex is long lasting.
## Ligand-induced conformational closure is initially
Triggered by the Binding of the P1' Group in the S1' Pocket Given the similarity between actinonin and natural substrate binding, the very slow kinetics of inhibitor binding (10-s time-scale) remains puzzling compared to the 10 ms required for catalysis (deduced from the k cat ). This finding could be explained as a conformational effect during the formation of the hydrogen bond, aligning the substrate as an additional beta-sheet and eventually stabilizing the entire enzyme-ligand complex. The significantly longer time needed to reach the most stable state compared to the substrate would most likely be due to the presence of the flexible and one carbon longer metal-binding group in actinonin (i.e., hydroxamate versus formyl, [fig_ref] Figure 1: Slow, tight-binding inhibition of PDF by actinonin induces conformational change in the... [/fig_ref]. This suggestion is in line with the overall data obtained when we investigated more deeply the role of the first carbonyl group of the ligand. This group is well known to exert a crucial effect in both productive and unproductive ligand binding (i.e., substrate and inhibitor) [bib_ref] Design and synthesis of substrate analogue inhibitors of peptide deformylase, Meinnel [/bib_ref]. In this respect, we studied the binding of compound 6b [fig_ref] Figure 5: Inhibition and enzymatic reactions progress through an induced fit pathway [/fig_ref] , a PDF ligand that does not exhibit a reactive group at this position [bib_ref] Discovery and refinement of a new structural class of potent peptide deformylase..., Boularot [/bib_ref]. We observed that this compound binds strongly to both EcPDF (K I* = 6366 nM) and AtPDF (K I* = 400635 nM) but, unlike actinonin, does not display slow, tight binding as K I* = K I . This impact on binding is consistent with the absence of the hydrogen bond involving the first carbonyl group of the ligand. The 3D structure of AtPDF was determined after soaking the compound in crystals of the free, open AtPDF form. Upon binding, 6b induced a complete conformational change, identical to that observed with actinonin [fig_ref] Figure 2: Four distinct conformational states of PDF enzymes [/fig_ref] ; ''O'' state). This result further suggests that the conformational change is not induced initially by the formation of this hydrogen bond and that the encounter complex is primarily driven by the fit within the S1' pocket. This also reveals that the timescale of the large conformational change is several orders of magnitude faster than the kinetics of slow binding and fully compatible with both the first step of actinonin binding (k 4 = 140 s 21 ; see [fig_ref] Table 1: Comparison of the main kinetic and thermodynamic parameters describing the inhibition of... [/fig_ref] and the catalytic rate of the substrate (k cat = 37 s 21 ; see [fig_ref] Table 1: Comparison of the main kinetic and thermodynamic parameters describing the inhibition of... [/fig_ref] and . The 3D structure also revealed that both the P1' and the hydroxamate groups are bound similarly to the corresponding groups of actinonin [fig_ref] Figure 6: Effect of 6b binding on the conformation of key residues of PDF [/fig_ref]. As expected, no additional bonding occurs, especially around the backbone nitrogen of Ile42 [fig_ref] Figure 6: Effect of 6b binding on the conformation of key residues of PDF [/fig_ref].
Taken together, these data allow us to conclude that the conformational change observed upon ligand binding is triggered primarily by binding in the S1' pocket. As revealed by the binding of 6b, the one carbon longer metal-binding group fits, immediately upon recognition of the P1' group, in the S1' pocket and forms a bidentate complex with the metal cation, mimicking the transition state as a result. Thus, the active site is very confined and rigid due to the presence and length of the hydroxamate group (compare right and left panels in [fig_ref] Figure 1: Slow, tight-binding inhibition of PDF by actinonin induces conformational change in the... [/fig_ref]. As a result, compared to the complex made with the substrate, it is likely that the formation of the hydrogen bond involving the carbonyl of actinonin and the backbone nitrogen of Ile42 becomes strongly rate-limiting (k 5 = 0.044 s 21 ; [fig_ref] Table 1: Comparison of the main kinetic and thermodynamic parameters describing the inhibition of... [/fig_ref]. Once this hydrogen link is locked, the uncleavable bond, mimicking the labile formyl group at the transition state, stabilizes the enzyme-inhibitor complex, making it long-lasting (k 6 = 0.0006 s 21 ; [fig_ref] Table 1: Comparison of the main kinetic and thermodynamic parameters describing the inhibition of... [/fig_ref] and providing a mechanistic explanation for the slow-binding effect that involves both large and fine conformational changes. The large conformational change is similar to the one occurring with the substrate, whereas the second is more subtle and locks the hydrogen bond involving the backbone nitrogen of Ile42. The second step is rate-limiting with some transition state analogs such as actinonin [fig_ref] Figure 5: Inhibition and enzymatic reactions progress through an induced fit pathway [/fig_ref] and C).
## Proper positioning of the carbonyl group is required to
Stabilize the Complex at S1' Compound 21 corresponds to another interesting derivative designed to probe the impact of the peptide bond in PDF binding [bib_ref] Discovery and refinement of a new structural class of potent peptide deformylase..., Boularot [/bib_ref]. In addition to the hydroxamate group, this compound features both a hydrophobic benzyl group at P1' and a reverse peptide bond. Compound 21 shows modest but significant inhibitory potency to AtPDF1B (K I * = 400637 nM), confirming the crucial role of the peptide bond in PDF binding. After soaking with crystals of apo-AtPDF, compound 21 could be detected in high-resolution electron density maps . Unlike 6b, 21 did not bind the active site of the enzyme but an alternative pocket at the surface of the protein . A docking study performed with EcPDF had previously revealed this alternative binding pocket ; [bib_ref] Docking studies of Nickel-Peptide deformylase (PDF) inhibitors: exploring the new binding pockets, Wang [/bib_ref].
The aforementioned data indicate that the occurrence of a S1'binding group placed in the unfavorable context of a reverse peptide bond does not stably promote binding at the active site of AtPDF. Upon binding of 21, the 3D structure of both molecules of the asymmetric unit remain in an O conformation (r.m.s.d. ,0.2 Å with respect to the apo-structures in the ''O'' state). This finding suggests that only the binding of compounds entering the S1' pocket, such as actinonin or 6b, induces conformational change, in keeping with the crucial role of the P1' group if located in the frame of a classic peptide bond. Moreover, we noticed that the binding pocket of 21 was located on the rear side of the true S1' pocket and induced a weak modification of the P1' hosting platform . Indeed, when crystals of the 21:AtPDF complex were soaked in actinonin, the final 3D structure no longer showed evidence of compound 21 occupancy greater than 5%. Instead, this structure revealed both actinonin and closing of the protein . The r.m.s.d. between this structure and that obtained directly with actinonin was less than 0.2 Å ; the actinonin position was virtually identical, indicating that the protein had retained full capacity for binding actinonin and closing despite the presence of compound 21. We conclude that actinonin does compete with 21 because of the overlap at P1' of AtPDF1B . As the actinonin S1' subsite strongly mimics that of a true substrate, this result also explains the inhibitory behavior of 21 towards AtPDF.
# Discussion
Although PDF catalysis has been extensively studied and the mechanism has been elucidated [bib_ref] Iron center, substrate recognition and mechanism of peptide deformylase, Becker [/bib_ref] , how the enzyme achieves the catalytically competent state remains unknown. Here, we provide insight on how the enzyme might reach a catalytically competent conformation, demonstrating that the reactive groups move into proximity to promote catalysis [fig_ref] Figure 2: Four distinct conformational states of PDF enzymes [/fig_ref]. We suggest that the motions of the catalytic centre starting with free ligand-PDF favor a final configuration that is optimal for binding and/or catalysis [fig_ref] Figure 1: Slow, tight-binding inhibition of PDF by actinonin induces conformational change in the... [/fig_ref] , and 5B and C). In our model, we propose that free PDF might exist in at least two conformational states, that is, open (O) or super-closed (S). The relative abundance of each conformation varies by enzyme type and incubation conditions, explaining why both conformations have not been trapped thus far. In the case of AtPDF, it is likely that the most abundant form corresponds to an O state, which is the form that leads to a productive complex. Indeed, in the NMR spectra for EcPDF, a few residues show exchange cross-peaks from an additional, alternative form [bib_ref] Solution structure of nickel-peptide deformylase, Dardel [/bib_ref]. The most strongly affected residues are Cys90, one of the metal ligands, its neighbor Leu91, as well as Ala47 and Ala48 on the facing strand. This suggests that EcPDF exists in at least two conformations (''S'' and ''C''; see [fig_ref] Figure 2: Four distinct conformational states of PDF enzymes [/fig_ref] , which undergo slow interconversion on the NMR timescale. The 3D structure of the major conformation (75%, lifetime 300 ms) could be solved at high resolution, but the structure of the minor form (25%, lifetime 100 ms), which exhibits very weak signals, could not be solved [bib_ref] Solution structure of nickel-peptide deformylase, Dardel [/bib_ref]. This conformation appears to correspond to that of the complex obtained with the product of the reaction (Met-Ala-Ser). A very similar situation-although more balanced between the two states-appears to occur in the case of variant G41M, suggesting that a mechanism involving conformational selection followed by induced fit is a general model for PDF and that AtPDF is a specific case where population shift virtually does not occur as the free enzyme is completely in the O conformation. This is also in line with data obtained with L. interrogans PDF (LiPDF), which reveal conformers in both the S and C states (see [fig_ref] Figure 2: Four distinct conformational states of PDF enzymes [/fig_ref] and suggest a population-shift mechanism [bib_ref] Novel conformational states of peptide deformylase from pathogenic bacterium Leptospira interrogans: implications..., Zhou [/bib_ref]. It is interesting to note that LiPDF is a poorly active PDF [bib_ref] Enzymatic properties of a new peptide deformylase from pathogenic bacterium Leptospira interrogans, Li [/bib_ref]. According to the representation shown in [fig_ref] Figure 2: Four distinct conformational states of PDF enzymes [/fig_ref] , Plasmodium falciparum PDF (PfPDF), a poorly active PDF [bib_ref] Characterization of an eukaryotic peptide deformylase from Plasmodium falciparum, Bracchi-Ricard [/bib_ref] , was retrieved only in the S state. Finally, weak decompaction of the structure of Bacillus cereus and Staphylococcus aureus PDFs in the presence of actinonin have been described [bib_ref] Crystal structure of peptide deformylase from Staphylococcus aureus in complex with actinonin,..., Yoon [/bib_ref] [bib_ref] Structure analysis of peptide deformylase from Bacillus cereus, Moon [/bib_ref]. These examples suggest that the enzyme is trapped in the S conformer in the free state and converts to the C conformer when bound to actinonin, suggesting that the S conformer is overrepresented in solution compared to the O state, unlike AtPDF.
This study of AtPDF-including 10 different crystal structures of apo-and complexed enzyme variants-reveals the 3D structure of a PDF in at least four distinct states. This includes the O form, the occurrence of which is crucial for catalysis, as it is the active form.
Here, we propose that the transition from the O to the C state is directly induced by the ligand. Indeed, the O form, which is captured in the crystal, undergoes closure directly upon ligand binding in our soaking experiments. Progression to this closure involves intermediary states (''I'') similar to those observed with variants G41Q and G41M in the presence of actinonin (see [fig_ref] Figure 2: Four distinct conformational states of PDF enzymes [/fig_ref]. Extrapolating the situation to catalysis, which occurs in the crystalline states of PDF, it is likely that hydrolysis of the substrate frees the enzyme in its S state, which in turn needs to open to accommodate a new substrate [fig_ref] Figure 2: Four distinct conformational states of PDF enzymes [/fig_ref]. This is well illustrated in the 3D structure of EcPDF complexed with a product of the reaction, obtained after co-crystallization of the enzyme with the substrate in a closed conformation [bib_ref] Iron center, substrate recognition and mechanism of peptide deformylase, Becker [/bib_ref]. The S free form is likely to exhibit a slower on-rate for the ligand (k 3 ) compared to the O form because of steric occlusion of the active site [fig_ref] Figure 1: Slow, tight-binding inhibition of PDF by actinonin induces conformational change in the... [/fig_ref]. In support of this hypothesis, recent data show that the 3D structure of a C-terminally truncated, poorly active version of AtPDF is in the C conformation in the unbound state, although crystallized under conditions identical to ours [bib_ref] Eukaryotic peptide deformylases. Nuclear-encoded and chloroplast-targeted enzymes in Arabidopsis, Dirk [/bib_ref] [bib_ref] Insights into the substrate specificity of plant peptide deformylase, an essential enzyme..., Dirk [/bib_ref]. This structure is similar to that of chain B, one of the two molecules of the asymmetric subunit of variant G41M [fig_ref] Figure 2: Four distinct conformational states of PDF enzymes [/fig_ref]. This suggests that alterations remote from the active site significantly unbalance the equilibrium between the two conformers, thus altering the efficiency of the reaction [fig_ref] Figure 5: Inhibition and enzymatic reactions progress through an induced fit pathway [/fig_ref]. As the S version corresponds to a significantly less active version of AtPDF compared to that reported in our present work, this further confirms that, compared to the O state, the S state has a significantly weaker propensity to bind substrate or a close mimic ligand, such as actinonin. Comparison of the 3D structures of the free-closed and the ligand-bound-closed forms reveals some differences responsible for the slight steric reduction of the active site of free-closed AtPDF1B with respect to that of the actinonin-AtPDF1B complex [fig_ref] Figure 1: Slow, tight-binding inhibition of PDF by actinonin induces conformational change in the... [/fig_ref] , including the side chain of Ile42 burying the S1' binding pocket [fig_ref] Figure 1: Slow, tight-binding inhibition of PDF by actinonin induces conformational change in the... [/fig_ref]. Overall, these data suggest that an S form might exist under the free state but that it would feature a k 3 value with respect to the ligand that is significantly weaker than that of the O form, which would strongly slow down the reaction or the binding as a result.
With the interaction scheme proposed in our model [fig_ref] Figure 5: Inhibition and enzymatic reactions progress through an induced fit pathway [/fig_ref] , the ligand/substrate binds more easily to the O form and induces the optimal conformation of the enzyme to reach the transition state, thus allowing the reaction to be efficiently catalyzed. In the final model [fig_ref] Figure 5: Inhibition and enzymatic reactions progress through an induced fit pathway [/fig_ref] , there is both conformational selection and induced fit subsequently involved in line with the recently proposed existence of such mixed mechanisms for other enzymes [bib_ref] Free-energy landscape of enzyme catalysis, Benkovic [/bib_ref] [bib_ref] Conformational selection or induced fit: a flux description of reaction mechanism, Hammes [/bib_ref]. Nevertheless, in our model [fig_ref] Figure 5: Inhibition and enzymatic reactions progress through an induced fit pathway [/fig_ref] , we suggest that induced fit is the primary mechanism, as it provides energy input from the ligand, which eventually drives the enzyme towards the productive key-lock complex. Unambiguous distinction between the relative contributions of the two mechanisms is deduced from the observation that k obs is a saturable function of actinonin with various PDF, including EcPDF, BsPDF, AtPDF [fig_ref] Figure 4: Evidence for an induced fit in crystalline and solution states of AtPDF [/fig_ref] , and S. aureus PDF [bib_ref] Residence time of receptor-ligand complexes and its effect on biological function, Tummino [/bib_ref] [bib_ref] Conformational selection or induced fit: a flux description of reaction mechanism, Hammes [/bib_ref] [bib_ref] Mechanism of time-dependent inhibition of polypeptide deformylase by actinonin, Van Aller [/bib_ref] [bib_ref] Discovery and refinement of a new structural class of potent peptide deformylase..., Boularot [/bib_ref].
Using crystallographic reconstruction analysis involving enzyme variants, motions of small mobile loops and movie reconstructions of snapshots of catalytic events have been previously documented [bib_ref] Enzyme catalysis: not different, just better, Knowles [/bib_ref] [bib_ref] Multiple conformational changes in enzyme catalysis, Hammes [/bib_ref] [bib_ref] A perspective on enzyme catalysis, Benkovic [/bib_ref] [bib_ref] Structural snapshots of MTA/AdoHcy nucleosidase along the reaction coordinate provide insights into..., Lee [/bib_ref] [bib_ref] Seeing the process of histidine phosphorylation in human bisphosphoglycerate mutase, Wang [/bib_ref] [bib_ref] Enzymatic capture of an extrahelical thymine in the search for uracil in..., Parker [/bib_ref] , often by visualizing the binding of unnatural inhibitors and not necessarily mimicking closely the substrate and transition state as actinonin does [bib_ref] ACE2 X-ray structures reveal a large hinge-bending motion important for inhibitor binding..., Towler [/bib_ref] [bib_ref] Implications of protein flexibility for drug discovery, Teague [/bib_ref]. However, only a few examples make use of soaking conditions of a crystal to promote the motion and show the importance of induced fit [bib_ref] Enzyme catalysis: not different, just better, Knowles [/bib_ref] [bib_ref] Enzymatic catalysis in crystals of Escherichia coli maltodextrin phosphorylase, Geremia [/bib_ref]. None of these data show a motion of the amplitude revealed here with PDF and a large stabilization of the complex involving the formation of the four-stranded b-sheet superstructure and the entire N-domain of the enzyme. Compared to previous crystallographic analyses, our work integrates biophysical, computational, and kinetic analyses to reconstruct the whole picture, allowing a better understanding of the slow-binding mechanism.
While our work primarily focused on an induced-fit mechanism of enzyme inhibition and catalysis, it should be emphasized that this phenomenon is also applicable to the broader area of receptor-ligand interactions. For example, in all cases where conformational change mechanisms have been proposed for kinase inhibitors without supporting experimental data [bib_ref] Residence time of receptor-ligand complexes and its effect on biological function, Tummino [/bib_ref] [bib_ref] Drug-target residence time and its implications for lead optimization, Copeland [/bib_ref] , further experimental work must be provided to clarify the precise mechanism. We expect this will have important implications on how one conducts future drug-discovery efforts against such enzymes [bib_ref] Inhibition of p38 MAP kinase by utilizing a novel allosteric binding site, Pargellis [/bib_ref].
# Materials and methods
## Protein expression and purification
Expression and purification of mature Arabidopsis thaliana PDF1B and all variants (i.e., AtPDF) were derived from the previously described protocol [bib_ref] Distinctive features of the two classes of eukaryotic peptide deformylases, Serero [/bib_ref] : the lysis supernatant after sonication was applied on a Q-Sepharose column (GE Healthcare; buffers A and B as described containing 5 mM NiCl 2 ) followed by Superdex-75 chromatography (GE Healthcare) using buffer C consisting of buffer A supplemented with 0.1 M NaCl. For crystallization experiments, the protein was purified further. The sample was concentrated on an Amicon Ultra-15 centrifugal filter unit (Millipore Corp.) with a 5-kDa cut-off and applied to a MonoQ HR5/5 column (GE Healthcare) previously equilibrated in buffer A (50 mM Hepes, pH 7.5, and 5 mM NiCl 2 ). Elution was performed with a 50-mL gradient from 0% to 100% buffer B. The buffer of the pooled purified AtPDF1B was exchanged using a PD-10 desalting column (GE Healthcare) to yield a protein solution in 50 mM Hepes, pH 7.5, 0.1 M NaCl, and 5 mM NiCl 2 (buffer C). The protein was concentrated on an Amicon Ultra-15 centrifugal filter unit. The resulting AtPDF1B preparation was frozen in aliquots and stored at 280uC (for crystallization purposes) or diluted 2-fold in 100% glycerol and stored at 220uC (for enzymatic purposes). The typical yield was 5-10 mg AtPDF per liter of culture. All purification procedures were performed at 4uC. Samples of the collected fractions were analyzed by SDS-PAGE on 12% acrylamide gels, and protein concentrations were estimated from the calculated extinction coefficients for each variant.
Site-directed mutagenesis of AtPDF sequence in plasmid pQdef1bDN [bib_ref] Identification of eukaryotic peptide deformylases reveals universality of N-terminal protein processing mechanisms, Giglione [/bib_ref] was carried out using the QuickChange Site-Directed Mutagenesis Kit (Stratagene).
## Enzymology
Assay of PDF activity was coupled to formate dehydrogenase, where the absorbance of NADH at 340 nm was measured at 37uC as previously described [bib_ref] Formate dehydrogenase-coupled spectrophotometric assay of peptide deformylase, Lazennec [/bib_ref]. For measurements of classical kinetic parameters (i.e., K m and k cat ), the reaction was initiated by addition of the substrate Fo-Met-Ala-Ser to the mixture containing purified enzyme in the presence of 1 mM NiCl 2 . The kinetics parameters were derived from iterative non-linear least square calculations using the Michaelis-Menten equation based on the experimental data (Sigma-Plot; Kinetics module). For determination of kinetic parameters related to actinonin, the reaction mixture contained 750 mM NiCl 2 . In some cases, the mixture containing PDF and actinonin was incubated for 15 min at 37uC before kinetic analysis, which was initiated by the addition of substrate. The same protocol was used to determine the dissociation constant of actinonin [K I * = k 4 /(k 3 +k 3 k 5 /k 6 )], but the initial reaction velocities were measured with varying concentrations of Fo-Met-Ala-Ser and actinonin. The data were then calculated according to the method of Henderson, which can be used to determine the dissociation constant of the tight-binding competitive enzyme inhibitor [bib_ref] Actinonin, a naturally occurring antibacterial agent, is a potent deformylase inhibitor, Chen [/bib_ref] [bib_ref] Discovery and refinement of a new structural class of potent peptide deformylase..., Boularot [/bib_ref] [bib_ref] A linear equation that describes the steady-state kinetics of enzymes and subcellular..., Henderson [/bib_ref] by varying both the inhibitor and substrate concentrations. To determine K I , k 5 , and k 6 , the reaction was initiated by the addition of enzyme as previously described [bib_ref] Mechanism of time-dependent inhibition of polypeptide deformylase by actinonin, Van Aller [/bib_ref] [bib_ref] Discovery and refinement of a new structural class of potent peptide deformylase..., Boularot [/bib_ref]. K I * app measurements were used for comparative studies of AtPDF variants at a concentration of 2 mM substrate by varying the concentration of actinonin. K I * app is the slope of the v [Actinonin] /v 0 line curve. k obs was fitted from the kinetic data without preincubation with v I = v s + (v 0 2 v s )e 2kobst where v I is the observed velocity at a given concentration of inhibitor I, v 0 is the velocity, and v s is the steady-state velocity [bib_ref] The behavior and significance of slow-binding enzyme inhibitors, Morrison [/bib_ref]. From the set of values obtained at various concentrations of I, k 5 and k 6 could be derived using k obs = k 6 + k 5 [I]/(K I + [I]). By choosing a set of values with k obs ..k 6 , 1/k obs = 1/k 5 (K I /[I] +1) and 1/k obs = f(1/[I]) is expected to be a straight line in case of induced fit whose positive slope corresponds to 1/k 5 . k 6 was derived from equation k 6 = k 5 / (K I /K I *21) [bib_ref] The behavior and significance of slow-binding enzyme inhibitors, Morrison [/bib_ref].
## Microcalorimetry
ITC experiments were performed using a VP-ITC isothermal titration calorimeter (Microcal Corp.). Experiments were performed at 37uC. For each experiment, injections of 10 mL actinonin (180 mM) were added using a computer-controlled 300 mL microsyringe at intervals of 240 s into the Ni-AtPDF variant solution (5 to 10 mM, cell volume = 2.1 mL) dissolved in buffer C with stirring at 310 rpm. A theoretical titration curve was fitted to the experimental data using the ORIGIN software (Microcal). This software uses the relationship between the heat generated after each injection and DHu (enthalpy change in kcal/ mol), K A (the association binding constant in M 21 ), n (number of binding sites per monomer), total protein concentration, and free and total ligand concentrations. The thermal stability of the WT and variants of Ni-AtPDF1B was studied by DSC using VP-DSC calorimetry (Microcal Corp.). DSC measurements were made with 10 mM protein solutions in buffer C. The actinonin concentration was 20 mM. The same buffer was used as a reference. All solutions were degassed just before loading into the calorimeter. Scanning was performed at 1uC/min. The temperature dependence of the partial molar capacity (Cp) was expressed in kcal/K after subtracting the buffer signal using Origin(R) software.
## Crystallization and soaking experiments
Crystallization conditions were screened by a robot using the sitting drop vapor diffusion method. Crystals were obtained and optimized at 20uC with 15%-20% PEG-3350 and either 0.1 or 0.2 M zinc acetate. The drops were formed by mixing 2 mL of a solution containing 2 to 4 mg/mL protein and 2 mL of the crystallization solution. Crystals were soaked for 24 h by adding actinonin to the crystallization drops at a final concentration of 5 mM. Cryoprotection was achieved by placing crystals for 30 s in a solution that was composed of 20% PEG-3350 and 0.2 M zinc acetate, supplemented with 5%, 10%, and 15% glycerol. Crystals were then directly flash frozen in liquid nitrogen using cryoloops (Hampton Research). Crystals were also grown under conditions described for the C-terminally deleted, weakly active version of AtPDF [bib_ref] Insights into the substrate specificity of plant peptide deformylase, an essential enzyme..., Dirk [/bib_ref].
## X-ray diffraction data collection
Data collections were performed at 100 K at the European Synchrotron Radiation Facility (Grenoble, France) on station ID29, FIP-BM30A, ID14-1, and ID23-2, and at SOLEIL (Gif-sur-Yvette, France) on station PROXIMA1. In each case, a single crystal was used to collect a complete dataset. Data were processed and scaled using XDS software [bib_ref] Automatic processing of rotation diffraction data from crystals of initially unknown symmetry..., Kabsch [/bib_ref]. Two crystal forms were encountered with different cell parameters. In each case, b parameter was nearly equal to a, and data could be indexed into two space groups, P2 1 2 1 2 1 or P4 3 2 1 2. The data are shown in .
## Structure determination and refinement
The structure of free AtPDF was solved by molecular replacement with Phaser [bib_ref] Likelihoodenhanced fast translation functions, Mccoy [/bib_ref] followed by a rigid-body refinement by CNS [bib_ref] Crystallography & NMR system: a new software suite for macromolecular structure determination, Brunger [/bib_ref] using coordinates from the Plasmodium falciparum PDF (PDB code 1RL4) [bib_ref] An improved crystal form of Plasmodium falciparum peptide deformylase, Robien [/bib_ref] as a search model. The structures of actinonin-bound proteins-that is, WT and mutants-were solved using rigid-body refinement by CNS of the free AtPDF structure. The ten final models were obtained by manual rebuilding using TURBO-FRODO [bib_ref] Silicon Graphics geometry partners directory Mountain View, CA, Roussel [/bib_ref] and combined with refinement of only calculated phases using CNS and Refmac [bib_ref] Refinement of macromolecular structures by the maximum-likelihood method, Murshudov [/bib_ref] software. No noncrystallographic symmetries were used. Quality control of the three models was performed using the PROCHECK program [bib_ref] Main-chain bond lengths and bond angles in protein structures, Laskowski [/bib_ref]. To probe for alternative conformers, Ringer was used [bib_ref] Automated electron-density sampling reveals widespread conformational polymorphism in proteins, Lang [/bib_ref]. Ringer is a program to detect molecular motions by automatic Xray electron density sampling, and can be accessed at http:// ucxray.berkeley.edu/ringer.htm.
## Accession numbers
PDB codes for the PDF structures presented within this manuscript are as follows: 3M6O, 3PN2, 3M6P, 3O3J, 3PN3, 3PN4, 3PN5, 3M6Q, 3PN6, and 3M6R. UniProtKB accession numbers for other PDF studied are P0A6K3 (EcPDF) and O31410 (BsPDF). [fig_ref] Figure 1: Slow, tight-binding inhibition of PDF by actinonin induces conformational change in the... [/fig_ref] Alignment of PDF sequences and secondary structures. (A) PDF1B from Arabidopsis thaliana (AtPDF1B) is compared with bacterial type 1B (EcPDF and LiPDF), pathogenic protozoa (PfPDF1B), eukaryotic mitochondrial PDF1A from A. thaliana (AtPDF1A), and bacterial type 2 (BsPDF2). This figure was created with ENDscript [bib_ref] ESPript: analysis of multiple sequence alignments in PostScript, Gouet [/bib_ref]. The sequence alignment was realized with the algorithm muscle included in ENDscript, and modified according to the superimposition of structures. The blue frames indicate conserved residues, white characters in red boxes indicate strict identity, and red characters in yellow boxes indicate homology. The secondary structures at the top (a-helices, 3 10 helices, b-strands, and b-turns are shown by medium squiggles, small squiggles, arrows, and TT letters, respectively) were predicted by DSSP [bib_ref] Dictionary of protein secondary structure: pattern recognition of hydrogen-bonded and geometrical features, Kabsch [/bib_ref]. Relative accessibility (acc) of subunit A is shown by a blue-colored bar below sequence. White is buried, cyan is intermediate, and blue with red borders is highly exposed. A red box means that relative accessibility is not calculated for the residue, because it is truncated. Hydropathy (hyd) is calculated from the sequence according to [bib_ref] A simple method for displaying the hydropathic character of a protein, Kyte [/bib_ref]. It is shown by a second bar below accessibility: pink is hydrophobic, grey is intermediate, and cyan is hydrophilic. Motifs 1 ( 41 GwGwAAXQ 48 ), 2 ( 89 EGCLS 93 ), and 3 ( 133 HEwDH 137 ), where w is a hydrophobic amino acid, are labeled by red stars below the sequence alignment. To simplify the nomenclature, AtPDF1B is referred to as AtPDF throughout the text. (B) Topology cartoon of AtPDF, free (left) or actinonin bound (right), in the same color code as (A). Actinonin (represented by the yellow arrow) binding to the ligand binding site allows the linkage of the two distinct b-sheets into one single b-sheet, by mimicking an additional b-strand. PDB sum (http://www.ebi.ac. uk/thornton-srv/databases/pdbsum/) was used. (C) 3-D structure of AtPDF is represented showing the position of the residues discussed in the text, indicated in red. (EPS) [fig_ref] Figure 2: Four distinct conformational states of PDF enzymes [/fig_ref] Microcalorimetric titration of AtPDF with actinonin. Data were obtained at 37uC by an automated sequence of 28 injections of 180 mM actinonin from a 300 ml syringe into the reaction cell, which contain 9.85 mM AtPDF. The volume of each reaction was 10 ml, and injections were made at 240 s intervals. Top, raw data from the titration. Each peak corresponds to the injection. Bottom, the peaks in the upper panel were integrated with ORIGIN software and the values were plotted versus injection number. Each point corresponds to the heat in mcal generated by the reaction upon each injection. The solid line is the curve fit to the data by the Origin program. This fit yields values for K d . Experiments were done with wild type protein and others variants, and gave similar raw data and curve fit. complex (left) and bound to actinonin (right) (B). Non-crystallographic contacts into asymmetric unit are not modified by closing movement of the protein due to actinonin binding, except for zinc atom number 6. This metal ion is coordinated by side chains of Asp40 and Glu63, and water molecules, Asp40 and Glu63 being hydrogen bonded by side chain of Lys38 of the other subunit of the asymmetric unit. With the closing movement of the protein into the crystal, Ca of Asp40 shifted by 3.1 Å and its side chain flipped by 90u. Therefore, it does no longer participate to the coordination shell of this Zn 2+ ion. However, it is still hydrogen bonded by Lys38 from chain B. (EPS) [fig_ref] Figure 4: Evidence for an induced fit in crystalline and solution states of AtPDF [/fig_ref] Binding of actinonin to AtPDF closely mimics both actinonin and product binding to EcPDF. Superimposition of EcPDF and AtPDF bound to either actinonin (1LRU PDB code, panel A) or Met-Ala-Ser (1BS6 PDB code, panel B), the product of the reaction. The r.m.s.d. value is 1.11 Å for 151 Ca superimposed.
## Supporting information
(EPS) [fig_ref] Figure 5: Inhibition and enzymatic reactions progress through an induced fit pathway [/fig_ref] The ligand binding site of AtPDF. This picture shows the residues of AtPDF that are in contact with actinonin (left) and 6b (right) according to the 3-D structure; this should be compared to the similar scheme shown in [fig_ref] Figure 1: Slow, tight-binding inhibition of PDF by actinonin induces conformational change in the... [/fig_ref] for EcPDF. (EPS) [fig_ref] Figure 6: Effect of 6b binding on the conformation of key residues of PDF [/fig_ref] Electronic densities of the moving side-chains and of actinonin at the binding site in some variants of AtPDF. Actinonin and selected residues (G/Q/M41, I42, F58, and I130) are drawn in stick and are shown in their F O -F C electron density omit maps contoured at 2s, in free wild-type AtPDF (two crystallization conditions, WT1 and WT2), and ligand-bound WT (actinonin, 6b and 21), G41Q, and G41M variants. (EPS) Only few residues show alternative conformation in AtPDF. Alternative conformers in the crystalline state of AtPDF. Ringer plots of electron density (r) versus x 1 angle for representative residues of the 3-D apostructure of AtPDF. Data were obtained with the 3M6O dataset (see [fig_ref] Table 1: Comparison of the main kinetic and thermodynamic parameters describing the inhibition of... [/fig_ref]. The secondary peaks in the Ile residues are observed because Ile is a branched amino acid. To evidence an alternative conformation with Ile, three peaks should be observed. (EPS) Impact of induced fit on the binding free energy of actinonin depends on the capacity to stabilize a hydrogen bond with PDF. (A) The gyration radii [bib_ref] A simplified representation of protein conformations for rapid simulation of protein folding, Levitt [/bib_ref] of the side chain occurring at position 41 is displayed with black squares and compared to the k cat /K m values (grey bars). (B) The distance between the NH of I42 and the CO of actinonin was measured in each case. The percentage of the distance required to make a hydrogen bond (2.8 Å ) is reported (dark squares). The difference of binding free energy (DDG binding ) between the open, free state and the variants closed complexes of the G41 variants are displayed as grey bars. The values were calculated as follows. For the WT, it corresponds to the RT ln(K I* /K I ) value [bib_ref] Mechanism of time-dependent inhibition of polypeptide deformylase by actinonin, Van Aller [/bib_ref] , where R is the ideal gas constant and T is the temperature in Kelvin. RT is 0.616 kcal.mol 21 at 37uC. For the G41M and G41Q variants, the DDG binding corresponds to RT ln(K I-G41variant /K D-WT ). The obtained values are similar to that obtained if the k cat /K m substitutes the K D value in the calculation (DDG binding = RT ln(k cat /K m -G41variant /k cat /K m -WT ). (EPS) Compound 21 does not bind AtPDF1B at S1'. (A) 21 is shown in ball-and-stick format in its F O -F C electron density omit map contoured at 2s. (B) Binding site of 21 into AtPDF1B is detailed. Red and blue residues indicate residues that accommodate the ''phenylalanine'' and ''trimethyl'' groups of 21, respectively. (C) Overall view of 21 binding site (left). Molecular surface of AtPDF is represented, as well as 21 in ball-and-stick format. Residues belonging to the 21 binding pocket are colored in orange. For comparison, molecular surface of EcPDF (PDB code 1G2A) in the same orientation is also represented, with residues forming the new ligand binding pocket colored in orange. Actinonin is represented in ball-and-stick format and is seen through the molecular surface of each PDF. (D) Ball-and-stick representation of the interaction network around compound 21.
The metal cation is shown as a grey sphere. (EPS) [fig_ref] Figure 1: Slow, tight-binding inhibition of PDF by actinonin induces conformational change in the... [/fig_ref] Poorly active versions of AtPDF are in a closed conformation incompatible with actinonin binding. (A) Free and close AtPDF were superimposed as in [fig_ref] Figure 1: Slow, tight-binding inhibition of PDF by actinonin induces conformational change in the... [/fig_ref] and are figured in brown and yellow, respectively. Both the G41M (chain B, shown in orange) and the free C-deleted weakly active AtPDF versions [bib_ref] Insights into the substrate specificity of plant peptide deformylase, an essential enzyme..., Dirk [/bib_ref] , colored in purple, PDB entry code 3CPM) were superimposed, to the two structures, showing that they both fit better to the ligand-bound full-length close form than to the free open form, but that the closure is further pronounced, burying the entrance to a ligand. (B) Close-up showing that the shape of the S1' pocket of the poorly active closed versions make it poorly available to P1' recognition (see circled Ile142 and Ile130 side chains).
## (eps)
Table S1 Catalytic properties of AtPDF. Nm, not measurable; ND, not determined; WT, is wild-type. a Kinetic constants were determined using the coupled assay as indicated in Materials and Methods with substrate Fo-Met-Ala-Ser, in the presence of 100 nM enzyme variant and 750 mM NiCl 2 , at 37uC. The relative value of k cat /K m for wild-type AtPDF was set at 100%. b Data correspond to the binding constant of actinonin as obtained either from ITC or from enzymatic analysis when indicated with an asterisk. c Data from . d Gyration radii are from [bib_ref] A simplified representation of protein conformations for rapid simulation of protein folding, Levitt [/bib_ref]. (DOC)
## Table s2
Crystallographic data and refinement statistics. Values in parentheses are for the outer resolution shell. a R sym (I) = S hkl S i |I hkl,i 2 ,I hkl .|/S hkl Si|I hkl,i |, where ,I hkl . is the mean intensity of the multiple I hkl,i observations for symmetry-related reflections. b R work = 1006(S hkl |Fobs 2 Fcalc|/S hkl |Fobs|). R free is a test set including ,5% of the data. c Percentage of residues in most-favored/additionally allowed/generously allowed/disallowed regions of the Ramachandran plot. d Compound 21 was added first, and actinonin afterwards. (DOC)
## Table s3
Kinetic parameters for inhibition of some AtPDF variants by actinonin. The enzyme concentration used in the assay was 100 nM. Prior to kinetic analysis for determination of K I*app values, actinonin was incubated in the presence of each variant set at the final concentration for 10 min at 37uC; kinetic assay was started by adding a small volume of the substrate. For determination of K I , k 5 , and k 6 values, actinonin was not preincubated with enzyme and kinetic assay was started by adding the enzyme. (DOCX)
Movie S1 Dynamics of actinonin binding to peptide deformylase and closure of the active site.
[fig] Figure 1: Slow, tight-binding inhibition of PDF by actinonin induces conformational change in the protein. (A) Inhibition by a two-step mechanism, involving a tightening of the initial enzyme-inhibitor complex (E?I) to form a more stable complex (E?I*), with the chemical structure of actinonin (I), the natural inhibitor of PDF enzymes (E). (B) Structures of EcPDF bound to actinonin (left) and to the transition state resulting from the cleavage of its substrate, Fo-Met-Ala-Ser (right) [34,35]. (C) Superimposition of free and actinonin-bound AtPDF indicated in green and purple, respectively. The three conserved motifs of the PDF enzymes family are indicated in orange and numbered I, II, and III. Molecules A of both models were superimposed, resulting in an r.m.s.d. of 0.9 Å for 100% of the Ca. Left inset, close-up comparison of the open and closed forms figured in the ribbon representation. (D) Baseline-corrected DSC thermograms of free and actinonin-bound WT AtPDF recorded under the same experimental conditions. doi:10.1371/journal.pbio.1001066.g001 [/fig]
[fig] Figure 2: Four distinct conformational states of PDF enzymes. (A) AtPDF and several other representative bacterial PDFs are superimposed. [/fig]
[fig] Figure 3: Effect of actinonin binding on the conformation of key residues in PDF. Conformation of key residues Ile42, Phe58, and Ile130 in the different complexes: (A) in unbound WT AtPDF, (B and C) in the structure of G41Q and G41M actinonin-bound variants, respectively, and (D) of actinonin-bound WT protein. In the final complex (D), a hydrogen bond is formed between actinonin and the peptidic bond, which links Gly41 and Ile42. During the deformylation reaction, which is catalyzed by the PDF enzyme, the N-terminal formyl-methionine fits into the S1' pocket.The solvent-accessible surface of this pocket is represented here, and only the aliphatic chain of actinonin is shown, mimicking the N-terminal methionine. (E) Free WT enzyme with the S1' pocket shown open in two orientations (top and bottom). (F and G) S1' pocket in the G41Q and G41M variant structures, respectively, shown in two orientations (top and bottom). (H) After the complete conformational modifications of actinonin-bound WT protein induced by actinonin binding, the S1' pocket is shown closed in two orientations (top and bottom). (I) The four models are superimposed; the ligand-binding site is magnified: unbound WT AtPDF; G41Q and G41M actinonin-bound enzyme; and WT actinonin-bound enzyme are indicated in brown, red, orange, and yellow, respectively. Actinonin is indicated by lines. (J) A detailed view of the AtPDF ligand-binding site for all the complexes, which are superimposed, as indicated in the same colors. Arrows indicate the direction of the closing movement within the enzyme, from the open, unbound state to the closed, bound state. doi:10.1371/journal.pbio.1001066.g003 [/fig]
[fig] Figure 4: Evidence for an induced fit in crystalline and solution states of AtPDF. (A) Absence of evidence for alternative conformers in the crystalline state of AtPDF. [/fig]
[fig] Figure 5: Inhibition and enzymatic reactions progress through an induced fit pathway. (A) The catalytic parameters K m and k cat, for all AtPDF variants are provided as a percentage of the wild-type values (WT). Detailed values are presented in [/fig]
[fig] Figure 6: Effect of 6b binding on the conformation of key residues of PDF. Superimposition of free, 6b-, and actinonin-bound AtPDF indicated in brown, red, and yellow, respectively. (A) Molecule A in the three models was superimposed, resulting in an r.m.s.d. of 0.9 Å for 100% of the Ca. Actinonin is shown in yellow and 6b in red. (B) Conformation of key residues Ile42, Phe58, and Ile130 in the different complexes and in unbound WT AtPDF. Actinonin is shown in yellow and 6b in red. (C) A detailed view of the AtPDF ligand-binding site for both actinonin and 6b complexes, which are indicated by sticks and are superimposed. The two ligands are colored in pale and dark grey, respectively. The hydrogen bond made by actinonin only is shown. doi:10.1371/journal.pbio.1001066.g006 [/fig]
[table] Table 1: Comparison of the main kinetic and thermodynamic parameters describing the inhibition of PDF by actinonin. [/table]
|
Controlling Spatial Distributions of Molecules in Multicomponent Organic Crystals, with Quantitative Mapping by Confocal Raman Microspectrometry
We report four experimental strategies for controlling the three-dimensional arrangement of molecules in multicomponent organic crystals, exploiting confocal Raman microspectrometry to quantify the three-dimensional spatial distributions. Specifically, we focus on controlling the distribution of two types of guest molecule in solid organic inclusion compounds to produce composite core−shell crystals, crystals with a homogeneous distribution of the components, crystals with continuous compositional variation from the core to the surface, and crystals with alternating shells of the components. In this context, confocal Raman microspectrometry is particularly advantageous over optical microscopy as it is nondestructive, offers micrometric spatial resolution, and relies only on the component molecules having different vibrational properties.
M ulticomponent crystals or cocrystals 1 are important in the pharmaceutical industry 2 and have the potential to be used as advanced functional materials, exhibiting new optical,electronic, and magnetic 4 properties. Many of these materials are based on the concept of isostructurality, whereby the chemically distinct subunits in the composite solid adopt very similar packing arrangements. It is particularly attractive to "tailor" the function of such materials simply by varying the spatial distribution of the components in the material,while maintaining the same crystal structure. Such multicomponent crystals span a wide range of chemical types from metal alloys 5 to hybrid inorganic/organic systems,organic solid solutions, [bib_ref] Growth Des, Mnyukh [/bib_ref] and crystals resulting from dyeing/zoning procedures. Recently, metal organic frameworks (MOFs) containing mixtures of components have also attracted attention in view of their potential as tunable multifunctional materials. Such mixed MOF crystals may be constructed either by altering the metal center 3,4,9a−c and keeping the organic linker constant or by altering the organic linker and keeping the metal center constant. 9d−g Strategies to synthesize mixed MOFs have generated composite core−shell crystals,9a−f and solid solutions. Although the literature is replete with examples of mixed MOFs, there are far fewer examples of purely organic multicomponent crystals with controlled spatial distributions (e.g., core−shell arrangements [bib_ref] Growth Des, Bardelang [/bib_ref]. With a few notable exceptions, 11 most studies of multicomponent crystals have relied heavily on optical microscopy to establish the spatial distribution of the components. However, this technique is limited by the requirement that the components have different optical properties, which may be particularly problematic for purely organic materials. Furthermore, this technique often requires the crystal to be cut prior to analysis.
Given these issues, we were motivated to develop ways of controlling the spatial distribution of components in purely organic multicomponent crystals and to demonstrate the advantages of using confocal Raman microspectrometry to quantify the three-dimensional spatial distribution of the components. Here we present four strategies for controlling the spatial distribution of guest molecules in solid organic inclusion compounds containing binary mixtures of guest molecules, including the formation of composite core−shell crystals, crystals with a homogeneous distribution of the components, crystals with continuous compositional variation from the core to the surface, and crystals with alternating shells of the components.
The specific inclusion compounds used as model systems for this work are urea inclusion compounds, 12 in which guest molecules (typically based on n-alkanes) are located within the one-dimensional tunnelsof a crystalline urea host structure. The guest molecules are densely packed along the host tunnels (diameter 13c ca. 5.5 Å) with a periodic repeat that is usually incommensurate 13b,14 with the periodic repeat of the host structure. For urea inclusion compounds containing binary mixtures of guest molecules, the host tunnel structure is independent of the relative proportions of the two types of guest and the material grows as a single crystal even though the guest composition inside the crystal may vary. Some of the strategies presented here rely on the fact that different guest molecules have different relative affinities for inclusion within the host tunnel structure. Specifically, under conditions of competitive co-inclusion from a crystallization solution containing a binary mixture of guest molecules, the molar ratio of guest molecules incorporated at the growing surfaces of the crystal at time t is m A (t) = χγ A (t), where χ depends on the relative affinity of the host tunnel for inclusion of guests of types A and B and γ A (t) is the molar ratio of the guest molecules in the crystallization solution. If inclusion of guests of type A is energetically favored over inclusion of guests of type B, then χ > 1 and hence m A (t) > γ A (t).
One of the four strategies for controlling the spatial distribution of components reported herethe natural growth strategy [fig_ref] Figure 2: The four different crystal growth strategies to control the spatial distribution of... [/fig_ref] has been discussed previously,and involves crystal growth under conditions of competitive coinclusion of two different types of guest (denoted A and B). If the relative affinity of incorporating guest A into the crystal is higher than that for guest B (i.e., χ > 1), then the initial stages of growth are dominated by incorporation of guest A. As a consequence, guest A is depleted from the solution more rapidly than guest B [i.e., γ A (t) decreases with time]; hence, as crystallization proceeds, the proportion of guest A incorporated into the crystal decreases [from the equation m A (t) = χ γ A (t), if γ A (t) decreases with time, then m A (t) also decreases with time]. This process leads to a monotonic variation in the composition of the crystal: the region with the highest proportion of guests of type A corresponds to the earliest stages of growth and the region with the lowest proportion of guests of type A corresponds to the latest stages of growth [fig_ref] Figure 3: Schematic representations of the types of mixed crystals produced from the crystallization... [/fig_ref].
In homogeneous growth [fig_ref] Figure 2: The four different crystal growth strategies to control the spatial distribution of... [/fig_ref] , the crystal is grown from a solution containing a mixture of two types of guest that have essentially equal affinity to be included in the crystal (i.e., χ ≈ 1). In this case, the ratio of the two types of guest in the crystal should be the same as the ratio in the crystallization solution [i.e., m A (t) ≈ γ A (t)] and remains essentially constant during crystal growth. The resultant crystal should contain a homogeneous spatial distribution of the two types of guest in the same ratio as the initial crystallization solution [fig_ref] Figure 3: Schematic representations of the types of mixed crystals produced from the crystallization... [/fig_ref].
In the transfer method [fig_ref] Figure 2: The four different crystal growth strategies to control the spatial distribution of... [/fig_ref] , a seed crystal is grown in a solution containing one type of guest (A) and then transferred to a second crystallization solution containing a different type of guest (B).The resultant crystal is expected to contain discrete regions with only guest A in the inner core and only guest B in the outer shell [fig_ref] Figure 3: Schematic representations of the types of mixed crystals produced from the crystallization... [/fig_ref]. In principle, the crystal could be transferred multiple times between different crystal growth solutions (involving two or more different types of guest) to produce different generations of heterogeneous composite crystal with sequences such as AB, ABA, ABAB, ABC, etc.
In the injection method [fig_ref] Figure 2: The four different crystal growth strategies to control the spatial distribution of... [/fig_ref] , crystallization begins with a single type of guest (B) in solution. After the first crystals of the inclusion compound appear, small aliquots of a different guest (A) are injected at regular time intervals into the solution. Guest A is selected as one with a significantly higher affinity for incorporation into the growing crystal. Hence, although present in a relatively small proportion in the crystallization solution, guest A is the predominant guest incorporated into the crystal in the period immediately following injection. Shortly thereafter, all the injected guest A is consumed and guest B again dominates the crystal growth until the next injection of guest A. The resultant crystal should comprise mainly guest B, but with "bands" rich in guest A associated with each injection [fig_ref] Figure 3: Schematic representations of the types of mixed crystals produced from the crystallization... [/fig_ref]. As the time between injections is known, the "bands" rich in guest A act as time-markers, and analysis of the distribution of these bands within the final crystal could potentially be exploited to yield kinetic information on the crystal growth process.
In all experiments discussed in the following sections, the two types of guest were an α,ω-dibromoalkane and an alkane, recognizing that the different Raman signatures of these molecules allow their spatial distribution in the crystals to be determined readily by confocal Raman microspectrometry. In the natural growth and transfer method experiments, the guests were 1,8-dibromooctane (1,8-DBrO) and pentadecane (PD). For homogeneous growth, 1,8-DBrO and undecane (UD) were used, and for the injection method, 1,8-DBrO and hexadecane (HD) were used. The affinity of the urea host structure for including PD and HD guests is significantly greater than that for 1,8-DBrO, whereas UD and 1,8-DBrO have essentially equal affinities for inclusion (and thus are appropriate for the homogeneous growth experiment).
After crystallization, a single crystal was selected and the guest composition was determined as a function of position in the crystal using confocal Raman microspectrometry. Previous studies [bib_ref] Guest composition is established from the ratio R = I(CBr)/I(CN) of the..., Marti-Rujas [/bib_ref] (with a different motivation) demonstrated that the spatial distributions of alkane and α,ω-dibromoalkane guests in urea inclusion compounds are readily quantified using this technique. Specifically, 19 the quantity R N established from analysis of the Raman micrographs indicates the relative amounts of α,ω-dibromoalkane and alkane guests as a function of threedimensional position in the crystal. Higher R N indicates a higher proportion of α,ω-dibromoalkane. By definition, 0 ≤ R N ≤ 1, with the limiting values being attained when only the α,ωdibromoalkane (R N = 1) or only the alkane (R N = 0) is present. The crystal morphology of conventional urea inclusion compounds is long needles with hexagonal cross-section. The host tunnels are parallel to the needle axis (Z-axis). Analysis by confocal Raman microspectrometry involved onedimensional "Y-scans" (with X and Z fixed) or two-dimensional "XY-scans" (with Z fixed at Z = 0 μm) within the crystal as depicted in. The incident laser was parallel to the Y-axis and Y = 0 μm represents the upper surface of the crystal. Along the Y-axis, the scans reported here typically extended from the upper surface to a region near the center of the crystal.shows results from Y-scans for crystals produced by natural growth and homogeneous growth. In each case, the crystal contains both types of guest molecule, but with a substantial difference in the spatial distribution. In the case of natural growth, R N varies continuously and monotonically as a function of depth on moving from the center of the crystal (R N ≈ 0.3; rich in PD guests) to the surface (R N ≈ 1.0; essentially only 1,8-DBrO guests). The observed variation of R N as a function of depth is entirely consistent with the expectation that the region around the center of the crystal (X ≈ 0 μm; Y ≈ 140 μm) was formed at the earliest stage (i.e., lowest R N ) and the regions near the surface (Y = 0 μm) were formed at the latest stage (highest R N ) of the crystal growth process. These observations are consolidated by the results of two-dimensional XY-scans (in a plane perpendicular to the tunnel direction). In the case of natural growth, the variation of guest composition reveals the development of the hexagonal cross-section of the crystal shape, with essentially equal rates of growth of the symmetry-related {100} faces.
In the case of homogeneous growth, on the other hand, there is no significant variation in R N as a function of depth (Y) in the crystal, with R N ≈ 0.5 throughout the crystal. The molar ratio of the two guests in the initial crystallization solution was 0.5, demonstrating that, as predicted above, the 1,8-DBrO/UD system is an example of a pair of guests for which the relative affinities for inclusion within the urea tunnel structure are essentially equal (i.e., χ ≈ 1).
For the crystal produced by the transfer method, the Y-scanexhibits some qualitative similarity to the Y-scan for the crystal prepared by natural growthin that R N increases monotonically from the interior of the crystal to the surface. However, for the crystal prepared by the transfer method, a large region (120 μm < Y < 200 μm) around the center of the crystal has R N = 0, indicating that only PD guest molecules (the guest in the first crystallization solution) are present, whereas the region near the surface of the crystal (0 μm < Y < 20 μm) has R N = 1, indicating that only 1,8-DBrO guest molecules (the guest in the second crystallization solution) are present. In the "intermediate region" (20 μm < Y < 120 μm), the rate of change of R N from R N = 0 to R N = 1 as a function of Y is much greater than for the crystal produced by natural growth. The presence of both types of guest in this region (albeit in substantially varying relative amounts) arises because some amount of the first crystallization solution (containing PD molecules) was transferred together with the seed crystal to the second solution. 20 The central core originates from the PD/urea seed crystal grown in the original crystallization solution, while the 1,8-DBrO/urea shell in the outer region of the crystal arises from the post-transfer crystal growth. Optimization of the experimental procedure for the transfer method may allow the thickness of the intermediate region to be reduced, ideally to achieve an abrupt boundary between the core containing only PD guests and the shell containing only 1,8-DBrO guests. The two-dimensional XY-scan inconfirms that the crystal produced from the transfer method exhibits the hexagonal crosssectional shape at each stage of the growth process, with the welldefined PD core in the central region and the 1,8-DBrO shell near the surface. For the experiment using the injection method, the original crystallization solution contained 1,8-DBrO guest molecules (guest B), and a solution containing HD (guest A) was injected periodically as the crystal growth proceeded. From the Y-scan, R N fluctuates as a function of depth in the crystal, with minima at Y ≈ 30 and 105 μm, and maxima at Y ≈ 5 and 55 μm. The minima correspond to regions of the crystal that are relatively rich in HD, whereas the maxima correspond to regions relatively rich in 1,8-DBrO. The corresponding regions are also observed in the XY-scan, in which two HD-rich regions (low R N ; green) and two 1,8-DBrO-rich regions (high R N ; blue/purple) are clearly identified. In the HD-rich regions, growth occurred immediately following injection of HD into the crystallization solution, and it is clear that the crystal actually started to grow immediately following the first injection of HD rather than in the period prior to the first injection (if crystal growth had started before the first injection, the core of the crystal would contain only 1,8-DBrO guests), as evidenced by the HD-rich region close to the center of the crystal (X ≈ 10 μm; Y ≈ 110 μm) and indicated by the blue arrow in [fig_ref] Figure 2: The four different crystal growth strategies to control the spatial distribution of... [/fig_ref]. In the period following the first injection of HD, R N increases with further growth, indicating that the HD introduced to the solution is consumed rapidly. Following the second injection of HD (30 min after the first injection), R N drops rapidly as HD dominates the growth process again (indicated by the red arrow in [fig_ref] Figure 2: The four different crystal growth strategies to control the spatial distribution of... [/fig_ref]. The fact that R N does not drop abruptly (e.g., vertically in [fig_ref] Figure 2: The four different crystal growth strategies to control the spatial distribution of... [/fig_ref] immediately following the injection of HD suggests that, after injection, diffusion of HD molecules to the surfaces of the growing crystal is not instantaneous. These phenomena result in the formation of an "onionskin" crystal composed of shells that are alternately rich in the two types of guest molecule (see [fig_ref] Figure 3: Schematic representations of the types of mixed crystals produced from the crystallization... [/fig_ref]. However, for reasons discussed above, the composition of the crystal does not alternate abruptly between shells of only HD and shells of only 1,8-DBrO.
The results reported here demonstrate several successful experimental strategies that have been designed for controlling the spatial distribution of binary mixtures of guest molecules in solid organic inclusion compounds. In principle, these strategies could be applied to a wide range of materials, including any crystal constructed from two or more isostructural subunits, for instance other inclusion compounds such as gas hydrates, zeolites and other microporous inorganic solids, and metal− organic framework materials. This paper also demonstrates the utility of confocal Raman microspectrometry as a noninvasive and nondestructive technique for mapping the composition of multicomponent crystals. Although the interpretations in this paper have been restricted to a qualitative level, our ongoing research is focused on further optimizing the injection method to gain deeper kinetic insights into the crystal growth by using the "bands" as "time-markers" for the growth process. Current research is also directed toward understanding the physical properties of multicomponent crystals by investigating, inter alia, phase transition behavior and diffraction properties of solid inclusion compounds prepared with different types of spatial distribution of binary mixtures of guest molecules.
## ■ associated content
## * s supporting information
Experimental procedures, typical Raman spectra, and more details on the quantification method. This material is available free of charge via the Internet at http://pubs.acs.org.
## ■ author information corresponding author
[email protected]; [email protected]
[fig] Figure 1: (a) Crystal structure of the hexadecane/urea inclusion compound viewed along the tunnel axis. (b) Schematic of a single crystal of a urea inclusion compound (needle morphology with hexagonal cross-section) with axis system defined. The Z-axis is parallel to the tunnel direction; the {100} faces are parallel to this axis. In the confocal Raman microspectrometry experiments, the incident laser was parallel to the Y-axis. Different types of mapping are indicated (red line, Y-scan; blue plane, XY-scan). [/fig]
[fig] Figure 2: The four different crystal growth strategies to control the spatial distribution of guest molecules: (a) natural growth, (b) homogeneous growth, (c) transfer method, and (d) injection method. [/fig]
[fig] Figure 3: Schematic representations of the types of mixed crystals produced from the crystallization strategies inFigure 2, with different spatial distributions of guest molecules: (a) natural growth, (b) homogeneous growth, (c) transfer method, and (d) injection method. [/fig]
[fig] Figure 4: (a) Y-scans for crystals produced by natural growth (black) and homogeneous growth (red) [total thickness of crystals along Y-axis: 250 μm (natural growth), 205 μm (homogeneous growth)], and (b) XY-scan for the crystal prepared by natural growth. (c) Y-scan and (d) XY-scan for a crystal produced by the transfer method (total thickness of crystal along Y-axis: 415 μm). (e) Y-scan and (f) XY-scan for a crystal produced by the injection method (total thickness of crystal along Y-axis: 220 μm). The color scheme for values of R N is defined in the inset. [/fig]
|
Stability and Feasibility of Dried Blood Spots for Hepatitis E Virus Serology in a Rural Setting
# Introduction
Hepatitis E virus (HEV) is a major cause of acute hepatitis around the world, but with a disproportionately high burden in low-income countries, where pregnant women and their offspring are at particular risk of severe disease and death due to HEV genotype 1 in particular [bib_ref] Hepatitis E virus: Advances and challenges, Nimgaonkar [/bib_ref]. This genotype is common in many low-income countries such as Bangladesh [bib_ref] A tightly clustered hepatitis E virus genotype 1a is associated with endemic..., Hoa [/bib_ref] , while genotypes 3 and 4 dominate in high-and middle-income countries [bib_ref] Hepatitis E virus: Advances and challenges, Nimgaonkar [/bib_ref]. Several public health interventions, including a new vaccine, could probably alleviate some of the burden. However, large studies of HEV transmission and immunity are needed to establish effective interventions.
Serosurveillance studies are important for understanding the effect of infection control measures, where factors like incidence, prevalence, and herd immune levels may be investigated [bib_ref] The role of seroepidemiology in the comprehensive surveillance of vaccine-preventable diseases, Wilson [/bib_ref]. Such studies usually depend on plasma or serum samples collected by venipuncture, requiring trained health personnel, cold chain for transportation, and storage for samples. These factors are likely to limit the number and sizes of serological studies, particularly in low-income countries.
An alternative, easy, and cost-effective sample type is dried blood spots (DBS), where capillary blood from a fingertip punction is collected on filter paper and dried. The method is minimally invasive and can be easily performed by anyone, even the donor. The cards require little space and can be stored at room temperature for longer periods. DBS have been used successfully for measuring antibodies against other viruses such as HIV, Hepatitis B and C, and SARS-CoV-2, and is well suited for field studies in low-income countries [bib_ref] Reliability of dried blood spot (DBS) cards in antibody measurement: A systematic..., Amini [/bib_ref]. However, the accuracy of measuring serological response to infection or vaccination from DBS samples seem to vary between different viruses [bib_ref] The use of dried blood spot cards to assess serologic responses of..., Reinhardt [/bib_ref] , and hence need to be evaluated for the virus in question. A previous small study provided promising results in detecting HEV-IgG from DBS samples of acute HEV cases. However, using DBS for IgG detection in people with previous HEV infection or vaccination has not been studied.
The main objective of this study was to optimize and validate DBS as sampling material, in a low-income country setting, to detect HEV IgG antibodies for use in serosurveys and vaccine trials. The validation included testing the effect of freeze-thawing of DBS, as well as long time storage at room temperature at 20 - C (normal) and 40 - C (tropical climate).
# Materials and methods
This study is part of a large randomized controlled trial evaluating the effectiveness of a hepatitis E vaccine (HEV 239) in Bangladeshi women. The vaccine is based on a 239 amino acid long recombinant HEV genotype 1 peptide, which encodes the capsid protein [bib_ref] Prophylactic Hepatitis E Vaccine, Zhang [/bib_ref]. The clinical trial is conducted by the Norwegian Institute of Public Health (NIPH) (Oslo, Norway) and the International Centre for Diarrheal Disease Research, Bangladesh (icddr,b).
# Sample material
Three sources of sample material were used [fig_ref] Figure 1: Flow chart of sample processing for the three different source materials used... [/fig_ref]
## Specimen preparation
DBS samples were prepared from venous whole blood (groups 1 and 2) or capillary blood directly applied from fingertip puncture (group 3). The whole blood samples were transported to the laboratory at a temperature between 2 °C and 8 °C and prepared within 24 h after venipuncture. Approximately 50 µL of whole venous or capillary blood was transferred to each circle on the Whatman 903 filter paper card (Sigma-Aldrich). The blood was allowed to air-dry flat, for at least half an hour, and then in a DBS drying rack until dry (at least four hours), then placed in individual zippered storage bags with two silica desiccant packs and a humidity indicator. Plasma was isolated from the remaining venous blood by centrifugation. Both DBS bags and plasma samples were stored at −80 °C until analysis.
## Dbs extraction
Circles of 3.2 mm in diameter (equivalent to approximately 1.5 µL plasma [bib_ref] Influence of Hematocrit and Total-Spot Volume on Performance Characteristics of Dried Blood..., Hall [/bib_ref] were Group 1: Whole blood samples were drawn by venipuncture into EDTA-coated tubes from 10 confirmed anti-HEV negative volunteers. The 10 whole blood samples were split into separate tubes and spiked with WHO reference reagent for HEV IgG antibody (NIBSC code: 95/584) to achieve the relevant anti-HEV IgG levels for further analysis.
Group 2: Whole blood samples were drawn from May to July 2017 from 100 healthy participants (50 men and 50 women) aged 16 to 39 from two villages (Sepaikandi and Naburkandi) in Matlab, Bangladesh, participating in a pilot study of the HEV 239 vaccine trial. The participants were vaccinated with HEV 239 or a control vaccine (Hepatitis B) as a 2-dose regimen at day 0 and 30 (described in ClinicalTrials.gov Identifier: NCT02759991 (pilot-study)). Blood samples were drawn before vaccination and 30 days after the second vaccine dose by venipuncture into EDTA-coated tubes. The samples were immediately refrigerated and kept at a temperature between 2 - C and 8 - C during transport to the laboratory (max 24 h transport time).
Group 3: Blood samples were drawn from 50 healthy women aged 16 to 39 randomly selected from a group of 19,640 women from 67 villages in Matlab, Bangladesh, participating in the main HEV 239 vaccine trial. The women received either HEV 239 or a control vaccine (HBV) at day 0, 30 and 180. Paired venous and capillary blood were collected from each participant before vaccination and 30 days after the third vaccine dose. Venous blood was drawn by venipuncture into EDTA-coated tubes. Capillary blood was collected by finger prick onto a Whatman 903 filter paper card (Sigma-Aldrich, St. Louis, MO, USA).
## Specimen preparation
DBS samples were prepared from venous whole blood (groups 1 and 2) or capillary blood directly applied from fingertip puncture (group 3). The whole blood samples were transported to the laboratory at a temperature between 2 - C and 8 - C and prepared within 24 h after venipuncture. Approximately 50 µL of whole venous or capillary blood was transferred to each circle on the Whatman 903 filter paper card (Sigma-Aldrich). The blood was allowed to air-dry flat, for at least half an hour, and then in a DBS drying rack until dry (at least four hours), then placed in individual zippered storage bags with two silica desiccant packs and a humidity indicator. Plasma was isolated from the remaining venous blood by centrifugation. Both DBS bags and plasma samples were stored at −80 - C until analysis.
## Dbs extraction
Circles of 3.2 mm in diameter (equivalent to approximately 1.5 µL plasma [bib_ref] Influence of Hematocrit and Total-Spot Volume on Performance Characteristics of Dried Blood..., Hall [/bib_ref] were punched out from one spot on a DBS card and diluted in Wantai ELISA sample diluent to obtain the desired dilution (for 1:40, 3 punches in 150 µL diluent was used). The samples were incubated at 4 - C overnight before being centrifuged at 698 g (2500 RPM) for 7 min the next day. The supernatant was transferred to a new tube and analyzed within 24 h (stored at 4 - C) or stored at −80 - C until analysis.
## Serology
DBS eluates and plasma samples were tested for anti-HEV IgG using Wantai HEV IgG ELISA (Beijing Wantai, Beijing, China). This enzyme-linked immunosorbent assay uses a peptide (PE2) encoded by a structural region of ORF-2 derived from a Chinese isolate of HEV genotype 1 [bib_ref] A comparison of two commercially available anti-HEV IgG kits and a re-evaluation..., Bendall [/bib_ref] and is validated for detecting IgG-class antibodies to hepatitis E virus in human serum or plasma. The original dilution is 10 µL plasma/serum added to 100 µL diluent. The optical density (OD) values are reported between 0-3.5, and the original cut-off value (CO) is calculated by adding 0.16 to the average OD of the negative internal controls according to the manufacturer's instructions.
The initial dilution step in the ELISA protocol was optimized for DBS as described above, while the remaining procedure for DBS eluates and the complete procedure for plasma samples was done according to the manufacturer's instructions. If the results were above the limit of quantification of the kit (OD > 3.5) the samples were rerun at higher dilutions (20, 160, 400, and 1000).
To quantify all IgG results in WHO international units (WU/mL), we made separate standard curves for plasma and DBS samples using spiked samples and used a 5 parametric logistic regression described by Gottschalk and Dunn to standardize the results [bib_ref] The five-parameter logistic: A characterization and comparison with the four-parameter logistic, Gottschalk [/bib_ref].
## Optimization (dilution and cut-off)
The standard dilution for Wantai HEV IgG ELISA using serum/plasma is 1:11. For DBS, dilutions between 1:11 and 1:200 (including the extraction process) were evaluated. The optimal dilution was determined by maximizing the difference in OD values between negative and weak positive samples, and the ease of processing.
From the results of paired samples from group 2, we evaluated three methods for calculating the optimal cut-off value above which DBS samples were considered positive: (1) the assay protocol by dividing the optical density value by the cut-off value (OD/CO) and values above one counted as positive, (2) two standard deviations (SD) above the mean DBS OD/CO of the negative samples, and (3) a Receiver Operating Curve (ROC) analysis by evaluating a range of possible cut-off values and their corresponding sensitivity and specificity.
## Precision, linearity and stability (group 1)
Inter and intra-laboratory coefficients of variation (%CV) were determined by running spiked samples (0.2-30.0 WU/mL) on nine ELISA plates in two different laboratories (NIPH and icddr,b). Precision limits of ≤25% were deemed acceptable.
Two DBS samples with 30 WU/mL (spiked) HEV-IgG were diluted 1.5, 3, 6, 12, 24, and 48 times, and the percentage recovery relative to the theoretical values linearity were calculated. DBS samples from spiked whole blood with levels of 2.5, 10, and 20 WU/mL were stored in separated zip bags with desiccant packs at temperatures −80, −20, 20, and 40 - C for up to 60 days (6 months for −20 - C and −80 - C). A separate set of samples with the same values underwent up to five freeze-thaw cycles. Each cycle consisted of a thaw period of 2 h at room temperature and at least 12 h subsequent storage at −80 - C. All paired samples were analyzed together, and HEV-IgG levels were compared to the matched samples stored at −80 - C.
## Clinical performance (groups 2 and 3)
Qualitative results from paired DBS and plasma samples from groups 2 and 3 were used to calculate the sensitivity and specificity using the optimized dilution and cutoff values.
The correlation between antibody concentration in DBS and plasma was calculated by Lin's concordance correlation coefficient as recommended by Watson and Petrie [bib_ref] Method agreement analysis: A review of correct methodology, Watson [/bib_ref] , using the "concord" module in Stata [bib_ref] CONCORD: Stata module for concordance correlation, Cox [/bib_ref] , and by linear regression analysis.
All statistical analyses were performed in STATA [bib_ref] Variability in the performance characteristics of IgG anti-HEV assays and its impact..., Kodani [/bib_ref]. The Pearson chi-squared test was used to check for significant differences between groups (binomial results).
# Results
Seronegative blood from 10 participants (group 1) was used for preclinical validation and optimization. For the clinical validation, 300 paired DBS and plasma samples from 150 participants were analyzed (groups 2 and 3). In group 2, the seroprevalence was 0.33 (95% CI: 0.24-0.43) and the geometric mean IgG titer (GMT) in positive samples before vaccination was 3.33 (2.05-5.40) WU/mL. In group 3, the seroprevalence was 0.38 (0.25-0.53) and the GMT in positive samples before vaccination was 0.74 (0.41-1.34).
## Optimization of elisa assay for dbs
By analyzing serial dilutions of spiked venous samples, we determined the laboratory specific preferred dilution to be 1:40 and 1:60 at our two laboratories. OD/CO results from paired plasma and DBS from group 2 were used to establish the DBS cut-off value. The area under the curve in the ROC analysis was 0.95 (0.90-0.97) [fig_ref] Figure 2: ROC analysis for ELISA results [/fig_ref] and a cut-off value of 1.6 (OD/CO) corresponding to 0.6 WU/mL (2 SD above mean value from negative samples), was deemed favorable for our purpose. The lower limit of quantification was 0.6 WU/mL in DBS and 0.1 WU/mL in plasma. The upper limit of quantification was 30 WU/mL for DBS and 8 WU/mL for plasma. from paired plasma and DBS from group 2 were used to establish the DBS cut-off value. The area under the curve in the ROC analysis was 0.95 (0.90-0.97) [fig_ref] Figure 2: ROC analysis for ELISA results [/fig_ref] and a cut-off value of 1.6 (OD/CO) corresponding to 0.6 WU/mL (2 SD above mean value from negative samples), was deemed favorable for our purpose. The lower limit of quantification was 0.6 WU/mL in DBS and 0.1 WU/mL in plasma. The upper limit of quantification was 30 WU/mL for DBS and 8 WU/mL for plasma.
## Precision, linearity and stability
Inter-laboratory precision was assessed using spiked samples (group 1) with IgG levels between 0.5 and 30 WU/mL in two parallels for 9 ELISA runs. The overall coefficient of variation (%CV) was 24%, with a higher %CV in weak positives (0.5-2) of 41%, and lower in strong positives (2-30) of 6.3%. In the same way, the intra-laboratory variability within our two laboratories was found to be 19.2%.
Serial dilutions of a high positive sample (30 WU/mL) yielded a fairly accurate linearity [fig_ref] Table 1: Dilution linearity of DBS samples starting with 30 WU/mL HEV IgG analyzed... [/fig_ref].
## Precision, linearity and stability
Inter-laboratory precision was assessed using spiked samples (group 1) with IgG levels between 0.5 and 30 WU/mL in two parallels for 9 ELISA runs. The overall coefficient of variation (%CV) was 24%, with a higher %CV in weak positives (0.5-2) of 41%, and lower in strong positives (2-30) of 6.3%. In the same way, the intra-laboratory variability within our two laboratories was found to be 19.2%.
Serial dilutions of a high positive sample (30 WU/mL) yielded a fairly accurate linearity [fig_ref] Table 1: Dilution linearity of DBS samples starting with 30 WU/mL HEV IgG analyzed... [/fig_ref]. For the stability testing, storage for up to 60 days at 20 - C and 40 - C did not result in any major (%CV above 10%) changes in OD/CO, except for the low sample at 40 - C Viruses 2022, 14, 2525 6 of 11 with a %CV of 18.5% after 30 days and 20.4% after 60 days [fig_ref] Figure 3: Stability of DBS stored at 40 °C [/fig_ref]. Further, for freeze-thawing stability testing we did not detect any sign of degradation of the analyte in DBS samples for up to five freeze-thaw rounds [fig_ref] Figure 3: Stability of DBS stored at 40 °C [/fig_ref]. The mean %CV was 2.26 after five cycles and the highest %CV was 10.1, occurring in the sample with low antibody levels after two cycles. Storage for 6 months at −20 - C did not result in any substantial alterations in OD/CO values compared to storage at −80 - C (%CV 2.1%). For the stability testing, storage for up to 60 days at 20 °C and 40 °C did not result in any major (%CV above 10%) changes in OD/CO, except for the low sample at 40 °C with a %CV of 18.5% after 30 days and 20.4% after 60 days [fig_ref] Figure 3: Stability of DBS stored at 40 °C [/fig_ref]. Further, for freezethawing stability testing we did not detect any sign of degradation of the analyte in DBS samples for up to five freeze-thaw rounds [fig_ref] Figure 3: Stability of DBS stored at 40 °C [/fig_ref]. The mean %CV was 2.26 after five cycles and the highest %CV was 10.1, occurring in the sample with low antibody levels after two cycles. Storage for 6 months at −20 °C did not result in any substantial alterations in OD/CO values compared to storage at −80 °C (%CV 2.1%).
## Clinical performance
The overall sensitivity of DBS compared to plasma in groups 2 and 3 combined was 81% (95% CI, Clopper-Pearson, 73.8-87.0) and specificity 97% (91.3-99.0). The sensitivity was lower for detection of past infection compared to recent vaccination (Tables 2 and 3) (p < 0.01). The difference in sensitivity of unvaccinated participants in groups 2 and 3 (68% vs. 55%) was not significant (p = 0.25). Comparing DBS to plasma, used as the gold
## Clinical performance
The overall sensitivity of DBS compared to plasma in groups 2 and 3 combined was 81% (95% CI, Clopper-Pearson, 73.8-87.0) and specificity 97% (91.3-99.0). The sensitivity was lower for detection of past infection compared to recent vaccination [fig_ref] Table 2: Qualitative results from paired plasma samples [/fig_ref] (p < 0.01). The difference in sensitivity of unvaccinated participants in groups 2 and 3 (68% vs. 55%) was not significant (p = 0.25). Comparing DBS to plasma, used as the gold standard, we detected 5 false positive and 29 false negative DBS results in a total of 300 samples [fig_ref] Table 2: Qualitative results from paired plasma samples [/fig_ref].
To evaluate the clinical significance of the reduced limit of detection, we calculated separate sensitivities for different IgG levels. The results were compared to two different IgG thresholds (WU/mL) of which: (1) 90% of people infected in the past two years will have IgG levels above, (2) 90% of people vaccinated two years prior will have IgG levels above [fig_ref] Figure 4: Positive correlation between IgG titer [/fig_ref]. The thresholds were estimated on results from the HEV vaccine pilot trial (group 2) (submitted for publication).
The seroprevalence estimated by plasma samples was 0.36 (95% CI, Clopper Pearson, 0.28-0.44) and 0.25 (0.18-0.33) for DBS (p = 0.04). A linear regression of DBS and plasma produced a coefficient of determination r2 of 0.90 [fig_ref] Figure 5: Correlation between IgG concentrations obtained from plasma and DBS, standardized to WU/mL,... [/fig_ref]. Lin's concordance correlation coefficient was 0.94 (95% CI 0.92-0.95), with similar results in vaccinated (0.92) and unvaccinated (0.94) participants. The average difference was 6.3 WU/mL higher for plasma samples, with a 95% limit of agreement (Bland and Altman) between −70.7 and 58.1. This indicates a good correlation of WU/mL between the two sampling methods, but with a high degree of difference between individual paired samples. The geometric mean of IgG levels for plasma was 14.95 WU/mL (95% CI 10.96-20.36), and for DBS 14.75 WU/mL (95% CI: 10.70-20.35).
## Figure 4.
Positive correlation between IgG titer (WU/mL) and the sensitivity from DBS sample (orange line). Dotted lines represent the level of IgG of which 90% of people with an infection (blue) or vaccination (green) will be above.
# Discussion
This study shows that DBS from capillary or venous blood is a reliable sampling alternative to plasma for HEV IgG measurement and can be used to detect vaccine responses and past infections for at least up to two years. We found DBS to be stable for two months at temperatures up to 40 - C and after five freeze-thaw cycles. Our study demonstrates that DBS can be successfully used to collect and transport blood samples in rural Bangladesh. A single fingerprick reduces the burden for blood donors and the demand for trained health personnel. The stability of IgG levels and small sample sizes reduces the cost and logistic hurdles of cold chain transport and storage [bib_ref] Reliability of dried blood spot (DBS) cards in antibody measurement: A systematic..., Amini [/bib_ref].
Our results show that HEV serosurveys and vaccine response studies can be added to the extensive and growing list of applications of DBS. Although we obtained a lower sensitivity (81%) compared to the mean sensitivity, in a recent review of similar studies (99%) [bib_ref] Reliability of dried blood spot (DBS) cards in antibody measurement: A systematic..., Amini [/bib_ref] , our specificity (97%) was higher compared to the same study (95%). The reduced sensitivity to detect low levels of IgG should be considered if this method is used for serosurveys, as DBS resulted in a significant reduction in estimated seroprevalence Viruses 2022, 14, 2525 9 of 11 compared to results from the plasma samples. Still, it is likely to provide useful data globally, if the results are carefully interpreted and adjusted, especially at a population level.
The observed increase in optical density from IgG negative samples is likely a matrix effect caused by hematocrit or substances in the filter paper [bib_ref] Hemato-critical issues in quantitative analysis of dried blood spots: Challenges and solutions, De Kesel [/bib_ref]. Such effects might be a more prominent issue when using ELISA assays that originally require a low dilution of plasma, such as ours. In order to avoid many false positive samples, we increased the dilution and cut-off value for DBS compared to plasma. This resulted in a higher limit of detection and a higher upper limit of quantification, thereby increasing the overall quantifiable measurement range of the assay.
Our study included DBS samples from both venous blood (group 2) and capillary blood (group 3). The lower (non-significant) sensitivity observed in group 3 could be a result of this difference, but the 4.5 times difference in GMT in positive samples prior to vaccination between the two groups is a more likely cause of this difference.
The inter-laboratory %CV of ELISA results from DBS reported in the literature varies from below 10 [bib_ref] Development of a high-throughput SARS-CoV-2 antibody testing pathway using dried blood spot..., Moat [/bib_ref] to above 40 [bib_ref] Use of a commercial ELISA for the detection of measles-specific immunoglobulin G..., Colson [/bib_ref]. The measured %CV of 24 in our study is relatively high for a quantitative method, but within expectations of a semi-quantitative method due to the sampling methodology used. We suggest that individual IgG level results from DBS should be interpreted with caution as laboratory precision for DBS samples was at the lower end of our defined limits (25% CV) and the relatively high degree of variance we observed between individual paired DBS and plasma samples. However, a high correlation and similar geometric means between DBS and plasma should allow for a high confidence in interpreting quantified IgG levels on group levels.
A limitation of our study is the lack of hematocrit measurements in DBS samples, which could have helped us explore its impact on the specificity, both between individual samples and on the mean values observed. It is possible that the specificity of the described ELISA method can be different in other populations with different average hematocrit levels. Another limitation of the study is that the stability testing was only conducted on one sample for every timepoint, and only lasted for two months. However, we see a trend showing that the DBS is stable throughout the 2-month period tested, but may decrease slightly over time in high temperatures, as expected.
Our validation study is limited to one type of filter paper and one diluent without evaluating alternatives. Other filter papers or diluents may provide better results and could be explored in future studies.
The procedure described in our paper can be considered a semi-quantitative method and is likely to produce reliable data on mean antibody titers in groups and help determine the prevalence of infections or vaccine responses within at least two years after the event.
# Conclusions
This study provides the first evidence that vaccine and infection-induced HEV IgG can be accurately detected from blood collected as DBS within a 2-year period after the event. DBS also proved to be a stable storage media for HEV IgG, withstanding temperature variation and longtime storage. Based on these findings, this DBS method was implemented in a large HEV vaccine trial currently ongoing in Bangladesh. DBS offers a robust and effective sample method for monitoring HEV prevalence, population immunity, and effectiveness of interventions such as vaccination. The observed decrease in sensitivity of weak positive DBS samples should be accounted for in serosurveys. By using DBS rather than plasma/serum, screening for HEV IgG would become more feasible and accessible, thus facilitating much needed research and surveillance programs for HEV. Institutional Review Board Statement: The study was approved by the icddr,b Research Review Committee (RRC) and the Ethical Review Committee (ERC), the Directorate General of Drug Administration (DGDA) in Bangladesh, and the Regional Ethics Committee (REC) in Norway.
# Informed consent statement:
Informed consent was obtained from all subjects involved in the study. For participants aged 16-17 years old, assent was obtained, and their parents signed the informed consent. The study was done by the principles of the Declaration of Helsinki, the standards of Good Clinical Practice, and the regulatory requirements of Bangladesh. ClinicalTrials.gov Identifier: NCT02759991.
Data Availability Statement: Not applicable.
[fig] Figure 1: Flow chart of sample processing for the three different source materials used in this study. [/fig]
[fig] Figure 2: ROC analysis for ELISA results (OD/CO) from DBS compared to plasma, based on results from 200 paired DBS and plasma samples. [/fig]
[fig] Figure 3: Stability of DBS stored at 40 °C (a) and 20 °C (b) for up to 60 days and a various number of freeze-thaw cycles (c) as measured by HEV IgG in international units (WU/mL) at various time points. [/fig]
[fig] Figure 4: Positive correlation between IgG titer (WU/mL) and the sensitivity from DBS sample (orange line). Dotted lines represent the level of IgG of which 90% of people with an infection (blue) or vaccination (green) will be above. [/fig]
[fig] Figure 5: Correlation between IgG concentrations obtained from plasma and DBS, standardized to WU/mL, based on the results from 300 paired DBS and plasma samples. FP = False positive, TP = True positive, TN = True negative, FN = False. [/fig]
[fig] Author: Contributions: Conceptualization, J.Ø., A.A., K.Z., C.H.J., F.Q., K.S.-J., R.B., S.I., T.R.B., W.H., S.S., J.L.D., S.D.; methodology, J.Ø., A.A., K.Z., C.H.J., F.Q., K.S.-J., R.B., S.I., T.R.B., W.H., S.S., J.L.D., S.D.; validation, J.Ø., A.A., K.Z., C.H.J., F.Q., K.S.-J., R.B., S.I., T.R.B., W.H., S.S., J.L.D., S.D. formal analysis, J.Ø.; investigation, J.Ø., A.A., K.Z., C.H.J., F.Q., K.S.-J., R.B., S.I., T.R.B., W.H., S.S., J.L.D., S.D.; resources, J.Ø., A.A., K.Z., C.H.J., F.Q., K.S.-J., R.B., S.I., T.R.B., W.H., S.S., J.L.D., S.D.; data curation, J.Ø., A.A., K.Z., C.H.J., F.Q., K.S.-J., R.B., S.I., T.R.B., W.H., S.S., J.L.D., S.D.; writing-original draft preparation, J.Ø, writing-review and editing, J.Ø., A.A., K.Z., C.H.J., F.Q., K.S.-J., R.B., S.I., T.R.B., W.H., S.S., J.L.D., S.D.; visualization, J.Ø.; supervision, K.Z, F.Q., S.D; project administration, J.Ø., A.A., K.Z., C.H.J., F.Q., K.S.-J., R.B., S.I., T.R.B., W.H., S.S., J.L.D., S.D.; funding acquisition, J.Ø., A.A., K.Z., C.H.J., F.Q., K.S.-J., R.B., S.I., T.R.B., W.H., S.S., J.L.D., S.D. All authors have read and agreed to the published version of the manuscript.Funding: The Norwegian Research Council funded the project through the GLOBVAC funds (project number: 248143). [/fig]
[table] Table 1: Dilution linearity of DBS samples starting with 30 WU/mL HEV IgG analyzed in parallel (DBS 1 and 2). Results are presented in IgG level (WU/mL) and the difference between the observed and expected IgG level in percent (% expected value). [/table]
[table] Table 2: Qualitative results from paired plasma samples (used as gold standard) and DBS in vaccinated and unvaccinated participants. (group 2 = 200 samples, group 3 = 100 samples). [/table]
[table] Table 3: Clinical [/table]
|
Exploring Cross-Sectoral Implications of the Sustainable Development Goals: Towards a Framework for Integrating Health Equity Perspectives With the Land-Water-Energy Nexus
Objectives: To assess existing evidence and identify gaps in the integrative framework of the Sustainable Development Goals (SDGs) for their potential to advance cross-sectoral perspectives and actions that connect health equity with the land-water-energy nexus in a watershed context.Methods: Five bibliographic databases were searched from 2016 to 2021. This yielded an initial 226 publications, which were screened for titles, abstracts, and full texts on DistillerSR; resulting in a final 30 publications that were studied. These keywords defined the search terms: "health equity," "SDGs," "watershed," "resource nexus," and "cross-sectoral."Results: Thematic syntheses of debates and gaps point to the relevance of the SDGs as a cross-sectoral, integrative platform for place-based programming of the land-waterenergy nexus, and to account for negative externalities and cascaded impacts on human and environmental health.Conclusion: For the purpose of monitoring health equity in the contexts of interactions of land, water, and energy in rural, remote, and Indigenous contexts, and on the basis of the SDGs, this paper generates evidence to inform health equity-oriented policies, programs and practices, and to enhance health for equity-seeking populations.
# Introduction
The World Health Organization's 2016 Shanghai Declaration on Health Promotion [bib_ref] A Watershed for Health Promotion, Kickbusch [/bib_ref] [bib_ref] One behind: A UNSDG Operational Guide for UN Country Teams, Unsdg [/bib_ref] stresses the imperative of leveraging interactions among the Sustainable Development Goals (SDGs) [bib_ref] A Nexus Approach to the Post-2015 Agenda, Weitz [/bib_ref] to promote health and well-being. Relatedly, an overlapping body of literature [bib_ref] A Review of the Current State of Research on the Water, Energy,..., Endo [/bib_ref] [bib_ref] Supporting Collaboration in Interdisciplinary Research of Water-Energy-Food Nexus by Means of Ontology..., Kumazawa [/bib_ref] [bib_ref] Resource Nexus Perspectives towards the United Nations Sustainable Development Goals, Bleischwitz [/bib_ref] [bib_ref] Pacific Connections for Health, Ecosystems and Society: New Approaches to the Land-Water-Health..., Parkes [/bib_ref] [bib_ref] Anatomy of a Buzzword: The Emergence of 'the Water-Energy-Food Nexus' in UK..., Cairns [/bib_ref] has identified the interactions at the nexus of land, water and energy as priority areas for progressing the SDGs, with far-reaching implications. [bib_ref] Towards Integrated Governance for Water, Health and Social-Ecological Systems: The Watershed Governance..., Parkes [/bib_ref] , [bib_ref] Supporting Collaboration in Interdisciplinary Research of Water-Energy-Food Nexus by Means of Ontology..., Kumazawa [/bib_ref] , and [bib_ref] Resource Nexus Perspectives towards the United Nations Sustainable Development Goals, Bleischwitz [/bib_ref] acknowledge that challenges emerging from the land-waterenergy nexus continue to impede the attainment of global sustainability goals, including health [bib_ref] Anatomy of a Buzzword: The Emergence of 'the Water-Energy-Food Nexus' in UK..., Cairns [/bib_ref] [bib_ref] Health and Sustainable Development: Challenges and Opportunities of Ecosystem Approaches in the..., Boischio [/bib_ref]. Conflicts emerging from the land-water-energy nexus [bib_ref] Mosquito Net Fishing Exemplifies Conflict Among Sustainable Development Goals, Trisos [/bib_ref] [bib_ref] Policy: Map the Interactions between Sustainable Development Goals, Nilsson [/bib_ref] [bib_ref] Planetary Health: Human Health in an Era of Global Environmental Change, Cole [/bib_ref] not only impact on the social and ecological determinants of health [bib_ref] Preparing for the Future of Public Health: Ecological Determinants of Health and..., Parkes [/bib_ref] [bib_ref] The Interacting Axes of Environmental, Health, and Social Justice Cumulative Impacts: A..., Gislason [/bib_ref] , but they also lead to the inequitable distribution of risks among social groups with pre-existing vulnerabilities [bib_ref] Addressing the Environmental, Community, and Health Impacts of Resource Development: Challenges across..., Parkes [/bib_ref] [bib_ref] Defining Equity in Health, Braveman [/bib_ref] [bib_ref] The Nexus Approach to Water-Energy-Food Security: an Option for Adaptation to Climate..., Rasul [/bib_ref]. The groups most affected by resource insecurities at the interfaces of land, water, energy, and health consist of those who expend the largest share of their income to secure basic needs of water, food, and energy as necessities for health.
Unfortunately, inequitable impacts on health are often overlooked in land, water, energy nexus programming, particularly, in rural and Indigenous contexts [bib_ref] Addressing the Environmental, Community, and Health Impacts of Resource Development: Challenges across..., Parkes [/bib_ref] [bib_ref] Storylines of Research on Resource Extraction and Health in Canada: A Modified..., Brisbois [/bib_ref]. Nexus studies of the land-water-energy domains [bib_ref] Anatomy of a Buzzword: The Emergence of 'the Water-Energy-Food Nexus' in UK..., Cairns [/bib_ref] [bib_ref] Local Community Perceptions toward Livelihood and Water-Energy-Food Nexus: A Perspective on Food..., Wolde [/bib_ref] [bib_ref] Public Health Guide to Field Developments Linking Ecosystems, Environments and Health in..., Buse [/bib_ref] are beginning to highlight the need to account for disproportionate health impacts on local-settings [bib_ref] Addressing the Environmental, Community, and Health Impacts of Resource Development: Challenges across..., Parkes [/bib_ref] already experiencing healthy inequities [bib_ref] Addressing the Environmental, Community, and Health Impacts of Resource Development: Challenges across..., Parkes [/bib_ref] [bib_ref] The Nexus Approach to Water-Energy-Food Security: an Option for Adaptation to Climate..., Rasul [/bib_ref]. A focus on health [bib_ref] Mosquito Net Fishing Exemplifies Conflict Among Sustainable Development Goals, Trisos [/bib_ref] [bib_ref] Measuring Health Inequalities in the Context of Sustainable Development Goals, Hosseinpoor [/bib_ref] [bib_ref] Nexus Approaches to Global Sustainable Development, Liu [/bib_ref] has the potential to overcome some of the limitations produced through traditional siloed approaches [bib_ref] Pacific Connections for Health, Ecosystems and Society: New Approaches to the Land-Water-Health..., Parkes [/bib_ref] [bib_ref] Managing Water, Soil and Waste Resources to Achieve Sustainable Development Goals: Monitoring..., Hülsmann [/bib_ref] to managing land, water and energy systems, especially, failures to address the interconnected challenges of land, water, and energy insecurities [bib_ref] Public Health Guide to Field Developments Linking Ecosystems, Environments and Health in..., Buse [/bib_ref] , as well as the cascade of implications for human and ecosystem health [bib_ref] Anatomy of a Buzzword: The Emergence of 'the Water-Energy-Food Nexus' in UK..., Cairns [/bib_ref] [bib_ref] Preparing for the Future of Public Health: Ecological Determinants of Health and..., Parkes [/bib_ref] [bib_ref] Sustainable Watersheds: Integrating Ecosystem Services and Public Health, Jordan [/bib_ref].
Responding to these gaps, this review explores the case for localizing the land-water-energy nexus-related SDGs in ways that resonate with the notion of "leaving no one behind"-LNOB [bib_ref] A Nexus Approach to the Post-2015 Agenda, Weitz [/bib_ref]. The LNOB approach is in keeping with sub-national efforts around the world to localize the SDG agenda and develop place-relevant capacities for the local monitoring of achievements on the goals [bib_ref] Watershed Management and Public Health: An Exploration of the Intersection of Two..., Bunch [/bib_ref].
Specifically, our review explores the relevance of watersheds as an appropriate localized setting [bib_ref] Health and Sustainable Development: Challenges and Opportunities of Ecosystem Approaches in the..., Boischio [/bib_ref] [bib_ref] Defining Equity in Health, Braveman [/bib_ref] to examine nexus interlinkages of land, water, and energy systems in relation to health equity [bib_ref] Health and Sustainable Development: Challenges and Opportunities of Ecosystem Approaches in the..., Boischio [/bib_ref]. Watersheds offer an integrative, ecologically coherent context to consider the land-waterenergy nexus, as well as a settings-based approachthat is consistent with the Ottawa Charter for Health Promotion [bib_ref] Watersheds in Planetary Health Research and Action, Jenkins [/bib_ref] , which emphasizes understanding health in the everyday, common-place contexts in which people live, work, learn and play. Watersheds also offer local representations of larger social, economic, and ecological processes and challenges [bib_ref] The Watershed as A Conceptual Framework for the Study of Environmental and..., Kolok [/bib_ref] ; as well as constitute a microcosmic unit of analysis to study the embeddedness of complex nexus challenges in sustainable development [bib_ref] Using Watershed Boundaries to Map Adverse Health Outcomes: Examples from Nebraska, USA, Corley [/bib_ref]. As a unit of an ecosystem, a watershed separates a larger ecosystem or landscape into interconnected geospatial units or settings, which can facilitate an inclusive mapping of environmental health inequities by supporting availability of denominator data that can unmask inequities in the social and ecological determinants of health [bib_ref] Using Watershed Boundaries to Map Adverse Health Outcomes: Examples from Nebraska, USA, Corley [/bib_ref] [bib_ref] A Typology of Reviews: an Analysis of 14 Review Types and Associated..., Grant [/bib_ref].
This paper responds to a growing need to connect health, equity, and place-based perspectives into land-water-energy nexus programming. It presents a narrative mapping review that aims to synthesize and visually represent available evidence and existing gaps in cross-sectoral applications of the SDGs in local nexus programming, and to explore the relevance of the SDGs as an integrative tool in this space, especially, in relation to health equity. Specific objectives are to: 1) map evidence on cross-sectoral potential of the SDGs and implications for advancing health equity within the land, water, energy nexus; 2) visualize the interlinkages of land, water and energy within the SDGs and connections to health at a watershed scale; and 3) identify knowledge gaps and integration lapses within the literature to inform the SDGs' cross-sectional potential for health equity with focus on how indigenous knowledge and decolonizing concepts have been integrated into the nexus.
Our paper begins by describing methods used to select and analyze the literature and is followed by a presentation of results, which understands the nexus to function as an analytical tool, a conceptual framework, and a discourse, in ways that have the potential to foster connections across the SDGs within the watershed context. The discussion synthesizes findings associated with cross-sectoral applications of the SDGs in linking health equity with the land-water-energy nexus. The final sections provide concluding reflections and future research considerations on fostering health equity within the land-water-energy nexus through the SDGs.
# Methods
## Selection of review approach
Mapping reviews are increasingly used to visually depict, categorize, and synthesize measures, features and patterns of evidence existing in the broader literature with the goal of determining knowledge gaps that can inform future research [bib_ref] An Evidence Map of Research Linking Dietary Sugars to Potentially Related Health..., Tybor [/bib_ref]. This approach is selected to employ a broader set of questions to complement, supplement and add structure to a preliminary literature review conducted as well as to capture newly emerging literature. In addition to visual representations, a narrative synthesis is used to thematically identify existing evidence and gaps, and qualitatively explore the relationships and patterns between the findings [bib_ref] Synthesis without Meta-Analysis (SWiM) in Systematic Reviews: Reporting Guideline, Campbell [/bib_ref] [bib_ref] The Role of Google Scholar in Evidence Reviews and its Applicability to..., Haddaway [/bib_ref].
## Search strategy
A literature search of Web of Science, MEDLINE, Science Direct, Google Scholar and Academic Search Complete was carried out with the assistance of two librarians. The date parameters were set between 2016 and 2021 to capture relevant studies published in English since September 2015, when the SDGs were launched. Keywords that made up the search terms were: "watershed," "health equity," "SDGs," "resource nexus," and "cross-sectoral." The full list of search terms and how they were combined (using AND/OR, etc.) are presented in Supplementary Appendix A. The database search returned both peer-reviewed articles and grey literature. Grey literature predominantly arose from title searches in Google Scholar, consistent with the finding that more grey literature is found using title searches rather than full text searches in this search engine [bib_ref] Beyond PICO: The SPIDER Tool for Qualitative Evidence Synthesis, Cooke [/bib_ref]. The last search was conducted on May 06, 2021. Citations of all papers identified were saved and stored in Zotero.
## Eligibility criteria
Inclusion and exclusion criteria for the study were developed using a PICO-adapted (Problem/Perspectives-Intervention-Context/ Setting-Outcome) framework [bib_ref] Comparison Study of Specificity and Sensitivity in Three Search Tools for Qualitative..., Methley [/bib_ref] [bib_ref] An Evaluation of DistillerSR's Machine Learning-Based Prioritization Tool for Title/abstract Screening -Impact..., Hamel [/bib_ref] , creating a flexible guide to focus on specific criteria of interest, and is presented in Supplementary Appendix B. An article was included if it satisfied at least one of the following three conditions: 1) discusses the SDGs and cross-linkages through the lens of a nexus approach; 2) focuses on the interplay of actors, sectors or interest and distribution of power, resources and impacts in a watershed context or in the context of land/food, water, energy nexus; 3) engages with integrative, cross-sectoral, or Indigenous perspectives in a watershed context. While "health equity" was part of the keywords searched, it was not an explicit consideration in the selection criteria. This was on account of the PICO guide being used for selection criteria, where the SDGs' integrative potential were regarded as "intervention," and cross-sectoral possibilities taken for "outcome." In this regard, health equity was considered as one of the potential "outcomes" of cross-sectoral possibilities arising at the land-water-energy nexus. Moreover, since previous studies already stressed that the land-water-energy nexus does not often account for health and equity concerns, we did not want to miss relevant articles that did not discuss health equity in explicit terms. Rather, literature that addressed health equity considerations were identified throughout the screening stages of the review.
## Screening, quality appraisal and data extraction
The initial selection of articles retrieved from the database searches were passed through title, abstract and full-text screening on DistillerSR. A list of the included and excluded studies can be found in Supplementary Appendices C, D respectively. To be included in the final subset of articles that met the inclusion criteria, the grey literature was appraised using the AACODS checklist, which is a standard quality appraisal tool that assesses on authority, accuracy, coverage, objectivity, date and significance (Supplementary Appendix E). After ascertaining the quality of included studies, key data for each study were extracted, using a PICO-adapted framework [bib_ref] Comparison Study of Specificity and Sensitivity in Three Search Tools for Qualitative..., Methley [/bib_ref] [bib_ref] An Evaluation of DistillerSR's Machine Learning-Based Prioritization Tool for Title/abstract Screening -Impact..., Hamel [/bib_ref] in Microsoft Excel, to collate categories of information, which were later distilled into relevant themes for discussion.
# Data analysis
Thematic analysis [bib_ref] The Role of Google Scholar in Evidence Reviews and its Applicability to..., Haddaway [/bib_ref] was used to identify themes, and capture patterns and trends with a focus on creating snapshot profiles of land, water and energy interlinkages in the SDGs, in relation to categorized themes, existing evidence and identified gaps.
Consistent with the review objectives, the first focus of the analyses was to identify trends and patterns by dividing the included studies into three nexus role categories, as proposed by [bib_ref] The Water-Energy-Food Nexus: A Systematic Review of Methods for Nexus Assessment, Albrecht [/bib_ref] and expanded by [bib_ref] Navigating towards Sustainable Development: A System Dynamics Approach, Hjorth [/bib_ref]. Analyzing the three nexus roles (as an analytical tool, as a conceptual framework and as a discourse) helped bring together theoretical, conceptual, and value-laden approaches within the nexus integration agenda. Specific attention was paid to cross-sectoral activity given its relevance for health equity integration into the land-water-energy nexus in that it fosters an understanding of the cross-sectoral capacities of the goals and their targets and potential opportunities that exist for promoting health equity when linking the land-water-energy nexus and SDGs indicators [bib_ref] Managing Water, Soil and Waste Resources to Achieve Sustainable Development Goals: Monitoring..., Hülsmann [/bib_ref]. The second analyses focused on mapping the interlinkages of land, water and energy within the SDGs and their connections to health at a watershed scale. The third thematically identified knowledge gaps and integration lapses to inform thinking about the SDGs' cross-sectional potential for linking health equity with the land-water-energy nexus in small, rural, and Indigenous contexts.
# Results
Our findings are separated into four sections starting with the results of the database search in bibliographic databases. This is followed by summary tables of the categorization of nexus roles describing how these roles are characterized in the literature and pathways of connections in relation to cross-sectoral potential of the SDG's and implications for health equity. Next, we depict a visual map of interlinkages of land, water and energy within the SDGs and implications for health at a watershed scale. These results further lead to a visual representation of high priority nexus gaps, underscoring integration lapses for health equity in rural, remote indigenous contexts.
## Results of database search
Of the 226 returned articles from the database search, 13 duplicates were removed, and 213 were screened. Level 1 (title and abstract) screening excluded 96 studies, and 117 were moved to level 2 (full texts) screening, which resulted in 30 articles being included in the study. All five grey literature items were included following AACODS quality appraisal checklist. [fig_ref] FIGURE 1 |: Prisma Flow Chart [/fig_ref] depicts the study selection and screening steps in a flowchart.
## Summary of included studies categorized by nexus roles
2) as a conceptual framework, and 3) as discourse. As an analytical tool, nexus approaches employ quantitative or qualitative methods, or a combination of both, to study interlinkages among water, energy, and land systems. As a conceptual framework, the nexus approach draws on interlinkages between land, water, and energy to advance policy coherence. As discourse, the nexus concept is used to frame inherent challenges in fostering cross-sectoral collaboration [bib_ref] Navigating towards Sustainable Development: A System Dynamics Approach, Hjorth [/bib_ref]. [fig_ref] TABLE 1 |: Key characteristics of identified Journal articles and Grey Literature [/fig_ref] below provides an overview of the key characteristics of the 30 studies arranged by publication dates, describing nexus roles (as an analytical tool, a conceptual framework, a discourse) across the 30 studies, as well as relationships with the SDGs and links to social determinants of health and health equity. These tables are structured to distinguish journal articles from grey literature and appreciate differences in publication trends between the two. [fig_ref] TABLE 1 |: Key characteristics of identified Journal articles and Grey Literature [/fig_ref] presents the range of approaches to the nexus role among the 30 included studies. Without exhibiting ties with other roles, ten focused on analytical approaches, four on conceptual frameworks, and seven employed a discourse. The remaining eight employed two roles, and one study connected across the three roles. From the sample, there appears to be an overall trend, over the past 5 years, towards the use of analytical approaches to understanding interlinkages of complex systems as well as the use of discourses to frame challenges in cross-sectoral collaboration within and across the nexus.
## Mapping evidence on cross-sectoral potential of the sustainable development goals and implications for advancing health equity
The nexus term and the SDGs have both been used to connote principles and processes of integration [bib_ref] Pacific Connections for Health, Ecosystems and Society: New Approaches to the Land-Water-Health..., Parkes [/bib_ref]. They both possess elements of an integrated human-environment frameworkand serve multiple and wide-ranging objectives that link one to the other. [fig_ref] TABLE 2 |: Characterizing Nexus-informed Cross-sectoral potential of the Sustainable Development Goals and Implications for... [/fig_ref] below cross-links the three identified roles of nexus framing to inform the evidence gathered on potential cross-sectoral applications of the SDGs in fostering sensitivity to health equity in the land-water-energy nexus.
Considering that the "nexus" term has been used interchangeably with the SDGs in recent literature, there is a burgeoning "vice versa" opportunity for using nexus variables to facilitate the localization of the SDGs and adapt SDGs target to local realities and priorities [bib_ref] Managing Water, Soil and Waste Resources to Achieve Sustainable Development Goals: Monitoring..., Hülsmann [/bib_ref]. [fig_ref] TABLE 2 |: Characterizing Nexus-informed Cross-sectoral potential of the Sustainable Development Goals and Implications for... [/fig_ref] provides a summary of SDGs' cross-sectoral roles cross-linked with nexus framings. The table also profiles examples of the pathways and relationships [56] The five-node resource nexus at sea 1 Explored the potential for developing conservation and policy interventions to preserve threatened ecosystem functions and services in the marine water-energy-biomass-minerals-land nexus through conservation of ecologically critical natural habitats that sustain these services -which determine the health and wellbeing of humans and ecosystems[83] Energy, water, and food: towards a critical nexus approach 3 Developed frames for the challenges within the water-energy-food nexus and articulated the conceptual, methodological, and practical solutions for advancing cross-sectoral integration [bib_ref] Water and Food: towards a Critical Nexus Approach, Leck [/bib_ref] [73] Beyond zero sum game allocations: expanding resources potentials through reduced interdependencies and increased resource nexus synergies 2 Explored how to deal with the interdependencies of water, energy, and food systems through use of improved policies, technologies, and adapted human behaviors; such that foster system resilience and cross-sectoral communication [bib_ref] Beyond Zero Sum Game Allocations: Expanding Resources Potentials through Reduced Interdependencies and..., Mohtar [/bib_ref] [54] The Water-Energy-Food Nexus: A systematic review of methods for nexus assessment 3
Use of frames to articulate methodological limitations of nexus analytical tools in achieving cross-connections for health and equity [bib_ref] The Water-Energy-Food Nexus: A Systematic Review of Methods for Nexus Assessment, Albrecht [/bib_ref] [61] Advancing the implementation of SDGs in Brazil by integrating water-energy nexus and legal principles for better governance 1,2 Explored interlinkages of water and energy nexus and drew on an understanding of interconnections to reveal challenges to integration within the SDGs and to policy coherence in nexus outcomes [bib_ref] Advancing the Implementation of SDGs in Brazil by Integrating Water-Energy Nexus and..., Carvalho [/bib_ref] [29] Nexus approaches to global sustainable development 2,3
In the use of nexus approaches for uncovering synergies and detecting tradeoffs, it is crucial for the nexus to internalize accounting for and reconciling spillover effects and cascaded impacts on human and environmental health externalized from trade-off interactions [bib_ref] Nexus Approaches to Global Sustainable Development, Liu [/bib_ref] [58] Assessing the State of the Water-Energy-Food (WEF) Nexus in South Africa 1 Employed the platform of an established link between WEF indicators and SDGs indicators to examine pathways between the water, energy, food (WEF) nexus and rural livelihoods, health, and well-being in southern Africa [bib_ref] Assessing the State of the Water-Energy-Food (WEF), Mabhaudhi [/bib_ref] [65] How extractive industries affect health: Political economy underpinnings and pathways 1, 3 Employed system thinking perspectives and frames to draw attention to pathways by which extractive industries affect health outcomes and engender health inequities [bib_ref] How Extractive Industries Affect Health: Political Economy Underpinnings and Pathways, Schrecker [/bib_ref] [75] Water-energy-food nexus: a platform for implementing the Sustainable Development Goals 1, 2, 3 Examined tight interconnections within and across water, energy, and food systems. Proposed that SDG criteria should be the baseline and minimum development goals to be pursued in implementation of the water-energy-land nexus at any scale [bib_ref] Waterenergy-food Nexus: a Platform for Implementing the Sustainable Development Goals, Stephan [/bib_ref] [69] Opportunities and Trade-offs among BECCS and the Food, Water, Energy, Biodiversity, and Social Systems Nexus at Regional Scales.
2 Developed a conceptual framework that incorporated biodiversity and social systems as part of the water-energy-food nexus. The framework was used as an interdisciplinary platform to analyze the trade-offs and opportunities among emerging policy strategies at a river basin scale [69]
[57] The Water-Food-Energy Nexus: Power, Politics, and Justice 3 Framed the nexus challenges to advancing cross-sectoral integration. This combined perspectives and concerns on the politics of the nexus, power sharing, equity, and justice[79] Complexity versus simplicity in water energy food nexus (WEF) assessment tools 1
Recognizing unique constraints and complexities across "resource hotspots," the authors developed a tool consisting of a simple-complex spectrum for assessing complexity and appropriation of nexus tools [bib_ref] Complexity versus Simplicity in Water Energy Food Nexus (WEF) Assessment Tools, Dargin [/bib_ref] [58] The water-energy-food nexus as a tool to transform rural livelihoods and well-being in southern Africa 1, 2 An analytical tool was armed with capabilities to interrogate complex systems for livelihood and health impacts of the resource nexus. The tool was later used as a conceptual framework to support decision making for coherent policies [bib_ref] The Water-Energy-Food Nexus as a Tool to Transform Rural Livelihoods and Well-Being..., Mabhaudhi [/bib_ref] [82] Structuring an integrated water-energy-food nexus assessment of a local wind energy desalination system for irrigation 1 Used a novel analytical approach for integrated assessment of water, energy, and food systems in a local desalination case study in the Canary Islands, Spain [bib_ref] Structuring an Integrated Water-Energy-Food Nexus Assessment of a Local Wind Energy Desalination..., Serrano-Tovar [/bib_ref] [77] From a few security indices to the FEW Security Index: Consistency in global food, energy, and water security assessment 1, 3 Analyzed the methodological inconsistencies associated with various indices used in nexus approaches and discussed underlying assumptions to identify and explain these inconsistencies [bib_ref] From a Few Security Indices to the FEW Security Index: Consistency in..., Venghaus [/bib_ref] [78] Linking Environmental Policy Integration and the Water-Energy-Land-(Food-)Nexus: A Review of the European Union's Energy, Water, and Agricultural Policies.
## 2
Used the nexus as one of the conceptual frameworks to evaluate European energy, water, and agricultural policies; and the extent to which integration was inculcated into the design and implementation of these policies [bib_ref] Linking Environmental Policy Integration and the Water-Energy-Land-(Food-)Nexus: A Review of the European..., Venghaus [/bib_ref] (Continued on following page) [64] Sustainable development as the ultimate target of adopting a nexus approach to resources management 1 Stressed the necessity of making nexus approaches more robust with innovative tools that will factor in ecosystem services pathways and make for comprehensive unravelling of interlinkages and cross-sectoral externalities-important for propelling resources management towards achieving Sustainable development [bib_ref] Sustainable Development as the Ultimate Target of Adopting a Nexus Approach to..., Caucci [/bib_ref] [76] Toward understanding the convergence of researcher and stakeholder perspectives related to water-energy-food (WEF) challenges: The case of San Antonio, Texas 2, 3 Evaluated levels of convergence in perspectives and challenges of cross-sectoral communication between water, energy, and food stakeholders and researchers [bib_ref] Toward Understanding the Convergence of Researcher and Stakeholder Perspectives Related to Water-Energy-Food..., Daher [/bib_ref] [26] Local community perceptions toward livelihood and water-energy-food nexus: A perspective on food security.
3
Examined the framing of nexus contributions to livelihoods in a local community. This was important to identify missing links on how nexus resources can enhance living conditions [bib_ref] Local Community Perceptions toward Livelihood and Water-Energy-Food Nexus: A Perspective on Food..., Wolde [/bib_ref] [68] Sustaining the ecological functions of the Litani River Basin, Lebanon.
1 Examined water quality and quantity indicators using Sustainable Development Goal 6 (SDG 6) to provide a guide on water availability and sustainable management of water and sanitation for all [bib_ref] Sustaining the Ecological Functions of the Litani River Basin, Lebanon, Darwish [/bib_ref] [67] Linking reservoir ecosystems research to the sustainable development goals Land-Water-Energy nexus role, relationships with the Sustainable Development Goals, and links to social determinants of health and health equity; ordered by publication date (oldest to newest). Notes: Role of Nexus Framing-(1) As an Analytical Tool: employs quantitative or qualitative methods to study nexus interactions. (2) As a Conceptual Framework: draws on interlinkages between land, water, and energy to advance policy coherence. (3) As a Discourse: frames inherent challenges in fostering cross-sectoral collaboration.
connecting the SDGs to the social determinants of health and equity. In relation to the SDGs' role as an analytical tool and usage in the included literature, the SDGs framework is used to unpack interlinkages and identify options for maximizing synergies and balancing trade-offs [bib_ref] Managing Water, Soil and Waste Resources to Achieve Sustainable Development Goals: Monitoring..., Hülsmann [/bib_ref]. This is often facilitated by linking nexus variables with SDG indicators and mapping areas of indicator overlap or correlations with the social and ecological determinants of health [bib_ref] Managing Water, Soil and Waste Resources to Achieve Sustainable Development Goals: Monitoring..., Hülsmann [/bib_ref].
Mabhaudhi et al. [bib_ref] Assessing the State of the Water-Energy-Food (WEF), Mabhaudhi [/bib_ref] employed the water-energy-food nexus as an analytical tool and linked nexus indices with SDGs' indicators, in a systems-thinking manner, to explore impacts on rural livelihoods, health, and well-being in Southern Africa. [bib_ref] The Potential of Water Security in Leveraging Agenda 2030, Taka [/bib_ref] and [bib_ref] A Nexus Approach to Water, Energy, and Food Security in Northern Canada, Natcher [/bib_ref] statistically assessed interlinkages among nexusrelated SDGs 2, 6 and 7 to uncover deep interconnections of resource insecurities and opportunities for unlocking synergies and balancing trade-offs. In [bib_ref] Managing Water, Soil and Waste Resources to Achieve Sustainable Development Goals: Monitoring..., Hülsmann [/bib_ref] and [bib_ref] A Nexus Approach to Water, Energy, and Food Security in Northern Canada, Natcher [/bib_ref] , application of the nexus, in an analytical role to the SDGs, entailed collapsing nexus variables to overlap with SDGs indicators in a way that drew correlations with the social determinants of health and furthered the SDGs.
As a conceptual framework, the SDGs can be promoted to resolve challenges often associated with the coordination of knowledge, interests, perspectives, and factors, and to address co-production failures in analyzing the nature and extent of trade-offs, which often produce injustices and inequities in socio-environmental outcomes. [bib_ref] Advancing the Implementation of SDGs in Brazil by Integrating Water-Energy Nexus and..., Carvalho [/bib_ref] and [bib_ref] Public Health Guide to Field Developments Linking Ecosystems, Environments and Health in..., Buse [/bib_ref] employed the SDGs as a normative, conceptual framework for sustainable development, drawing on an appreciation of the complex relationships between water and energy, and multiple crosscutting targets that cater to more than one goal. As a conceptual framework, the SDGs are promoted to resolve inherent challenges often associated with interlinkages, tending to stem from inadequate considerations of the potentials for synergies, and failures to analyze the nature and extent of tradeoffs. The SDGs, as a conceptual framework, can bring into perspective prospective challenges and coordinate mechanisms needed for coherent nexus solutions, which include navigating different management approaches and tackling bottlenecks which also reinscribe injustices and inequities in socio-environmental outcomes [bib_ref] Advancing the Implementation of SDGs in Brazil by Integrating Water-Energy Nexus and..., Carvalho [/bib_ref]. In the light of the power relations and coordination challenges associated with the transdisciplinary character of cross-sectoral nexus framing, the SDGs as a discourseuses value-laden judgements to draw attention to governance, power sharing, distributive justice and equity concerns around access to and use of resources as well as the burden of resource use impacts.
## Mapping interlinkages of land, water and energy within the sustainable development goals and connections to health at a watershed scale
In [fig_ref] FIGURE 2 |: Resource dynamics of land, water and energy at a watershed scale [/fig_ref] below, the SDGs are ascribed with goals and targets for each of land, water, and energy resources, and [fig_ref] FIGURE 2 |: Resource dynamics of land, water and energy at a watershed scale [/fig_ref] depicts that land (SDGs 2, 15 and 12), water (SDGs 6, 14 and 15), and energy (SDG 7) are closely interlinked. The contextual boundary is a watershed, which is considered a resource nexus hotspot where interaction dynamics around the use of land, water and energy resources for production and consumption are very tangible. In a watershed context, numerous factors are implicated in driving the dynamics [fig_ref] FIGURE 2 |: Resource dynamics of land, water and energy at a watershed scale [/fig_ref] around the use of land, water and energy resources for production and consumption. These drivers, shown in [fig_ref] FIGURE 2 |: Resource dynamics of land, water and energy at a watershed scale [/fig_ref] , can be as direct as increased demands for food, water, and energy or as indirect as climate change, increasing human population, urbanization, globalization and human civilization. The responses to these drivers are encapsulated in the resource dynamics of production and consumption, which consists of activities of dam construction and hydropower generation, coal mining, crop diversification, expanded irrigation, biofuel production, biomass generation and desalination. [fig_ref] FIGURE 3 |: Schematic diagram of interactions among land, water and energy systems, pathways to... [/fig_ref] below depicts potential pathways [bib_ref] Sustainable Development as the Ultimate Target of Adopting a Nexus Approach to..., Caucci [/bib_ref] by which resource nexus projects can impinge on the health of humans [bib_ref] How Extractive Industries Affect Health: Political Economy Underpinnings and Pathways, Schrecker [/bib_ref] biodiversity, and natural ecosystems [bib_ref] Linking Reservoir Ecosystems Research to the Sustainable Development Goals, Guo [/bib_ref] as well as impact on determinants of health [bib_ref] A Nexus Approach to Water, Energy, and Food Security in Northern Canada, Natcher [/bib_ref] [bib_ref] Sustaining the Ecological Functions of the Litani River Basin, Lebanon, Darwish [/bib_ref] through disrupting the ecosystem services of provisioning, regulating, supporting, and preserving culture [bib_ref] Sustainable Development as the Ultimate Target of Adopting a Nexus Approach to..., Caucci [/bib_ref].
Desalination, for instance, is an energy demanding process that removes salt from sea water to make it potable for drinking. The process is, however, complicated by leeching of the chemicals used into the soil, thus contaminating water storage in aquifers, thereby lowering water quality [bib_ref] Desalination and Environment: A Critical Analysis of Impacts, Mitigation Strategies, and Greener..., Ihsanullah [/bib_ref]. There is also a possibility of brine dumping that contaminates the
## Nexus-informed cross-sectoral potential of the sustainable development goals (sdgs) current evidence for advancing health equity
The SDGs as an Analytical tool Uses a systems-thinking basis to unpack interlinkages and draw correlations with the social determinants of health [bib_ref] The Potential of Water Security in Leveraging Agenda 2030, Taka [/bib_ref] [bib_ref] A Nexus Approach to Water, Energy, and Food Security in Northern Canada, Natcher [/bib_ref] [bib_ref] Beyond Zero Sum Game Allocations: Expanding Resources Potentials through Reduced Interdependencies and..., Mohtar [/bib_ref] The SDGs as a Conceptual Framework A paradigm to navigate coordination challenges in analyzing the nature and extent of trade-offs which breed injustices and inequities in socio-environmental outcomes [bib_ref] Assessing the State of the Water-Energy-Food (WEF), Mabhaudhi [/bib_ref] [bib_ref] Advancing the Implementation of SDGs in Brazil by Integrating Water-Energy Nexus and..., Carvalho [/bib_ref] [bib_ref] The Water-Energy-Food Nexus as a Tool to Transform Rural Livelihoods and Well-Being..., Mabhaudhi [/bib_ref] The SDGs as a Discourse Uses value-laden judgements and frames to draw attention to governance, power sharing and equity concerns in terms of these lines of questioning: "Integration for whom? Who leads the coordinated efforts? Whose interests are integrated? Whose are traded-off?" [bib_ref] Local Community Perceptions toward Livelihood and Water-Energy-Food Nexus: A Perspective on Food..., Wolde [/bib_ref] [bib_ref] How Extractive Industries Affect Health: Political Economy Underpinnings and Pathways, Schrecker [/bib_ref] [bib_ref] Global Governance and the SDGs, Ivanova [/bib_ref] Public Health Reviews | Owned by SSPH+ | Published by Frontiers May 2022 | Volume 43 | Article 1604362
food chain for both marine life and human consumption. These contamination chains will impact food and water security and cause disparities in access to safe drinking water, sanitation services, and affordable, nutritious food. The pollution chains reveal how nexus activities often follow ecosystem pathways in how they influence the social determinants of health. These connections between the resource nexus and the ecosystem service pathways which influence health and wellbeing have not been factored into nexus assessments [bib_ref] Sustainable Development as the Ultimate Target of Adopting a Nexus Approach to..., Caucci [/bib_ref].
## Mapping nexus gaps and lapses in relation to the sustainable development goals: implications for advancing health equity
The results presented here are based on thematic grouping of knowledge gaps and integration lapses within nexus literature and explain how the SDGs' framework in its cross-sectoral strengths accounts for the nexus gaps. [fig_ref] FIGURE 4 |: A framework for addressing knowledge gaps and cross-sectoral health equity impacts of... [/fig_ref] is a visual map of knowledge gaps within the land, water-energy nexus and it represents a proposed simple framework of often overlooked dynamics within the nexus to which attention should be accorded for addressing health equity concerns in rural, remote, and Indigenous contexts.
Resource insecurities disproportionately impact rural and remote communities and impinge on Indigenous Peoples' selfdetermining goals, socio-environmental values that connect land to health, and ways of knowing and being. In relation to the gaps grouped under decolonial perspectives and place-based contexts in [fig_ref] FIGURE 4 |: A framework for addressing knowledge gaps and cross-sectoral health equity impacts of... [/fig_ref] , our analysis has highlighted ways in which literature focused on nexus approaches have not tended to be inclusive of theoretical perspectives grounded in decolonial scholarship [bib_ref] A Nexus Approach to Water, Energy, and Food Security in Northern Canada, Natcher [/bib_ref] , local and place-based approaches, and Indigenizing ideas [bib_ref] Managing Water, Soil and Waste Resources to Achieve Sustainable Development Goals: Monitoring..., Hülsmann [/bib_ref] [bib_ref] A Nexus Approach to Water, Energy, and Food Security in Northern Canada, Natcher [/bib_ref] [bib_ref] Sustainable Development as the Ultimate Target of Adopting a Nexus Approach to..., Caucci [/bib_ref]. Addressing these gaps and engaging meaningfully with these literatures could enhance options for the SDG's to provide a basis from which to encourage effective cross-sectoral engagement, bridge epistemological divides and address the strengths and limitations of different knowledge domains and approaches [bib_ref] Advancing the Implementation of SDGs in Brazil by Integrating Water-Energy Nexus and..., Carvalho [/bib_ref] [bib_ref] Sustainable Development as the Ultimate Target of Adopting a Nexus Approach to..., Caucci [/bib_ref].
Another integration lapse within the nexus is related to failures to promote participatory approaches. Nexus tools are largely quantitative; thus, there is a need to consider qualitative, participatory approaches [bib_ref] Managing Water, Soil and Waste Resources to Achieve Sustainable Development Goals: Monitoring..., Hülsmann [/bib_ref] [bib_ref] Navigating towards Sustainable Development: A System Dynamics Approach, Hjorth [/bib_ref] , which highlight sociopolitical nuances, such as power, politics, equity, distributional justice, identities, and emotions in nexus framing and conceptualization of challenges and solutions [bib_ref] Navigating towards Sustainable Development: A System Dynamics Approach, Hjorth [/bib_ref] [bib_ref] Advancing the Implementation of SDGs in Brazil by Integrating Water-Energy Nexus and..., Carvalho [/bib_ref]. Moreover, as a result of a missing focus on social and environmental dimensions of sustainability within the nexus, there are significant gaps in understanding ecosystem service pathways of land, water and energy interactions that underlie health inequities [bib_ref] Managing Water, Soil and Waste Resources to Achieve Sustainable Development Goals: Monitoring..., Hülsmann [/bib_ref] [bib_ref] Navigating towards Sustainable Development: A System Dynamics Approach, Hjorth [/bib_ref] [bib_ref] Sustainable Development as the Ultimate Target of Adopting a Nexus Approach to..., Caucci [/bib_ref]. However, the SDGs, which offer a way to address human-nature coupled systems, also offer the integrative capacity to redress these missing sustainability dimensions.
Additionally, there are scalar lapses associated with the nexus in connection with failures to account for health and equity externalities from cross-scale and cross-regional interactions of land, water, and energy systems [bib_ref] Managing Water, Soil and Waste Resources to Achieve Sustainable Development Goals: Monitoring..., Hülsmann [/bib_ref]. Current nexus tools often focus on a specific place or context [bib_ref] Managing Water, Soil and Waste Resources to Achieve Sustainable Development Goals: Monitoring..., Hülsmann [/bib_ref] [bib_ref] Navigating towards Sustainable Development: A System Dynamics Approach, Hjorth [/bib_ref]. This precludes considerations for cross-sectoral and cross regional interactions that often result in leakages or spillover effects. Attending to cross-scalar issues will support the potential of the SDG's to provide a meta-coupling framework [bib_ref] Managing Water, Soil and Waste Resources to Achieve Sustainable Development Goals: Monitoring..., Hülsmann [/bib_ref] that facilitates the integration of human-nature interactions (people and ecosystems) across spatial scales and builds on concepts such as globalization to address multi-scalar, socio-ecological challenges occurring between adjacent or distant systems at local, regional, and global scales.
# Discussion
This discussion considers findings from the mapping review in relation to the challenges of localising and understanding health equity dynamics of the SDG's as well as recommendations for future studies investigating possible pathways to center health equity in the land, water, energy nexus at a watershed scale. Our review identified numerous ways that water, energy, and land insecurities drive disparities in the determinants of health for human wellbeing and ecosystems [bib_ref] Assessing the State of the Water-Energy-Food (WEF), Mabhaudhi [/bib_ref] [bib_ref] A Nexus Approach to Water, Energy, and Food Security in Northern Canada, Natcher [/bib_ref] [bib_ref] Linking Reservoir Ecosystems Research to the Sustainable Development Goals, Guo [/bib_ref] [bib_ref] Sustaining the Ecological Functions of the Litani River Basin, Lebanon, Darwish [/bib_ref] [bib_ref] The Water-Energy-Food Nexus as a Tool to Transform Rural Livelihoods and Well-Being..., Mabhaudhi [/bib_ref] , particularly, in rural, remote, and Indigenous contexts where the use and extraction of natural resourcescreate downstream health inequities. For example, Natcher and Ingram report incidences of higher rates of water, energy, and food (WEF) insecurity in rural and remote communities in Northern Canada, where residents struggle with wide disparities in access to safe drinking water, sanitation services, as well as affordable energy and nutritious food [bib_ref] A Nexus Approach to Water, Energy, and Food Security in Northern Canada, Natcher [/bib_ref]. [fig_ref] FIGURE 2 |: Resource dynamics of land, water and energy at a watershed scale [/fig_ref] depicts examples of land-water-energy interactions unfolding within watersheds, which serve as resource nexus hotspots [bib_ref] Beyond Zero Sum Game Allocations: Expanding Resources Potentials through Reduced Interdependencies and..., Mohtar [/bib_ref] , as multi-sectoral and multi-stakeholder spaces [bib_ref] Sustainable Development as the Ultimate Target of Adopting a Nexus Approach to..., Caucci [/bib_ref] and as sites of interdependent land, water and energy resource insecurities [bib_ref] Sustainable Development as the Ultimate Target of Adopting a Nexus Approach to..., Caucci [/bib_ref]. Issues of resource insecurities associated with the nexus of land, water, and energy illustrate why it is impossible to consider one dimension without taking into consideration the others. One example is derived from the Kootenai River basin, a transboundary river bordering some parts of British Columbia in Canada and the United States. Complex systems linkages of land, water and energy resources take shape through intensive resource development activities such as agricultural expansion, dam construction and hydropower generation and open pit coal mining. Land, water, and energy are inputs into these resource activities, and efforts to address challenges in one sector impacts as well as produces emergent issues in other sectors [bib_ref] A Nexus Approach to Water, Energy, and Food Security in Northern Canada, Natcher [/bib_ref] [bib_ref] Beyond Zero Sum Game Allocations: Expanding Resources Potentials through Reduced Interdependencies and..., Mohtar [/bib_ref]. Hence, nexus actors are compelled to make an increasing number of tradeoff decisions that produce health externalities and present communities with conflicts that impact health outcomes. A few studies [bib_ref] Public Health Guide to Field Developments Linking Ecosystems, Environments and Health in..., Buse [/bib_ref] [bib_ref] Assessing the State of the Water-Energy-Food (WEF), Mabhaudhi [/bib_ref] [bib_ref] A Nexus Approach to Water, Energy, and Food Security in Northern Canada, Natcher [/bib_ref] [bib_ref] The Water-Energy-Food Nexus as a Tool to Transform Rural Livelihoods and Well-Being..., Mabhaudhi [/bib_ref] traced these health conflicts to resource insecurity challenges with concerns about disproportionate environmental burdens and the uneven distribution of health risks and environmental exposures in social groups (humans and biodiversity) with pre-existing vulnerabilities. Some studies [bib_ref] A Nexus Approach to Water, Energy, and Food Security in Northern Canada, Natcher [/bib_ref] [bib_ref] Waterenergy-food Nexus: a Platform for Implementing the Sustainable Development Goals, Stephan [/bib_ref] employed an environmental justice lens to analyze socioenvironmental conflicts emanating from trade-off interactions among land, water, energy resources and explored gendered and intersectional implications of resource use among social groups. These analyses dovetail with social, economic, and environmental determinants of health and health equity at a watershed scale and raise questions about the use of these resources in terms of: "Who has access to the resources? For what purposes? At what cost? With what impacts? And who bears the burden of the impacts? [bib_ref] Waterenergy-food Nexus: a Platform for Implementing the Sustainable Development Goals, Stephan [/bib_ref]."
In line with the socio-environmental lapses discussed above, and depicted in [fig_ref] FIGURE 2 |: Resource dynamics of land, water and energy at a watershed scale [/fig_ref] , many of the studies [bib_ref] Managing Water, Soil and Waste Resources to Achieve Sustainable Development Goals: Monitoring..., Hülsmann [/bib_ref] [bib_ref] Navigating towards Sustainable Development: A System Dynamics Approach, Hjorth [/bib_ref] [bib_ref] Toward Understanding the Convergence of Researcher and Stakeholder Perspectives Related to Water-Energy-Food..., Daher [/bib_ref] [bib_ref] From a Few Security Indices to the FEW Security Index: Consistency in..., Venghaus [/bib_ref] [bib_ref] Linking Environmental Policy Integration and the Water-Energy-Land-(Food-)Nexus: A Review of the European..., Venghaus [/bib_ref] [bib_ref] Complexity versus Simplicity in Water Energy Food Nexus (WEF) Assessment Tools, Dargin [/bib_ref] [bib_ref] Gateway to the Perspectives of the Food-Energy-Water Nexus, Proctor [/bib_ref] [bib_ref] Global Governance and the SDGs, Ivanova [/bib_ref] [bib_ref] Structuring an Integrated Water-Energy-Food Nexus Assessment of a Local Wind Energy Desalination..., Serrano-Tovar [/bib_ref] [bib_ref] Water and Food: towards a Critical Nexus Approach, Leck [/bib_ref] pointed to a lack of attention within existing nexus methods to systematically unpacking synergies and trade-offs between social and environmental issues across a range of contexts and scales. Our findings [fig_ref] FIGURE 4 |: A framework for addressing knowledge gaps and cross-sectoral health equity impacts of... [/fig_ref] , characterize these as scalar and theoretical lapses and link them to a lack of attention to decolonial perspectives that are needed to address impacts on those rural, remote, and Indigenous communities most affected by resource development activities. We underscore that the sustainability framework of the SDG agenda can contribute to the nexus agenda by addressing sustainability dimensions emerging between social and environmental domains through engaging principles of integration such as the notions of intergenerational equity, environmental protection, and the linking of the economic, social, cultural, and environmental dimensions of sustainability.
When considering the cross-sectoral relevance of the SDG's, our review identifies delays and barriers to engagement within rural and Indigenous contexts which also further entrench continued epistemological divides between western and Indigenous conceptualizations of sustainability and development [bib_ref] Watershed Management and Public Health: An Exploration of the Intersection of Two..., Bunch [/bib_ref]. A pertinent question is whether (and how) the SDGs can help to bridge the two ideologies and inculcate Indigenous Peoples' selfdetermining goals, socio-environmental values, ways of knowing and being into nexus issues. Findings presented in Mapping Evidence on Cross-Sectoral Potential of the Sustainable Development Goals and Implications for Advancing Health Equity section, illustrate ways that the SDGs have the potential to offer an integrative socio-ecological framework that can accommodate the strengths and limitations of different knowledge domains, approaches, and perspectives, while also offering a cross-sectoral platform that can foster communication and co-production among diverse actors and interest groups [bib_ref] Toward Understanding the Convergence of Researcher and Stakeholder Perspectives Related to Water-Energy-Food..., Daher [/bib_ref].
Finally, studying the links between the land-water-energy nexus and SDGs indicators underscores the importance of future studies exploring the cross-sectoral impacts on health equity [bib_ref] Assessing the State of the Water-Energy-Food (WEF), Mabhaudhi [/bib_ref] [bib_ref] Waterenergy-food Nexus: a Platform for Implementing the Sustainable Development Goals, Stephan [/bib_ref]. Connecting SDG indicators relating to land, water and energy, and making clear links to health indicators at a watershed scale are areas for future work that highlight the potential value of an SDGsdata-driven approach within watersheds that centres health equity in this nexus. These future studies will also need to address datarelated challenges at a watershed scale [bib_ref] Using Watershed Boundaries to Map Adverse Health Outcomes: Examples from Nebraska, USA, Corley [/bib_ref] [bib_ref] A Typology of Reviews: an Analysis of 14 Review Types and Associated..., Grant [/bib_ref]. The land-waterenergy-health equity nexus has the potential to be strengthened through increased emphasis on ecosystems services and related pathways [bib_ref] Sustainable Development as the Ultimate Target of Adopting a Nexus Approach to..., Caucci [/bib_ref] by which the nexus of land, water and energy have impacts on livelihoods, human well-being and the other species that depend on these services [bib_ref] Linking Reservoir Ecosystems Research to the Sustainable Development Goals, Guo [/bib_ref] [bib_ref] Global Governance and the SDGs, Ivanova [/bib_ref]. This emphasis may also enhance the emergence of community-based application of the SDGs focused on this nexus [bib_ref] Public Health Guide to Field Developments Linking Ecosystems, Environments and Health in..., Buse [/bib_ref] [bib_ref] Assessing the State of the Water-Energy-Food (WEF), Mabhaudhi [/bib_ref] [bib_ref] A Nexus Approach to Water, Energy, and Food Security in Northern Canada, Natcher [/bib_ref] [bib_ref] Sustaining the Ecological Functions of the Litani River Basin, Lebanon, Darwish [/bib_ref] [bib_ref] The Water-Energy-Food Nexus as a Tool to Transform Rural Livelihoods and Well-Being..., Mabhaudhi [/bib_ref] [bib_ref] Waterenergy-food Nexus: a Platform for Implementing the Sustainable Development Goals, Stephan [/bib_ref] , with the potential to inform decisions around land, water, and energy insecurities as well as how these underlie the social and ecological determinants of health inequities. The land, water, energy nexus can also be applied to analyze specific resource issues such as forestry, hydropower [bib_ref] Linking Reservoir Ecosystems Research to the Sustainable Development Goals, Guo [/bib_ref] , and desalination [bib_ref] Water and Food: towards a Critical Nexus Approach, Leck [/bib_ref] , with an emphasis on understanding impacts on health outcomes, such as the distribution of the impacts of socioecological determinants of health. If future research is to be effective in supporting greater equity and sustainability within nexus systems of land, water, and energy [bib_ref] Waterenergy-food Nexus: a Platform for Implementing the Sustainable Development Goals, Stephan [/bib_ref] , the scale and geography of the nexus will be important to consider along with the socio-political nuances and particular socio-ecological dimensions contouring these settings.
## Strengths and limitations of the study
The findings from this review need to be considered with reference to both limitations and strengths. One limitation is the use of only studies published in English and between 2016 and Pertaining to the interests of this research in how indigenous contexts have been advanced into land-water-energy nexus programming, the term-"Indigeneity"-was employed as a keyword along with other keywords introduced in the search strategy to explore how nexus approaches have integrated Indigenous ways of knowing and being. "Indigeneity"-as a keyword-was broken into search terms that included a range of possible synonyms in the literature: (indigen* OR decoloniz* OR aboriginal OR "* ecological knowledge" OR "first nations" OR metis OR Inuit OR "native people"). However, combining these "Indigeneity" search terms with the search terms of other keywords used in the search strategy, using "AND", returned "0" for most of the database searches, except for Google scholar, which returned a few relevant articles. One interpretation of this outcome is that this indicates limited research conducted within the nexus domain that considers the integration of Indigenous socio-environmental outcomes and self-determination goals. To ensure uniformity of keywords considered across the search databases, "Indigeneity" and synonyms were explicitly removed from the search terms but considered as an implicit inclusion criterion.
# Conclusion
There is a wealth of overlapping literature [bib_ref] Managing Water, Soil and Waste Resources to Achieve Sustainable Development Goals: Monitoring..., Hülsmann [/bib_ref] [bib_ref] Advancing the Implementation of SDGs in Brazil by Integrating Water-Energy Nexus and..., Carvalho [/bib_ref] [bib_ref] Sustainable Development as the Ultimate Target of Adopting a Nexus Approach to..., Caucci [/bib_ref] [bib_ref] Water and Food: towards a Critical Nexus Approach, Leck [/bib_ref] on how prioritizing a focus on the nexus of land, water and energy systems can accelerate progress in meeting the SDGs. The water-energy-land nexus offers a potential tool for centering health across scales and contexts, in keeping with the WHO's Shanghai's Declaration on leveraging interactions to promote health in the SDGs [bib_ref] One behind: A UNSDG Operational Guide for UN Country Teams, Unsdg [/bib_ref]. However, several existing nexus approaches do not actively foster cross-sectoral integration [bib_ref] Linking Environmental Policy Integration and the Water-Energy-Land-(Food-)Nexus: A Review of the European..., Venghaus [/bib_ref] [bib_ref] Complexity versus Simplicity in Water Energy Food Nexus (WEF) Assessment Tools, Dargin [/bib_ref] , system thinking, transdisciplinarity [bib_ref] Managing Water, Soil and Waste Resources to Achieve Sustainable Development Goals: Monitoring..., Hülsmann [/bib_ref] [bib_ref] Waterenergy-food Nexus: a Platform for Implementing the Sustainable Development Goals, Stephan [/bib_ref] , nor place-based considerations in support of promoting the equitable integration of multiple perspectives, knowledges, and needs within decision-making and policy outcomes [bib_ref] From a Few Security Indices to the FEW Security Index: Consistency in..., Venghaus [/bib_ref]. This review has addressed an important knowledge gap by mapping interlinkages of land, water, energy, and health equity within the SDGs at a watershed scale. It has characterized the cross-sectoral potential of the SDGs to advance health equity within the land, water, energy nexus, and illustrated how integration lapses may compromise the capacity of the land-water-energy nexus to address health equity considerations in rural, remote, and Indigenous contexts. Our review [bib_ref] Beyond Zero Sum Game Allocations: Expanding Resources Potentials through Reduced Interdependencies and..., Mohtar [/bib_ref] also identifies the SDGS as a promising driver for progressing the nexus integration agenda to foster sensitivity to health equity within the nexus of land, water and energy. It focuses on the cross-sectoral potential of the SDGs to foster an appreciation of the impacts of the land water, energy nexus on health outcomes, such as the distribution of the socio-ecological determinants of health. For the purpose of monitoring health equity in the context of interactions of land, water and energy systems, this paper generates evidence to inform health equity-oriented policies, programs, and practices, and to enhance health for equity-seeking populations. Our work identifies further research needs to address knowledge gaps regarding health equity and the SDGs at the scale of watersheds, including closer attention to unmet needs of equity-seeking populations, priorities for Indigenous communities, and a closer focus on health equity as an integral dynamic within the land-water-energy nexus.
# Author contributions
All Authors contributed to conception and design of the study. CO conducted the literature search for the mapping review with input from the university library. CO led the thematic analyses with input from NA, MG, HH, and MP. CO wrote the first draft of the manuscript with input from MP, and contributions on specific sections from NA, MG, and HH. All authors contributed to manuscript revision, read, and approved the submitted version.
[fig] FIGURE 1 |: Prisma Flow Chart: A flow chart representing the study selection and screening steps Creative by DistillerSR (c) (Prince George, Canada, 2021). [/fig]
[fig] FIGURE 2 |: Resource dynamics of land, water and energy at a watershed scale (Prince George, Canada, 2021). [/fig]
[fig] FIGURE 3 |: Schematic diagram of interactions among land, water and energy systems, pathways to health equity impacts and corresponding implicated targets within the Sustainable Development Goals' framework (Prince George, Canada, 2021). Public Health Reviews | Owned by SSPH+ | Published by Frontiers May 2022 | Volume 43 | Article 1604362 8 [/fig]
[fig] FIGURE 4 |: A framework for addressing knowledge gaps and cross-sectoral health equity impacts of the Land, Water, Energy nexus at the Watershed Scale (Prince George, Canada, 2021). Public Health Reviews | Owned by SSPH+ | Published by Frontiers May 2022 | Volume 43 | Article 1604362 [/fig]
[table] TABLE 1 |: Key characteristics of identified Journal articles and Grey Literature (Prince George, Canada, 2021). Description: Key connections with Land-Water-Energy nexus and the SDGs, and examples of the pathways and relationships connecting the SDGs to the Social determinants of Health and Equity. [/table]
[table] TABLE 2 |: Characterizing Nexus-informed Cross-sectoral potential of the Sustainable Development Goals and Implications for Advancing Health Equity (Prince George, Canada, 2021). [/table]
|
Illicit Tobacco in Lithuania: A Cross-Sectional Survey
Taxation policies are the most cost-effective measure to reduce overall tobacco consumption. However, cigarettes in Lithuania are among the cheapest in the European Union. The threat of the illicit trade is often used to compromise evidence-based policies, pricing policies particularly. The aim of this study was to determine the extent of illicit cigarette consumption in Lithuania and identify the main characteristics of illicit cigarette smokers. The national cross-sectional survey with direct observation of the latest purchased pack of cigarettes was conducted between August and September 2019. In total, 1050 smokers aged ≥18 were interviewed face-to-face. The illicit share of the total consumption of cigarettes per year was 10.7% with 9.7% of smokers showing or describing illicit cigarette packs compared to 17% reported by industry-funded studies. Older smokers, smokers with lower education and heavy smokers were more likely to regularly purchase illicit cigarettes. The average price of an illicit pack was almost two times lower than licit. Although the illicit trade of tobacco products is a serious policy challenge, the threat of an increase in illicit trade should not delay tobacco taxation improvements. Author Contributions: Conceptualization, V.L.-G., J.V., L.G., J.P., and M.Š.; methodology, V.L.-G. and L.G.; formal analysis, J.V., J.P., and V.L.-G.; data curation, J.V. and M.Š.; writing-original draft preparation, V.L.-G., J.V., and L.G.; writing-review and editing, J.P., M.Š., L.M., and A.M.; supervision, M.Š. All authors have read and agreed to the published version of the manuscript.
# Introduction
World Health Organization (WHO) defines illicit trade in tobacco products as any practice or conduct relating to the production, shipment, receipt, possession, distribution, sale, or purchase that is prohibited by law and includes any practice or conduct intended to facilitate such activity. It is estimated that the global illicit cigarette market accounts for about 11.6%, resulting in approximately $40.5 billion USD tax revenue lost per year.
Illicit cigarette trade poses a serious threat to public health because it undermines tobacco control policies and government's authority; causes substantial loss in tax revenues; increases availability of tobacco products and results in cheaper prices, which can increase consumption; contributes to the funding of international criminal activities; allows tobacco industry (TI) to engage with policy makers and helps them to get media attention; and last, but not least, helps them to be perceived as a legitimate stakeholder in a policy debate.
Lack and even non-existence of public transparent data on illicit tobacco is a global challenge for tobacco control research. The nature of the illicit trade makes it hard to monitor and track; therefore, reliable data is often very scarce. Such circumstances give the opportunity to the TI to fill existing gaps with their conducted or commissioned research on the topic. Though such research is long being questioned in Europe, the United States, Asia, Australia, Brazil, and South Africa, TI is still a major funding source of the data on illicit trade. Philip Morris International (PMI) alone has pledged $48 million USD to fund research on illicit tobacco trade and granted 61 projects worldwide in 2016-2017. This includes two projects in Lithuania. Given that many countries lack independent data on the illicit tobacco trade, an urgent need for counterbalance is evident.
In Lithuania, there is no regular monitoring of the extent of illicit trade, apart from the data on the seizure of cigarettes provided by enforcement authorities. Local policy makers, following the example of the European Commission, frequently use data compiled by the Klynveld Peat Marwick Goerdeler International Cooperative (KPMG), which has a longstanding relationship with the TI . KMPG reports have always listed Lithuania among countries with the highest prevalence of illicit cigarettes. According to the latest (2018) report, Lithuania has one of the highest illicit cigarette consumption rates (17%) in the European Union (EU). Meanwhile, The Pricing Policies and Control of Tobacco in Europe Project (PPACTE) found that KPMG overestimated illicit trade in 11 out of 16 analysed countries.
It is known that illicit trade is more frequent in countries with a land or sea border with Non-European Economic Area countries. Besides, Ukraine, Russia, Moldova, and Belarus are considered to be major suppliers of cheap and illicit cigarettes in Europe. Lithuanian border with Belarus and Russia accounts for 60% of state borders. Due to the favourable geographical and economic (price differences in the region) conditions, Lithuania is a country of transit, transhipment, temporary storage, and preparation for further shipments of the contraband cigarettes to Western Europe. Together with Latvia, Croatia, Greece, and Bulgaria, Lithuania is listed among countries where illicit cigarettes have the highest penetration. This adds to the evident urgent need of methodologically transparent independent estimates of illicit tobacco in Lithuania.
The aim of this study was to determine the extent of illicit cigarette consumption in Lithuania using direct methods and identify the main characteristics of illicit cigarettes smokers.
# Materials and methods
## Study sample and data collection
The survey was performed by the market research company Rotelas S.A. Face-to-face interviews including an on-site cigarette pack observation were conducted in the households between August and September 2019. A multistage, stratified, proportional sampling method was used. The sample was designed to represent the smoking population aged 18 and over in Lithuania. The sample size of 1000 and the ratio of men and women (3:1) in the sample were based on the research data of smoking prevalence in Lithuania. The survey covered the whole country and was performed in all ten administrative regions of Lithuania. The number of respondents in each region was calculated according to national population statistical data, taking into consideration the population density in the regions. Furthermore, the sample was proportional to the number of people living in big cities, cities, and villages. When calculating the sample size, sampling design was taken into account.
In each location, where the study was performed, the multiple starting points were randomly selected. The number of starting points depended on the size of the location. In total, 105 starting points were selected. The interviewer would visit every third house. After visiting every third house in the street, they would turn right, and again would visit every third house. From each house, the interviewer would choose one stairwell, usually the first (or where one would manage to enter) and visit every third apartment (household). On average, 10 smokers were interviewed in one cluster.
Within each randomly selected household, the interviewer informed the first adult individual about the purpose of the study, and all current adult smokers were identified. If no adult smokers resided, the interviewer continued to the next household. Otherwise, an adult smoker who had the most recent birthday was selected to complete the questionnaire. Only one individual per household was sampled. In total, 64% of selected households participated in the survey.
In total, 1050 smokers participated in the survey. The characteristics of the participants are presented in. According to the Order V-28/2016 of the Minister of Health on the procedure for obtaining the informed consent, written consents are only required for the surveys falling under the biomedical research category. Additionally, according to the Bioethics Center at Lithuanian University of Health Sciences, if the questionnaire is fully anonymous, respondents are not identified, and if objective health information is not collected, written consent is not required.
## Questionnaire
The Questionnaire consisting of 17 questions was designed by the researchers of Health Research Institute and was compiled on the basis of previously published smoker survey. Socio-demographic data collected included age, gender, highest education received, area of residence, and current average monthly income. Respondents were classified into two age groups using the median age as the cut-off point (median age = 44). Respondents were divided into higher and lower income groups according to their average monthly income being below (or equal) or above national minimum net wage in Lithuania (€395 as of 2019).
The use of tobacco products was determined using question: "Please list all tobacco or related products that you used during the past 12 month", with answer options: (1) cigarettes, (2) e-cigarettes, (3) heat-not-burn tobacco products, (4) roll-your-own cigarettes, (5) hookah, (6) cigars and cigarillo, (7) chewing or snuff tobacco, (8) other (specify).
The smoking intensity was measured with the question: "How many cigarettes do you usually smoke per day?" Smokers were divided into two groups by the number of smoked cigarettes: light smokers (1-19 cig/day) and heavy smokers (≥20 cig/day). Smoking intensity classification was based on previously published literature.
Purchasing illicit tobacco was inquired with the question: "Have you ever bought illicit tobacco products?" Possible answer choices were the following: (1) no; (2) yes, but only sometimes; (3) yes, regularly. Respondents were also asked: "Do you know where you can buy illicit tobacco products?" Answer options were the following: (1) I do not know; (2) I do not know, but it is easy to find out; (3) yes, I know specific places or people I could buy from. The opinion about illicit trade was determined by the question: "What is your overall opinion on the purchase of illicit tobacco products?", with four response options: (1) positive, I do not see anything wrong with that; (2) negative, but I understand why people buy and use illicit tobacco products; (3) very negative, I would never buy illicit tobacco products myself; (4) I have no opinion.
## Identification of illicit packs
Identification of illicit cigarette packs was based on the information provided by smokers who either agreed to show the latest pack purchased to the interviewer or provided information about it if they did not have the pack or did not agree to show it.
In both cases, interviewers filled in observational questions to record pack characteristics: (1) the package type (standard 20 cigarette package, roll your own tobacco package, e-device package);
(2) health warnings (in Lithuanian, in a foreign language, no health warnings); (3) tax stamps (in Lithuanian, in a foreign language, too damaged to identify language but present, no tax stamp/duty-free pack); (4) price of the pack; and (5) type of outlet/source from where the pack was purchased. A pack could be bought from: (1) national legal shop, kiosk, or catering company;
(2) specialized tobacco shop; (3) over the internet; (4) from legal shops in other countries; (5) from duty-free shops; (6) fom individuals selling cigarettes independently at local markets; (7) from individuals when meeting at the time and place agreed; (8) delivered to the house; (9) offered by peers, friends. All abovementioned characteristics later helped to indicate if the pack was illicit.
Following the existing practice from previous study, pack was classified as illicit if it had at least one of the four characteristics: (1) pack had no appropriate health warnings-health warning was provided in a foreign language or was not existent (unless the pack had been bought in other country or in a duty-free shop); (2) pack had no appropriate tax stamp-another language than Lithuanian on a tax stamp (unless the pack had been bought in other country or in a duty-free shop), but if the stamp was torn off or damaged it was not considered as illicit; (3) pack with a price lower than the 70% of the lowest price of cigarettes (€3.15) as listed in the WHO country profile 2019(if not bought over the internet, in other countries or in duty-free shops or offered); (4) pack was bought from an illicit source. In this study, packs purchased over the internet, from individuals at local markets, place agreed or delivered to the house were considered as illicit packs. Regarding the price for a pack of cigarettes, items costing €2.21 and less were considered illicit.
## Data and statistical analysis
Statistical data analysis was performed using the statistical package IBM SPSS Statistics for Windows, Version 20.0 (2011 version, IBM Corp, Armonk, NY, US) and Microsoft Excel 2019 (Microsoft, Redmond, WA, US). Statistical significance level (p values) of 0.05 was chosen to test the hypotheses. The qualitative variables were presented as percentages. Differences in qualitative variables were compared using Chi-squared (χ2) and Z test with Bonferroni correction. Student's t-test was used to compare the mean values of a number of cigarettes smoked per day.
Multivariable logistic regression analysis was used to assess the associations of self-reported buying of illicit cigarettes with socio-demographic variables.
The following criteria were taken into account when approximating the number of illicit cigarettes smoked per year: (1) respondent was smoking cigarettes; (2) respondent reported the number of cigarettes smoked per day; (3) respondent indicated whether he/she has ever bought illicit tobacco products; (4) respondent stated the proportion of illicit cigarettes smoked in the last 12 months. Number of cigarettes smoked per year was calculated by multiplying number of cigarettes smoked per day by 365. Then, number of illicit cigarettes smoked per year was calculated by multiplying number of cigarettes smoked per day by 365 and by the self-reported proportion of illicit cigarettes smoked in the last 12 months. In total, 731 cigarette smokers met abovementioned criteria.shows the characteristics of the study population. A total of 1050 smokers aged 18-82 years participated in the survey. The mean of respondents' age was 41.98 . The most of respondents were younger than 45 years, had higher than secondary education, were living in big cities and had monthly income higher than the minimum net wage. Unfortunately, not all the study respondents were willing to provide personal information about the age, education, or monthly income.
# Results
Of all current smokers, 73.7% reported smoking only cigarettes, 5% only e-cigarettes, and 10.7% only heat-not-burn tobacco products. Although cigarette use was more prevalent among men (75.8%), e-cigarettes and/or heat-not-burn tobacco products were more common among women (23.7%).shows the distribution of illicit tobacco purchase by socio-demographic characteristics. Compared to men, a significantly larger proportion of women responded that they had never bought illicit tobacco products. Besides, 13.4% of ≥45 years old individuals regularly buy illicit tobacco products, which is significantly higher proportion compared to 18-44 years old smokers. Respondents with secondary or lower education were more likely to buy illicit tobacco products on a regular basis. A larger proportion of people living in towns or villages regularly bought illicit tobacco products compared to respondents living in big cities in Lithuania. The purchase of illicit tobacco products was more common among lower-income respondents. Five percent of light smokers reported using illicit tobacco products regularly. However, among heavy smokers, the proportion was 3 times higher-15.3%.
The multivariable logistic regression analysis revealed that the odds of self-reported purchase of illicit tobacco products sometimes or regularly were more than two times higher for smokers with secondary or lower education compared to higher than secondary education. Likewise, heavy smokers were two times more likely to buy illicit tobacco products sometimes or regularly compared to light smokers. The odds of regular purchase of illicit tobacco products were 3.04 times higher for 45 and older smokers compared to younger smokers. Having secondary or lower education increased the likelihood of buying illicit tobacco products regularly four times compared to smokers with higher than secondary education. Heavy smokers were 3.61 times more likely to regularly buy illicit tobacco products than light smokers.
Illicit share of the total consumption of cigarettes is based on the self-reported intensity of smoking. The estimated proportion of illicit cigarettes was 10.7%. Smoking illicit cigarettes was more common among men. In addition, older age, smoking intensity, lower education, and income level, as well as living in a smaller living place have been associated with consumption of a larger proportion of illicit cigarettes (p < 0.001). Significantly more illicit cigarettes users than non-users (84.4% compared to 28.2%) knew where to acquire illicit tobacco products. Attitudes towards purchasing illicit tobacco products differed between those who do or do not purchase it themselves. Significantly fewer smokers of legal tobacco products found nothing wrong with purchasing illicit products (11.3% compared to 53.7%), whereas 3% of smokers who have purchased illicit cigarettes had very negative opinion on the illicit tobacco products compared to more than one third of legal tobacco smokers.
Among the 1050 respondents, the percentage of smokers who showed their packs was relatively low (35.2%, n = 370). More than third of respondents (39.5%, n = 415) agreed to provide self-reported information on the latest pack bought. Overall, 785 smokers have showed or described the latest pack; however, some packs were excluded from the analysis because either the respondent smoked only other types of tobacco products than cigarettes or did not provide information about all four pack characteristics. Thus, pack observation analysis was performed for 473 packs. As current cigarette smokers reported, 90.9% of those packs were bought from legal tobacco shop in Lithuania, 0.7% from a legal shop in another country, 0.4% from duty-free shops, 7.6% from illicit sources, and 0.4% were offered/gifted by others. Seven percent of smokers reported an extremely low price of the latest pack. The average price of a legal cigarette pack was €3.77 ± 0.39 (with lowest price €2.90 and highest €6.50). The average price of a self-reported illicit cigarette pack was €2.12 ± 0.92 (with lowest price €1.10 and highest €5.10). Furthermore, 9.1% smokers showed or reported packages with and inappropriate health warnings, 7.8%-inappropriate tax stamp. Overall, the prevalence of illicit packs was 9.7%. Most respondents who reported regularly buying illicit tobacco products and only few of those who bought them sometimes have shown or described illicit cigarette packs . . Proportion of smokers who showed or described illicit cigarette packs according to self-reported frequency of purchasing illicit tobacco products.
## Variables
Have
# Discussion
Lithuania has long been listed among countries with the highest prevalence of illicit cigarettes in the EU. To our knowledge, this paper is the first to provide industry-independent nationally representative estimates on the illicit cigarette market in Lithuania. The results of our survey indicate that illicit share of the total consumption of cigarettes per year was 10.7% with 9.7% of smokers showing or reporting illicit cigarette packs compared to 17% reported by industry-funded studies.
Smokers with lower education, 45 years old or older, and heavy smokers were more likely to regularly purchase illicit cigarettes. Additionally, older age, being male, smoking intensity, lower education, and income level, as well as living in a smaller living place have been associated with the consumption of a larger proportion of illicit cigarettes. The average price of illicit packs of cigarettes was almost two times lower than licit.
The illicit market of tobacco products in Lithuania has always been a heated argument between public health society and TI allies. Nevertheless, industry-independent studies on illicit tobacco in Lithuania are scarce. Most often estimates of the illicit trade are provided by TI-related and/or funded organizations (i.e., KPMG, Euromonitor, Transcrime, Lithuanian Free Market Institute).
Evaluation of industry-funded studies conclude that there are some major issues regarding the quality of the estimates, including lack of methodological transparency; problems with data collection, analytical methods and result interpretation; and that they are not peer-reviewed. Some studies show that industry-driven data exaggerate the extent of illicit trade. It appears that our findings support this observation.
As previously mentioned, the latest (2018) KPMG report shows that Lithuania had one of the highest shares (17%) of the counterfeit and contraband cigarettes in the EU. According to our estimates, share of illicit cigarettes is 1.5 times lower than industry estimates. Thus, a direct comparison of these estimates is very complex.
PMI IMPACT (an initiative of Philip Morris International Inc) funded survey performed by Sprinter and commissioned by the Lithuanian Free Market Institute showed that 17% of respondents self-reported purchase of cigarettes from illicit sources in the past year in 2018. Thus, we found that 11.6% of smokers reported purchasing illicit cigarettes only sometimes in their lifetime, and 6.8% reported doing so regularly. Again, it is hard to directly compare estimates, as methodological aspects of industry-funded studies mostly remain unknown.
According to another survey performed by Sprinter and commissioned by the State Tax Inspectorate, Lithuanians assume that the share of the illicit tobacco trade accounted for 16.3% in 2017. Our estimates are lower. Arguably, surveys that analyse attitudes and perceptions of citizens and do not ask about their specific smoking habits cannot be compared and should not be considered the same as methods directly evaluating the illicit share of the tobacco.
At the same time, our estimates of the share of illicit packages (9.7%) are comparable with the overall proportion (6.5%) of illicit cigarettes and hand-rolled tobacco in 18 European countries, as estimated in the PPACTE survey in 2010. Some previous research stated that 9.8% of cigarette consumption in high-income countries is illicit.
Findings from our study confirmed that, as in the previous studies, older smokers, smokers who smoke more intensivelyand have lower educationare more likely to use the illicit cigarettes.
Our analysis barely detected any cases of tax-avoidance. Only a few (1%) smokers in Lithuania reported buying tax-avoided cigarettes from duty-free shops or legal shops abroad. This is consistent with overall patterns in the EU, in an average EU Member State only a small portion of cigarette sales can be explained by cross-border purchasing; besides, in most EU countries only a small portion of all smokers reported frequent cross-border cigarette purchasing. Evidence suggests that smokers living in regions bordering countries where cigarettes were at least €1 per pack lower than in their home country had significantly higher odds of purchasing cheaper out-of-country cigarettes. However, due to the tax harmonization policy in the EU, the weighted average price of cigarettes in Lithuania, Latvia, and Poland are almost the same, €3.18, €3.20, and €3.26 per pack respectively. At the same time, in Russia and Belarus, cigarette prices are almost 3 times lower. Nonetheless, cross-border shopping even for the smokers living close to the physical borders with the later countries might be inconvenient due to required visas.
This study provides industry-independent estimates on the illicit cigarette consumption in Lithuania and is the first to counterbalance existing TI-related evidence. The results of this study may assist policy makers to revise tobacco taxation and approaches for controlling illicit tobacco trade. Another strength of the analysis is that we followed the recent practiceto validate self-reported information with observational pack data by asking to show the latest purchased cigarette pack and inspecting it.
Unfortunately, the percentage (35.2%) of smokers who showed their packs was low. Compared to other studies, the percentage was higher than in Mexico (29%)but smaller than in Turkey (76%), Georgia (71%), and average from PPACTE study of 18 European countries (74%). It is difficult to explain the reasons for such a low proportion of shown packages. Smokers' unwillingness to show pack may be related to the desire to complete the survey as soon as possible and simply choosing a more convenient option-to describe the package rather than to go and look for it. On the other hand, smokers who reported never purchasing illicit cigarettes were more likely to show the illicit pack. This might indicate that smokers who had illicit packs did not feel comfortable to admit it. In this case, we could have underestimated the proportion of illicit packs.
There is a long and well-described tradition of the TI to use the threat of the illicit trade to compromise such evidence-based policies as tobacco tax and price regulations, plain packaging, pictorial warnings, point-of-sale display bans, marketing regulations, flavour bans, pack size restrictions. TI in particular often claims that taxes on tobacco products drive the illicit trade, although there is a vast amount of evidence showing a limited relationship between higher tobacco prices and size of the illicit market. In fact, illicit trade can be minimized even in the presence of increased taxes if effective monitoring and enforcement is in place.
Furthermore, Shafferer et al.estimated that tax-induced price hikes of 10% would increase illicit cigarette use the most (by 11%) in Czech Republic, Latvia, Lithuania, Poland, and Slovakia out of the 36 European countries. At the same time, the consumption of licit cigarettes in these countries would be reduced by 18.4%. The average tax loss due to price increases in all the observed 36 European countries was estimated to account for about $377.46 million USD, whereas average total tax increases could be as high as $14.69 billion USD. This shows that tax losses would be marginal in comparison with the total gains in taxes.
Lowering overall demand is an evidence-based strategy of reducing illicit trade and taxation is one of the most effective measures to reduce demand. According to the European Commission, the weighted average price of cigarettes in Lithuania is still one of the lowest in the EU (€3.18 per pack; cheaper cigarettes can be found only in Bulgaria-€2.57). What is more, among 36 European countries, Lithuania scores 29th in the Tobacco Control Scale 2019. The Tobacco Control Scale report states that in Lithuania "fear of illicit supply of the cigarettes from neighbouring countries and tobacco industry pressure contribute to tax levels remaining low". Given the fact that tobacco taxation is the most effective measure to reduce smoking and the growth of the illicit market is the most common contra-argument, lack of reliable data on illicit tobacco allows speculative positions and threatens the country's efforts to reduce smoking.
The study was subject to limitations. First, the selected study design might have caused selection bias. However, the distribution of the selected smokers by sociodemographic characteristics did not differ from that in the Health Behaviour among Lithuanian Adult Population study where all participants were randomly selected from the National population registry. We did not account for clustering in the data analysis. Therefore, some p-values might be too low, and the risk of a false-positive error increased. However, a high number of clusters and a small number of respondents in the clusters, as well as a long distance between households where smokers were interviewed might have reduced both the interaction between respondents and the cluster effect.
Second, our survey provides most estimates based on self-reporting, and consequently likely to be under-estimated. However, direct observation of smokers' packs or asking concrete questions about the pack is likely to elicit responses that are both honest and less subject to respondent error. There may be errors when respondents do not accurately remember the price paid or other details.
Smokers may not answer truthfully about the number of cigarettes smoked. This is especially valid if smokers are questioned in the household where other family members might be present. Following the example of other studies, the latter situations might be avoided if smokers are questioned in the street.
Third, we assume that the latest pack purchased reflects characteristics of the usual pack. This assumption is standard and relies on smokers' loyalty to particular brands. Nevertheless, in the future surveys it is possible to ask to show all packages that smokers have at home. Besides, the previous study in Poland, showed that if asked to show all packs in their possession, smokers show both, taxed and non-taxed packs.
Fourth, we cannot differentiate between packs intended and not intended for the domestic market. Fifth, our data on the last purchased pack did not include a brand name, which could have allowed us to better validate self-reported information and identify "cheap whites" in the Lithuanian market. Sixth, the relatively small size of the database and sometimes lack of complete survey response could have had an impact on the results. More extensive data could have enabled a more rigorous statistical analysis to have been undertaken from which more reliable conclusions could have been generated. The current study deserves further confirmatory investigations with a major focus on data collection.
Survey of smokers' packs could be improved by increasing the sample size, assuring better training for interviewers, and having the possibility to take pictures of the pack that actually would strengthen cross-validation of data.
The International Agency for Research on Cancer identified three methods that are most used to measure illicit trade: (1) comparison of tax paid sales and individually reported consumption measures, also called gap analysis; (2) survey of tobacco users' purchase behaviours; and (3) observational data (empty packs) collection. This paper describes only one of the methods used to determine the illicit trade of cigarettes, though some authorssuggest that the most optimal practice for researchers could be cross validation of estimates using both, survey and the littered pack survey.
Due to the low number of packs observed, there was no comprehensive analysis of regional differences in illicit cigarettes consumption. Previous research shows that smokers in border regions are more likely to engage in cross-border purchases from neighbouring countries with lower tobacco prices. Similarly, more illicit cigarettes are found in the living places close to borders. Current evidence does not allow us to evaluate if illicit packs were more common in the areas closer to borders.
Therefore, further independent, and comprehensive data on illicit tobacco could address the data gaps by applying mixed methods, providing data across population groups and geographies, as well as estimate taxes lost through illicit trade. Besides, the survey showed that the proportion of smokers who smoke only e-cigarettes or heat-not-burn cigarettes, as well as the proportion of dual smokers, is relatively high. Further research on the consumption of other tobacco products is needed.
# Conclusions
This study provides TI-independent and methodologically transparent estimates of illicit trade in Lithuania, showing the lower share of illicit trade than reported by industry-funded studies. Although the illicit trade of tobacco products is a serious policy challenge in Lithuania, the threat of an increase in illicit trade should not delay reinforcement of the implementation of the WHO recommended tobacco control policies (MPOWER), and tobacco taxation policies in particular. Further research employing mixed methods would be needed to better assess the magnitude of the illicit tobacco trade. |
Occupational cancer in central European countries.
The countries of central Europe, including Poland, the Czech Republic, Slovakia, Hungary, Romania, and Bulgaria, suffer from environmental and occupational health problems created during the political system in place until the late 1980s. This situation is reflected by data on workplace exposure to hazardous agents. Such data have been systematically collected in Slovakia and the Czech Republic since 1977. The data presented describe mainly the situation in the early 1 990s. The number of workers exposed to risk factors at the workplace represent about 10% of the working population in Slovakia and 30% in Poland. In Slovakia in 1992 the percentage of persons exposed to chemical substances was 16.4%, to ionizing radiation 4.3%, and to carcinogens 3.3% of all workers exposed to risk factors. The total number of persons exposed to substances proven to be carcinogens in Poland was 1.3% of the employees; 2.2% were exposed to the suspected carcinogens. The incidence of all certified occupational diseases in the Slovak Republic was 53 per 100,000 insured employees in 1992. Cancers certified as occupational cancers are skin cancer caused by occupational exposure to carcinogens, lung cancer caused by ionizing radiation, and asbestosis together with lung cancer. Specific information on occupational cancers from Romania and Bulgaria was not available for this paper. It is difficult to predict a trend for future incidences of occupational cancer. Improved control technology, governmental regulatory activity to reduce exposure, surveillance of diseases and risk factors, and vigilant use of preventive measures should, however, ultimately reduce occupational cancer -Environ Health Perspect 107(Suppl 2): 279-282 (1999). http.//ehpnet1.niehs.nih.gov/ docs/1999/Suppl-2/279-282fabianova/abstract.html
The countries of central Europe, including Poland, the Czech Republic, Slovakia, Hungary, Romania, and Bulgaria, are currently in a transition period with serious economic problems. These countries suffer from many environmental and occupational health problems created during the previous political system. The situation of the work environment is unsatisfactory because of the economic recession accompanied by insufficient resources for improvement of old-fashioned technologies and replacement of worn-out equipment. This unfavorable situation is reflected in the data on exposures to hazardous agents in the workplace. Generally, the legislation in these countries has followed the recommendations of international agencies such as the International Labor Organization and the World Health Organization. For this article, no specific information was available from Romania and Bulgaria, and only sparse information from Hungary could be included. Parts of the present article have been published previously [bib_ref] Occupational cancer in central Europe, Fabianova [/bib_ref].
## Exposure to occupational carcinogens
Data on workplace exposure to noxious substances and factors have been collected systematically, with yearly updating, in the former Czechoslovakia since 1977; this practice has been continued in the Czech Republic and Slovakia. In 1992, the total number of workers classified as exposed to any occupational hazard in Slovakia was 193,000, representing approximately 10% of the working population. Of these workers, 4.3% were classified as exposed to ionizing radiation and 3.3% were exposed to chemical carcinogens . In 1991 about 32% of all employed persons, or 1,753,000 workers were classified as being exposed to occupational hazards in Poland. Of these, 5% were exposed to chemical carcinogens .
In the former Czechoslovakia, a computerized system of classification of hazardous working operations had been established, which is still in operation in the Czech Republic and in Slovakia. The system was based on data obtained from monitoring hazardous chemical, physical, and biologic agents in the workplace. Of four categories, workplaces with exposure to chemical factors were classified in categories 3 and 4, depending on the concentration of harmful chemicals in the working environment (up to 50% higher than maximum allowable concentration for category 3 and even higher for category 4). The sparse data available on estimates of the number of workers exposed to specific known or suspected carcinogens are summarized in . For Slovakia, the agents with the highest prevalence of exposure are tar and other polycyclic aromatic hydrocarbon (PAH) mixtures (1,937 exposed in total, 199 women), cytostatics (1,407 exposed, 897 women), and vinyl chloride (800 exposed, 96 women). Only 376 workers (78 women) were classified as being exposed to asbestos, which is likely to represent a gross underestimation. The total number of workers exposed to carcinogens was estimated to be 6,792 in 1988 (3,981 females), (1,444 females) and . In 1997 the total number of workers exposed to carcinogens had decreased to an estimated 3,845. The most common carcinogens are now cytostatics, asbestos, PAH, vinyl chloride, and arsenic. Most recently, a new hygienic code with an enlarged list of carcinogens based on the International Agency for Research on Cancer (IARC) documents is being prepared for approval by the state authorities.
In the Czech Republic 2,152 workers, 52 of whom were women, were classified as exposed to known carcinogens. An additional 7,300 workers were employed in occupations and industries entailing exposure to carcinogens; a further 5,800 workers were classified as exposed to suspected carcinogens, among whom the largest group comprised 2,300 health workers exposed to cytostatics.
The carcinogenic factors exist: a) the list of definite carcinogens (41 chemical substances, 1 physical factor, 2 biologic factors, and 11 production processes), and b) the list of possible carcinogens (47 chemical substances and 1 production process).
The rules of exposure monitoring in the factories and enterprises in Poland are regulated in the ordinance of the Council of Ministers (2). According to this ordinance, each entrepreneur is obliged at the start of the activity to send information concerning the manufacturing process to the State Sanitary Inspection and the State Labor Inspection. On the basis of this information, the State Sanitary Inspector determines the list of factors to be monitored, the frequency of measurements, and the mode of sample collecting. The entrepreneur is obliged to submit the results to the State Sanitary Inspection and the State Labor Inspection. Cancer Registration Reporting of cancer incidence has been compulsory in Poland since 1952, and in the former Czechoslovakia, Hungary, and Romania since 1964. The completeness of the registration, however, was questionable in most regions in these countries for a long period of time. In Poland four regional cancer registries were established during the 1960s. It was estimated that in the mid-1980s only 70% of new cancer cases were registered. Later, the completeness 280 improved and registration of new cases was estimated to be 95% in 1993. The descriptive epidemiologic data on cancer incidence and mortality are now published annually, allowing detection of trends in incidence. In the former Czechoslovakia, underestimation of cancer incidence resulted in the founding of two cancer registries in 1976 covering the later Czech Republic and the Bratislava region in later Slovakia, respectively. The Bratislava registry was expanded in 1980 to cover all Slovakia. The registration is obligatory and the quality has been high, with an estimated completeness of 95%. Certain regions in Romania and Hungary had cancer registries of reasonable quality during the 1970s and 1980s.
## Compensation of occupational cancer
The occurrence of occupational diseases and poisonings is an important index of the standard of working conditions. In Slovakia in 1992 there were a total of 1,056 cases of newly diagnosed occupational diseases and poisonings, of which 40% were in women. This figure is 16% lower than the corresponding figure for 1991. The total number of occupational diseases hasbeen dedining further in the last 5 years. In 1997 only 697 new cases were counted. This decline is most probably related to the disintegration of the dominantly curative occupational health services in many plants and the slow creation of the new preventive occupational health services. In the Czech Republic, the number of recognized occupational diseases was 8603 in 1991.
As for cancer, only skin cancer, lung cancer from radioactive materials, and lung cancer following asbestosis, are certified as distinct entities in the Czech Republic and in Slovakia. Other cases of occupational cancer caused by chemical exposures may be recognized and included in the group of poisonings; their number is not known. The yearly increase in incidence of occupational cancer in former Czechoslovakia was from 0 to 2 cases of skin cancer , and 66 to 1 15 cases of lung cancer from ionizing radiation . The only sizeable group of compensated cancers was lung cancer from radioactive materials, mainly among uranium and metal miners; in the Czech Republic, approximately 100 such cases are compensated each year, half of whom are below 65 years of age. In contrast, the number of other types of neoplasms is very low. In particular, only a few cases of lung cancer associated with asbestosis in the Czech Republic and Slovakia, and none in Hungary, are compensated [fig_ref] Table 3: Number of cases of occupational cancer compensated in selected countries [/fig_ref]. In Hungary the total number of registered occupational cancer cases was 107 during a 10-year period . Most of these were malignant lung tumors in miners (103 cases).
In Poland, occupational diseases have been registered since 1971. During the following 25 years the overall number of registered occupational diseases and poisonings was 228,744, induding 1,408 cases of malignant neoplasms. The majority of occupational cancers (93%) were diagnosed in men. The numbers of occupational diseases registered yearly increased from 5,008 in 1971 to 11,318 in 1996. In the latter year, 144 cases of occupational cancers were reported. According the ordinance of the Council of Ministers, a disease is determined to be occupationally related if it can be judged to be caused by a health hazard occurring in the work environment and the disease is included in the list of occupational diseases. In the assessment of health hazard, the following factors are taken into account: the degree and duration of occupational exposure and the mode of work. In all cases of possible work-related cancers, an investigation at the workplace is conducted by the sanitary and epidemiologic stations to assess whether contact with carcinogens present in the obligatory list of carcinogens has taken place. Every cancer case that has been recognized and assessed Abbreviations: NR, not recognized as specific occupational disease; -, no data available.
according to the above-mentioned rules and acknowledged as work related is compensated and reported to the Polish Register of Occupational Diseases. The rules for evaluation of causality of occupational diseases are very similar in Poland, Slovakia, and the Czech Republic. Most likely, the registered figures do not represent all occupational cancers. A number of problems affect the certification of occupational cancers: the long period of latency, the retirement status of most patients when they are diagnosed, the possible role of smoking habits and other lifestyle factors, and the long duration of the evaluation and certification process. Generally, these problems contribute to underdiagnosis of cases of occupational cancer.
## Research on occupational cancer
Most studies on occupation and cancer risk in the central European countries have been published in the countries of their origin. Some studies, however, have also been made available in the international literature, more frequently in recent years with increasing collaboration with scientists and organizations outside this particular area. The following are only a few examples of research activities in the central European countries published in English and they do not constitute an extensive or necessarily representative review. As is the case in other parts of the world, the most studies concern the risk of lung cancer.
An elevated risk of lung cancer among uranium miners has been observed in central Europe since before World War II [bib_ref] Cancer in the lung in the miners of Joachimsthal, Pirchan [/bib_ref]. More recently, the lung cancer mortality experience in a cohort of miners was presented in a series of articles, with increasing duration of follow-up and increasingly refined methods of follow-up and exposure estimates [bib_ref] Patterns of lung cancer mortality among uranium miners in West Bohemia with..., Tomasek [/bib_ref] [bib_ref] Recent results from the study of West Bohemian uranium miners exposed to..., Tomasek [/bib_ref]. These studies confirmed the elevated risk of lung cancer associated with exposure to radon and its progeny in uranium mines, and demonstrated the importance in this case of considering additional radon exposure from employment in other mines.
Further analysis of this cohort of uranium miners gave no consistent indication that other cancers are associated with work in uranium mines, with the possible exception of gallbladder, extrahepatic bile ducts, and multiple myeloma [bib_ref] Radon exposure and cancers other than lung cancer among uranium miners in..., Tomasek [/bib_ref]. This lack of a consistent association of cumulative radon exposure with cancers other than lung cancer was also found in an international collaborative study of 11 cohorts of underground miners [bib_ref] Radon and cancers other than lung cancer in underground miners: a collaborative..., Darby [/bib_ref].
The risk of lung cancer associated with radiation and dust exposure among iron ore miners (relative risk = 3-4) was investigated in Slovakia [bib_ref] Lung cancer among iron ore miners in East Slovakia: a casecontrol study, Lcso [/bib_ref]. In Crakow, Poland, a population-based case-control study indicated a 4-fold excess risk of lung cancer among the most highly exposed foundry workers when controlling for smoking habits and partly for exposure in other industries [bib_ref] Lung cancer, smoking, and employment in foundries, Becher [/bib_ref].
A mortality study in a cohort ofworkers in the rubber industry in Poland found the expected elevation of bladder cancer and leukemia in the subcohort employed during the period when ,-naphthylamine was in use and hypothesized an association of cancer of the gallbladder with occupational activity in the rubber industry [bib_ref] Cancer mortality among male workers in the Polish rubber industry, Szeszenia Dabrowska [/bib_ref].
In two small cohorts of chrysotile asbestos workers, an elevated risk of lung cancer was found among the men but not among the women [bib_ref] A mortality study among male workers occupationally exposed to asbestos dust in..., Szeszenia Dabrowska [/bib_ref] [bib_ref] Mortality among female workers in an asbestos factory in Poland, Szeszenia Dabrowska [/bib_ref]. For men employed between 1945 and 1973 and followed until 1985, the standardized mortality ratio (SMR) for lung cancer was 183.3; no cases of mesothelioma were reported. For women employed during the same period, the SMR for lung cancer was 95.6; SMRs for cancer of digestive organs, peritoneum, liver, and pancreas were significantly elevated.
An investigation of the relationship of skin cancer with occupational exposure to PAHs found that only mineral oils elevated risk [bib_ref] Case-referent study on skin cancer and its relation to occupational exposure to..., Kubasiewicz [/bib_ref] , thus confirming a classical epidemiologic association. An analysis of a Polish cohort of pulp and paper workers, set up as a part of an international collaborative study, indicated no excess of leukemia, lymphoma, or lung cancer [bib_ref] Cancer mortality among pulp and paper workers in Poland. A cohort study, Stanczyk [/bib_ref] , but showed an elevated risk of cancer of the peritoneum and prostate. As the cohort was relatively young, the results were based on a small number of cases, and the relationship to occupational exposure was difficult to determine.
Recently, case-control studies and cohort studies have been established to investigate the association between angiosarcoma of the liver and exposure to vinyl chloride monomers, between skin and lung cancer and arsenic exposure, and between lung cancer and exposure to the PAHs in aluminum production.
# Conclusions
It is difflcult to predict a trend for future incidences of occupational cancers, but data seem to indicate that the incidence of cancer has increased remarkably during the last 5 to 10 years in central European countries. However, improved control technology, governmental regulatory activity to reduce exposure, surveillance of diseases, and risk factors and use of preventive measures would ultimately reduce occupational cancer.
[table] Table 3: Number of cases of occupational cancer compensated in selected countries. [/table]
|
Explaining Racial Inequality in Food Security in Columbus, Ohio: A Blinder–Oaxaca Decomposition Analysis
Food insecurity is a leading public health challenge in the United States. In Columbus, Ohio, as in many American cities, there exists a great disparity between Black and White households in relation to food insecurity. This study investigates the degree to which this gap can be attributed to differences in food shopping behavior, neighborhood perception, and socioeconomic characteristics. A Blinder-Oaxaca decomposition method is used to analyze a household survey dataset collected in 2014. We find a 34.2 percent point difference in food security between White and Black households. Variables related to food shopping behavior, neighborhood perception, and socioeconomic characteristics explain 13.8 percent, 11.6 percent, and 63.1 percent of the difference, respectively. These independent variables combined can explain 68.2 percent of the food security gap between White and Black households. Most of this is attributable to socioeconomic variables. Sense of friendship in neighborhood, use of private vehicles, and satisfaction of neighborhood food environment also partially contribute to the food security gap.
# Introduction
## Overview of food (in)security
Food security is a condition under which a person can access adequate, safe, and nutritious food for an active, healthy life without any barriers . To achieve food security, food should be stably available over time and people can easily access food for one's nutritional use/utilization [bib_ref] Climate change and food security, Gregory [/bib_ref] [bib_ref] Global food security under climate change, Schmidhuber [/bib_ref]. The United States Department of Agriculture (USDA) uses four labels to characterize the food security status of households (i.e., all the people who occupy a housing unit regardless of family relations) . Along a continuum, these are high food security (i.e., no reported food procurement issues), marginal food security (i.e., some anxiety reported on sufficient food procurement, but little or no change in diet or food intake), low food security (i.e., reduced diet quality, but little or no reduction in food intake), and very low food security (i.e., negative, reduced changes in both diet and food intake) [1]. In 2016, about 12.3 percent of the U.S. population were food-insecure, meaning they had difficulty procuring enough food for all household members at least sometime during the year. Food insecurity may directly result in an unhealthy diet, immediate hunger, and nutritional deficiencies. It is also associated with negative health outcomes in the long term, potentially impairing mobility and ability to work [bib_ref] Food insecurity: Consequences for the household and broader social implications, Hamelin [/bib_ref] [bib_ref] Functional limitation and chronic diseases are associated with food insecurity among US..., Venci [/bib_ref]. For adults, food insecurity is associated with diet-related chronic diseases including hypertension, diabetes, and obesity, as well as higher mortality [bib_ref] Food insecurity status and mortality among adults in Ontario, Gundersen [/bib_ref] [bib_ref] Food insecurity is positively related to overweight in women, Townsend [/bib_ref]. Mental health conditions associated with adult food insecurity include depression, anxiety, and social isolation [bib_ref] Food insecurity and mental health among females in high-income countries, Maynard [/bib_ref] [bib_ref] Poor nutritional status among low-income older adults: Examining the interconnection between self-care..., Jung [/bib_ref] [bib_ref] Associations between food insecurity and the severity of psychological distress among African-Americans, Allen [/bib_ref]. For pregnant women, maternal food insecurity is associated with increased risk of certain birth defects and prenatal and postpartum depression [bib_ref] Food insufficiency and the physical and mental health of low-income women, Siefert [/bib_ref] [bib_ref] Maternal food insecurity is associated with increased risk of certain birth defects, Carmichael [/bib_ref] [bib_ref] An examination of prenatal and postpartum depressive symptoms among women served by..., Sidebottom [/bib_ref]. Food insecurity during childhood and adolescence also contributes to poor health, increased susceptibility to disease, impeded cognitive and physical development, and poor school performance [bib_ref] Food security, poverty, and human development in the United States, Cook [/bib_ref] [bib_ref] Association of food insecurity with children's behavioral, emotional, and academic outcomes: A..., Shankar [/bib_ref] [bib_ref] Household food insecurity and early childhood health and cognitive development among children..., Huang [/bib_ref].
Importantly, racial and ethnic minority groups in the United States are more vulnerable to household food insecurity [bib_ref] Racial and ethnic disparities in food insufficiency: Evidence from a statewide probability..., Nam [/bib_ref] [bib_ref] Prevalence of perceived food and housing security-15 states, Njai [/bib_ref] [bib_ref] Household food security in the United States, Nord [/bib_ref]. In a 2017 USDA report using a nationally representative sample, the percentage of food-insecure Black households (21.8 percent) was 2.5 times higher than that of White households (8.8 percent) [bib_ref] Food insecurity research in the United States: Where we have been and..., Gundersen [/bib_ref]. Using 2013 Behavioral Risk Factor Surveillance System collected in 15 U.S. states, Njai et al. [bib_ref] Prevalence of perceived food and housing security-15 states, Njai [/bib_ref] found that 68.5 percent (95% confidence interval (CI): 66.3-70.7) of Blacks were food-secure, whereas 81.8 percent (95% CI: 81.2-82.4) of Whites were food-secure. Another study using 2011-2014 National Health and Nutrition Examination Survey found that 17.8 percent (95% CI: 14.3-22.0) of non-Hispanic Blacks were in low food security, while 6.3 percent (95% CI: 5.0-8.0) of non-Hispanic Whites experienced low food security.
## Risk factors of food (in)security
Multiple factors, independently and interactively, determine food security. First, food security is associated with the socioeconomic characteristics of each person or household. Poverty is closely related to food insecurity: nearly one-third (31.6 percent) of low-income households below 185 percent of the U.S. federal poverty guideline were reported food-insecure in 2018 [bib_ref] Understanding the link between poverty and food insecurity among children: Does the..., Wight [/bib_ref]. In 2018, the federal poverty level (FPL) was $25,100 for a family of four, so 185% of the FPL was $46,435. Although the U.S. Supplemental Nutrition Assistance Program (SNAP), formerly known as the Food Stamp Program, provides a monthly allotment of money for food purchases from authorized retailers to eligible participants, SNAP participants are often at a higher risk for food insecurity [bib_ref] The food insecurity-obesity paradox: A review of the literature and the role..., Dinour [/bib_ref].
Lower educational attainment is also associated with a higher chance of food insecurity [bib_ref] Food insecurity research in the United States: Where we have been and..., Gundersen [/bib_ref]. Single-parent households, living alone, and having children in the household are also related to a higher risk for food insecurity [bib_ref] Understanding the link between poverty and food insecurity among children: Does the..., Wight [/bib_ref] [bib_ref] How much does the Supplemental Nutrition Assistance Program reduce food insecurity?, Ratcliffe [/bib_ref] [bib_ref] Economic determinants and dietary consequences of food insecurity in the United States, Rose [/bib_ref]. While older adults (≥65 years old) reported being more food-secure in general, additional healthcare costs could put them at risk for being food-insecure [bib_ref] Food insecurity, social capital and perceived personal disparity in a predominantly rural..., Dean [/bib_ref] [bib_ref] Factors associated with food insecurity among US elderly persons: Importance of functional..., Lee [/bib_ref].
Second, food shopping behaviors and characteristics such as shopping frequencies, shopping distance, and the access to a private vehicle correlate with food security. Food-insecure households are likely to shop for groceries less frequently [bib_ref] New neighborhood grocery store increased awareness of food access but did not..., Cummins [/bib_ref]. When shopping for their food, food-insecure households are more likely to walk and less likely to use their own vehicles [bib_ref] Neighborhood racial composition, neighborhood poverty, and the spatial accessibility of supermarkets in..., Zenk [/bib_ref]. Limited modes of transportation may restrict the total volume of food purchased because it is difficult to carry bulky or heavy items [bib_ref] The association between food security and store-specific and overall food shopping behaviors, Ma [/bib_ref]. Food-insecure households often bypass nearby smaller grocery stores and travel farther to large supermarkets that accept SNAP benefits and usually offer lower prices and promotional offers as well as convenient one-stop services like bank tellers, prepared and takeout foods, and pharmacies [bib_ref] A model of household grocery shopping behavior, Bawa [/bib_ref] [bib_ref] Differences in food environment perceptions and spatial attributes of food shopping between..., Sohi [/bib_ref] [bib_ref] Food store choices of poor households: A discrete choice analysis of the..., Taylor [/bib_ref]. However, they may purchase more nonperishable food items than fresh produce, especially when they make a long food shopping trip relying on other persons' vehicles or public transportation [bib_ref] A tale of two food environments: Differences in food availability and food..., Kaiser [/bib_ref].
Finally, social perception of neighborhood or community is associated with food security. A study conducted in Montreal, Canada, reported that food-insecure households were less satisfied with their food environment in terms of available food options and affordability [bib_ref] Do residents of food deserts express different food buying preferences compared to..., Walker [/bib_ref]. The satisfaction of the neighborhood food environment was also related to more fruit and vegetable consumption in predominantly minority neighborhoods in Chicago [bib_ref] Associations between the local food environment and the severity of food insecurity..., Pérez [/bib_ref]. Perceived informal social bonds and networks in neighborhoods can reduce the risk of food insecurity by providing additional supports and resources to households living in poor local food environments [bib_ref] Perceptions of the food shopping environment are associated with greater consumption of..., Blitstein [/bib_ref] [bib_ref] The effects of the local food environment and social support on rural..., Garasky [/bib_ref]. However, impoverished neighborhoods, especially Black neighborhoods, had relatively weak social bonds and networks in the U.S. [bib_ref] Informal assistance to urban families and the risk of household food insecurity, King [/bib_ref] [bib_ref] Racial differences in networks: Do neighborhood conditions matter?, Small [/bib_ref] [bib_ref] Neighborhoods, social support, and African American adolescents' mental health outcomes: A multilevel..., Hurd [/bib_ref]. [fig_ref] Figure 1: A conceptual model of food security and its determinants [/fig_ref] illustrates a conceptual model of food security based on the literature. and resources to households living in poor local food environments [bib_ref] Perceptions of the food shopping environment are associated with greater consumption of..., Blitstein [/bib_ref] [bib_ref] The effects of the local food environment and social support on rural..., Garasky [/bib_ref]. However, impoverished neighborhoods, especially Black neighborhoods, had relatively weak social bonds and networks in the U.S. [bib_ref] Informal assistance to urban families and the risk of household food insecurity, King [/bib_ref] [bib_ref] Racial differences in networks: Do neighborhood conditions matter?, Small [/bib_ref] [bib_ref] Neighborhoods, social support, and African American adolescents' mental health outcomes: A multilevel..., Hurd [/bib_ref]. [fig_ref] Figure 1: A conceptual model of food security and its determinants [/fig_ref] illustrates a conceptual model of food security based on the literature. Most quantitative studies using regression models report the roles of individual risk factors on the odds of a person or household being food-secure. A typical use of race as an independent variable in such studies is to simply show how much a racial group is more likely to be food-secure compared with a reference group. To address between-groups differentials, however, it would be beneficial for researchers and policymakers to identify how much of the current between-group gap is attributed/not attributed to known risk factors. The Blinder-Oaxaca (BO) decomposition is a wellestablished analysis method to quantify how much between-group difference (e.g., male-female wage gaps, Black-White obesity gaps) can be attributed to known risk factors and unobserved factors, respectively [bib_ref] Racial and ethnic disparities in food insufficiency: Evidence from a statewide probability..., Nam [/bib_ref]. While originally developed to examine gender and racial group differences in wages, BO analysis has recently gained popularity in population health studies [bib_ref] Social isolation of disadvantage and advantage: The reproduction of inequality in urban..., Krivo [/bib_ref].
Using BO approach, this study aims to divide the Black-White food security differential into a part that is "explained" by group differences in socioeconomic characteristics, food shopping behaviors, and neighborhood perception and a remaining part that cannot be accounted for by such differences in the known determinants of food security in Columbus, Ohio [bib_ref] Unexplained gaps and Oaxaca-Blinder decompositions, Elder [/bib_ref]. Only one study, to our knowledge, has used this method to investigate the racial food security gap in the U.S. Using a sample obtained in Oklahoma, Nam et al. found that 87 percent of racial inequality in food security can be explained using a set of socioeconomic variables (e.g., income, homeownership, bank account/credit card ownership) and some environmental variables (i.e., metropolitan residency, unemployment rates). Nam et al., however, focused on households with infants [bib_ref] Racial and ethnic disparities in food insufficiency: Evidence from a statewide probability..., Nam [/bib_ref]. Our study, therefore, is the first to use BO method to investigate racial inequality in food insecurity between a diverse composition of Black-White households from socioeconomic, behavioral, and neighborhood perspectives in a metropolitan U.S. city. Most quantitative studies using regression models report the roles of individual risk factors on the odds of a person or household being food-secure. A typical use of race as an independent variable in such studies is to simply show how much a racial group is more likely to be food-secure compared with a reference group. To address between-groups differentials, however, it would be beneficial for researchers and policymakers to identify how much of the current between-group gap is attributed/not attributed to known risk factors. The Blinder-Oaxaca (BO) decomposition is a well-established analysis method to quantify how much between-group difference (e.g., male-female wage gaps, Black-White obesity gaps) can be attributed to known risk factors and unobserved factors, respectively [bib_ref] Racial and ethnic disparities in food insufficiency: Evidence from a statewide probability..., Nam [/bib_ref]. While originally developed to examine gender and racial group differences in wages, BO analysis has recently gained popularity in population health studies [bib_ref] Social isolation of disadvantage and advantage: The reproduction of inequality in urban..., Krivo [/bib_ref].
Using BO approach, this study aims to divide the Black-White food security differential into a part that is "explained" by group differences in socioeconomic characteristics, food shopping behaviors, and neighborhood perception and a remaining part that cannot be accounted for by such differences in the known determinants of food security in Columbus, Ohio [bib_ref] Unexplained gaps and Oaxaca-Blinder decompositions, Elder [/bib_ref]. Only one study, to our knowledge, has used this method to investigate the racial food security gap in the U.S. Using a sample obtained in Oklahoma, Nam et al. found that 87 percent of racial inequality in food security can be explained using a set of socioeconomic variables (e.g., income, homeownership, bank account/credit card ownership) and some environmental variables (i.e., metropolitan residency, unemployment rates). Nam et al., however, focused on households with infants [bib_ref] Racial and ethnic disparities in food insufficiency: Evidence from a statewide probability..., Nam [/bib_ref]. Our study, therefore, is the first to use BO method to investigate racial inequality in food insecurity between a diverse composition of Black-White households from socioeconomic, behavioral, and neighborhood perspectives in a metropolitan U.S. city.
# Material and methods
## Data and study area
We used the Ohio State University Food Mapping Team Survey (FMTS) for this analysis. The FMTS was conducted in 2014 via in-person interviews or online surveys in 10 diverse and representative zip code areas in Columbus, Ohio. The study area covers most socioeconomically disadvantaged neighborhoods in Columbus, Ohio. [fig_ref] Figure 2: Study area [/fig_ref] illustrates the study area. The six-part, 88-item questionnaire collected data on food access, food shopping patterns, neighborhood environment, health conditions, food security, and sociodemographic background of respondents and their households [bib_ref] A Stata implementation of the Blinder-Oaxaca decomposition, Jann [/bib_ref] [bib_ref] Community-University Engagement via a Boundary Object: The Case of Food Mapping in..., Clark [/bib_ref]. The total sample size qualified for our study was 586 out of 809 respondents who completed the FMTS. The survey was approved by the Behavioral and Social Sciences Institutional Review Board at The Ohio State University. All respondents were 18 years of age or older and gave informed consent.
# Material and methods
## Data and study area
We used the Ohio State University Food Mapping Team Survey (FMTS) for this analysis. The FMTS was conducted in 2014 via in-person interviews or online surveys in 10 diverse and representative zip code areas in Columbus, Ohio. The study area covers most socioeconomically disadvantaged neighborhoods in Columbus, Ohio. [fig_ref] Figure 2: Study area [/fig_ref] illustrates the study area. The six-part, 88-item questionnaire collected data on food access, food shopping patterns, neighborhood environment, health conditions, food security, and sociodemographic background of respondents and their households [bib_ref] A Stata implementation of the Blinder-Oaxaca decomposition, Jann [/bib_ref] [bib_ref] Community-University Engagement via a Boundary Object: The Case of Food Mapping in..., Clark [/bib_ref]. The total sample size qualified for our study was 586 out of 809 respondents who completed the FMTS. The survey was approved by the Behavioral and Social Sciences Institutional Review Board at The Ohio State University. All respondents were 18 years of age or older and gave informed consent.
## Variables
## Measure of food security
Using the Six-Item Household Food Security Survey Module designed by USDA, the FMTS elicited responses to five questions and statements about respondents' food procurement experience in the last 12 months [bib_ref] Finding our direction: The process of building a community-university food mapping team, Kaiser [/bib_ref] : (Q1) "The food that I bought just didn't last, and I didn't have money to get more." Is this statement often, sometimes, or never true for you/your household in the last 12 months? (Q2) "I couldn't afford to eat balanced meals." Is this statement often, sometimes, or never true for your situation in the last 12 months? (Q3) In the last 12 months did you or other adults in your household ever cut the size of your meals or skip meals because there wasn't enough money for food? (Q4) In the last 12 months, did you ever eat less than you felt you should because there wasn't
## Variables
## Measure of food security
Using the Six-Item Household Food Security Survey Module designed by USDA, the FMTS elicited responses to five questions and statements about respondents' food procurement experience in the last 12 months [bib_ref] Finding our direction: The process of building a community-university food mapping team, Kaiser [/bib_ref] : (Q1) "The food that I bought just didn't last, and I didn't have money to get more." Is this statement often, sometimes, or never true for you/your household in the last 12 months? (Q2) "I couldn't afford to eat balanced meals." Is this statement often, sometimes, or never true for your situation in the last 12 months? (Q3) In the last 12 months did you or other adults in your household ever cut the size of your meals or skip meals because there wasn't enough money for food? (Q4) In the last 12 months, did you ever eat less than you felt you should because there wasn't enough money for food? Finally, (Q5) in the last 12 months, were you ever hungry but didn't eat because there wasn't enough money for food? and (Q5-1) if yes, how often did this happen"? A raw score of 1 was given to each response when answered "often" or "sometimes" on questions (Q1) and (Q2), "yes" on (Q3)-(Q5), and "almost every month" and "some months but not every month" on (Q5-1). Households with a row score of 2 or greater (maximum of 6) were considered as food-insecure.
## Independent variables
We used three types of independent variables from FMTS for analysis: First, a set of food shopping behaviors of household (i.e., modes of transportation, monthly shopping frequencies, and mean shopping distance to a food store) were included in the model. Second, the variables describing the social perception of neighborhood reported by the FMTS respondents were included in the model. Nasar and Julian's questionnaire surveying sense of community was used to obtain scaled responses to four survey questions and statements as follows: (1) "How satisfied are you with the ease of which you can access the food you want to eat in your neighborhood?" (2) "My friends in the neighborhood are part of my everyday activities", (3) "People here know they can get help from others in the neighborhood if they are in trouble", and (4) "I have NO friend in the neighborhood on whom I could depend." And finally, socioeconomic characteristics related to households' food security (i.e., educational attainment, household income, the number of children in the household, and participation in the SNAP) were included in the model.
# Statistical analysis
We conducted a BO decomposition analysis to investigate the contribution of each independent variable to racial differences in food security [bib_ref] The psychological sense of community in the neighborhood, Nasar [/bib_ref]. The mean group-specific food security rates for Whites (FS W ) and Blacks (FS B ) can be denoted as follows:
[formula] FS W = E ss W Y i W X i W (1) FS B = E ss B Y i B X i B(2) [/formula]
where Y i is a regression models to estimate food security, X i is a vector of i covariates, and E ß (Y i |X i ) is the conditional expectation of Yi estimated at the parameter vector ß. The food security gap between Whites and Blacks, ∆, can be expressed as follows:
[formula] ∆= FS W −FS B = E ss W Y i W X i W − E ss B Y i B X i B = E ss W Y i W X i W − E ss W Y i B X i B + E ss W Y i B X i B − E ss B Y i B X i B(3) [/formula]
where E ß W (Y i B |X i B ) refers to a hypothetical term with the mean X values of the Blacks and the coefficient ßs of the Whites. The term assumes a hypothetical condition that Blacks have the same characteristics in all covariates (i.e., the effect of education is the same in both racial groups). In Equation (3), the first term on the right side is considered the "explained" portion of the gap, because the differential of the two groups is due to group differences in levels of covariates. In contrast, the second term is the "unexplained" portion due to differences in coefficients between the two groups [bib_ref] Social isolation of disadvantage and advantage: The reproduction of inequality in urban..., Krivo [/bib_ref] [bib_ref] Unexplained gaps and Oaxaca-Blinder decompositions, Elder [/bib_ref] [bib_ref] On discrimination and the decomposition of wage differentials, Oaxaca [/bib_ref]. In other words, the mean values of covariates and their regression coefficients for each group are involved in the decomposition [bib_ref] A Stata implementation of the Blinder-Oaxaca decomposition, Jann [/bib_ref]. For analysis, we constructed a set of four logit models. In all four, the status of being food-secure (high or marginal food security = 1, low or very low food security = 0) was the dependent variable. Model 1, 2, and 3 analyzed shopping behavior, neighborhood perception, and socioeconomic variables, respectively; Model 4 incorporated all three. Analyses were conducted using "Oaxaca" command in Stata SE 14.1 [bib_ref] A Stata implementation of the Blinder-Oaxaca decomposition, Jann [/bib_ref] [bib_ref] An extension of the Blinder-Oaxaca decomposition to nonlinear models, Bauer [/bib_ref]. The categorical variables in the model were transformed as a set of dummy variables for the use of "Oaxaca" command since the decomposition results could change with the choice of reference categories [bib_ref] A Stata implementation of the Blinder-Oaxaca decomposition, Jann [/bib_ref].
# Results
Table 1 provides the descriptive statistics of the two racial groups. Compared with Black households, White households used their own cars more often and walk less to buy food, shopped 1.4 times more frequently, and travelled 0.6 miles less to acquire food in one month. Regarding neighborhood perception, White households were more satisfied with food accessibility in their neighborhood, and had more perceived connections with friends and neighbors in their communities. In terms of socioeconomic characteristics, White respondents were younger, possessed higher educational and household income levels, had fewer children in their homes, and were less likely to participate in SNAP. Model 1 shows racial inequality in food security explained by shopping behavior and characteristics. Only 13.8 percent (95% CI: 4.2-23.3) of the gap was explained by the shopping behavior variables-monthly shopping frequency, mode of travel (use of own vehicle, walking), and mean distance to shopping destination in combination; the remaining 86.2 percent (95% CI: 59.5-113.0) remained unexplainable by differences in shopping behavior. Shopping by own vehicle was the only variable independently explaining the racial gap in food security with statistical significance.
Model 2 presents how much of the racial gap in food security can be explained with perception of neighborhood. Only 11.6 percent (95% CI: 3.7-19.5) of the gap was explained with the satisfaction of food environment in neighborhood, sense of connection with friends in the neighborhood, of having help from neighbors, and of having no friends in the neighborhood. Notably, sense of having no friends in the neighborhood was the most statistically significant in explaining the racial food security gap.
Model 3 summarizes the extent to which socioeconomic variables explain racial inequality in food security. These variables explained 63.1 percent (95% CI: 46.2-80.0) of the gap. Educational attainment level, household income, the number of children in the household, and participation in SNAP were all statistically significant in explaining racial inequality in food security. Model 4 reports the results when using all three types of variables together. While 68.2 percent (95% CI: 50.4-85.9) of the racial food security gap was explained by these variables, 31.8 percent (95% CI: 3.8-60.0) remained unexplained. Satisfaction with neighborhood food environment, sense of no friends in the neighborhood, education, household income, number of children, and participation in SNAP were statistically significant in this model.
# Discussion
As summarized in [fig_ref] Table 1: Descriptive statistics [/fig_ref] , Black households in the study area have more disadvantaged socioeconomic and environmental conditions. These respondents are older, less educated, and have a lower income compared with White respondents. While traditionally the study area has been predominantly Black, neighborhood change through gentrification and the growth of a local state university has drawn in young White residents. Differences in socioeconomic conditions may further explain differences in shopping behavior. Higher levels of education and income in White households may contribute to easier access to private vehicles in which to go shopping and higher shopping frequency. Easier access to personal vehicles could also mean that White households are able to choose to shop at supermarkets or grocery stores with consistently lower prices, discounts, and deals and, therefore, are less likely to report being food-insecure compared with Black households [bib_ref] Stata Statistical Software: Release 14, Statacorp [/bib_ref]. There are also differences in food environment and perceptions of food environment. Black households travel further to purchase food, implying there may be fewer food stores with healthy food in their neighborhood. Therefore, it is unsurprising that Black households are less satisfied with their food environment in the study area. The additional distance travelled by Black households to acquire food would also increase the total cost of food shopping (i.e., additional gas, public transportation fare, time spent traveling), which is yet another example of how the food environment may be associated with the racial gap in food insecurity rates [bib_ref] Fresh produce consumption and the association between frequency of food shopping, car..., Gustat [/bib_ref].
Among the three categories of variables, the majority of the food security gap between White and Black households is attributable to differences in individual socioeconomic variables. We find a distinct inequality in household income and educational attainment between Black and White households, and this is the primary factor driving the food security gap between the two racial groups. When examining such racial disparities, it is critical to consider the historical impact of policy in driving the disparities. Columbus ranks 78 out of the 318 most racially segregated cities in the U.S. [bib_ref] Mapping the cost of a balanced diet, as a function of travel..., Hilbert [/bib_ref]. The racial segregation of American cities is a cumulative result of policies enacted throughout the twentieth century. For example, most historically Black communities received "D" ratings from the Home-Owners Loan Corporation (HOLC), redlining these communities [fig_ref] Figure 3: A Home-Owners' Loan Corporation [/fig_ref]. Redlining meant that these communities were ineligible to receive federally insured mortgages for the purchase or refinancing of property, resulting in disinvestment and "blight." Black people in Columbus and across the U.S. were barred from purchasing property in communities with higher HOLC grades due to the prolific use of racially restrictive covenants, conventions written into home deeds which prohibited the sale of properties to Black people and other ethnic and religious minority groups [bib_ref] Redlining and the home owners' loan corporation, Hillier [/bib_ref]. In Ohio, schools are funded by local property taxes, resulting in unequal funding allocations to school districts in wealthier versus less wealthy neighborhoods. Due to policies like redlining, many communities of color have experienced decades of disinvestment, resulting in lower property values and less money to fund public education. Policies that fund public schools equitably by either using revenue from sources other than property taxes or aggregating and redistributing property taxes at the state level could address some of the educational disparities found in the study area. The impact of redlining and racially restrictive covenants on wealth generation in redlined communities is still visible today through the racial gap in food security. and less money to fund public education. Policies that fund public schools equitably by either using revenue from sources other than property taxes or aggregating and redistributing property taxes at the state level could address some of the educational disparities found in the study area. The impact of redlining and racially restrictive covenants on wealth generation in redlined communities is still visible today through the racial gap in food security. The long-term impacts of practices like redlining which led to disinvestment in many minority urban communities are found not only in blighted built environments, but also in the social fabric of communities. We find that social networks in neighborhoods are a strong predictor of the Black-White gap in food security. Satisfaction with neighborhood food environment also partially explains the racial gap in food security when considering all variables. These findings imply that improving neighborhood conditions may be an effective strategy to improve food security by building social networks and self-supports [bib_ref] Associations between the local food environment and the severity of food insecurity..., Pérez [/bib_ref]. For example, implementing community-based gardening projects could be an effective neighborhood-based intervention to provide additional access to fresh vegetables and build healthier family and social relationships. Also notable is the "corner store intervention", a response to major supermarket chains' tendency to relocate existing stores from impoverished inner-city neighborhoods to affluent suburbs or unwillingness to open new stores in poor areas [bib_ref] Impact of a community gardening project on vegetable intake, food security and..., Carney [/bib_ref]. The goal of the intervention is to provide additional fresh fruit and vegetables and The long-term impacts of practices like redlining which led to disinvestment in many minority urban communities are found not only in blighted built environments, but also in the social fabric of communities. We find that social networks in neighborhoods are a strong predictor of the Black-White gap in food security. Satisfaction with neighborhood food environment also partially explains the racial gap in food security when considering all variables. These findings imply that improving neighborhood conditions may be an effective strategy to improve food security by building social networks and self-supports [bib_ref] Associations between the local food environment and the severity of food insecurity..., Pérez [/bib_ref]. For example, implementing community-based gardening projects could be an effective neighborhood-based intervention to provide additional access to fresh vegetables and build healthier family and social relationships. Also notable is the "corner store intervention", a response to major supermarket chains' tendency to relocate existing stores from impoverished inner-city neighborhoods to affluent suburbs or unwillingness to open new stores in poor areas [bib_ref] Impact of a community gardening project on vegetable intake, food security and..., Carney [/bib_ref]. The goal of the intervention is to provide additional fresh fruit and vegetables and other healthy food to local stores and upscale the store arrangement and display to promote increased consumption of healthy food [bib_ref] Spatial Supermarket Redlining and Neighborhood Vulnerability: A Case Study of Hartford, Zhang [/bib_ref]. The findings of our study suggest that the corner store intervention would be more effective if it also aimed to enhance the sense of community among residents by increasing daily contact and friendships between neighbors through shopping in these spaces.
When considering only shopping behavior variables, it is notable that using one's own vehicle to go shopping is attributable to the racial gap in food security. This finding reiterates the lower socioeconomic status of Black households and also supports the food security literature stressing the importance of transportation mode [bib_ref] New neighborhood grocery store increased awareness of food access but did not..., Cummins [/bib_ref]. However, this effect becomes statistically not significant when analyzed with all the variables in this study (Model 4), implying that the differences in transportation mode alone may not lead to a large racial gap in food security when considering a variety of the other predictors of food security.
Receiving SNAP is an important factor in explaining the racial gap in food security because more Black households (46.6%) than White (13.2%) participate in SNAP in our study. This implies that SNAP may be insufficient to make a household food-secure. Leung et al. also found that SNAP participation does not necessarily lead to food security, compared with non-participants. While SNAP substantially helps food-insecure households, there is criticism regarding its benefits for individual health and well-being. For example, while the average meal cost in the U.S. is $2.36, the SNAP benefit per meal is $1.86 [bib_ref] Few changes in food security and dietary intake from short-term participation in..., Leung [/bib_ref]. For households with children, it is calculated even lower, at $1.40. Further research, therefore, is needed to focus on the dynamics of disposable income and food purchase for food security by incorporating other sources of income, social welfare programs, and spending for other purposes such as medical treatment, housing, and education as well as the appropriate level of the SNAP benefit and the budget.
While our study illuminates the drivers of the racial food gap in Columbus, Ohio, it has some limitations. First, although the study uses comprehensive information on household food security, this information was collected from a cross-sectional survey. Regularly performed surveys in a wider range of geographical areas would help to inform a more insightful analysis. Second, the data used in the study was collected via self-reporting. Future research is needed to design more comprehensive methods to quantitatively measure household food security and neighborhood characteristics for analysis. Leroy et al. [bib_ref] Measuring the food access dimension of food security: A critical review and..., Leroy [/bib_ref] suggest that more robust food security indicators based on accessibility, details of food consumption, and more information about nutritional value may reveal more relevant findings for food security inequalities. Third, while this study identifies the extent to which socioeconomic variables, neighborhood perception, and shopping behavior contribute to the racial gap in food security, it does not aim to find new contributing variables. More research is needed to focus on finding new variables that could potentially explain the substantial portion of the racial food security gap that remains unexplained. For example, future studies can utilize more socioeconomic (e.g., amount of SNAP benefit, number of employed people per household, amount of healthcare spending) and/or environmental (e.g., availability of food items or community food assistance programs, walkability) determinants of households. It is also possible that structural racism could help explain the unexplained portion of the racial food security gap. Qualitative research exploring this notion could shed light on these disparities by providing more insight into the extent of barriers to social capital, such as discrimination and racism in the food environment.
# Conclusions
This study uses the BO decomposition method to partition the gap in food security between Black and White households in Columbus, Ohio, into explained and unexplained portions with independent variables. We analyze a survey on food security status, shopping behavior, neighborhood, and socioeconomic characteristics, finding that the majority of the racial gap in food security is attributable to socioeconomic characteristics of households. Poor neighborhood social network is an important factor in explaining the gap. We also find that use of a private vehicle for food shopping and satisfaction with neighborhood food environment partially contribute to the racial food gap. However, a substantial portion of the racial food security gap remains unexplained by the variables used in this study. To reduce racial disparities in food security, future policy interventions should address differences in variables, especially levels of educational attainment and household income. Specifically, community leaders can seek opportunities to distribute free and affordable produce through the local foodbanks and expand SNAP and WIC (Special Supplemental Nutrition Program for Women, Infants, and Children) benefits to lessen the financial burden on low-income households. Future studies are also needed to explore additional variables that potentially contribute to the racial gap in food security.
[fig] Figure 1: A conceptual model of food security and its determinants. Note. Authors created based on the literature. [/fig]
[fig] Figure 2: Study area. Note. CBUS: The boundary of the City of Columbus, Ohio. [/fig]
[fig] Figure 3: A Home-Owners' Loan Corporation (HOLC) redlining map in Columbus, Ohio, in 1936. Note. CBUS: the boundary of the City of Columbus, Ohio. Source: Redlining Maps-Ohio Cities: Ohio State University Libraries: http://guides.osu.edu/maps-geospatial-data/maps/redlining/. Authors recreated. [/fig]
[table] Table 1: Descriptive statistics.Table 1. Cont. [/table]
[table] Table 2: shows the explained and unexplained portions of the food security gap between Black and White households. There was a 34.2 percent point (95% CI: 25.4-43.1) difference in food security between White (75.6%, 95% CI: 71.9-79.9) and Black (41.6%, 95% CI: 33.7-49.5) households. [/table]
|
Hemophagocytic Lymphohistiocytosis Complicated by Acute Respiratory Distress Syndrome and Multiorgan Failure
Hemophagocytic lymphohistiocytosis (HLH) is a rare and life-threatening condition that is characterized by an overactive response of the immune system with excessive production of proinflammatory cytokines. Initial presentation of this condition often mimics and overlaps with many diseases including infections, sepsis, and multiorgan failure syndrome, which makes diagnosis the diagnosis of HLH challenging. Herein is described a case of a patient who developed acute respiratory distress syndrome and multiple organ failure related to HLH in a setting of probable viral pneumonia. The diagnosis was established based on laboratory and bone marrow biopsy findings. This patient was treated with the standard chemotherapy regimen of intravenous dexamethasone, etoposide in addition to intrathecal methotrexate for central nervous system involvement.
# Introduction
Hemophagocytic lymphohistiocytosis (HLH) is a rare, lifethreatening condition that is characterized by an overactive and abnormal response of the immune system with excessive production of proinflammatory cytokines. [bib_ref] Clinical characteristics, prognostic factors, and outcomes of adult patients with hemophagocytic lymphohistiocytosis, Otrock [/bib_ref] The diagnosis of this condition is challenging because of its nonspecific signs and symptoms that often mimic other disease processes, for instance, infectious diseases. [bib_ref] The unique aspects of presentation and diagnosis of hemophagocytic lymphohistiocytosis in adults, Nikiforow [/bib_ref] Acute respiratory distress syndrome (ARDS) is a rare and life-threatening presentation of HLH and only a few cases have been reported in the literature to date. Herein is described a case of a patient who developed ARDS and multiple organ failure related to HLH in a setting of probable viral pneumonia. This patient was treated with the standard chemotherapy regimen.
## Case presentation
A 74-year-old female with a past medical history significant for ulcerative colitis s/p remote colectomy and ileostomy and hypertension presented to the emergency department (ED) with fatigue, myalgia, and confusion. On presentation, she had a temperature of 100°F, blood pressure of 108/60 mmHg, heart rate of 106 beats per minute (bpm), and respiratory rate of 20 breaths per minute. She was disoriented to the surrounding environment, but her neurologic examination was otherwise unremarkable. Upon cardiac auscultation, she had normal S1 and S2 without murmurs, gallops or rubs. Lung auscultation revealed decreased breath sounds on lung bases without crackles or rales. Initial lab analysis was significant for sodium of 128 (136-145 mEq/L), creatinine of 1.21 (0.51-0.95 mg/dL), and elevated lactate at 3.0 (0.4-2.0 mEq/L) with urinalysis suggestive of urinary tract infection (UTI). Complete blood count (CBC) showed a white blood cell count of 8.8 (3.98-10.04 k/cmm), hemoglobin of 15.1 (11.2-15.7g/dL), and platelet count of 131 (147-409 k/cmm). Computed tomography (CT) of the brain and initial chest X-ray were unremarkable for any acute abnormality. She was resuscitated with intravenous fluid and started on IV ceftriaxone for suspected UTI, and admitted for treatment of sepsis secondary to UTI and delirium. The patient was discharged after 3 days of hospitalization following improvement of mentation to baseline and resolution of her acute kidney injury (AKI), hyponatremia, and lactic acidosis.
Two days later, the patient returned to the ED with complaint of significant weakness and inability to ambulate. Upon return to the ED, her vital signs were notable for an elevated temperature of 103°F, heart rate of 115 bpm, blood pressure of 90/65 mmHg, and respiratory rate of 24 breaths per minute with normal mental status. Lab analysis was significant for a white blood cell count of 3.31 (3.98-10.04 k/ cmm), hemoglobin of 13.2 (11.2-15.7 g/dL), platelet count of 60 (147-409 k/cmm), and lactic acid of 5.4 (0.4-2.0 mEq/L). Other labs showed aspartate transaminase (AST) of 237 (15-37 U/L), alanine transaminase (ALT) of 189 (12-78 U/L), and alkaline phosphatase of 146 (45-117 U/L. Chest X-ray showed right middle lobe infiltrates; hence, she was started on vancomycin and piperacillin-tazobactam for possible hospital-acquired pneumonia and admitted to the medical floor. Four days after admission, she developed worsening respiratory status requiring noninvasive mechanical ventilation and was transferred to the medical intensive care unit. She was found to have bilateral pleural effusions on CT of the chest, and subsequently underwent a thoracentesis resulting in the removal of 750cc of exudative fluid. She became progressively hypoxemic with serial chest X-rays [fig_ref] Figure 1: Patient's chest radiograph demonstrating bilateral diffuse infiltrates consistent with acute respiratory distress... [/fig_ref] showing bilateral interstitial infiltrates consistent with ARDS and ultimately required intubation. Despite empiric treatment with broad-spectrum antibiotics, the patient's fever persisted with a continuous decline in her state. Multiple blood, sputum, urine, and stool cultures were negative for bacterial, fungal, mycobacterial, and parasitic infections. Follow-up labs showed a white blood cell count of 1.4 (3.98-10.04 k/ cmm), hemoglobin of 8.5 (11.2-15.7g/dL), platelet count of 45 (147-409 k/cmm) along with worsening kidney functions tests, with a creatinine of 2.1 (0.51-0.95mg/dL), and BUN of 65 (7-18mg/Dl). Moreover, liver transaminases had continued to rise to thousands and her jaundice worsened with remarkable coagulopathy and an elevated international normalized ratio (INR) at 8 (0.9-.1.3), consistent with acute liver injury (ALI). Moreover, Factor V and factor VIII activity assays were normal which ruled out disseminated intravascular coagulopathy (DIC). Abdominal ultrasound was unremarkable and magnetic resonance imaging (MRI) of the liver and pancreas showed no evidence of acute cholecystitis or biliary ductal dilation. Additional investigation showed significantly elevated LDH at 1367 (84-246U/L), elevated triglycerides at 808 (0-149mg/dL), fibrinogen of 240 (200-400mg/dL) and elevated ferritin at 40,000 (8-252 ng/ mL). Secondary HLH was suspected with significantly elevated ferritin. Due to suspicion of HLH, a bone marrow biopsy was performed, which showed normocellular marrow with myeloid predominance, mild granulocyte atypia, and increased histiocytes [fig_ref] Figure 2: Wright stain aspirate smear of bone marrowHistiocytes are noted to be increased [/fig_ref]. CD163 immunostaining of the core biopsy revealed markedly increased histiocytes with some cells showing ingested cellular material, which is consistent with HLH [fig_ref] Figure 3: H&E and CD163 immunohistochemical stain of core bone marrow biopsy [/fig_ref]. Then, the patient was immediately started on an 8-week-regimen of dexamethasone and etoposide with subsequent improvement of her clinical status, cell counts and liver functions. To further investigate potential secondary causes of HLH, viral serologies were obtained, which were negative for human immunodeficiency virus (HIV), cytomegalovirus (CMV), Herpes simplex virus (HSV) type 1 and 2, Epstein-Barr virus (EBV), hepatitis B virus, and hepatitis C virus. Also, antinuclear antibody, double-stranded DNA antibody, rheumatoid factor, and autoimmune hepatitis workup were negative, ruling out the most common disorders associated with macrophage activation syndrome (MAS). Hereditary hemochromatosis was ruled out as the DNA assay for C282Y, H63D, and S65C loci of the gene HFE showed no mutations. Given these results, the patient was diagnosed with HLH as she met 5 out of the 8 criteria (HLH-2004): fever, cytopenia of 2 cell lines, elevated ferritin>500 ng/mL, elevated fasting triglyceride >265 mg/ dL and hemophagocytosis in bone marrow. Ultimately, patient's respiratory status improved and was weaned off mechanical ventilation. At discharge, her blood counts improved dramatically with normalization of liver enzymes. One month later, the patient was readmitted to the hospital with persistent confusion and was eventually diagnosed with HLH with central nervous system involvement based on worsening cognitive impairment and increased protein level and pleocytosis on cerebrospinal fluid. She was started on intrathecal methotrexate with not much improvement in her symptoms. The patient's family changed her goals of care to comfort care. The patient passed away 2 weeks later.
# Discussion
Our patient initially presented with clinical signs and symptoms suggestive of infection. Despite treatment for severe sepsis with underlying possible pneumonia and urinary tract infection, the patient's clinical status deteriorated rapidly and she developed ARDS and ALI. She was later diagnosed with HLH, with no obvious infectious etiology, based on laboratory and bone marrow biopsy findings.
The true incidence and prevalence of HLH is unknown and remains difficult to establish accurately. [bib_ref] Hemophagocytic lymphohistiocytosis: review of etiologies and management, George [/bib_ref] The diagnosis of HLH is challenging and often delayed because of its inconsistent presentation and the many nonspecific clinical features it shares with other disease processes. Currently, HLH is well described in the pediatric population, but less is known about the incidence and prevalence, appropriate workup and treatment, and pathophysiological mechanisms of the disease in adults. Most reported cases of HLH in adults have been described in association with viral infections, hematological malignancy; most commonly lymphoma, and autoimmune disease, with few reports documenting bacterial infection as the etiology. [bib_ref] Reactive hemophagocytic syndrome in adult systemic disease: report of twenty-six cases and..., Dhote [/bib_ref] HLH is classified into primary and secondary forms. Primary HLH, or familial HLH, present in childhood, and is associated with certain gene mutations. [bib_ref] The unique aspects of presentation and diagnosis of hemophagocytic lymphohistiocytosis in adults, Nikiforow [/bib_ref] Secondary HLH or acquired HLH usually presents in adulthood and is acquired as a complication due to underlying conditions. [bib_ref] The unique aspects of presentation and diagnosis of hemophagocytic lymphohistiocytosis in adults, Nikiforow [/bib_ref] Left untreated, both forms of HLH may eventually progress to multi-organ failure and death. Without any treatment, the median survival for primary HLH is less than 2 months following diagnosis, while the prognosis for secondary HLH varies depending on the etiology. [bib_ref] Hemophagocytic lymphohistiocytosis: review of etiologies and management, George [/bib_ref] The pathophysiology of HLH, whether primary or secondary, includes the defective function of natural killer (NK) cells and cytotoxic lymphocytes in eliminating activated macrophages. [bib_ref] Hypercytokinemia in familial hemophagocytic lymphohistiocytosis, Henter [/bib_ref] This ultimately leads to the overactivity of macrophages with resulting excessive secretion of cytokines and interferon-gamma, which manifests with nonspecific signs, and symptoms of inflammation and even multiorgan failure, like our patient. [bib_ref] Advances in the pathogenesis of primary and secondary haemophagocytic lymphohistiocytosis: differences and..., Brisse [/bib_ref] However, the immune system activation, in both primary and secondary forms, is intact with impaired the regulatory mechanisms responsible for the termination of the activated immune system. In our case, the patient initially had signs and symptoms suggestive of pneumonia, which triggered HLH and resulted in progression to severe systemic inflammatory response syndrome and multiorgan failure.
HLH often presents with nonspecific manifestations that mimic infections; however, our patient developed multiorgan failure with ARDS, ALI, and AKI, which is considered a rare presentation of HLH. HLH is often left out of the differential due to its nonspecific clinical presentation. As in our patient, the recognition of the disease can be challenging since its manifestation is often akin to septicemia. As such, it is important to include HLH in the differential diagnosis for patients with fever of unknown origin, ARDS, acute hepatitis or acute liver failure. Furthermore, a suspected case of HLH is difficult to confirm because there are currently no gold standard confirmatory tests. Therefore, it is important for systematic evaluation for both infectious and noninfectious agents in these patients.
Once other potential causes are ruled out, the diagnosis of HLH can be made utilizing the HLH-2004 criteria [fig_ref] Table 1: HLH-2004 Diagnostic Criteria [/fig_ref].In our case, the patient fulfilled 5 out of the 8-point diagnostic criteria for HLH including a fever of >38.3°C, 2 cellline-cytopenia, hypertriglyceridemia, hyperferritinemia, and hemophagocytosis on bone marrow biopsy. It should be noted that HLH is an evolving diagnosis and that not all of the HLH diagnostic criteria may be met initially. Therefore, it is important to revisit the diagnosis frequently and to follow clinical signs and laboratory markers of pathologic inflammation, such as sCD25 and ferritin, repeatedly to diagnose potential HLH. It is also worth noting that the absence of hemophagocytosis in bone marrow does not rule out HLH in the presence of other signs, symptoms and laboratory findings suggestive of this condition. [bib_ref] Marrow assessment for hemophagocytic lymphohistiocytosis demonstrates poor correlation with disease probability, Ho [/bib_ref] There was no suspicion for genetic predisposition in our patient. Therefore, we classified the patient as a case of secondary HLH. While she had a history of ulcerative colitis, she had a prophylactic colectomy and had not complained of any related symptoms before or during her hospital course. In addition, rheumatoid factor and autoimmune hepatitis workup had been negative, which ruled out autoimmune-associated HLH. All patients with suspected HLH should have a bone marrow biopsy, both to determine the cause of cytopenia as well as to detect hemophagocytosis. Our patient's bone marrow biopsy simultaneously demonstrated hemophagocytosis and ruled out malignancy. Lastly, we considered an infectious etiology. Among the infectious agents that trigger HLH, EBV is the most frequently reported virus that is associated with HLH. [bib_ref] Hemophagocytic lymphohistiocytosis: a case series, Fatima [/bib_ref] Other infectious etiologies for secondary HLH described in the literature include CMV, HIV, human herpes virus, varicella-zoster, parainfluenza, and measles virus. Bacterial etiologies have been noted less commonly, but include mycobacteria, Brucella, Rickettsia, Serratia, and Haemophilus influenza, which were ruled out in our patient as well. [bib_ref] Hemophagocytic lymphohistiocytosis: a case series, Fatima [/bib_ref] Although the viral serology in this patient was negative, serologic studies for viral infection in patients with the picture of a hemophagocytic syndrome may not be reliable because the impaired immune function may preclude a detectable antibody response. [bib_ref] Epstein-Barr virus-associated hemophagocytic syndrome: virological and immunopathological studies, Sullivan [/bib_ref] Once the cause is determined, therapy should be initiated promptly due to the high mortality rate associated with this condition if left untreated. Due to the life-threatening nature of the condition, timely diagnosis and treatment of HLH are paramount. While important, identifying and targeting the inciting infectious organism in infection-associated HLH is not adequate; these cases require aggressive management via standardized HLH protocols. Treatment according to HLH protocol involves targeting the excessive inflammation and immune dysregulation seen in the condition. Currently, the mainstay of treatment for HLH is immunosuppression with an 8-week-induction regimen, which includes etoposide and dexamethasone, which we initiated in our patient. [bib_ref] Chemoimmunotherapy for hemophagocytic lymphohistiocytosis: long-term results of the HLH-94 treatment protocol, Trottestam [/bib_ref] After eight weeks, many patients require treatment with pulses of dexamethasone. In many cases, hematopoietic stem cell transplantation is needed if an adequate response is not achieved with the above regimen. [bib_ref] Improved outcome in haemophagocytic lymphohistiocytosis after bone marrow transplantation from related and..., Dürken [/bib_ref] As it was true in our case, patients with central nervous system involvement require treatment with intrathecal methotrexate. 13
# Conclusions
In conclusion, this case illustrates various and rare presentations of HLH including ARDS and multiorgan failure. The true incidence of ARDS in adults diagnosed with HLH is still unknown. Hence, it is important to keep HLH in the differential diagnosis in patients presenting with acute respiratory failure and not responding adequately to appropriate antibiotic therapy. Early diagnosis and prompt treatment of this condition are crucial to prevent life-threatening organ failure. This patient was treated with a standard regimen of immunosuppressive therapy with resulting initial and dramatic improvement in her clinical condition and resolution of respiratory insufficiency.
## Authors' note
This case was previously published in abstract form and presented at the CHEST medical student and resident case report virtual conference October 18 to 21, 2020.
## Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
# Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
## Ethics approval
Our institution does not require ethical approval for reporting individual cases or case series.
## Informed consent
Verbal informed consent was obtained from the patient(s) for their anonymized information to be published in this article.
[fig] Figure 1: Patient's chest radiograph demonstrating bilateral diffuse infiltrates consistent with acute respiratory distress syndrome. [/fig]
[fig] Figure 2: Wright stain aspirate smear of bone marrowHistiocytes are noted to be increased (black arrows) demonstrating evidence of hemophagocytosis, with one histiocyte showing ingestion of an immature granulocyte (red arrow). [/fig]
[fig] Figure 3: H&E and CD163 immunohistochemical stain of core bone marrow biopsy. H&E stain showing normocellular marrow for age (approximately 30-35% cellularity). CD163 stain highlighting numerous histiocytes throughout the marrow space. [/fig]
[table] Table 1: HLH-2004 Diagnostic Criteria.The diagnosis of HLH is established based on fulfilling one of the following criteria A molecular diagnosis consistent with HLH Five out of the following 8 criteria are fulfilled: Hemophagocytosis in bone marrow, spleen, lymph nodes, or liver Abbreviation: HLH, Hemophagocytic lymphohistiocytosis. [/table]
|
Unexpected pain with electrocortical stimulation in a teenager with temporal encephalocele
a b s t r a c tTemporal lobe encephalocele has emerged as a potentially unrecognized cause of drug-resistant temporal lobe epilepsy (TLE) that can be effectively treated with epilepsy surgery. Here we present a case in which a 17-year-old male with drug-resistant epilepsy and left temporal encephalocele underwent workup for epilepsy surgery, and experienced unexpected pain with electrocortical stimulation. Stimulation of stereo-EEG electrodes in the left temporal pole resulted in severe, unilateral left-sided facial pain due to inadvertent stimulation of the trigeminal nerve. Stereo-EEG showed seizure onset adjacent to encephalocele with no involvement of mesial temporal structures. A temporal pole resection sparing the mesial temporal structures and repair of the sphenoid bone defect was performed. The patient experienced post-operative seizure freedom, with no loss of language function or sensory deficits in the distribution of the trigeminal nerve. This case highlights temporal encephalocele as a potentially overlooked cause of TLE and underscores the anatomical proximity of the trigeminal nerve to the temporal pole, an important consideration for surgical planning.
# Introduction
Encephalocele is defined as a protrusion of brain tissue and its membranes through defects in the skull. As resolution of imaging has improved, temporal lobe encephalocele has been increasingly identified as a cause of drug-resistant temporal lobe epilepsy (TLE) [bib_ref] Occult middle fossa encephaloceles in patients with temporal lobe epilepsy, Byrne [/bib_ref] [bib_ref] Epilepsy surgery for skull-base temporal lobe encephaloceles: should we spare the hippocampus..., Bannout [/bib_ref] [bib_ref] Small temporal pole encephaloceles: a treatable cause of ''lesion negative" temporal lobe..., Abou-Hamden [/bib_ref] [bib_ref] Small temporal pole encephalocele: a hidden cause of ''normal" MRI temporal lobe..., Toledano [/bib_ref] [bib_ref] Temporal anteroinferior encephalocele, Saavalainen [/bib_ref]. In such cases, surgical intervention often reduces seizure burden or eliminates seizures entirely [bib_ref] Epilepsy with temporal encephalocele: characteristics of electrocorticography and surgical outcome, Panov [/bib_ref] [bib_ref] Spontaneous temporal pole encephalocele presenting with epilepsy: report of two cases, Shimada [/bib_ref]. Several different surgical approaches may be employed [bib_ref] Epilepsy with temporal encephalocele: characteristics of electrocorticography and surgical outcome, Panov [/bib_ref] [bib_ref] Temporal encephalocele: a novel indication for magnetic resonance-guided laser interstitial thermal therapy..., Ranjan [/bib_ref] ; as in all cases of epilepsy surgery, an optimal surgical approach maximizes the likelihood of seizure freedom while minimizing the chance of postoperative neurologic or cognitive deficits. Stereo-EEG (sEEG) may be used to select an optimal surgical approach by localizing the seizure onset zone and identifying adjacent eloquent cortex through electrocortical stimulation [bib_ref] The usefulness of stereo-electroencephalography (SEEG) in the surgical management of focal epilepsy..., Souza [/bib_ref] [bib_ref] Stereoelectroencephalography in epilepsy, cognitive neurophysiology, and psychiatric disease: safety, efficacy, and place..., Youngerman [/bib_ref]. Here, we present a case in which a 17-year-old male with drugresistant TLE underwent sEEG for pre-operative planning. During electrocortical stimulation with sEEG, he experienced unexpected pain that had significant implications for his surgical management.
## Case report
## Patient presentation
A 17-year-old right-handed male with high-functioning autistic spectrum disorder presented with a two-year history of focal impaired awareness seizures. His seizures lasted 1-2 min. The semiology was characterized by loss of awareness at seizure onset, behavioral arrest, facial grimace to the right, and dystonic stiffening of right upper extremity with hand in a C-shape, or at times stiffening of bilateral upper and lower extremities resulting in falls. Seizures had no associated aura or post-ictal symptoms, and the patient did not remember events due to seizures. His seizures occurred 2-10 times per month, and did not resolve with appropriate doses of clobazam, oxcarbazepine, lacosamide, levetiracetam, carbamazepine, and valproic acid. He was born full term following an uncomplicated pregnancy and had no epilepsy risk factors including trauma, CNS infections, or a family history of epilepsy. His neurological examination was normal. Neuropsychiatric testing was notable for a full-scale IQ of 114, with a strong verbal com- The patient was initially reluctant to consider epilepsy surgery, but after two additional years of uncontrolled seizures he agreed to a phase one epilepsy surgery evaluation.
## Scalp eeg findings
Video-EEG showed rare spikes in the left anterior temporal region. Over 20 typical seizures were recorded, generally lasting 30-60 s. Most were electrographically nondescript, predominantly characterized by left hemisphere or diffuse delta activity often obscured by EMG artifact, followed by left temporal post-ictal slowing. Some seizures had similar onset but become focal to bilat-eral tonic-clonic seizures. Some seizures were better localizing, with sharply contoured theta activity in the left temporal head region.
## Neuroimaging findings
His initial 3T MRI brain showed mild gliosis and ex vacuo dilation of the left greater than right temporal horn. MRI brain was repeated during his phase one evaluation and showed a small left anterior temporal encephalocele with subtle cortical protrusion into the posterior aspect of the left sphenotemporal buttress. No other structural abnormalities were observed. Paired ictal/interictal single-photon emission computed tomography (SPECT) imaging showed a corresponding area of focal hyperperfusion for a seizure with bilateral upper extremity stiffening. PET scan did not show focal changes in cerebral metabolism.
## Pre-surgical seeg mapping
The identification of the left temporal encephalocele that was concordant with the electrographic onset of seizures that localized well on scalp EEG provided a promising surgical target for lesionectomy. However, the patient was hesitant to proceed with epilepsy surgery due to the potential for ongoing seizures with lesionectomy or verbal memory deficits with left anterior temporal lobectomy and amygdalohippocampectomy (ATLAH). To better localize seizure onset zone, the patient underwent phase 2 epilepsy surgery workup with sEEG. Nine depth electrodes were implanted in the left temporal lobe to determine whether mesial-temporal structures were part of his seizure onset zone. Nine spontaneous seizures were recorded on sEEG with no involvement of the mesial temporal structures. Extra-operative electrocortical stimulation was also performed to map language and stimulate typical seizures. Stimulation was delivered in five second trains at 50 Hz, beginning with electrode nearest the encephalocele. On stimulation of the second contact pair with relatively low current (2 mA), the patient experienced severe 7/10 burning, lancinating pain in the left cheek for the duration of the stimulus train, accompanied by subtle rapid twitching of the left lower eyelid. The patient understandably refused to continue unless providers could explain the cause of his discomfort and guarantee that it would not recur.
After some thought, the team explained that the most likely cause of his pain was unintentional stimulation of the trigeminal nerve, which lies just inferior to the encephalocele. Stimulation of the temporal cortex is not associated with pain, and in our experience electrocortical stimulation of electrodes adjacent to the dura at much higher currents has not resulted in pain. Thus, we reasoned his pain was likely to arise from stimulation of another adjacent structure. The quality of the patient's pain was akin to that of trigeminal neuralgia, defined as a sudden, usually unilateral, severe, brief, stabbing, recurrent pain in the distribution of one or more branches by the trigeminal nerve [bib_ref] Trigeminal neuralgia and facial nerve paralysis, Borges [/bib_ref]. Although his pain that was time locked to electrocortical stimulation differs from trigeminal neuralgia in that it is not an ongoing disorder, the similar character of pain implicated the trigeminal nerve as the likely source of the patient's discomfort. While the trigeminal nerve does not provide motor innervation to the eyelid, the patient's unilateral eyelid twitching was thought to result from stimulation of the left orbicularis oculi via connections between distal branches of the trigeminal and facial nerve. The patient was satisfied with this explanation and completed mapping in contacts remote to the encephalocele.
## Intraoperative findings
Based on the sEEG findings, the patient later underwent resection of the left temporal pole sparing the mesial temporal structures, as well as repair of the skull base defect with autologous temporalis fascia graft. Intraoperatively, the V2 segment of the trigeminal nerve was visualized passing through the foramen rotundum adjacent to the encephalocele. Intra-op electrocorticography showed frequent spikes in the left anterior temporal region with no spikes at post-op resection margins.
## Post-operative course
The patient recovered well following the surgery and was discharged on postoperative day three. He described left-sided facial numbness beneath his eye during the three months following sur- gery which subsequently resolved. He remained seizure-free at his last follow-up one year after surgery.
# Discussion
One of the most common localizations of focal epilepsy is the temporal lobe. Individuals with TLE may present with a wide variety of symptoms, including visceral symptoms (rising epigastric sensation), olfactory or gustatory hallucinations, autonomic symptoms, psychic sensations (déjà vu, jamais vu, depersonalization, panoramic visions), affective symptoms (dysphoria, euphoria, fear, terror, impending doom), behavioral arrest with loss of awareness and motor automatisms, or better lateralizing motor symptoms such as dystonic posturing of contralateral limbs. Temporal lobe seizures may be of structural, genetic, autoimmune, or infectious etiology. As structural causes are most amenable to treatment with surgical resection, imaging plays an essential role in the workup of patients with TLE.
Improved resolution of neuroimaging has led to increased recognition of temporal encephalocele as a potentially curable cause of drug-resistant TLE [bib_ref] Epilepsy surgery for skull-base temporal lobe encephaloceles: should we spare the hippocampus..., Bannout [/bib_ref] [bib_ref] The usefulness of stereo-electroencephalography (SEEG) in the surgical management of focal epilepsy..., Souza [/bib_ref]. Temporal encephaloceles are easily overlooked, and a high index of suspicion should be maintained in cases of non-lesional TLE [bib_ref] Small temporal pole encephaloceles: a treatable cause of ''lesion negative" temporal lobe..., Abou-Hamden [/bib_ref] [bib_ref] Small temporal pole encephalocele: a hidden cause of ''normal" MRI temporal lobe..., Toledano [/bib_ref] [bib_ref] Anterior temporal encephaloceles: elusive, important, and rewarding to treat, Tse [/bib_ref]. This case shows that even with 3T MRI, a temporal encephalocele may not be initially identified, and underscores the utility of repeat MRI imaging in cases of drug-resistant focal epilepsy. When identified, radiologic features of the encephalocele do not predict the seizure onset zone, which may not be adjacent to the encephalocele [bib_ref] Anterior temporal encephaloceles: elusive, important, and rewarding to treat, Tse [/bib_ref]. Additionally, a recent study suggests that the majority of patients with temporal encephalocele may not have epilepsy at the time of imaging [bib_ref] MR imaging features of middle cranial fossa encephaloceles and their associations with..., Pettersson [/bib_ref]. Therefore, it should not be assumed that a temporal encephalocele identified on MRI is the underlying cause of TLE. Instead, clinical electrophysiology studies such as sEEG should be used together with neuroimaging findings to localize the seizure onset zone and to determine an appropriate surgical approach.
Temporal encephalocele has been treated with a variety of surgical approaches ranging from lesionectomy to ATLAH [bib_ref] Epilepsy surgery for skull-base temporal lobe encephaloceles: should we spare the hippocampus..., Bannout [/bib_ref] , which have different risk/benefit profiles. The utility of sEEG in surgical planning has been previously described [bib_ref] The usefulness of stereo-electroencephalography (SEEG) in the surgical management of focal epilepsy..., Souza [/bib_ref]. In this case, there was concern that, given this patient's excellent verbal function, resection of the dominant mesial temporal structures could lead to deficits in verbal memory. sEEG established that the seizure onset zone was proximal to the encephalocele and did not include the mesial temporal structures, which were consequently spared to reduce the risk of postoperative deficits in verbal memory. The patient was very concerned about verbal memory deficits; it was only because this risk was minimized that the patient chose epilepsy surgery as a treatment option.
In this case, stimulation of cortex adjacent to encephalocele elicited ipsilateral left-sided facial pain for the duration of the stimulation. Pain is not evoked by electrocortical stimulation with a few exceptions: painful sensations have been described with stimulation of primary sensory areas or posterior insula [bib_ref] On the origin of painful somatosensory seizures, Montavont [/bib_ref]. Mechanical or electrical stimulation of dura has been shown to cause pain, but to our knowledge has not been described in the sEEG literature despite routine stimulation of electrodes proximal to dura at much higher amplitudes [bib_ref] Electrical stimulation mapping of brain function: a comparison of subdural electrodes and..., Grande [/bib_ref] [bib_ref] Dural and pial pain-sensitive structures in humans: new inputs from awake craniotomies, Fontaine [/bib_ref] [bib_ref] Referral of pain from dural stimulation in man, Wirth [/bib_ref]. Additionally, dural stimulation does not explain associated unilateral eyelid movements that were time locked to stimulation. Thus, given the distribution and lancinating quality of the pain, it likely was caused by inadvertent stimulation of the V2 branch of the left trigeminal nerve through the adjacent temporal encephalocele.
Finally, it is important to consider the risk of injury to the trigeminal nerve when planning a surgical approach for temporal encephalocele. While there have not been reported cases of trigeminal nerve damage specifically from temporal encephalocele resection, a recent report described two cases of trigeminal neuropathic pain resulting from anterior temporal lobectomies for drugresistant epilepsy [bib_ref] Trigeminal neuropathic pain as a complication of anterior temporal lobectomy: report of..., Gill [/bib_ref]. Additionally, cases of trigeminal neuralgia secondary to an unresected temporal lobe encephalocele have previously been reported, highlighting the anatomical proximity of temporal encephaloceles and the trigeminal nerve [bib_ref] Unusual presentations of middle fossa encephaloceles: report of two case, Shafa [/bib_ref]. Recently a novel case series was published that described the ablation of temporal encephaloceles with magnetic-resonance guided laser interstitial thermal therapy (MRgLITT) and reported excellent post-operative seizure outcomes [bib_ref] Temporal encephalocele: a novel indication for magnetic resonance-guided laser interstitial thermal therapy..., Ranjan [/bib_ref]. The two patients described in this series were similar to the patient in this report: they presented with drug-resistant focal impaired awareness seizures secondary to an anterior temporal encephalocele and underwent epilepsy surgery during adolescence. While MRgLITT is an attractive option that is less invasive than lesionectomy or ATLAH, this case highlights the proximity of the trigeminal nerve to the encephalocele and the potential for thermal damage to the nerve during the procedure.
# Conclusion
Temporal lobe encephaloceles are an important and potentially unrecognized cause of drug-resistant temporal lobe epilepsy that is treatable with epilepsy surgery. Lesionectomy, ATLAH and MRgLITT have all demonstrated favorable post-operative seizure outcomes. This case highlights the anatomical proximity of the trigeminal nerve to the temporal encephalocele, an important consideration in the selection of a surgical approach.
# Ethical statement
LCS, DH, RA, JB and ATK have no relevant financial or nonfinancial relationships to disclose. Informed consent was obtained by the individual described in the case report.
## Declaration of competing interest
[fig] Figure 1: (A) Scalp EEG recording of the patient's typical electrographic seizure, showing sharply contoured theta activity in the anterior left temporal head region (T1). (B) Representation of the patient's typical electrographic seizure captured by stereo EEG. Phase 1 consists of fast (20 Hz) activity in the deep contacts of 1MTP (top three waveforms, see black arrow). The patient was asymptomatic throughout phase 1 evaluation. Phase 2 monitoring of seizure onset began with the simultaneous onset of 4-8 Hz evolving spikes in the temporal pole, amygdala and hippocampus, clinically accompanied by unresponsiveness and bilateral stiffening of the arms and hands. [/fig]
[fig] Figure 2: (A) Sagittal view of the left temporal pole encephalocele (red arrows) as demonstrated by pre-operative MRI. (B) Sagittal view of anatomical position of the sEEG electrodes. (C) Pre-operative coronal-oblique reformatted CT image demonstrating the spatial relationship between the vidian canal (red arrow), foramen rotundum (yellow arrow), and temporal pole encephalocele (green circle). (D) Intra-operative extradural exposure showing defects in the sphenoid bone and encephalocele observed during resection. (E) Uncovered V2 segment of trigeminal nerve passing through foramen rotundum in extradural subtemporal exposure. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) [/fig]
|
Unique Transcriptome Patterns of the White and Grey Matter Corroborate Structural and Functional Heterogeneity in the Human Frontal Lobe
The human frontal lobe has undergone accelerated evolution, leading to the development of unique human features such as language and self-reflection. Cortical grey matter and underlying white matter reflect distinct cellular compositions in the frontal lobe. Surprisingly little is known about the transcriptomal landscape of these distinct regions. Here, for the first time, we report a detailed transcriptomal profile of the frontal grey (GM) and white matter (WM) with resolution to alternatively spliced isoforms obtained using the RNA-Seq approach. We observed more vigorous transcriptome activity in GM compared to WM, presumably because of the presence of cellular bodies of neurons in the GM and RNA associated with the nucleus and perinuclear space. Among the top differentially expressed genes, we also identified a number of long intergenic non-coding RNAs (lincRNAs), specifically expressed in white matter, such as LINC00162. Furthermore, along with confirmation of expression of known markers for neurons and oligodendrocytes, we identified a number of genes and splicing isoforms that are exclusively expressed in GM or WM with examples of GABRB2 and PAK2 transcripts, respectively. Pathway analysis identified distinct physiological and biochemical processes specific to grey and white matter samples with a prevalence of synaptic processes in GM and myelination regulation and axonogenesis in the WM. Our study also revealed that expression of many genes, for example, the GPR123, is characterized by isoform switching, depending in which structure the gene is expressed. Our report clearly shows that GM and WM have perhaps surprisingly divergent transcriptome profiles, reflecting distinct roles in brain physiology. Further, this study provides the first reference data set for a normal human frontal lobe, which will be useful in comparative transcriptome studies of cerebral disorders, in particular, neurodegenerative diseases.
# Introduction
The human cerebrum is extraordinarily complex and is composed of billions of neurons and trillions of synaptic connections. Neurons are organized into circuit assemblies that are modulated by specific interneurons and non-neuronal cells. The frontal lobe is often considered the most highly developed and most human featured brain region. As the prefrontal and frontal cortexes exert executive type control over other structures, they are expected to show significant connectivity to other brain regions. This part of the brain manages the most complex thought, decision making, planning, conceptualization, attention control, and working memory [bib_ref] A unified account of cognitive impairments following frontal lobe damage: the role..., Kimberg [/bib_ref] [bib_ref] The prefrontal cortex: categories, concepts and cognition, Miller [/bib_ref].
Broadly speaking, the human cortex can be divided into the phylogenetically older allocortex and the newer neocortex. The neocortex has expanded the most in humans and is composed of six superimposed layers with distinct cellular compositions; the thickness of each layer differs in each cortical lobe [bib_ref] Architecture of cerebral cortex, Ziles [/bib_ref]. The most prominent distinction of the cerebrum is its division into outer cortical grey matter (GM) and inner white matter (WM). GM consists of neural cell bodies, their dendrites, and parts of their axons, as well as glial cells, mainly astrocytes. In contrast, WM is mainly aggregations of myelinated and non-myelinated axons linking different cortical and subcortical regions [bib_ref] Development and evolution of the human neocortex, Lui [/bib_ref]. Of long-neglected significance, WM has recently started to be a subject of intensive studies due its involvement in the development of working memory capacity and reading ability [bib_ref] Maturation of white matter is associated with the development of cognitive functions..., Nagy [/bib_ref]. Recently, the frontal WM has been proposed as a major contributor to human brain enlargement and higher structural connectivity, as compared to other primates [bib_ref] Prefrontal white matter volume is disproportionately larger in humans than in other..., Schoenemann [/bib_ref] [bib_ref] Frontal white matter volume is associated with brain enlargement and higher structural..., Smaers [/bib_ref].
Despite several RNA-Seq-based studies investigating whole transcriptome profiles of cerebral tissue in pathological conditions such as Alzheimer's disease (AD) [bib_ref] RNA-Seq analysis of the parietal cortex in Alzheimer's disease reveals alternatively spliced..., Mills [/bib_ref] [bib_ref] Whole transcriptome sequencing reveals gene expression and splicing differences in brain regions..., Twine [/bib_ref] or autism [bib_ref] Transcriptomic analysis of autistic brain reveals convergent molecular pathology, Voineagu [/bib_ref] there has been no systematic attempt to investigate transcriptomic landscape of normal brain tissue with resolution of RNA sequencing. Hawrylycz and colleagues recently published a microarray-based transcriptome profile of the distinct brain regions in two individuals [bib_ref] An anatomically comprehensive atlas of the adult human brain transcriptome, Hawrylycz [/bib_ref]. They provided important insights into the spatial distribution of expression across well-defined neuroanatomical regions. The critical conclusion of this study was the importance of local gene expression patterns for the maintenance of physiological uniqueness within these regions. Because this analysis was performed using microarray technology, it does not provide further information on posttranscriptional control in the brain, namely, alternative splicing.
Here, we report, for the first time, detailed transcriptome profiles of GM and WM of the human lobe using RNA-Seq. Comparative analysis of the gene and isoform expression, combined with the pathway analysis, revealed surprisingly distinct transcriptome patterns reflecting the contribution of different glial cell types and neuronal structures to WM and GM, respectively. Moreover we observe elevated expression of lincRNAs in WM, as well as isoform switching between WM and GM for genes encoding DNA binding proteins and proteins involved in signal transduction.
# Materials and methods
## Human brain tissue
Human brain tissues were obtained from the Sydney Brain Bank and NSW Tissue Resource Centre, part of The Australian Brain Bank Network funded by the National Health and Medical Research Council of Australia. Ethics approval was from the University of New South Wales Human Research Ethics Committee. Frozen brain tissue samples from superior frontal gyrus (SFG) GM and WM were collected from three individuals aged 79, 94 and 98. The PMI of samples ranged 8-24 hrs and pH 5.77-6.65. All three brains were pathologically diagnosed and were free of any pathology or neurodegeneration.
## Rna isolation, library preparation and sequencing
Total RNA was isolated using RNeasy Lipid Tissue Midi Kit (Qiagen) followed by RNase-free DNase treatment to remove traces of genomic DNA. The RNA quality of the total RNA was assessed using the Agilent 2100 Bioanalyser RNA Nano Chip and the RIN values ranged between 6.0 and 7.0. This RIN range was previously shown to have a little effect on relative gene expression ratios [bib_ref] Reverse transcription-quantitative polymerase chain reaction: description of a RIN-based algorithm for accurate..., Ho--Pun-Cheung [/bib_ref] (and our unpublished observations). Six RNA samples (three WM and three GM) were prepared for sequencing according to the Illumina TruSeq RNA sample preparation guide and subjected to 100 bp pairedend sequencing using Illumina HiSeq1000. The sequence data have been submitted to the NCBI Short Read Archive with accession number SRA091951.
## Mapping of rna-seq reads using tophat
Bioinformatics analysis was carried out using Galaxy; an open access web-based program that contains a variety of next-generation sequencing analysis tools including, TopHat and the Cufflinks package [bib_ref] Galaxy: a web-based genome analysis tool for experimentalists, Blankenberg [/bib_ref] [bib_ref] Galaxy: a comprehensive approach for supporting accessible, reproducible, and transparent computational research..., Goecks [/bib_ref]. The Galaxy server was based at the Garvan Institute, Sydney, Australia. Using TopHat the reads were processed and aligned to the H. sapiens reference genome (build hg19). TopHat utilizes the ultra highthroughput short read aligner Bowtie to align the RNA-Seq reads, the reads are then analyzed and splice junctions between the exons are identified [bib_ref] TopHat: discovering splice junctions with RNA-Seq, Trapnell [/bib_ref].The default parameters for TopHat were used. Subsequently the aligned reads from each sample were analyzed for 5'-3' end bias using RSeQC [bib_ref] RSeQC: quality control of RNA-seq experiments, Wang [/bib_ref].
## Transcript assembly with cufflinks
The aligned reads were processed with Cufflinks. Cufflinks assembles the RNA-Seq reads into individual transcripts, inferring the splicing structure of the genes [bib_ref] Transcript assembly and quantification by RNA-Seq reveals unannotated transcripts and isoform switching..., Trapnell [/bib_ref]. Cufflinks assembles the data parsimoniously giving a minimal set of transcripts that fits the data. Cufflinks normalizes the RNA-Seq fragment counts to estimate the abundance of each transcript. Abundance was measured in the units of fragments per kilobase of exon per million fragments mapped (FPKM). For this analysis a .GTF annotation file (iGenomes UCSC hg19 gene annotation) was used to guide the assembly.
## Differential analysis with cuffmerge and cuffdiff
The alignment files produced by TopHat are merged for CuffDiff processing so that combinatorial pairwise sample comparison is performed. The output GTF files from each of the Cufflinks analysis and the .GTF annotation file were sent to Cuffmerge [bib_ref] Transcript assembly and quantification by RNA-Seq reveals unannotated transcripts and isoform switching..., Trapnell [/bib_ref]. Cuffmerge takes these files and amalgamates them into a single unified transcript catalog; it also filters out any transcribed fragments that may be artifacts. The inclusion of the reference annotation allows gene names and other details such as, transcript ID, exon number, transcription start site ID and coding sequence ID to be added to the merged transcript catalogue. It also allows for the gene and transcripts to be classified as known or novel. The merged GTF file was then fed to Cuffdiff along with the original alignment files produced from TopHat. Cuffdiff takes the replicates from each condition and looks for statistically significant changes in gene expression, transcript expression, splicing and promoter use. Cuffdiff uses a corrected p-value, known as the q-value to determine if the differences between the two groups are significant (q-value<0.05).
## Visualization with cummerbund and interactive genome viewer
The resultant Cuffdiff output files were fed into CummeRbund. CummeRbund is an R package that is designed to simplify the analysis of the Cuffdiff outputs. CummeRbund is user friendly and allows for easy data exploration and figure generation [bib_ref] Differential gene and transcript expression analysis of RNA-seq experiments with TopHat and..., Trapnell [/bib_ref]. The Broad Institutes Integrative Genome Viewer (IGV) (http:// www.broadinstitute.org/igv/), was used to visualize Cufflinks GTF outputs, this allowed for comparisons to be made between genes of known structure and the gene structure of novel transcripts identified by Cufflinks [bib_ref] Integrative genomics viewer, Robinson [/bib_ref] [bib_ref] Integrative Genomics Viewer (IGV): high-performance genomics data visualization and exploration, Thorvaldsdóttir [/bib_ref].
## Gene-set enrichment analysis with david
The gene list of differential expressed genes was split into two groups; those up-regulated in GM and those up-regulated in WM. Only annotated genes can be utilized by enrichment tools, all novel genes and indecisively annotated genes were removed. Each of these lists was fed into the Database for Annotation, Visualization and Integrated Discovery (DAVID) (http://david.abcc.ncifcrf.gov/) [bib_ref] Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources, Huang [/bib_ref]. DAVID tested the gene ontology (GO) terms for over representation in each of the gene lists. The GO terms list produced by DAVID, were processed using the 'Enrichment Map' plug in for Cytoscape (http://www.cytoscape.org/) [bib_ref] Integration of biological networks and gene expression data using Cytoscape, Cline [/bib_ref]. This produces a visual output of the text based GO term lists.
## In situ hybridization validation
Our RNA-Seq expression data were compared with in situ hybridization (ISH) data from the Allen brain atlas database (http://www.brain-map.org/), which uses RNA probes to measure gene expression in normal human dorsolateral frontal cortex.
# Results
## Total transcription in gm and wm
An expression catalogue containing 32,740 genes was created by the Cufflinks package. Of this catalogue, 18,362 of the genes were considered expressed in the analyzed brain samples. This number includes 3,615 unannotated genes, 44 small nucleolar RNAs (snoRNAs), 67 micro RNAs (miRNAs), 40 lincRNAs and 52 RNAs of uncharacterized function (locRNAs). Short non-coding RNA sequences were presumably derived from their polyadenylated long non-coding RNA parent transcript or might be carryovers captured during library preparation. The cumulative coverage of the expressed transcripts in GM and WM was approximately 37% of the entire human genome; this figure includes the transcribed introns that were spliced out. There was no significant difference between the total human genome coverage of GM and WM.
The developed isoform catalogue assembled by Cufflinks contained over 85,000 distinct isoforms (splice variants). Among these isoforms, 36,306 were identified as being expressed in the selected regions of brain tissue, including 3844 unannotated isoforms, which equates to approximately two isoforms per gene. The actual distribution of isoforms per gene reveals a different picture; there were 9,427 genes with 1 isoform and 8,653 genes with between two and seven isoforms. There were only 107 genes with eight or more isoforms. This suggests that almost 50% of genes undergo alternative splicing.
Analysis using RSeQC showed that the aligned reads from each RNA sample had a similar degree of bias towards the 3' end of each transcript . This result was expected due to the poly-A selection of RNA for sequencing. Of note, the Cufflinks package, utilized in this study, has been designed to correct for any sequence bias that occurs due to steps undertaken during template preparation [bib_ref] Improving RNA-Seq expression estimates by correcting for fragment bias, Roberts [/bib_ref].
## Differentially expressed genes and isoforms
Within in the set of 18,362 expressed genes, a total of 1,652 were identified as differentially expressed between the two conditions (q-value<0.05) . This included 1,218 that were up-regulated in GM and 434 that were up-regulated in WM [fig_ref] Table 1: Top 10 up-regulated genes in GM and WM [/fig_ref]. Overall, there was a transcription bias toward GM. To preclude possible discordant results the gene lists were filtered so that any gene with an FPKM<1 for both conditions was excluded. An FPKM<1 means that there is less than one fragment per million aligned fragments mapped onto a 1-kb exon; this can be considered the result of background noise arising from erroneous sequencing or statistical errors during mapping. This reduced the overall list of differentially expressed genes to 1,591, of which 1,162 genes were upregulated in GM and 429 genes were up-regulated in WM, the respective top 10 up-regulated genes in GM and WM by foldchange are shown in [fig_ref] Table 1: Top 10 up-regulated genes in GM and WM [/fig_ref]. Of the top 10 up-regulated genes in GM, there were two unannotated genes, and among the top 10 up-regulated genes in WM, there were three unannotated genes. On further inspection of the size and location of these genes, it was established that these genes fit the criteria for lincRNAs. Further, there was one annotated lincRNA (LINC00162) in the top 10 up-regulated genes in WM.
Comparing the entirety of the lists, there were more lincRNAs in WM than GM (3 to 1). While there were more unannotated up-regulated genes in GM (80) than WM (57), the unannotated genes in WM represented a greater proportion of the up-regulated genes in WM than the proportion of upregulated genes in GM represented by unannotated genes. A chi-squared statistical test was performed to determine if there was any statistically significant difference between the proportions of unannotated genes in each of the data sets. The chi-squared test conclusively showed that the difference in the proportions of unannotated genes in each data set was greater than what would be expected by chance alone (p<0.00005), suggesting that, overall, there was a statistically significant higher proportion of unannotated genes identified in WM.
Cuffdiff identified 36,306 isoforms in both GM and WM, including 3844 unannotated isoforms [fig_ref] Figure 2: Volcano plot of isoform expression in GM and WM [/fig_ref]. Of this set, 882 isoforms were significantly differentially expressed between GM and WM (q-value<0.05), including 681 isoforms up-regulated in GM and 201 up-regulated in WM [fig_ref] Table 2: Top 10 up-regulated isoforms in GM and WM [/fig_ref]. When the criteria of having at least one condition with an FPKM>1 was applied this list was reduced to a total of 856 isoforms, with 657 isoforms up-regulated in GM and 199 up-regulated in WM. The top ten differentially expressed isoforms sorted by fold-change from GM and WM respectively are listed in [fig_ref] Table 2: Top 10 up-regulated isoforms in GM and WM [/fig_ref]. This list shares almost no commonality with [fig_ref] Table 1: Top 10 up-regulated genes in GM and WM [/fig_ref] ; only C1QL3 and FLJ41278 (up-regulated in GM) appear in both tables. All of the top 10 up-regulated isoforms in WM are unique to WM and have no expression in GM [fig_ref] Table 2: Top 10 up-regulated isoforms in GM and WM [/fig_ref]. In fact, the top 24 up-regulated isoforms in WM were not detected at all in GM; in contrast, there was only one unique up-regulated isoform present in GM [fig_ref] Table 2: Top 10 up-regulated isoforms in GM and WM [/fig_ref].
## G protein-coupled receptor 123
The G protein-coupled receptor 123 (GPR123) gene had one such isoform that was uniquely expressed in WM. However, when analyzed as an entire gene, GPR123 was found to be 5fold up-regulated in GM, with overall expression levels of 10 FPKM in GM and 2 FPKM in WM. Further investigation revealed unique splicing patterns that underline the importance of analyzing genes at the isoform level.
Overall, there were four splice variants identified in both GM and WM, including two previously identified splice variants-GPR123-002 (ENST00000392607) and GPR123-003 (ENST0000039606)-and two novel splice variants-GPR123-004 and GPR123-005 [fig_ref] Figure 3: Splice variants of GPR123 [/fig_ref]. The two previously identified isoforms (GPR123-002 and GPR123-003) are both protein coding; GPR123-003, with its extended N-terminal, can be considered a full-length, fully functional protein. By comparison, GPR123-002 was missing the first four exons. Its first exon is an untranslated alternate exon. The translation of GPR123-002 results in a 97-amino acid N-terminal truncated protein. The exon structure of the novel isoform GPR123-004 was most similar to GPR123-003, but it lacks the 5th exon. The lack of this exon is predicated to result in a 199-amino acid truncation at the C-Terminal end of the GPR123-004 protein when compared to GPR123-003. GPR123-005 was most similar to GPR123-002, albeit with a slight modification to the 5' region. Translation of GPR123-005 was predicted to result in a protein similar to GPR123-002. Among the four GPR123 isoforms, three different transcription start sites (TSS) were identified and GPR123-003 and GPR123-004 had the same TSS, while the two N-terminal truncated isoforms (GPR123-002 and GPR123-005) both utilized unique TSS.
In GM, the dominant isoform was found to be the novel truncated isoform GPR123-005. This isoform contributed to approximately 70% of the total GPR123 expression in GM, and the full length protein coding isoform GPR123-003 and the truncated protein coding isoform GPR123-002 contributed to approximately 20% and 10% of the total GPR123 expression seen in GM, respectively [fig_ref] Figure 4: Expression levels of the GPR123 isoforms [/fig_ref]. These three isoforms were all expressed at a higher level in GM when compared to WM, with the 13-fold up-regulation of the GPR123-005 considered to be statistically significant. The TSS of GPR123-005 was also up-regulated in GM. GPR123-004 was not expressed at all in GM. In WM, the three GM expressed isoforms were expressed at very low levels (<0.6 FPKM). In GPR123-004, the isoform that was not expressed in GM becomes the dominant isoform, contributing to 70% of all expression of GPR123 in WM. In this case, the full length protein coding isoform GPR123-003 contributed to only 1.5% of the total GPR123 expression.
## Figure 1. volcano plot of gene expression in gm and wm.
The fold-change of the genes was relative to their expression in WM. Those genes with a negative fold-change were up-regulated in GM (down-regulated in WM) and those genes with a positive fold-change were up-regulated in WM (down-regulated in GM). Genes that were statistically significant (q-value<0.05) are shown in red and were listed in [fig_ref] Table 1: Top 10 up-regulated genes in GM and WM [/fig_ref]. This figure demonstrates that a larger number of genes were significantly up-regulated in GM (1218) than were up-regulated in WM (434). Overall there was a greater spread of data for the genes that are up-regulated in GM. Between GM and WM, there was a switch in the dominant GPR123 isoforms from GPR123-005 to GPR123-004. Interestingly, although total GPR123 was expressed at a lower level in WM than GM, the dominant isoform contributed to the same percentage of the overall expression.
The isoform expression patterns of the genes that had the top three annotated differently expressed isoforms in GM and the top three annotated differently expressed isoforms in WM were also analyzed [fig_ref] Figure 2: Volcano plot of isoform expression in GM and WM [/fig_ref] -S7) and shown that the total expression levels of many genes result from dominant expression of one of the isoforms whereas the remaining splice variants are marginally present. Thus it is important to examine gene expression with resolution to individual transcriptional isoforms.
## Expression of well-known cell type markers
While few studies explore the transcriptome profiles of GM and WM, there are a number of well-characterized cell type markers [bib_ref] A transcriptome database for astrocytes, neurons, and oligodendrocytes: a new resource for..., Cahoy [/bib_ref]. As the cellular make up of GM and WM is considerably different, certain cell type markers were used to reflect the composition of the respective regions; in this case, established neuronal markers were used as surrogate markers for GM and established oligodendrocyte markers were used as surrogate markers for WM [fig_ref] Table 3: Cell type markers. [/fig_ref]. While oligodendrocytes are not entirely specific to WM, they would be expected to appear at a higher expression level in WM than GM.
Seven neuronal markers were chosen, all of which were expressed at higher levels in GM than WM. For five of these genes (NEFL, GABRA1, SYT1, SLC12A5, SV2B), the upregulation in GM was statistically significant (q-value<0.05) [fig_ref] Figure 5: Heatmap of gene expression of neuronal cell markers [/fig_ref]. The neuronal markers clearly correlate with GM, showing limited to no expression in WM. Ten different oligodendrocyte markers were chosen; all were expressed at a higher level in WM than in GM. Seven of these genes (SOX10, GJC2, MOG, MAG, MAL, GAL3ST1, and UGT8) were considered differentially expressed, being up-regulated in WM (q-value<0.05) [fig_ref] Figure 6: Heatmap of gene expression of oligodendrocyte cell markers [/fig_ref]. These results demonstrate that a correlation exists between the cellular composition of GM and WM and the results produced by the transcriptome sequencing.
## Validation with in situ hybridization
Six genes from the RNA-Seq dataset were selected for further validation through the use of the Allen Brain Atlas ISH database. The first two selected genes were neurofilament heavy polypeptide (NEFH) and myelin oligodendrocyte glycoprotein (MOG). NEFH encodes for neurofilament-heavy polypeptides, which are present in the chains that form one of the integral components in neuronal cytoskeleton neurofilaments, making it a relevant gene for the neuron rich GM [bib_ref] Neurofilament heavy polypeptide regulates the Akt-beta-catenin pathway in human esophageal squamous cell..., Kim [/bib_ref] [bib_ref] Neuronal intermediate filaments, Lee [/bib_ref]. NEFH had a FPKM of 57.76 in GM and a FPKM of 0.95 in WM. MOG is myelin specific in the central nervous system, and its expression level parallels the myelination of axons, suggesting that it plays an integral role in WM [bib_ref] The structure and function of myelin oligodendrocyte glycoprotein, Johns [/bib_ref]. MOG had a FPKM of 8.33 in GM and 49.08 in WM. The ISH results [fig_ref] Figure 7: Allen Human Brain Atlas in situ hybridisation for NEFH and MOG genes [/fig_ref] correlate strongly with the RNA-Seq data. As expected, the NEFH ISH slide shows a higher level of expression in GM, while there is almost no expression in WM. Conversely, the ISH slides of MOG confirm that MOG is expressed at its highest level in WM. The other four genes (RGS4, CAMK2A, SLC17A7, NEFM) which were selected for validation with in situ hybridization also correlated well with the RNA-Seq data [fig_ref] Figure 8: Heatmap of high abundant genes specifically expressed in GM and WM [/fig_ref]. Again, these results demonstrate the accuracy of the RNA-Seq results.
## Novel gene markers for grey matter and white matter
Validation of the RNA-Seq results via in situ hybridization and correlation with well-known cell type makers demonstrated that RNA-Seq is biologically accurate and thus a useful tool for transcriptome profiling of the human brain. Strict selection criteria were applied to all genes that were up-regulated in GM to define a normal healthy GM transcriptome. The criteria included genes that were >20-fold up-regulated in GM, had an FPKM>5 in GM, and had an FPKM<1 in WM. The aim of this selection stringency was to find genes that were expressed at significantly high levels in GM while not being expressed at all in WM. The criteria reduced the number of genes differentially expressed in GM from 1,218 to 145 [fig_ref] Table 3: Cell type markers. [/fig_ref]. As there is less transcriptional activity in WM, the selection criteria were relaxed to the following: >4 fold up-regulation in WM, an FPKM>5 in WM, and a FPKM <5 in GM. These criteria reduced the list of significantly expressed genes in WM from 434 to 76 [fig_ref] Table 3: Cell type markers. [/fig_ref]. The top 40 genes from both GM and WM are shown in the form of a heat map [fig_ref] Figure 8: Heatmap of high abundant genes specifically expressed in GM and WM [/fig_ref]. The heat map shows that a marked contrast exists between the refined expression profiles of the two tissue types. In the genes selected as GM markers, the differences between the two tissue types were more distinct than the expression differences seen in the WM. As WM is a less transcriptionally active tissue when compared to GM, WM may experience a flow of RNA from the more transcriptionally active GM. These two gene lists set a baseline for the expression profiles for healthy GM and WM.
## Enrichment map of pathway analysis
The gene lists of differentially expressed genes from GM and WM were fed into DAVID. DAVID sorts genes by gene ontology (GO) terms. The GO terms loosely define the functional relevance of the gene; a gene may belong to numerous ontologies. DAVID then collates the GO terms and determines which ontologies are enriched in the gene list. For GM, a total of 516 different GO terms were considered enriched [fig_ref] Table 4: Top 10 GO terms in GM and WM [/fig_ref]. The top 10 GO terms in GM sorted by p-value are listed in [fig_ref] Table 4: Top 10 GO terms in GM and WM [/fig_ref]. For GM, the top 10 GO terms related predominately to synapses and various transport activities. For WM, 284 GO terms were identified as enriched [fig_ref] Table 4: Top 10 GO terms in GM and WM [/fig_ref]. The top 10 GO terms in WM sorted by p-value are listed in [fig_ref] Table 4: Top 10 GO terms in GM and WM [/fig_ref]. Of note in this list are the 9th and 10th GO terms, which relate to the ensheathment of neurons and axons. The list of GO terms from GM and WM were then fed into Cytoscape and used to modify the total gene list, creating an enrichment map . The enrichment map revealed several large and distinct clusters related to vesicles and membrane: ion-gated channels, transporters and receptors, neuron morphogenesis, ensheathment and myelination, neuron projection, synaptosomes, neuron morphogenesis, transmission of nerve impulse, and plasticity and axon/dendrite projection. From the enrichment map, it can be seen that there are more GO terms enriched in GM. Also, there are more GM connections between GO terms, which reinforce the observed trend of higher levels of transcriptional activity in GM than in WM. The major clusters of GO terms for GM related to vesicles and membranes are ion-gated channels, transporters and receptors. These groups are related to the transport of substances such as ions throughout cells. Clusters of GO terms related to neuron morphogenesis and neuron projection were enriched by genes from WM and GM. Neuron morphogenesis refers to changes in the underlying neuronal cytoskeleton and its interaction with the plasma membrane [bib_ref] Actin cytoskeleton regulation in neuronal morphogenesis and structural plasticity, Luo [/bib_ref]. It can involve processes pertaining to axon initiation, growth, guidance and branching; dendritic growth, guidance, and branching; and synapse formation and stability. Similarly, neuron projection relates to any process involved in the initiation of neurite protrusion, and subsequent elongation often involving axons and dendrites [bib_ref] Signaling involved in neurite outgrowth of postnatally born subventricular zone neurons in..., Khodosevich [/bib_ref]. Neuron morphogenesis and neuron projection are related to improving communication between neurons and hence will involve a complex interaction between GM and WM. Only one cluster was clearly dominated by WM and was related to ensheathment and myelination, a process performed by glia enriched in WM.
# Discussion
This study is the first comparative transcriptome analysis of GM and WM from the human brain using RNA-Seq. It has shown that, overall, there are high levels of transcription in the human brain, with the identified transcripts covering approximately 37% of the genome. This number is close to the figure described by the ENCODE consortium [bib_ref] Landscape of transcription in human cells, Djebali [/bib_ref] ; the ENCODE study reported that, among 15 different human cell lines, the mean coverage of the human genome by primary transcripts was 39%. The ENCODE study included nonpolyadenylated RNAs, while the current study utilized poly-T oligo-attached magnetic beads for the selection of the RNA fraction for analysis. If the RNA selection for this study also included non-polyadenylated RNAs, it would be expected that the coverage of the genome would be much higher than the 39% suggested by ENCODE. This would suggest that transcription is pervasive in the human brain. Furthermore, this study has also shown that a number of differentially expressed genes and transcripts exist across GM and WM from the same anatomical region of the human brain, namely, the superior frontal gyrus. These results highlight the complexity and variability of the transcriptome and the presence of numerous non-coding elements. Importantly, the transcriptome profiles for each region, and the GO enrichment map correlated well with what is known about the function and composition of WM and GM. This previous information has been predominately established through the use of other techniques, such as histopathology and in situ hybridization.
The human brain is an extremely complex organ. This complexity is not only derived from the sheer number and variety of cells present in the brain, but also through the heterogeneity of the brain, with cellular composition and density having the propensity to vary over short distances. The tissue samples used for the RNA-Seq analysis were taken from two different tissue types (GM and WM) from adjacent regions of the superior frontal gyrus. While the samples were taken from regions of close proximity, their transcriptome profiles were distinct, suggesting that each of the two tissue types have separate functions and further highlights the heterogeneity of the brain tissue across small distances. The heterogeneity of the brain becomes an issue when attempting to gauge differing gene expression levels between case and control samples. If the brain tissue is not selected and matched properly, differences highlighted in the transcriptome profiles may be differences resulting from variation in the composition and function of different brain regions rather than disease-related changes in the transcriptome. While the heterogeneity of brain tissue could make it difficult to study, the problem can be overcome through appropriate experimental design. The brain regions being compared must have the same cellular composition and be of the same functional capabilities. Laser capture of specific cell populations could be used to create homogenous samples for RNA-Seq analysis [bib_ref] Laser capture microdissection, Emmert-Buck [/bib_ref].
Several previous attempts were made to establish specific gene expression patterns as markers for GM and WM [bib_ref] Brain perihematoma genomic profile following spontaneous human intracerebral hemorrhage, Rosell [/bib_ref] [bib_ref] Large-scale estimates of cellular origins of mRNAs: enhancing the yield of transcriptome..., Sibille [/bib_ref]. These studies, performed using microarrays, provided only partially overlapping results in terms of GM-and WM-specific genes. These discrepancies are due to the different cortical regions analyzed between studies; thus, the contribution of genes specifically expressed in neurons and oligodendrocytes variably contributed to the calculated GM/WM ratio [bib_ref] Architecture of cerebral cortex, Ziles [/bib_ref]. Another difficulty stems from the microarray technique itself, which is limited in dynamic range to quantify gene expression [bib_ref] Transcriptome profiling in neurodegenerative disease, Courtney [/bib_ref]. In the present study, we provide a range of over 40 new gene marker candidates based on their unequivocal abundance in expression in GM and WM. Further, we propose that cumulative gene expression signatures, as now commonly used in cancer research [bib_ref] RNA-Seq and human complex diseases: recent accomplishments and future perspectives, Costa [/bib_ref] , rather than individual genes should be used as markers for GM and WM. The utility of such marker signatures should however be further evaluated using larger cohort of samples.
Within the last few years, the RNA-Seq technique became widely used in transcriptome research, progressively replacing microarray techniques. Surprisingly, little has been done in the field of brain diseases using this sequencing approach. This can be partially caused by the lack of comprehensive reference data sets that might be used for comparative transcriptome analysis. Wu and colleagues recently reported comparative analysis of the transcriptome profiles derived from the superior temporal gyrus (STG) [bib_ref] Transcriptome sequencing revealed significant alteration of cortical promoter usage and splicing in..., Wu [/bib_ref]. This study was however limited to the GM of STG and performed in the context of schizophrenia. Our study provides the first well-defined data sets that can be used to explore transcriptome aberrations affecting the frontal lobe and, in particular, frontal WM as in the case of multiple system atrophy and bipolar disorders [bib_ref] White matter pathology--an endophenotype for bipolar disorder, Borgwardt [/bib_ref] [bib_ref] Neuropathology underlying clinical variability in patients with synucleinopathies, Halliday [/bib_ref]. By combining differential gene and isoform expression analysis with pathway analysis we were able to demonstrate that biological processes specific for WM and GM, such as myelination and ions transportation, respectively are reflected on the transcriptional level. Thus the transcriptome profiling may be effectively used to investigate cellular physiology, while a more global view of genomic expression may be used to study phenotypic features of complex structures such as human cortex.
This study identified high levels of various non-coding RNA classes, including snoRNAs, miRNAs, and lincRNAs; further, there were also high levels of locRNAs, novel genes, and novel isoforms produced from splicing events. It is possible that these last three classes could contribute further to the number of noncoding RNAs found in the human brain. Interestingly, while higher levels of transcriptional activity were found in GM, WM was identified as having higher levels of lincRNAs and a higher proportion of unannotated isoforms. Non coding RNAs (ncRNAs) were previously suggested to be widely expressed across the brain, where they often have cell specific regulatory functions [bib_ref] Long noncoding RNAs in neuronal-glial fate specification and oligodendrocyte lineage maturation, Mercer [/bib_ref]. These ncRNAs may play a role in regulating myelination patterns of axon in WM, a dynamic process that can be altered by experience and can carry on for decades in the human brain [bib_ref] White matter in learning, cognition and psychiatric disorders, Fields [/bib_ref]. The functionality of the unannotated RNAs will need to be explored further in the future. It should also be underlined that possible influence of age on the transcriptional patterns presented here cannot be excluded and will require utilization of larger number of samples to enable regression analysis of covariates.
GPR123 is a member of the human G protein-coupled receptor (GPCR) family. GPCRs play important roles in a variety of sensory systems and help modulate blood pressure, food intake, immune responses, and development. It has been suggested that GPR123 is expressed specifically in the central nervous system (CNS); it also shows high levels of conservation across the vertebrate lineage [bib_ref] The evolutionary history and tissue mapping of GPR123: specific CNS expression pattern..., Lagerström [/bib_ref]. These factors suggest that GPR123 expression may be a fundamental component of the vertebrate CNS.
The expression patterns of GPR123 between GM and WM reveal an interesting individual insight into the complexity of the transcriptome. At the gene level, GPR123 was up-regulated 5fold in GM, suggesting that GPR123 is of greater functional relevance in GM; however, when analyzed at the isoform level, it was shown that the full length isoform GPR123-003 was not the major contributor to the GPR123 overexpression. Instead, the dominant isoform was the C-terminal truncated GPR123-004, which contributed to approximately 70% of total GPR123 expression in GM. The level of GPR123-004 FPKM expression in GM and the use of an alternate TSS suggest that it may be relevant to functional GM. In contrast, the dominant isoform in WM was the N-terminal truncated isoform GPR123-005, which contributed approximately 70% of the total GPR123 expression in WM. It is known that the domain structure of GPR123-004 and GPR123-005 would differ, which would lead to distinct functions for proteins encoded by each isoform. This fact points to the possibility that the GPR123 proteins may carry out distinct tissue-specific functions while being transcribed and translated from the same genomic locus.
While more research needs to be directed at elucidating the role of the differing GPR123 isoforms in the human brain, this gene does demonstrate the complexity of the transcriptome and also shows how the repertoire of transcripts is greatly increased by splicing events. It also illustrates that it is not just total gene expression levels that are important, but which isoforms are contributing to the expression levels. Unfortunately, current pathway analysis bioinformatic tools, such as DAVID, have not yet been set up to delineate between different splice variants.
# Conclusions
This study identified a large number of differentially expressed genes between GM and WM that matched up well with previous studies and the known biology of the human cerebrum, underlining the accuracy and the advantages of using RNA-Seq for the analysis of complex transcriptomes. This set of results form a baseline expression database for healthy GM and WM and can be used as a resource to help detect abnormal gene expression between GM and WM tissues from sufferers of neurodegenerative diseases or psychiatric disorders. This is particularly relevant to WM which may play a major role in AD and multiple system atrophy [bib_ref] Is Alzheimer's a disease of the white matter?, Sachdev [/bib_ref] [bib_ref] Multiple system atrophy with severe involvement of the motor cortical areas and..., Wakabayashi [/bib_ref]. . Gene Ontology terms enrichment map for GM and WM transcriptomes. Each node represents a different GO term, the size of the node relates to the level of enrichment of each term. Red in the centre of the node represents up-regulation in GM, red on the edge of the node represents up-regulation in WM. The connections between each GO term are either green or blue. A green connection between node means that the both GO terms are in the GM lists, blue connections represent appearance of the GO terms in the WM lists. The more closely related GO terms are, the closer they appear on the enrichment map. A large number of closely related GO terms forms a cluster. Each cluster has been labelled with a general terms that captures all GO terms.
doi: 10.1371/journal.pone.0078480.g009
Supporting Information [fig_ref] Table 1: Top 10 up-regulated genes in GM and WM [/fig_ref]. Full list of significant DEGs between GM and WM.
(XLSX) There were two TF splice variants expressed across GM and WM. Both splice vairants contributed almost equal levels of expression to both GM and WM. TF-001 was upregulated 7x in WM when compared to GM. TF-202 was a novel splice varaints, it was expressed at higher levels in WM than in GM, however the changes in expression was not considered to be statistically significant. Error bars are ± standard error. (TIFF) [fig_ref] Figure 6: Heatmap of gene expression of oligodendrocyte cell markers [/fig_ref].
Expression levels of myelin associated glycoprotein (MAG) isoforms. There were four MAG splice variants expressed across GM and WM. MAG-001 was the dominant isoform and was up regulated 8x in WM when compared to GM. The three other splice variants (MAG-002, MAG-003, MAG-009) were expressed at low levels in both GM and WM. MAG-009 was a novel splice variant. Error bars are ± standard error. (TIFF) [fig_ref] Figure 7: Allen Human Brain Atlas in situ hybridisation for NEFH and MOG genes [/fig_ref].
Expression levels of MARCKS-like 1 (MARCKSL1) isoforms. There were two MARCKSL1 splice variants expressed across GM and WM. Both splice vairants contributed high levels of expression to both GM and WM. MARKSL1-001 was upregulated 7x in WM when compared to GM. MARCKSL1-002 was a novel splice varaints, it was also expressed at higher levels in WM than in GM, however the changes in expression was not considered to be statistically significant. Error bars are ± standard error. (TIFF)
[fig] Figure 2: Volcano plot of isoform expression in GM and WM. The fold-change of the isoforms was relative to their expression in WM. Those isoforms with a negative fold-change were up-regulated in GM (down-regulated in WM) and those isoforms with a positive fold-change were up-regulated in WM (down-regulated in GM). Isoforms that were statistically significant (q-value<0.05) are shown in red and were listed inTable S2. This figure demonstrates that a larger number of isoforms were significantly up-regulated in GM (681) than were up-regulated in WM (201). Overall there was a greater spread of data for the isoforms that were up-regulated in GM, both in terms of fold-change and p-values. doi: 10.1371/journal.pone.0078480.g002 [/fig]
[fig] Figure 3: Splice variants of GPR123. The intron/exon structure of the four GPR123 isoforms. GPR123-002 and GPR123-003 have been identified previously. GPR123-003 is considered the full-length isoform. GPR123-004 and GPR123-005 were previously unannotated. doi: 10.1371/journal.pone.0078480.g003 [/fig]
[fig] Figure 4: Expression levels of the GPR123 isoforms. GPR123-004 was not expressed at all in GM, but became the dominant isoform in WM. GPR123-002, GPR123-003 and GPR123-005 all had reduced expressed in WM. The TSS for GPR123-005 was also up-regulated in GM. The changes in expression of GPR123-003 and GPR123-003 were statisitically significant (q-value<0.05). Error bars are ± standard error. doi: 10.1371/journal.pone.0078480.g004 [/fig]
[fig] Figure 5: Heatmap of gene expression of neuronal cell markers. Shows the expression profile of neuronal cell markers in GM and WM. There was an expression bias towards GM with the up-regulation of NEFL, GABRA1, SYT1, SLC12A5, SV2B being statistically significant (q-value<0.05). doi: 10.1371/journal.pone.0078480.g005 [/fig]
[fig] Figure 6: Heatmap of gene expression of oligodendrocyte cell markers. Shows the expression profile of oligodendrocyte cell markers in GM and WM. There was an expression bias towards WM with the up-regulation of SOX10, GJC2, MOG, MAG, MAL, GAL3ST1, UGT8 being statistically significant (q-value<0.05). doi: 10.1371/journal.pone.0078480.g006 [/fig]
[fig] Figure 7: Allen Human Brain Atlas in situ hybridisation for NEFH and MOG genes. A. NEFH: Neurofilament, Heavy Polypeptide GM FPKM: 57.76 WM FPKM: 0.95. Slide from the dorsolateral cortex of a healthy 20-year-old male. The slide shows high levels of expression in GM B. MOG: Myelin Oligodendrocyte Glycoprotein GM FPKM: 8.33 WM FPKM: 49.08. Slide from the dorsolateral cortex of a healthy 20-yearold male. The slide shows high level of expression in WM. Source: Allen Human Brain Atlas (Hawrylycz et al. 2012 and ©2012 Allen Institute for Brain Science. Allen Human Brain Atlas [Internet]. Available from: http://human.brain-map.org/). doi: 10.1371/journal.pone.0078480.g007 [/fig]
[fig] Figure 8: Heatmap of high abundant genes specifically expressed in GM and WM. The following cut-offs were applied to the WM differentially expressed gene list; >4 fold up-regulation in WM, an FPKM>5 in WM and a FPKM <5 in GM. The following cut-offs were applied to the GM differentially expressed gene list; >20-fold up-regulated in GM, had an FPKM>5 in GM and an FPKM<1 in WM. This heatmap shows the top 40 genes from each list. The top half of the heatmap show genes deemed as WM specific, while the lower half shows genes expressed specifically in GM. [/fig]
[fig] Figure S1, Figure S2, Figure S3, Figure S4: Gene body coverage for GM and WM RNA samples used in this study. A -C. GM samples D -F. WM samples. Each figure shows the number of reads that map to particular portions of the gene body. The x-axis starts at the 5' end of the transcript and moves towards the 3' end (left to right). The y-axis represents the average wigsum. The wigusm is a normalised 'total read count' where a wigsum of 100,000,000 is equal to the coverage achieved by 1 million 100 base reads or 2 million 50 base reads. All figures are skewed to the 3' end of the transcripts, showing a 3' bias, caused by poly-A selection of the RNA fraction. (TIFF) Expression levels of visinin-like 1 (VSNL1) isoforms. There were two VSNL1 splice variants expressed across GM and WM. VSNL1-001 was up regulated 34x in GM when compared to WM. The second splice variant VSLN1-007 was novel and was expressed at low levels across both GM and WM. Error bars are ± standard error. (TIFF) Expression levels of syniclein, beta (SNCB) isoforms. There were two SNCB splice variants expressed across GM and WM. SNCB-002 was the dominant isoform and was up regulated 29x in GM when compared to WM. The second splice variant SNCB-001 was expressed at low levels across both GM and WM. Error bars are ± standard error. (TIFF) Expression levels of reticulon 1 (RTN1) isoforms. There were four RTN1 splice variants expressed across GM and WM. RTN1-202 was the dominant isoform and was up regulated 10x in GM when compared to WM. The splice variant RTN1-201 was expressed at approximately 20 FPKM in both conditions. RTN1-203 and RTN1-204 were expressed at low levels in both conditions. All four identified splice variants were novel. Error bars are ± standard error. (TIFF) Figure S5. Expression levels of transferrin (TF) isoforms. [/fig]
[fig] Figure S8: Allen Human Brain Atlas in situ hybridisation for the RGS4, CAMK2A, SLC17A7 and NEFM genes. A. RGS4: Regulator of G-protein signalling 4 GM FPKM: 100.75 WM FPKM: 1.52. Slide from the dorsolateral cortex of a healthy 20-year-old male. The slide shows high levels of expression in GM. B. CAMK2A: Calcium/calmodulin-dependent protein kinase II alpha GM FPKM: 233.16 WM FPKM: 5.27. Slide from the dorsolateral cortex of a healthy 20-year-old male. The slide shows high levels of expression in GM. C. SLC17A7: Solute carrier family 17 (sodium-dependent inorganic phosphate cotransporter), member 7 GM FPKM: 270.79 WM FPKM: 7.39. Slide from the dorsolateral cortex of a healthy 20-year-old male. The slide shows high levels of expression in GM. D. NEFM: Neurofilament, medium polypeptide GM FPKM: 224.01 WM FPKM: 4.09. Slide from the dorsolateral cortex of a healthy 20-year-old male. The slide shows high levels of expression in GM. Source: Allen Human Brain Atlas (Hawrylycz et al. 2012 and ©2012 Allen Institute for Brain Science. Allen Human Brain Atlas [Internet]. Available from: http://human.brain-map.org/). (TIFF) [/fig]
[table] Table 1: Top 10 up-regulated genes in GM and WM. [/table]
[table] Table 2: Top 10 up-regulated isoforms in GM and WM. [/table]
[table] Table 3: Cell type markers. [/table]
[table] Table 4: Top 10 GO terms in GM and WM. doi: 10.1371/journal.pone.0078480.t004 [/table]
[table] Table S2: Full list of significant DEIs between GM and WM. (XLSX)Table S3. Novel gene markers for GM and WM. (XLSX) Table S4. Gene Ontology terms enrichment analysis for DEGs between GM and WM. (XLSX) [/table]
|
Proteomic Approaches to Defining Remission and the Risk of Relapse in Rheumatoid Arthritis
Objectives: Patients with Rheumatoid Arthritis (RA) are increasingly achieving stable disease remission, yet the mechanisms that govern ongoing clinical disease and subsequent risk of future flare are not well understood. We sought to identify serum proteomic alterations that dictate clinically important features of stable RA, and couple broad-based proteomics with machine learning to predict future flare.Methods:We studied baseline serum samples from a cohort of stable RA patients (RETRO, n = 130) in clinical remission (DAS28<2.6) and quantified 1307 serum proteins using the SOMAscan platform. Unsupervised hierarchical clustering and supervised classification were applied to identify proteomic-driven clusters and model biomarkers that were associated with future disease flare after 12 months of follow-up and RA medication withdrawal. Network analysis was used to define pathways that were enriched in proteomic datasets.Results:We defined 4 proteomic clusters, with one cluster (Cluster 4) displaying a lower mean DAS28 score (p = 0.03), with DAS28 associating with humoral immune responses and complement activation. Clustering did not clearly predict future risk of flare, however an XGboost machine learning algorithm classified patients who relapsed with an AUC (area under the receiver operating characteristic curve) of 0.80 using only baseline serum proteomics.Conclusions:The serum proteome provides a rich dataset to understand stable RA and its clinical heterogeneity. Combining proteomics and machine learning may enable prediction of future RA disease flare in patients with RA who aim to withdrawal therapy.
# Introduction
Rheumatoid Arthritis (RA) is a systemic autoimmune disease that is characterized by inflammation of synovial joints. Modern RA therapy is initiated early and escalated aggressively using a treat-to-target approach to try an obtain disease remission. The development of both targeted treatments and combination regimens continues to improve expected outcomes for patients. Encouragingly, clinical remission, defined by multiple measures of disease activity (3), has become a realistic expectation for most patients with RA. Recent registry data of RA cohorts consistently show that DAS28 (Disease Activity Score) remission is achieved in about 50% of patients (4), a number that may be increasing over time.
Patients with RA who are able to achieve disease remission using standard therapy are not well studied, given their lack of disease activity and need for treatment changes. The main issue facing these patients is whether or not to remain on their treatment, or risk withdrawal and the potential for disease flare.
There are many prospective studies that have demonstrated successful Disease Modifying Anti-Rheumatic Drugs (DMARD) withdrawal in patients in clinical remissionbut the determinants of maintaining remission status after medication withdrawal are poorly defined. Unfortunately, given the limited understanding of the pathological mechanisms that drive subclinical disease, clinicians are left to guess which of their patients might sustain remission using less aggressive therapy.
Technological advances in high-throughput proteomics have allowed for an improved understanding of disease processes and biomarker discovery. Although mass spectrometry tends to dominate this evolving field, broad-based targeted proteomics has its own advantages, including simplified sample preparation and user-friendly output data. Our group has previously defined protein sets that are associated with future disease flare from pre-clinical RA by coupling machine learning with proteomic approaches. Indeed, leveraging omics approaches to resolve heterogeneity in common diseases remains a distinct challenge in clinical medicine, though this has not been systematically undertaken in a stable RA cohort.
The RETRO (15) (Reducing therapy in rheumatoid arthritis patients in ongoing remission) study is a prospective randomized trial which enrolled patients who had achieved disease remission with conventional RA therapy. One of the aims of this study is to define disease recurrence in patients with RA when either continuing or reducing their medications. It was previously shown in this trial that positive anti-citrullinated antibody (ACPA), and other biomarkersare associated with an increased likelihood of disease relapse. In spite of these studies, there is little understanding of the underlying biological mechanisms that are active in stable RA. If differences within RA patients in remission can be more clearly defined, there may be an enhanced understanding of the spectrum of RA pathogenesis, along with improved personalized clinical approaches surrounding the withdrawal of therapy. We hypothesized that high-throughput proteomics (18) would help identify underlying biological heterogeneity that might provide insights into mechanisms underpinning future disease flare. Our aim was to explore how the serum proteome shapes the underlying clinical experiences of stable RA patients.
# Methods
## Patients and inclusion criteria
RETRO is a multicentre, randomized, open, prospective, controlled parallel-group study. Details of the study are described in the original publication. The objective of the study is to evaluate tapering or discontinuation of DMARDs in patients with RA. All enrolled patients fulfilled the 2010 American College of Rheumatology (ACR) criteria for RA. Patients had to have sustained clinical remission defined by the Disease Activity score (DAS28 < 2.6) criteria for at least 6 months. Ethics committee of the Friedrich-Alexander-University of Erlangen-Nuremberg approval was granted.
## Treatment and follow-up
Patients were randomized to one of three arms: Arm 1 continued with existing DMARD regimen at full dose for 12 months, arm 2 reduced the dose of all DMARDs by 50%, while arm 3 reduced the dose of all DMARDs by 50% in the first 6 months, then discontinued all medications. Relapse of disease was defined as a DAS28-ESR score greater than 2.6. Participants were assessed for clinical disease activity every 3 months until month 12.
## Assessment of demographic and disease-specific parameters
Age and sex were recorded in all patients. Disease duration, tender joint count (68), swollen joint count (66), patient visual analogue scale (VAS) for pain and patient global were assessed and recorded. C-reactive protein (CRP), ESR, Rheumatoid Factor (RF), ACPA, DAS28-ESR and Health Assessment Questionnaire (HAQ-DI) were recorded.
## Somascan
SOMAscan is a proteomics assay that measures 1307 proteins using an aptamer library. This high-throughput proteomics assay has been used in recent publications to study the aging proteomealong with other human diseases. 130 baseline serum samples were available from the RETRO study. Briefly, a library of aptamers were incubated with serum, and those that bind are isolated and hybridized to DNA microarray for detection. The identity and relative concentration of the detected proteins are revealed by localization and fluorescence intensity. Protein quantification is reported as relative fluorescence units (RFU), an arbitrary value. In general, agreement between aptamer and antibody-based assays is high. Further details regarding the SOMAscan assay are available.
# Statistical analysis
Descriptive resultsare stated in means and standard deviation. SOMAmer protein expression RFU values for the study patients were transformed into a log2 scale for differential analysis (Supplemental File 2). Batch effect was removed in our SOMAmer data using internal controls within each plate to adjust proteomic intensity as per standard SOMAscan protocols. Batch effect was assessed between plates and determined to require no further correction. Data was loaded and analyzed in the R (v3.5.3) environment unless otherwise stated. Missing clinical data was imputed using multiple imputation by chained equations (MICE). Differential analysis between groups was undertaken using linear modeling with the package LIMMA. GO pathways analysis was performed using clusterprofiler. Graphs were generated using the ggplot2 package. Correlation analyses were performed for select proteins using Pearson correlation. Multidimensional scaling (MDS) was used for dimension reduction on all SOMAscan proteins.
The 200 most variable proteins measured by coefficient of variation were used to determine optimal number of clusters ranging from k = 2 to 10 and identify sample clusters using the R package Consensusclusterplus. We used 80% protein resampling and 80% patient resampling and selected Pearson as our distance function. Multinomial logistic regression implemented in R package glmnet was used to identify clinical variables that are independently associated with cluster assignment. Sliding window analysis of DAS28 scores and protein expression was performed using a previously published algorithm, DE-SWAN. Briefly, this algorithm analyzed serum protein expression across quintiles of DAS28 scores using linear modeling, while adjusting for baseline demographic factors, in this case age and sex. A protein expression score was developed on proteins that correlated with DAS28 which were identified by DE-SWAN. We filtered the proteomic data on the 34 score members, scaled the data by the mean and standard deviation, and multiplied by 1 (positively associated with DAS28) or -1 (negatively associated with DAS28) for each protein. The final score was the mean expression of all 34 proteins for each patient. We randomly generated 5000 data sets with 34 randomly selected proteins in each set to evaluate the significance of the association score.
## Machine learning classification algorithm
We applied two supervised machine learning (ML) techniques to develop algorithms to classify flare or remission based on serum proteomics. The first approach we used is XGBoost (Extreme Gradient Boosting), which employs a regularization term to overcome the overfitting. The second approach is the LASSO model, which was used as a baseline to compare its performance with that of XGBoost. Data was loaded into Python, and samples were randomly split into a training (n = 104, 80% of the samples) and test (n = 26) set. The training set was used to train and tune the parameters in the two models and the test set was used evaluate the models' performance, which were measured by the area under of the curve (AUC) of receiver operating characteristic (ROC), accuracy, sensitivity and specificity. To increase the interpretability of the XGBoost model to predict the flare status of a given sample, we used SHAP values (Shapley Additive Explanation). A higher SHAP value of a given feature in the model represents its strong influence on the model output. The final model parameters we used in the XGBoost are as follows: learning_rate = 0.01, max_depth = 3, subsample = 0.6, colsample_bytree = 0.7, n_estimators = 100, gamma = 0.0, reg_alpha = 0.5, the parameters used in the LASSO model is as follow: cost=1.17 and max_iterations = 5000. We used 5-fold cross-validation to get the optimal hyperparameters.
# Results
## Hierarchical clustering on serum proteins identifies heterogeneity amongst stable ra patients
Given the paucity of data aimed at understanding subclinical disease activity in RA patients who achieve remission, we sought to explore underlying heterogeneity using serum proteomics in this established cohort. We quantified over 1300 serum proteins from 130 RETRO patients at their baseline visit, all of whom were in stable clinical remission (DAS28 < 2.6). We hypothesized that despite the clinical similarities amongst individuals within this cohort, proteomic differences may provide important insights by identifying sub-clusters of patients. We applied consensus clustering to assign individuals to one of the 4 clustersand clustered scaled protein expression by hierarchical clustering, which can be seen in. MDS analysis revealed separation of the hierarchical clusters. Baseline characteristics split by cluster are listed in. We found no differences in sex, age, biologic use, or serological status across our 4 proteomic clusters. Cluster 4 had significantly lower DAS28 scores compared to the remaining clusters. With respects to future flare, Cluster 2 trended toward lower rates relative to the remaining clusters, however this did not reach statistical significance (16.7% vs 39.8%, p = 0.21). To assess this association by multinomial regression, we assigned Cluster 2 as the reference cluster and found that that Cluster 3 had higher odds of flare (OR 5.6, 0.97 to 33.06, p = 0.05), relative to Cluster 2 with similar trends observed for Cluster 1 and Cluster 4. Indeed, no clear distinction between individuals who developed future flare was observed in the MDS plot. Overall, these results suggest that global proteomic clusters within a clinically homogenous cohort can be identified but are associated with current clinical status rather than future outcomes.
## Machine learning classifies future flare using baseline serum proteomics
We next aimed to use the serum proteome to identify biomarkers associated with future disease flare in stable RA, given that clustering did not clearly associate with risk of flare. We identified DEP's between these groups and observed. However, after adjustment for multiple comparisons, none of the differentially expressed proteins reached statistical significance. This suggests that although subtle differences exist in the serum proteome between individuals who experience future flare, it's unlikely that singular biomarkers accurately predict this outcome in this population.
To test this hypothesis, we explored the use of two machine learning algorithms, LASSO and XGBoost, to build predictive models that classify future flare using baseline serum proteomics. We generated two models, both of which were validated on a test . XGboost delivered a model with higher specificity (0.78) than sensitivity (0.63) and an overall accuracy of 73.1% with an AUC 0.80. Therefore, we found that the XGboost model outperformed the LASSO model by the metric area under the curve, AUC, 0.80 vs 0.58), accuracy (73.1% vs 69.2%) and sensitivity (0.63 vs. 0.5).
To interpret this XGBoost model, we explored the impact of essential features in terms of SHAP values on the classifier's output for a single prediction which are shown in. We identified Interleukin 17F (IL17F) and Myc proto-oncogene protein (MYC) expression as indicators of future flare, while Roundabout homolog 3 (ROBO3), Synuclein alpha (SNCA), complement factor B (CFB) and vascular endothelial growth factor A (VEGF-A) expression were indicators of sustained remissionand. Given the small number of proteins that derived our boosted model, we next explored whether there were any functional links between these proteins. We developed gene concept plots to identify potential protein interactions and found that SNCA, MYC, VEGFA and PLAUR were connected by a single pathway, endopeptidase mediated apoptosis. We then analyzed IL17F restricted networks, given its conflicting role in RA, and that its expression was associated with future flare in our model. We found that IL17F interacted with VEGFA though growth factor function, and with SNCA through common effects on the inflammatory response. IL17F independently regulated GM-CSF production, a key driver of RA disease activity through the recruitment of neutrophils. Overall, this network analysis suggests that cellular apoptosis and GM-CSF production may be associated with future disease flares in RA patients who are otherwise stable.
## Disease activity in stable ra is reflected in the serum proteome
In our hierarchical clustering, we observed a lower mean DAS28 score in Cluster 4 compared to the remaining 3 clusters. RA patients who achieve DAS28 defined remission often have residual disease activity, however, since this population is not typically the focus of translational studies, little is known regarding biomarkers that are reflective of ongoing disease activity. Indeed, we found that several protein members correlated with DAS28 score, including Integrin alpha 2B (ITGA2B), Bactericidal permeabilityincreasing protein (BPI) and chemokine ligand 2 (CXCL2, Pearson R, all p value < 0.01). Further, complement proteins (C3, C4A, C1S) were all negatively associated with DAS28 score, suggesting activation and consumption of complement proteinswere indicators of disease. There was no indication that these parameters varied based on ACPA status .
To further explore the relationship between disease activity and serum biomarkers, we developed a sliding window model (SWAN) which examined protein variability across DAS28 quintiles, after controlling for Age and Sex. Across DAS28 a total of 34 proteins varied significantly with disease activity. We used these 34 protein members to annotate a meta-protein expression score (the mean expression profiles of the 34 proteins), which correlated with DAS28 (R = 0.45, p < 0.001, . To test the robustness of this finding, we sampled 5000 random sets with 34 proteins in each set and correlated their mean expression with DAS28 scores. We found a range of -0.37 -0.27, associated with a low probability (0) that the correlation of 0.45 would occur by chance . This protein disease activity score was significantly lower in Cluster 4 compared to the remaining 3 clusters, concordant with their lower DAS28 scores. Gene concept plots revealed that these 34 proteins interacted through nodes that included humoral immunity, apoptosis, and complement activation. Overall, these data suggest that stable RA disease activity is marked by complement consumption and humoral immune responses, which is reflected in a serum protein signature that is detectable in patients with stable RA.
# Discussion
With the advent of multiple targeted therapies, alone and in combination, most RA patients should reasonably expect to achieve low disease activity or clinical remission status. To date, few studies have sought to understand the heterogeneity of the biological pathways underpinning clinical remission in this rapidly expanding population of RA patients. When, and in whom to attempt withdrawal of therapy has become a compelling clinical question. On the one hand, there is justifiable concern regarding reactivation of systemic and articular inflammation. On the other hand, ongoing use of DMARDs and/or biologics is associated with increased risk of infectious complications, malignancyand cost. Strategies for successful taper however remain ill-defined and, importantly, lack precision. The RETRO clinical trial has previously generated predictive models that were based on clinical parameters and serum studies. ACPA seropositivity appears to be an important indicator for increased risk of future relapse, while clinical parameters have modest predictive value even when combined with advanced machine learning techniques. It remains unclear if these indicators are clinically applicable and generalizable to a wide range of RA patient populations. The focus of this study was to use a broad-based serum proteomic approach to better understand the underlying heterogeneity amongst RA patients who are in sustained clinical remission, prior to their participation in a clinical trial of therapy withdrawal. Our results identify proteomic signatures reflecting biological mechanisms that are associated with ongoing disease stability off therapy, or alternatively, the risk of future disease relapse. Our XGboost model suggested that individual circulating serum biomarkers are unlikely, on their own, to be predictive of future stability or relapse after therapy withdrawal. In spite of this, combinations of proteomic biomarkers identified by the machine learning achieved relatively high AUC and accuracy in predicting outcomes. Indeed, this is a testament to the power of rapidly evolving machine learning algorithms that are being developed for many clinical problems. Network analysis of proteins derived from machine learning suggested that inflammatory forms of cellular death was an indicator for risk future disease flare. Indeed, apoptosis is escaped by pathogenic fibroblast-like synoviocytes and likely contributes to their aggressive and hyperplastic phenotype in RAand this may point to systemic FLS as a potential source of these proteins. Hierarchical clustering identified proteomic clusters which were defined, in part, by clinical characteristics. Cluster 4 represented a patient group with lower DAS28 scores amongst the remaining patient cohort. Our disease activity signature, found to be lower in Cluster 4, suggested that elements of humoral immunity and complement activation might facilitate disease activity in otherwise stable RA patients. This suggests that activating pathways differentially regulate disease activity in this subset of RA patients, as many of the well-known disease activity markers in RA suggest that innate immune responses associate with DAS28 scores. The analyses we undertook utilized the SOMAscan aptamerbased technology to interrogate 1307 distinct serum proteins. Although this provided us with a robust array of biomarkers, it is well recognized that these represent only a fraction of the human proteome, and that larger arrays that span a larger proportion of the circulating proteome may help generate even more accurate predictive algorithms. Moreover, there remains an incomplete understanding of how aptamer-based detection of each individual analyte correlates with other detection methodologies such as those that are antibody based. Due to our modest sample size, we observed strongly statistically significant association although R values related to DAS28 scores were all below 0.3. However, we expect the R values may increase using a larger sample size while the significant association will be still held. Finally, SOMA proteins may bias over-representation analysis based on the selected proteins which are included in the set. Notably, network analysis was used in this study to connect proteins of interest through biological nodes. These results would not be impacted by inherent bias in the SOMA protein set.
In conclusion, we applied an unsupervised, high-throughput proteomics assay to delineate biomarkers and pathways that reflect the biological heterogeneity present in RA patients who are collectively deemed to be in stable clinical remission. Based on this, we used supervised machine learning to develop robust models that predicted ongoing disease stability after therapy withdrawal as opposed to future disease flare. Although it is premature to try and define the potential clinical utility of these models, they do provide an important impetus for further studies that aim to further define a biological definition of remission in RA patients that can ultimately guide clinical decision making.
# Data availability statement
The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding author.
# Ethics statement
The studies involving human participants were reviewed and approved by Friedrich-Alexander-University of Erlangen-Nuremberg. The patients/participants provided their written informed consent to participate in this study.
# Author contributions
LO'N: data analysis, manuscript construction. PH: data analysis. QL: data analysis. MI: data analysis. VS: data analysis. JR: data collection and clinical trial. AH: data collection and clinical trial. VA: manuscript, data analysis. IS: manuscript, data analysis. HE-G: manuscript, data analysis. GS: data collection and clinical trial. JW: oversight, manuscript, data analysis. All authors contributed to the article and approved the submitted version. |
Protocol for the economic evaluation of the InTENSE program for rehabilitation of chronic upper limb spasticity
Background: Assessment of the costs of care associated with chronic upper-limb spasticity following stroke in Australia and the potential benefits of adding intensive upper limb rehabilitation to botulinum toxin-A are key objectives of the InTENSE randomised controlled trial. Methods: Recruitment for the trial has been completed. A total of 139 participants from 6 stroke units across 3 Australian states are participating in the trial. A cost utility analysis will be undertaken to compare resource use and costs over 12 months with health-related quality of life outcomes associated with the intervention relative to a usual care comparator. A cost effectiveness analysis with the main clinical measure of outcome, Goal Attainment Scaling, will also be undertaken. The primary outcome measure for the cost utility analysis will be the incremental cost effectiveness ratio (ICER) generated from the incremental cost of the intervention as compared to the incremental benefit, as measured in quality adjusted life years (QALYs) gained. The utility scores generated from the EQ-5D three level instrument (EQ-5D-3 L) measured at baseline, 3 months and 12 months will be utilised to calculate the incremental Quality Adjusted Life Year (QALY) gains for the intervention relative to usual care using area-under the curve methods. Discussion: The results of the economic evaluation will provide evidence of the total costs of care for patients with chronic upper limb spasticity following stroke. It will also provide evidence for the cost-effectiveness of adding evidence-based movement therapy to botulinum toxin-A as a treatment, providing important information for health system decision makers tasked with the planning and provision of services.
# Background
People with spasticity following stroke have significantly higher care costs (particularly direct healthcare costs, and aged care costs) and lower quality of life than those survivors without spasticity. Therefore, identifying effective therapies to reduce upper-limb spasticity and improve function are an important target for research.
International clinical guidelines support the use of botulinum toxin-A in conjunction with active rehabilitation as the preferred treatment. However, the optimum rehabilitation strategy remains undetermined. There are a lack of adequately powered randomised controlled trials evaluating the effect of botulinum toxin-A injections alone, compared to the injection plus active rehabilitation. However, consideration of the costs of providing care for these patients and ultimately consideration of the cost effectiveness of new therapies (namely, whether they are a worthwhile spend of the constrained resources of the healthcare budget as compared to other potential therapies) is another important factor.
There have been few studies of the economic impact of upper-limb spasticity following stroke. Lundström et al.evaluated the healthcare costs for the year following stroke in those with and without spasticity in Sweden, and identified that direct health care costs were four times higher in those with spasticity compared to those without, predominantly due to increased costs of hospital care and post hospital community care (i.e. home help services, residential care etc). However, this study only included hospitalised patients and was based on only 25 participants with spasticity. More recently in the UK, Raluy-Calladoevaluated costs of care in over 2900 post-stroke spasticity patients and found that those with spasticity following stroke had double the healthcare costs of those without spasticity with increased hospital care contributing to increased costs in this group, but were not able to include information on home and community care in their estimate. In addition, the potential economic impacts of spasticity following stroke are broad ranging, with loss of workforce productivity among patients and their caregivers which persisit after the event. However, the potential cost-effectiveness of therapies is under-researched, with no economic evaluations to date evaluating the impact of evidence-based movement training combined with botulinum toxin-A injections.evaluated the impact for the health care costs and quality of life of botulinum toxin-A treatment vs usual care without botulinum toxin-A. The study showed a significant improvement in the physical and mental health status of participants over the follow up period. Increased healthcare costs were evident for the participants who received the treatment, but despite higher incremental costs (driven by higher pharmaceutical and nursing home care costs) the study authors concluded the intervention was very likely to be considered cost effective due to the large gains in quality of life attributed to the intervention group compared to usual care. However a key limitation of this study was that it was not randomised and the results may have been influenced by confounding factors in the treatment and usual care groups. Conversely, the BoTULS trial evaluated the clinical and cost effectiveness of treating upper-limb spasticity with botulinum toxin-A plus physical therapy vs physical therapy alone over a 4 week intervention period. The study authors concluded that the intervention had a low probability of cost-effectiveness compared to usual care using the UK reference care willingness to pay threshold of £20,000 for an additional QALY gained.
In addition, there is an absence of studies from an Australian perspective. Makino et al. 2018have published the only Australian based study which evaluated the cost-effectiveness of extending botulinum toxin-A therapy beyond the four treatments currently supported by the Pharmaceutical Benefits Scheme. This study was undertaken from the health-care payer perspective, and therefore included direct healthcare costs in the Markov-state transition model that was developed. It was found that extending the number of treatments beyond four was likely to be considered cost effective. However, the study authors didn't include costs or benefits from rehabilitation or physical therapy in addition to the botulinum toxin-A in their analysis.
The cost of botulinum toxin-A injections is significant, calculated as $1673 Australian Dollars per treatment cycle and patients may receive multiple cycles of treatment. The InTENSE trialaims to determine the clinical and cost effectiveness of including evidencebased movement training with botulinum toxin-A injections. Therefore, interventions to improve the long-term effect of botulinum toxin-A injections in this group could assist in improving quality of life of patients and reducing their healthcare and broader community care costs. Here we describe in detail the protocol for the economic evaluation to occur alongside the evaluation of clinical effect for the InTENSE trial.
# Methods
The clinical protocol for the trial has been described in detail previously. In summary, this Phase III Clinical Trial aims to determine clinical effectiveness of undertaking evidence-based movement training following a botulinum toxin-A injection in adults with neurological spasticity. The study is a national, multicentre randomized clinical trial with concealed allocation, blinded assessment and intention-to-treat analysis. The sites for the study can be found listed on the entry on the Australian and New Zealand Clinical Trial Registry for the Phase III Clinical Trial (ANZCTR12615000616572, date registered 12/06/2015). Recruitment for the study has been completed, with a total of 139 participants from six stroke units across three Australian states participating in the trial. The rational for the sample size for the study has been described in the main clinical protocol paperand was calculated to detect a difference of seven points in the main clinical outcome of the Goal Attainment Scale T-score between the groups with 80% power with a two-tailed significance level of 0.05.
The main objective of the economic evaluation is to undertake a cost-utility analysis of the InTENSE evidence-based movement training with botulinum toxin-A injections as compared to botulinum toxin injections alone. A secondary aim is to undertake a costeffectiveness analysis of the same using Goal Attainment Scaling (GAS) as the measure of benefit. A third aim, is to determine the costs associated with care for people with spasticity following stroke in a 12 month period in an Australian context.
## Measurement and valuation of costs
A summary of the sources of cost data and valuations are shown in. The total costs associated with delivering the intervention, including therapist intervention and travel time, consumables, and overheads are collected using a specially designed intervention tracking sheet. Costs of care collected include medical care costs (for example hospitalisation, doctors visits, allied health professional attendances) and non-medical care costs (for example home assistance). Resource use will be collected via a number of methods, using routine administrative datasets where possible. Data on tertiary health service utilisation, including in-patient hospital stays, emergency department presentations, and outpatient presentations are collected from participants via monthly diary over the 12 month follow up period. Use of out-ofpocket or privately funded health and community care services are also collected via this method. The majority of primary health care (such as visits to General Practitioners, and access to prescription pharmaceuticals) in Australia are delivered by the federal government through the Department of Human Services. Data on the use of the Medicare Benefits Schedule (MBS) and the Pharmaceutical Benefits Scheme (PBS)will be collected for consenting participants. Costs for relevant healthcare resources will be sourced from publicly available published data where available and as recommended by guidelines for conducting economic evaluations, including the National Hospital Cost Data Collection, MBS and PBS. All costs will be updated to a standard reference year for analysis. Discounting will not be necessary as the follow up period for the study will be 1 year.
# Cost-analysis
The total costs of care for participants in the intervention and control groups over the 12 month follow up period will be calculated. This will be aggregated into the health care costs (including costs of hospitalisation, emergency department presentations, doctors attendances, community nursing care attendances, and visits to allied health professionals). In addition, non-medical care costs will be collected and documented, including domestic assistance at home. Costs associated with providing the intervention will be recorded and reported separately. The proportion of participants who utilize the various health services and non-medical care services will be presented, along with the mean and median number of utilizations over the 12 month period. Both the mean and median total costs over the 12 month follow up period for health services and non-medical care will be presented, given cost data is generally of a skewed distribution. The costs of health services and non-medical care will be combined to provide a total cost of care for people with upper-limb spasticity following stroke.
## Measurement and valuation of benefit
The benefit of the intervention will primarily be measured by comparing the quality adjusted life years (QALY) gained in the intervention group as compared to the control group over the 12 month follow up period. Change over time in Health-related quality of life (HrQOL) will be measured using the EQ-5D-3 L. This five-item instrument measures HrQOL across five dimensions (mobility, self-care, usual activities, pain/discomfort, and anxiety/depression) and includes a global health rating using the visual analogue scale (EQ-VAS). It has been widely shown to have excellent validity, including in people with stroke. Participants will complete the EQ-5D-3 L to describe their current HrQOL at baseline (prior to randomisation), at 3 months and again at 12 months. EQ-5D-3 L responses will be converted into a utility score anchored on a scale between 0 (indicating a health state equivalent to death) and 1 (indicating near-perfect health) using the available preference-weights generated from an Australia general population sample. These utility scores will be converted to quality-adjusted life year (QALY) gains for each individual participant over the 12 month follow-up period in by combining data regarding the EQ-5D-3 L health-states of participants with information about the time spent in those health states, using area under the curve methods. In addition, as a secondary analysis, we will use the primary clinical outcome for the trial, the Goal Attainment Scale (GAS). GAS is a validated measure of achievement of rehabilitation specific goals and measured at the three and 12 month trial timepoints. Goals for their rehabilitation are identified by the participants themselves, as they relate to activity and participation in meaningful tasks. Participants score themselves for current and expected levels of performance (ranging from − 2 to + 2) and t-scores calculated as described by Kiresuk et al..
# Statistical analysis
Descriptive statistics will be presented along with simple statistical tests of difference. Given cost data is generally of a skewed distribution, statistics for normally and nonnormally distributed data will be used as appropriate. To adjust for any differences in baseline covariates in the sample (for example in symptom severity), regression analysis of the costs and QALY will be applied. Standard linear regression approaches will be explored, as well as generalized linear regression models which can account for skewed distribution and heteroscedasticity in the data while maintaining the original scale of the data. Where significant levels of missing data occur (5% or greater of the observations), approaches to account for missingness will be undertaken in the analysis. Multiple imputation will be undertaken to account for data missing at random or missing completely at random.
# Cost-effectiveness analysis
The primary economic analysis will be carried out on an intention-to-treat basis within a Cost Utility Analysis framework. Additional analyses using a clinical subgroup (those who had mobility through the arm at the beginning of the study assessed as those who could move one or more blocks on the Box and Block Testwill also be investigated. The mean differences in costs and outcomes for the two groups will be presented. The analysis will present the additional resources used (i.e. costs) for an improvement in the outcomes (i.e. QALYs) associated with a new health intervention are compared to usual care (i.e the control group). The result will then be presented as an incremental cost effectiveness ratio (ICER) which is a measure of the additional cost for each unit of improvement in the outcome, and the main outcome will be the incremental costs divided by the incremental QALY gain. The fundamental calculation for the ICER comparing the intervention and control groups will be ICER = (Ca − Cb) ÷ (Ea − Eb), where Ca is the cost of the intervention, Cb is the cost of the control, Ea is the effectiveness of the intervention and Eb is the effectiveness of the control measured in QALYs).
There are multiple potential scenarios that may occur when the costs and benefits of the intervention are combined into an ICER which can be represented by the four quadrants on a cost-effectiveness plane. The intervention may be more effective than the control and be less costly (scenario A), the intervention may be more effective than the control and be more costly (scenario B), the intervention may be less effective than the control (or no different in effect) and be more costly (scenario C), or the intervention may be less effective than the control and more costly (scenario D). Therefore, in the case that the intervention is found to be more effective than the control, the analysis of the costs associated and calculation of the ICER will occur to determine whether the intervention falls into scenario A or B. By comparison, if the intervention is found to be less effective (or no different in effect) to the control, the analysis of costs associated and calculation of the ICER will occur to determine whether the intervention falls into scenario C or D.
For economic evaluations alongside clinical trials, it is recommended to undertake sensitivity analyses to determine the reliability of the results from the analysis, to give an estimate of the level of uncertainty in the findings to take into account in policy decision making. Sensitivity analysis will be undertaken using non-parametric bootstrapping to provide the confidence ellipse, which reflects the uncertainty in the estimate of the ICER. The ellipse provides a region on the cost-effectiveness plane that should contain x% (e.g. 95%) of the uncertainty.
Uncertainty regarding the cost-effectiveness of the intervention will be summarized using a cost-effectiveness acceptability curve (CEAC). This curve provides a graphical presentation of the probability that the intervention is cost-effective (has an ICER below the cost-effectiveness threshold) compared with the alternative intervention, given the data, for a range of values for the costeffectiveness threshold. The CEAC therefore will represent the likelihood of the intervention being cost-effective at a range of ceiling willingness-to-pay thresholds for an additional QALY.
# Discussion
This study represents one of the few analyses conducted internationally of the costs of care associated with chronic upper-limb spasticity post stroke, and the first cost-utility study to assess the economic benefit of evidence-based movement training in conjunction with botulinum toxin-A for treatment of upper-limb spasticity post stroke. There is increasing demand for information on not only the clinical effectiveness of healthcare interventions but also their cost-effectiveness, to provide policy makers with critical information regarding the best value for money spend of finite budgets. As such, randomised controlled trials (RCT) of the effectiveness of interventions can provide good opportunities to conduct an economic evaluation alongside the trial, provided the appropriate steps are taken from the outset to ensure that the design of the RCT is fit for this purpose. The economic evaluation for the InTENSE has been considered from the inception of the study, thus allowing an appropriate design and measurement and valuation of costs and benefits to be undertaken within the clinical trial. Our protocol has been planned using available national and international guidelines for conducting economic evaluations, promoting greater transparency in the methods undertaken and increasing the rigor and validity of the findings.
This study also represents a unique opportunity to evaluate the economic impact of upper-limb spasticity in a sample of participants post-stroke. There have been few studies of the costs of care for those with upperlimb spasticity following stroke, but those that have been undertaken have identified two to four fold increases in health care expenditures relative to those without spasticity, largely driven by increased hospital and community-based care costs. Accurate information on the costs of care for those with spasticty following stroke is essential for decision-makers in planning for continuing care services available and to understand the potential value-for-money of strategies to support these individuals. In conclusion, this study will provide essential policy-relevant information for decision makers regarding the value of evidence-based movement training for those undergoing an expensive treatment for this life-changing condition. |
A qualitative examination of the usability of a digital cognitive behavioral therapy for insomnia program after stroke
Objective: Sleep is commonly impaired after stroke. Cognitive Behavioral Therapy for Insomnia (CBT-I) is the first-line recommended treatment for sleep difficulty. "Sleepio" is a digital CBT-I program, allowing delivery of this treatment at scale. However, Sleepio has not yet been tested specifically in people with stroke. Before doing so, we wanted to explore the experience of people with stroke using the program, and potential barriers to completion. Method: Community dwelling survivors of stroke (n = 11, 41-78 years of age, 6 male) were given access to Sleepio. Participants discussed their experiences with the program during a semi-structured interview, which was analyzed using thematic analysis. Results: We found four common themes: (1) positive and negative experiences impacted engagement with the program, (2) motivation to follow the program was proportional to perceived severity of sleep problem, (3) impractical advice for people with stroke, (4) difficulty operating the program. Conclusion: Sleepio can be used by some people at the chronic stage of stroke. However, some barriers to completion were highlighted, and not all suggestions were deemed practical for everyone. We therefore suggest possible adaptations which may make the program more easily usable and engaging for survivors of stroke with varying impairments.ARTICLE HISTORY
# Introduction
Stroke, a type of acquired brain injury that occurs when blood supply to the brain is interrupted, is a leading cause of disability worldwide [bib_ref] Global, regional, and national burden of stroke, 1990-2016: a systematic analysis for..., Co [/bib_ref]. Survivors of stroke commonly experience disrupted sleep [bib_ref] Circadian and homeostatic changes of sleep-wake and quality of life in stroke:..., Cavalcanti [/bib_ref] , with both self-reported and actigraphy measures showing poorer sleep for people with stroke than age-matched controls [bib_ref] Self-reported and objective sleep measures in stroke survivors with incomplete motor recovery..., Fleming [/bib_ref]. A recent systematic review estimates the prevalence of insomnia at 15-60% [bib_ref] Incidence and prevalence of post-stroke insomnia: a systematic review and meta-analysis, Baylan [/bib_ref]. Poor sleep is frequently accompanied by depression and anxiety [bib_ref] Incidence and prevalence of post-stroke insomnia: a systematic review and meta-analysis, Baylan [/bib_ref] , and may impact on engagement in activities of daily living [bib_ref] Evolution of sleep and sleep EEG after hemispheric stroke, Vock [/bib_ref] and quality of life long-term after stroke [bib_ref] Insomnia and health-related quality of life in stroke, Tang [/bib_ref]. It is therefore important to investigate treatments to improve sleep in this population.
Cognitive behavioral therapy for insomnia (CBT-I) is the first-line recommended treatment for insomnia [bib_ref] British association for psychopharmacology consensus statement on evidence-based treatment of insomnia, parasomnias..., Wilson [/bib_ref]. There are, however, limited studies investigating CBT-I (or adapted CBT-I) after acquired brain injury. Nevertheless, recent research suggests preliminary efficacy at improving sleep quality and insomnia severity in this population (10,11). However, adaptations to standard CBT-I techniques may be required for some due to complications arising from the lesion, such as reduced mobility [bib_ref] Development and evaluation of a cognitive behavioural intervention for chronic post-stroke insomnia, Herron [/bib_ref] , as well as typically older age which may impact on ability to travel to appointments and comprehend the supplied information [bib_ref] Cognitive behaviour therapy for insomnia and depression: qualitative reflections from older adults..., Sadler [/bib_ref]. Widespread access to inperson CBT-I is typically limited due to a lack of trained professionals and geographical distance [bib_ref] Increasing access to and utilization of cognitive behavioral therapy for insomnia (CBT-I):..., Koffel [/bib_ref]. Digital CBT-I (web and/or mobile applications) provides an effective option for increasing access to this treatment to meet the demand [bib_ref] Internet-delivered cognitive behavioral therapy to treat insomnia: a systematic review and meta-analysis, Seyffert [/bib_ref].
Sleepio (www.sleepio.com) is a fully automated digital CBT-I program. Large effects on a range of sleep outcomes across diverse populations have been demonstrated [bib_ref] Helping employees sleep well: effects of cognitive behavioral therapy for insomnia on..., Barnes [/bib_ref] [bib_ref] Sleep and productivity benefits of digital cognitive behavioral therapy for insomnia, Bostock [/bib_ref] [bib_ref] Efficacy of digital CBT for insomnia to reduce depression across demographic groups:..., Cheng [/bib_ref] [bib_ref] Is digital cognitive behavioural therapy for insomnia effective in treating sub-threshold insomnia:..., Denis [/bib_ref] [bib_ref] A randomized, placebo-controlled trial of online cognitive behavioral therapy for chronic insomnia..., Espie [/bib_ref] [bib_ref] Effect of digital cognitive behavioral therapy for insomnia on health, psychological well-being,..., Espie [/bib_ref] [bib_ref] Efficacy of digital cognitive behavioral therapy for the treatment of insomnia symptoms..., Felder [/bib_ref] [bib_ref] A randomized controlled trial of digital cognitive behavioral therapy for insomnia in..., Kalmbach [/bib_ref] [bib_ref] The effects of digital cognitive behavioral therapy for insomnia on cognitive function:..., Kyle [/bib_ref] [bib_ref] Sleep to lower elevated blood pressure: a randomized controlled trial (SLEPT), Mcgrath [/bib_ref] [bib_ref] The anxiolytic effects of cognitive behavior therapy for insomnia: preliminary results from..., Pillai [/bib_ref]. However, to our knowledge, digital CBT-I has yet to be examined in people with stroke. As many people experience long term difficulties with movement, language and cognition after a stroke, a digital program may need to be adapted in order to accommodate these difficulties.
We therefore sought to undertake a qualitative study to explore the experience of people with stroke using Sleepio, to gain an indication of usability and guide the design of future randomized controlled trials. Specifically, we aimed to examine 1) the general experience of using the program and 2) whether there were any barriers to completion.
# Methods
## Participants
A convenience sample of participants was recruited through stroke user group meetings and a database of past research participants. Inclusion criteria were: aged ≥18 years, >3 months post-stroke, interest in improving sleep.
Exclusion criteria were: current, frequent travel between time-zones, current ongoing shift work. We chose to include people with stroke who did not necessarily present with insomnia, as we felt that they would still have valuable contributions regarding usability of the program in relation to post-stroke difficulties, and a recent study suggests that Sleepio is similarly effective and acceptable in people with sub-clinical insomnia [bib_ref] Is digital cognitive behavioural therapy for insomnia effective in treating sub-threshold insomnia:..., Denis [/bib_ref].
Twenty people with stroke were emailed the participant information sheet and invited to participate; 2 were ineligible (1 had no interest in improving sleep, 1 was unable to provide consent), 5 didn't respond to the invitation, 2 declined (1 due to caring responsibilities, 1 didn't want to take part) and 11 provided written informed consent. A maximum variation sampling approach was used with a focus on finding participants of different ages, stroke outcomes and self-reported familiarity with digital technology. Data collection took place over 7 months, and recruitment stopped once data saturation was reached. The study was approved by the local Research Ethics Committee (R61184/RE001).
## Procedure
Participants completed two visits (baseline and follow-up), at the University or in their homes (based on participant preference). At the baseline visit, they completed the questionnaires then watched a 2-minute introductory video, created an account (with a code that enabled researchers to track progress) and completed the first digital CBT-I session. They then used the Sleepio program at home before returning for the follow up interview.
## Assessments
Participants completed the Sleep Condition Indicator (SCI-8 (28)), to determine self-reported sleep quality. The SCI is scored from 0 to 32, where lower values are indicative of greater insomnia symptoms and scores ≤ 16 indicate probable insomnia. We also used the Montreal Cognitive Assessment (MoCA (29)) at baseline to gain an indication of cognitive, language and perceptual difficulties which may influence ability to use the program. The MoCA is scored from 0 to 30, and higher values indicate fewer cognitive difficulties.
## During the program
Sleepio is delivered through a web-based platform or iOS app and is structured around 6 CBT-I sessions (each ~20 minutes). Sessions are available 7 days after completing the previous session. Content is based on cognitive (e.g., cognitive restructuring, paradoxical intention etc.) and behavioral (e.g., stimulus control therapy, sleep restriction therapy and relaxation) techniques, and is delivered by an animated therapist "The Prof." Treatment is personalized based on questionnaire responses and daily sleep diaries.
A researcher (TS) contacted participants weekly to monitor for adverse effects and to ask how they were finding the program [fig_ref] Table 1: Weekly checkup questions [/fig_ref]. Responses were used to help guide the discussion at the follow up interview, but were not recorded or analyzed specifically. TS also contacted participants when more than four days had passed without accessing an available session to prompt continuation and provide assistance when needed. Although this is a deviation from how Sleepio would ultimately be used, it was felt to be important to ensure that all issues with program use were identified and could be discussed at the final interview.
## Follow-up interview
Following completion of Sleepio (or earlier if they did not want to complete) participants attended a face-to-face semistructured, audio recorded interview (at the University or in their home, depending on participant preference). The same researcher (TS) conducted all participant interactions in an attempt to maximize participant comfort and openness during interview. The interviewer (TS) is a male postgraduate research assistant with an undergraduate degree in Psychology who was completing this study as part of a Master's degree in Clinical Neuroscience. He has worked extensively with people following stroke and his research interests are focused around seeking to understand recovery and rehabilitation after stroke. The topic guide included questions aiming to explore experience and ease/difficulty of usage . The first participant acted as the pilot for the topic guide, but no changes were deemed necessary prior to continuation of the study. Followup probes were used to encourage elaboration on points noted during the weekly contacts or if they mentioned elements that had not been addressed by the structured questions. In an attempt to reduce social desirability and positivity bias the researcher consciously avoided positive framing of questions. Participants were told that TS was not affiliated with Sleepio, and that his goal was to understand their experience of the program to help guide future studies. - How are you finding the sessions? - How are you finding keeping the diary?
- Have you made any changes suggested by the program?
- Does the weekly schedule suit you? - How difficult are you finding it to keep up with the program (scale of 1-10)? . Semi-structured interview topic guide.
- How did you find using the Sleepio program?
- Was there anything you found particularly difficult?
- Did you find it easier to use the phone or laptop/ why did you use the modality you used?
- Were you able to remember all the different times asked for in the diaries?
Did you find it difficult?
- How motivated were you to stick with the program? - Do you feel like it helped? Have you noticed a change?
- Did you learn anything new about the way we sleep?
- What was your favorite thing about the program?
- What was your least favorite thing?
- Were any problems you found with the program related to your stroke, difficulty using the program/ difficulty remembering to complete the sessions?
- Would you suggest anything to change within the program to make it better or easier to use?
- Did you use all the functions: diary, checklist etc. if so, were they easy to use/ helpful? If not, why not?
- Did you use the library articles or community?
- Were you able to implement any of the changes? - Do you have anything else you would like to talk about that we touched on, or do you have any questions for me?
# Analysis
Interview data were analyzed using a reflexive thematic analysis approach as outlined by Braun and Clarke (30), following six phases: familiarization, generating initial codes, searching for themes, reviewing themes, defining and naming themes, and producing a report. An inductive approach was used as there were no preexisting reports on the usage of Sleepio by people with stroke. Audio files were transcribed verbatim by a researcher (TS). Transcripts were not provided to participants, and we did not seek their feedback on the findings. Transcription and initial coding were conducted continuously during the study, while participant recruitment was ongoing, but creation of the themes occurred after all data collection had stopped.
A second researcher (MF) coded four transcripts, chosen pseudorandomly (i.e. one from the first 3 participants, one from the last 3 participants and the remainder from the middle participants). If discrepancies were found, the two researchers discussed this and agreed on a consensus, adapting themes if required. [fig_ref] Table 3: Participant characteristics, time to complete Sleepio, and interview duration [/fig_ref] shows the participant demographics, MoCA and SCI scores. Based on the baseline SCI score, five participants would meet the criterion for probable insomnia disorder. Three participants chose to have a partner help them, who also attended the qualitative interview. One of these participants had a low MoCA score, suggestive of some cognitive impairments. The other two had MoCA scores comparable to other participants who did not seek to have assistance. The researcher (TS) knew seven of the participants from previous studies. The remaining participants were met in person for the first time at the baseline assessment.
# Results
All participants completed the final interview, including the two who did not finish the Sleepio program. Interviews lasted 42 minutes on average (range 17-75 min; [fig_ref] Table 3: Participant characteristics, time to complete Sleepio, and interview duration [/fig_ref]. Four themes were identified [fig_ref] Figure 1: Coding tree. [/fig_ref].
## Theme 1: positive and negative experiences impacted engagement with the program
Participants reported feeling positive and actively engaged with Sleepio when the program was found to be effective. P3 "All it took was for one good night sleep and then I was just 'Oh my god I've had a good night's sleep' and then I was motivated." P11 "I tried every technique under the sun and nothing else had worked . . . well this did." Participants also reported feeling more in control of their sleep: P7 "If you're having trouble it makes you feel a bit more in control of what's happening to you." Even in cases where participants reported not having much of a sleep problem, enjoyment was mentioned in regards to the educational nature. Similarly, the partners aiding participants often commented on their own enjoyment:
P4 "Very informative actually, good ideas, very good ideas . . . I enjoyed the whole lot." P5 (Partner) "We didn't use the sessions [with the Prof] enough. They were really good, we really liked those." Once familiar with the program, participants reported daily usage becoming part of their routine.
P3 "The sleep diary wasn't a problem at all . . . It's just one of those other apps you check in the morning." However, the inclusion of advice that some participants found to be impractical or unfeasible led to feelings of disconnect at times, or feeling that the course might not be for them.
P1 "I just didn't feel it was really aware of, of the sort of living situation that people with disabilities have." It was frequently mentioned that hearing advice that didn't apply to them was frustrating:
P6 "The way they spoke as if most people's problems was thinking about not sleeping that's not my case. I can't sleep because my mind's active on thinking about completely different things." P8 "It assumes you're in perfect, in inverted commas 'health' all the time but when I was ill that seriously upset my routine." This disconnect also led to the perception of the program feeling long or being too demanding:
P3 "There was just too much information in all the sessions and some of it applied to me and some of it didn't." P1 "I got slightly flippant by the end because it's like the same question for 30 days in a row." Although the design of the program is to mimic personal CBT-I, at times when advice or information was not well suited to the individual, a disconnection was felt quite strongly.
P1 "I mean, I sort of lost faith in the whole thing. I continued to do my recording because as I said it was a good process for me just to do the recording but it felt they'd lost me." P11 "I came so close to stopping, so close to stopping because I was entirely convinced they were wrong."
## Sub-theme 1a: suggested changes to improve the program
Many participants suggested changes that they thought could be beneficial. Most commonly, they wanted to be able to input more information about their situation. P10 "he's asked rhetorically 'so how are you doing this week?' . . . and I thought well actually I wish I could answer that . . .."
P11 "The one big issue that I had with it, is that there was no . . . 'why?,' you know, I don't get up at 6 o'clock because I want to, but because I have to . . . " Participants also mentioned wanting additional advice in relation to other factors that were affecting their sleep.
P1 "possible suggestions for managing pain when you're trying to go to sleep . . . it didn't feel as if it came into the app at all."
## Theme 2: motivation to follow program was proportional to perceived severity of sleep problem
Participants reported some aspects as challenging, and their motivation to follow advice was proportional to their perception of need for improvement. P1 didn't perceive their sleep difficulties to be problematic enough to fully engage with all of the suggestions, despite having an SCI score indicative of probable insomnia (9 out of 32). P1 "I didn't find them [suggestions to change the room] hugely useful but I guess it's not that I feel I need radical changes or that I'm so worried about my sleep that I'll do whatever it takes." Sleep Restriction Therapy (SRT) was identified as a challenge. P11, who reported six years of sleep difficulty, felt a strong desire to try to improve despite the challenges:
P11 "I had to stay up till midnight and it nearly bloody killed me and I'm just sort of thinking 'this is ridiculous' but I thought 'no I said I would do it' so I forced myself through." However P8 reported that by the time they reached the sleep restriction week they felt that their sleep had improved enough that they didn't need to engage with more challenging techniques:
P8 "I had got into a routine and was beginning to crack it."
## Theme 3: impractical advice for people with stroke
Participants reported that advice could be impractical or unsuitable based on their stroke outcomes. Frequently it was reported that advice to avoid daytime naps was problematic. P6 "After a stroke you do get more tired . . . very often I have to have a little snooze during the day . . . this tends to say 'oh don't do that' . . . So, I'm not sure if that's necessarily good advice." P10 (Partner) "It's very difficult to control going to bed and getting up at a certain time and staying awake." Participants also reported difficulty with with aspects of Stimulus Control Therapy, particularly the suggestion, when struggling to sleep, to get out of bed and go to another room until sleepy. Several participants had concerns about limited mobility:
P3 "I'm just worried about being a bit too wobbly in the middle of the night . . . and with my mobility problems I don't really want to start moving around the house." Difficulties were also reported in regards to Progressive Muscle Relaxation (PMR), which involves systematically tensing and relaxing different body parts. This technique was difficult for participants who experience high muscle tone as a result of one-sided weakness.
P1 "So I can tense and then yeah I mean it is a problem getting it to relax."
## Theme 4: difficulty operating the program
A variety of difficulties operating the program were reported. These related to difficulties using computer-based programs due to stroke-related impairment, limited computer literacy or due to elements that were not seen as intuitive.
## Sub-theme 4a: stroke related accessibility
Outcomes of stroke were mentioned as the cause of difficulty for some. Three participants had a partner to assist as they felt their cognitive difficulties meant they would not have been able to access the program without help. P5 (Partner) "You could not have done it alone darling" P5 "No, indeed, yes. Remember the first time you had to talk to me many times, sometimes even more than 2, 3 times." P8 "Since my stroke I have lost the ability, attitude or what I use to use the computer and I don't use it now." Even in cases where participants were completing Sleepio alone, it was mentioned that the stroke had had an impact:
P9 "I mean before the stroke I would have thought I was much more competent in operating a program." In contrast, even though five participants had one-sided weakness leading to the functional usage of only one hand (determined through observation and discussion), no issues were reported:
P2 "I have only one hand, sadly, but everything that I could do, could be done with one hand." However, there was a potential issue with accessing some of the written parts of the program. P2 noticed that although sessions were largely audio driven, some questions did not have an accompanying audio.
P2 "Personally I would like to see all audios so the . . . what's it called? 'Think about the typical night of the last month.' It was readable but not speaking . . . So I could read that but some people can't."
## Sub theme 4b: initial difficulty navigating program
Six participants struggled to find the prompt to start the second digital CBT-I session, requiring contact with the researcher. For some, this early delay was a result of not noticing that a new session was available: P1 "It didn't say "click here . . . it's assuming a lot that someone should just assume that you know without a prompt." P7, using the iOS app, described the fact that she had not noticed the button used to access the sessions when using the iPad as it was usually "greyed out" when filling in the sleep diary:
P7 "I didn't really even notice it because when you first get it you press everything to see what it does (laughs) and that didn't do anything and then whenever I then looked at it, it was always grey and I couldn't do anything with it." One of the common reasons given by the older participants was a lack of computer experience:
P4 "I'm not very good with computers. Once I'm set up I'm fine but before that I'm a bit confused you know."
On the other hand, some participants cited their prior experience with computers as one of the reasons they found the program easy to navigate: P11 "I use computers all day, every day so I'm pretty tech savvy and I could see that if somebody wasn't they might get confused . . . "
# Discussion
We sought to explore the experience of people with stroke using the digital CBT-I program "Sleepio" and perceptions of usability. There were no adverse events reported and 9 out of 11 participants were able to complete the program. Many participants found Sleepio educationally enjoyable and useful in improving their sleep. However, there were some clear considerations for future studies, especially if seeking to include people with mobility issues, aphasia and/or cognitive impairment.
Although participants discussed their enjoyment and learning, some specific aspects were reported as not being particularly "stroke friendly." Many participants struggled to stop napping due to fatigue. This difficulty is not necessarily unique to people with stroke (31), but post-stroke fatigue is highly prevalent and debilitating [bib_ref] The prevalence of fatigue after stroke: a systematic review and meta-analysis, Cumming [/bib_ref]. Adapting this advice to recommend reducing the number or duration of naps, and to nap only at particular times of day, may help the feasibility for participants with fatigue and increase the feeling of inclusion and personalization.
We were pleased to find no clear physical issues associated with operating the program for participants with one-sided arm weakness. However, advice to leave the bed when unable to sleep was considered unsafe for participants with reduced mobility. This has also been noted for in-person CBT-I in this population as well as in people with traumatic brain injury (10,11). This could be adapted by advising users to sit on the edge of a bed, if leaving the room is not feasible, or using relaxation techniques such as those implemented in the study by . This may still help to break the bedinsomnia association. Similarly, it was noted that progressive muscle relaxation was deemed to be unsuitable for people with high tone (a common complication associated with motor deficits). It is worth mentioning that all participants are provided information on progressive muscle relaxation, however, following this "the Prof" instructs anyone who experiences muscular or joint problems to use autogentic relaxation instead. Progressive muscle relaxation could therefore be avoided if needed.
Only two participants failed to complete the program. One was unable to identify when new sessions were available, even with support from the researcher and a partner. The second participant was delayed by not having internet access whilst traveling. He reported struggling to follow the program upon return. Whilst this completion rate is promising, future larger studies are needed to examine attrition rate in comparison with other populations, given that in the current study the researcher provided additional contact compared with the standard operating of the Sleepio program.
Three participants reported requiring help from a partner to operate the program. This is an important consideration for future studies, suggesting that a wider group of people may be able to use digital CBT-I if given support from a partner or carer, than if studies are restricted to those who can use it alone. Although it is not possible in this study to determine to what extent these difficulties are due to the stroke per se, or due to the older age of the participants (the mean age of the sample was over 60 years), further instructional content such as a navigation tutorial to help novice users, and a more obvious prompt when the digital CBT-I sessions become available, could also prove helpful for this population. Sleepio does offer users the option of receiving e-mail and SMS reminders when new sessions are available, however most participants in this study did not opt to have reminders. This may be have been influenced by the knowledge that a researcher would be checking on them, but may also be a result of not knowing that this option was available. Even with additional navigation aids, it should be considered that individuals with more severe cognitive impairments and technological novices may require external assistance or may not benefit to the same extent. In cases where users need additional support, without access to a friend or relative to provide this help, Sleepio offers the option to connect your profile with your clinician. This feature was used in the present study to allow researchers to monitor participant progression. It is conceivable that outside of research, a hybrid approach could be useful in this population, where Sleepio is used in conjunction with input from a clinician where appropriate. A qualitative study examining in-person CBT-I in older adults with co-morbid insomnia and depression also found that participants expressed a desire for a multimodal delivery approach [bib_ref] Cognitive behaviour therapy for insomnia and depression: qualitative reflections from older adults..., Sadler [/bib_ref].
# Limitations
One limitation of this study is in our choice to include participants who did not report symptoms of insomnia. This may have led to more negative impressions of participant engagement than would have been observed otherwise, as participants may not have felt that certain techniques were warranted if they had mild symptoms. However, given that recent evidence suggests that Sleepio may also be useful in people with more sub-clinical insomnia symptoms [bib_ref] Is digital cognitive behavioural therapy for insomnia effective in treating sub-threshold insomnia:..., Denis [/bib_ref] , we feel that the inclusion of people with varied severities of sleep difficulty and wide ranging outcomes after stroke helped to ensure a breadth of viewpoints. Similarly, we did not exclude participants experiencing symptoms of other sleep disorders. This was partly because this study was focused on the experience of using the program rather than its efficacy. However, for future efficacy studies it will be important to consider excluding participants with other sleep disorders (such as sleep apnea) given that this will likely influence the level to which a participant can benefit from CBT-I.
Another limitation of this study is that all participants were in the chronic stage of stroke. This was intentional, as this is the time when the majority of rehabilitation input has finished, and ongoing consequences may be realized. It was mentioned by a couple of participants that their ability to complete the program was only possible given their recovery level. Therefore, more research is needed early after stroke to determine whether usability differs at this timepoint or whether there are additional barriers to consider.
The researcher (TS) knew seven of the participants from previous studies, which potentially introduces bias. Additionally, it was clear that some participants felt they should continue with the Sleepio program, having agreed to participate in a study, rather than having a desire to complete Sleepio itself.
In an attempt to reduce social desirability and positivity bias the researcher consciously avoided positive framing of questions, and participants were told that TS was not affiliated with Sleepio.
## Summary and implications for future research
Participants found Sleepio to be enjoyable and some reported it as being effective in improving their sleep quality. However, there were some consistently reported issues that could put people with stroke at a higher risk for dropout or incompletion, and so the provision of supplementary information alongside the program may prove useful.
Future research should investigate the efficacy of Sleepio in people with stroke as well as potential predictors of treatment response. If effective, providing Sleepio as part of clinical care could assist in increasing general well-being, reducing the risk of further health problems, and improving quality of life.
any Author Accepted Manuscript version arising from this submission.
# Data availability statement
Data is available upon reasonable request to the corresponding author (MF).
[fig] Figure 1: Coding tree. [/fig]
[table] Table 1: Weekly checkup questions. [/table]
[table] Table 3: Participant characteristics, time to complete Sleepio, and interview duration. MoCA = Montreal Cognitive Assessment (higher values indicate better cognition), SCI = sleep condition indicator (higher values indicate better self-reported sleep quality). Mo = months. N/C = not completed. [/table]
|
Applications of Deep Learning for Dense Scenes Analysis in Agriculture: A Review
Deep Learning (DL) is the state-of-the-art machine learning technology, which shows superior performance in computer vision, bioinformatics, natural language processing, and other areas. Especially as a modern image processing technology, DL has been successfully applied in various tasks, such as object detection, semantic segmentation, and scene analysis. However, with the increase of dense scenes in reality, due to severe occlusions, and small size of objects, the analysis of dense scenes becomes particularly challenging. To overcome these problems, DL recently has been increasingly applied to dense scenes and has begun to be used in dense agricultural scenes. The purpose of this review is to explore the applications of DL for dense scenes analysis in agriculture. In order to better elaborate the topic, we first describe the types of dense scenes in agriculture, as well as the challenges. Next, we introduce various popular deep neural networks used in these dense scenes. Then, the applications of these structures in various agricultural tasks are comprehensively introduced in this review, including recognition and classification, detection, counting and yield estimation. Finally, the surveyed DL applications, limitations and the future work for analysis of dense images in agriculture are summarized.Sensors 2020, 20, 1520 2 of 33 meaningful and challenging[2]. With the significant reduction of computer hardware cost and the significant improvement of GPU computing power, the computer vision task has become an important area of machine learning and artificial intelligence. In general, machine learning combined with high-performance computing guarantees the ability to process large amounts of image data effectively.Early solutions to computer vision tasks depended on traditional machine learning methods, i.e., feature-based manual method. Common features include Deformable Part-Based Model (DPM) [3], Histogram of Oriented Gradient (HOG) [4] and Scales-Invariant Feature Transformation (SIFT) [5], Speeded Up Robust Features (SURF) [6] and Haar-like features[7]. They were usually combined with classifiers such as Support Vector Machine (SVM) to classify each image pixel. Although the traditional methods are easy to understand and many improvements have been done to them, most of them are verified in low and medium density images and they usually need to be changed according to the specific situations[8,9]. Moreover, most traditional methods either ignore the problems in dense scenes, namely, there is no discussion on dense scenes, or use simple heuristic methods based on shape and size, these methods are very ineffective in natural environments with severe occlusion and large scale changes[10]. Therefore, traditional machine learning methods are not appropriate for dense images. It has been demonstrated that in many applications, features extracted by deep learning are more effective than these hand-crafted feature[11]. Moreover, deep learning solves various challenges in dense images.Deep Learning (DL) [12] belongs to machine learning, based on representation learning of data, which realizes artificial intelligence by means of artificial neural networks with many hidden layers and massive training data. DL has been successful in computer vision [13], natural language processing [14], bioinformatics [15], automatic control [16], machine translation [17], automatic driving [18] and other practical problems[19]. The reason for the success of DL lies in its unique characteristics of network structure: deep neural network can acquire high-level features by learning local features from the bottom and then synthesizing these features at the top. DL uses multi-level abstraction to learn complex feature representations from raw data and generate components automatically[20]. Different features at different levels can correspond to different tasks. Deep learning is a technology that uses deep network structure to learn features. Deep learning emphasizes the depth of the model structure, highlights the importance of feature learning and proposes various techniques to learn more and higher-level features better and faster. Strong learning ability enables them to implement various kinds of problems especially well and flexibly adapt to numerous highly complex problems. Monitoring and studying a large number of interesting objects in videos or images is an important task for the macro-world and micro-field, for instance, the research on crowding traffic and microscopic microorganisms[21,22].Usually, advances in one area are driven more by some combination of expertise, resources, and application requirements in other areas[23]. Similarly, applications of DL in analyzing dense scenes spread to the agricultural sector after advances in medical diagnostics and population analysis. More and more scenes in agriculture produce a lot of high-density images, and they are becoming more and more attention. At present, agricultural tasks have been transformed into tasks of agrovision (computer vision in agriculture). There have been some reviews on the applications of DL in agriculture[24,25]and some reviews pertinent to the use of DL in computer vision. They either gave a comprehensive overview of DL methods applied throughout the agricultural field or the latest research of DL technology in a certain agricultural field and also reviewed the application of DL methods in general computer vision tasks. However, none of them involved how DL works in dense agricultural scenes. Koirala et al. [23] summarized the use of DL in fruit detection and yield estimation, including the problem of occluded fruit in imaging and the solutions. However, they were only concerned with the detection and yield estimates while ignoring other agricultural tasks containing a large number of objects. Thus, the motivation for preparing this review stems from the need to summarize the applications of DL in agriculture with the increase of dense scenes and images.
# Introduction
Nowadays, smart agriculturehas become increasingly popular because it can adjust various agricultural management measures accurately according to the specific conditions of each unit of agricultural operation. In order to optimize the input of agricultural production links, almost all the management models in agriculture are based on the models of high-density planting and mass production, thus achieving the goal of obtaining the maximum economic and environmental benefits. However, in actual production and life, it is inevitable to produce a lot of dense images. It is worth taking note that the complexity of the problem increases exponentially with the number of objects and the interaction of the objects. Especially, the pattern of batch production and high-density planting determines the high density of agricultural scenes, and high density makes it difficult to meet the actual demand, such as detection, counting or yield estimation of agricultural products.
In recent years, due to the tremendous progress and popularity of image acquisition equipment, the image has become a huge amount of data and easy to obtain, which makes image analysis very Sensors 2020, 20, 1520 3 of 33
## Types of dense scenes in agriculture
Dense scenes mentioned in most papers refer to the videos or images about the crowds, and a lot of studies have been done in the field of automatic monitoring and crowd analysis. To the best of our knowledge, No one seems to define dense scenes in agriculture so far. In this section, by observing various scenes in agriculture, we distinguish whether the scene is dense or not according to the number of interesting objects with specific tasks, namely, there are two conditions to judge whether a scene is dense. One is that there are a large number of interested objects in the image. The other is the case that the research task is to detect/classify each object. Based on these available options, two types of dense scenes in agriculture were formulated:
(1) Quantity-dense scenes Quantity-dense scenesrefer to images or videos with a very large quantity of objects, which are often accompanied by severe occlusion and overlap. Although many systems can assess agricultural crops under controlled conditions, they ignore the interaction between crops and the environment, and many crops have to be evaluated directly in their growth environment. For example, fruit in orchards, grain in fields, vegetation in grasslands, etc. While in practice, producers usually adopt the way of high-density planting for agricultural production. In actual orchards, fruit is densely distributed and usually overlapped by each other, the ubiquitous branches and leaves even shade the fruit. Grain fields usually have very narrow row spacing when planting high-density crops, and overlapping plants often appears.
However, it should be noted that the images acquired in the above scenes are not necessarily what we call dense scenes. We should judge the scenes according to the specific tasks, such as in, which is also to study farmland, but the task is to classify different farmland types, and it does not care whether the grain distribution in farmland is dense or not, so we do not divide such scenes into dense scenes. The problems in analyzing this type of dense scenes are usually the same and the solutions are similar.
(2) Inner-dense scenes
The second type of dense scenes refers to some objects in an image which have dense attributes themselves, i.e., the objects in the images are the same kind of objects that are conglutinated with each other, such as wheat, grapes, hives, etc. The study of these objects, especially a wheat spike or an ear of rice, is often a study of dense objects. In this case, separating adjacent objects is the biggest challenge. That is to say, because of the large number and the interaction between adjacent objects, data labeling is a huge problem, not to mention the use of DL methods for subsequent processing. Due to the nature of those objects, these phenomena exist in every image of them. Therefore, the research on the second type of dense scenes is less than that of the first type.
With the deep study of agricultural production, it has been recognized that most scenes in agriculture are dense or crowded scenes. Therefore, analysis of agricultural problems mostly needs to analyze the dense scenes. As shown in, the high-density objects in specific agricultural tasks pose severe challenges to the analysis of agricultural dense scenes, especially for underwater fish or other aquatic objects. Consequently, more and more attention should be paid to the research of dense scenes in agriculture.
## Challenges in dense agricultural scenes
Research has found that there are a variety of thorny problems in dense scenes, as illustrated in. First of all, there are some challenges in both quantity-dense and inner-dense scenes in agriculture. Sa et al.pointed out that occlusion is a key challenge in dense scenes. There are two types of occlusion: scene occlusion and inter-object occlusion. Scene occlusion refers to the occlusion of the detected object by other non-objects; inter-object occlusion means that the detected objects in the image are occluded from each other, which is at the heart of the occlusion problem. It should be noted that in the quantity-dense scenes, there are two kinds of occlusion. However, for the innerdense scenes, there is often no occlusion of other objects in the image, so the main challenge is interclass occlusion. In addition, high-density objects will lead to a small size of a single object and the object resolution is relatively low in both types of agricultural dense scenes. Moreover, because of the small size of the object, it is easily affected by noise and leads to inaccurate detection, and there is a general limitation in agricultural applications, namely, sensitivity to changing illumination conditions. Although it has been proved that LIDAR sensors are reliable for solving this problem, they usually require higher investment costs. Besides, objects at different distances from the camera may vary greatly in scale, bringing about constant changes in imaging features, especially in object size.
Meanwhile, there are some unique challenges for the quantity-dense scenes in agriculture, i.e., complex background. The complex background includes two situations, one is that the background is similar to the object, and the noise is intense, which makes it difficult to distinguish the object from the background. The other is the case that the background may be dynamic, which changes with the moving of the object. Images obtained from realistic agricultural scenes such as farmland or orchard
## Challenges in dense agricultural scenes
Research has found that there are a variety of thorny problems in dense scenes, as illustrated in. First of all, there are some challenges in both quantity-dense and inner-dense scenes in agriculture. Sa et al.pointed out that occlusion is a key challenge in dense scenes. There are two types of occlusion: scene occlusion and inter-object occlusion. Scene occlusion refers to the occlusion of the detected object by other non-objects; inter-object occlusion means that the detected objects in the image are occluded from each other, which is at the heart of the occlusion problem. It should be noted that in the quantity-dense scenes, there are two kinds of occlusion. However, for the inner-dense scenes, there is often no occlusion of other objects in the image, so the main challenge is inter-class occlusion. In addition, high-density objects will lead to a small size of a single object and the object resolution is relatively low in both types of agricultural dense scenes. Moreover, because of the small size of the object, it is easily affected by noise and leads to inaccurate detection, and there is a general limitation in agricultural applications, namely, sensitivity to changing illumination conditions. Although it has been proved that LIDAR sensors are reliable for solving this problem, they usually require higher investment costs. Besides, objects at different distances from the camera may vary greatly in scale, bringing about constant changes in imaging features, especially in object size.
Sensors 2020, 20, 1520
## Of 33
Sensors 2020, 20, x FOR PEER REVIEW 5 of 32 tend to have high background noise, due to the nature of this challenge which is an unpredictable event, the methods are so varied and there is no robust solution that works in all conditions. These challenges may occur in other computer vision tasks, but more frequently or simultaneously in dense scenes. Thus, methods specifically for dense scenes in agriculture need to be proposed.
## Outline
As far as we know, this is the first such overview of the applications of DL in dense scenes of agriculture. We summarize various DL algorithms used in dense scenes, which will contribute to the application of DL algorithms in a natural agricultural environment. The rest of the paper is organized as following:
In Section 2, two categories of DL models, the most widely used models in dense scenes are reviewed: Convolutional Neural Networks (CNN) and Autoencoder, as well as the basic architectures, contributions, and limitations of each category. Then we introduce the achievements of those DL algorithms above to various agriculture applications in Section 3, including recognition, classification, detection, counting and yield estimation. Emphasis is placed on how the DL model handles dense scenes. Finally, we terminate this paper with a discussion and a conclusion. We hope this review can provide new ideas and suggestions researchers who are engaged in dense scenes analysis in agriculture and be satisfactorily used to develop expert or intelligent systems in other related fields.
# Deep learning methods
DL has shown an incomparable advantage for computer vision tasks in the context of big data. Especially in the tasks of detection, classification, and identification, DL technology occupies a core position at present. However, these methods are difficult to apply directly to dense scenes. Therefore, in this section, we focus on two DL methods that work well in dense agricultural scenes and introduce the popular backbone networks among them. As illustrated in, two DL methods are Meanwhile, there are some unique challenges for the quantity-dense scenes in agriculture, i.e., complex background. The complex background includes two situations, one is that the background is similar to the object, and the noise is intense, which makes it difficult to distinguish the object from the background. The other is the case that the background may be dynamic, which changes with the moving of the object. Images obtained from realistic agricultural scenes such as farmland or orchard tend to have high background noise, due to the nature of this challenge which is an unpredictable event, the methods are so varied and there is no robust solution that works in all conditions.
These challenges may occur in other computer vision tasks, but more frequently or simultaneously in dense scenes. Thus, methods specifically for dense scenes in agriculture need to be proposed.
## Outline
As far as we know, this is the first such overview of the applications of DL in dense scenes of agriculture. We summarize various DL algorithms used in dense scenes, which will contribute to the application of DL algorithms in a natural agricultural environment. The rest of the paper is organized as following:
In Section 2, two categories of DL models, the most widely used models in dense scenes are reviewed: Convolutional Neural Networks (CNN) and Autoencoder, as well as the basic architectures, contributions, and limitations of each category. Then we introduce the achievements of those DL algorithms above to various agriculture applications in Section 3, including recognition, classification, detection, counting and yield estimation. Emphasis is placed on how the DL model handles dense scenes. Finally, we terminate this paper with a discussion and a conclusion. We hope this review can provide new ideas and suggestions researchers who are engaged in dense scenes analysis in agriculture and be satisfactorily used to develop expert or intelligent systems in other related fields.
# Deep learning methods
DL has shown an incomparable advantage for computer vision tasks in the context of big data. Especially in the tasks of detection, classification, and identification, DL technology occupies a core position at present. However, these methods are difficult to apply directly to dense scenes. Therefore, in this section, we focus on two DL methods that work well in dense agricultural scenes and introduce the popular backbone networks among them. As illustrated in, two DL methods are respectively: Convolutional Neural Networks (CNN) and Autoencoder.displays the characteristics of various backbone networks and their highlights in dense scenes. There are many other deep learning structures that have been applied in agriculture, but have not been used in agricultural dense scenes. For example, Recurrent Neural Network (RNN) and Long Short-Term Memory (LSTM) are very useful in processing time-series data, which are typically used in agriculture for time prediction. Generative Adversarial Network (GAN) can be used to enrich datasets and it has been applied in agriculture, but it has not been widely used in agricultural dense scenes. In addition,shows the uses of surveyed DL methods in a recent study of dense scenes in agriculture. This chart information is very valuable to new researchers since it can help them find the most effective DL algorithm and further research.
Sensors 2020, 20, x FOR PEER REVIEW 6 of 32 respectively: Convolutional Neural Networks (CNN) and Autoencoder.displays the characteristics of various backbone networks and their highlights in dense scenes. There are many other deep learning structures that have been applied in agriculture, but have not been used in agricultural dense scenes. For example, Recurrent Neural Network (RNN) and Long Short-Term Memory (LSTM) are very useful in processing time-series data, which are typically used in agriculture for time prediction. Generative Adversarial Network (GAN) can be used to enrich datasets and it has been applied in agriculture, but it has not been widely used in agricultural dense scenes. In addition,shows the uses of surveyed DL methods in a recent study of dense scenes in agriculture. This chart information is very valuable to new researchers since it can help them find the most effective DL algorithm and further research.
## Convolutional neural networks (cnns)
## Review of cnns
CNN is a feedforward neural network proposed by Fukushimafor the first time, the essence of which is a Multi-Layer Perceptron (MLP). Despite some modifications, the core concept of a neural network has not changed much. It consists of the input layer, hidden layers, and output layer, each layer has hundreds of nodes. The nodes in the previous layer are connected to other nodes in the next layer by weight, which reflects the strength of the connection between the two nodes.shows Sensors 2020, 20, 1520 8 of 33 the overall architecture of the CNNs: as a whole, the CNNs can be divided into three main types of neural layers, namely, convolutional layer, pooling layer, and fully connected layer. Various CNNs have different implementation details. The function of the convolutional layer is to extract features from input data, which contains multiple kernels. Because the spatial connection of the image is local, each neuron only needs to feel the local image area, and CNN can get the global information at a higher level. The values in the convolution kernel is the weights, and the weights of each local area in the same convolution layer are shared. There are many initialization methods for weight, such as random initialization, pre-training, etc., and the weights are updated during model training. Using different kernels, various feature maps can be generated. The reason for the success of convolutional layer and popularity to date lies in the way of its local connectivity and weight sharing, which reduce the number of weights to reduce the complexity of the model and the risk of over-fitting, as well as making the network easy to optimize. The pooling layer, also known as the sub-sampling layer or down-sampling layer, is often behind the convolutional layer, which is the operation of feature selection and filtering for the input feature map. On the one hand, it makes the feature map smaller and simplifies the complexity of network computing; on the other hand, it compresses the feature, extracts the main features, and ensures the invariance of feature position and rotation. Average pooling and max pooling are the most commonly used strategies, among which max pooling is used more because of its better performance. The fully connected layer is usually the last part of the hidden layer of CNNs. Its function is to integrate the highly abstract features which have been convoluted and pooled many times before, that is, to connect all the features into one vector, then, the expanded vector is passed to the classifier through the activation function for classification or further processing. Research has shown that the representation learned by CNN is rich in an amazing amount of image information, including position, size and pose of objects. The training process of CNNs can be divided into two stages. The first stage is called forward propagation stage. Data flows from the input to the output, and the network extracts the features from the low level to the high level. Another stage is called the back-propagation stage, which aims to adjust the weight and bias to minimize the overall cost (error). Follow-up works improved CNNs to enable it to be more effectively learn and extract the features, which have great advantages in the process of 2-D image processing. Now, 3D convolution characterized by 3D convolution kernels for the processing of videos is commonly used. In the analysis of dense scenes, most of the research works use CNNs as the framework.
## Cnns' backbone network
There are many popular and advanced backbone networks of CNNs. Although most deep CNNs are composed of a set of base layers introduced above, each network has its characteristics and suitable application scenes; not all networks are suitable for dense scenes. Therefore, in this section, we will focus on the backbone networks which are most used and work well in dense scenes. Including classification networks: VGGNet, GoogLeNet, ResNet, detection networks: DetectNet, YOLO, and semantic segmentation networks: FCN, SegNet and U-Net. Most of the works reviewed in this paper are based on these network structures for combination or fine-tuning, making it more suitable for analyzing dense scenes in agriculture.
Visual Geometry Group Network (VGGNet)explores the relationship between the depth and performance of a convolution neural network. VGGNet inherits some frameworks of LeNetand AlexNet. However VGGNet uses the small convolution filter size. The outstanding contribution of VGGNet is to prove that using a very small convolution filter size and increasing the depth of the network can effectively improve the performance of the model. At the same time, it has strong expansibility, and good generalization ability for other datasets makes it a valuable network for transfer learning.
Different from previous structures of AlexNet and VGGNet, which increase the depth of the network to improve training results, GoogLeNetis a novel DL structure. To build a sparse and high computing performance network structure, an inception module is proposed in the GoogLeNet inspired by that sparse matrices can be clustered into more dense sub-matrices to improve computing performance. Convolution and reassembly of multiple sizes can extract the features of different scales and accurately gather the features with a strong correlation in advance, thus accelerating the convergence of the network. DetectNet is a GoogLeNet-based network for optimizing object detection. It consists of two parts: fully convolution neural network and clustering function. The fully convolution neural network is based on the GoogLeNet structure without the input layer, the final pool layer, and the internal product layer. Allowing the use of pre-trained GoogLeNet to initialize DetectNet reduces training time and improving the accuracy of the final model. DetectNet not only predicts the existence of the object but also predicts the position of the boundary angle of the object relative to the center of the grid square, this is the main advantage of using DetectNet for object detection.
The original inspiration of Residual Network (ResNet)is a problem called degradation, that is, accuracy will rise first and then reach saturation, and continuous increase in depth of the net will lead to a decline inaccuracy. As the convergence of deep network and accuracy reaches saturation, the network begins to degenerate. This problem is solved in ResNet by a structure consisting of several layers of networks containing a shortcut connection called a residual block. Mask R-CNNuses ResNet and Feature Pyramid Networks (FPN)to fuse and extract multi-layer features. Shortly after ResNet was launched, Google borrowed the essence of ResNet and proposed Inception-v4and inception-residual-v1/v2. Inception-residual network combines the inception idea, which links the results of convolutional layers with different filter sizes through inception module to capture the characteristics of multiple sizes.
The two-stage method represented by R-CNN has higher and higher detection accuracy in computer vision tasks, but it is sometimes difficult to meet the real-time requirements in some scenes. One-stage method represented by YOLOintegrates feature extraction, candidate frame regression, and classification into a network, which improves the detection speed. YOLO is improved based on the GoogLeNet architecture, and the difference is that YOlO uses 1 × 1 + 3 × 3 convolution kernels to replace the inception module. Nevertheless, it has a problem in detecting small objects since the YOLO model can only detect one object class in each cell. So, when using this model in dense images, different changes are usually made to the initial model. Later, YOLOv2and YOLOv3are therefore proposed to improve it. In addition, new classification models, DarkNet-19 and DarkNet-53, are proposed in YOLOv2 and YOLOv3. Using ResNet for reference, both of them are mainly composed of a series of convolution kernels of 1 × 1 and 3 × 3 to avoid gradient dispersion and explosion of depth network.
Fully Convolutional Networks(FCN)is a kind of network that uses the convolutional layer instead of the fully connected layer in the previous classification network. In dealing with dense images, the segmentation of objects is the basic work of many tasks. FCN is one of the first successful semantically segmentation method. There are three main technologies used in FCN: convolutional, up-sample and skip layer. The existence of a fully connected layer will lead to a large storage overhead and constrain local features. Then FCN transforms the last three layers of CNN into three layers of the convolutional layer, which is used for end-to-end and point-to-point semantic segmentation.
As an effective image segmentation method, FCN has some limitations. For example, the blurring of the underlying features makes the network insensitive to the details of the image, and often ignores the relationship between the local and the whole. In addition, the pooling layer of CNNs discards part of the location information, semantic segmentation needs to retain the location information discarded in the pooling layer. As a result, networks based on encoder-decoder are proposed to solve this problem. SegNetand U-Netare the most prevalent semantic segmentation networks based on encoder-decoder, which are also used widely in dense scenes.
SegNet is modified by the VGG-16 network based on FCN. The encoder used in SegNet the first 13 layers of the convolution network of VGG-16. Each encoder layer corresponds to a decoder layer. In the encoder process and decoder process, the convolution method utilized by SegNet is the same Sensors 2020, 20, 1520 10 of 33 convolution. The difference is that in the decoder process, convolution is used to make information lost in the pooling process available through the decoder.
U-Net is an improvement on FCN, consisting of a contraction path and an expansion path. The contraction path is used to get context information, and an expansion path is used to locate accurately. Different from FCN, U-Net uses skip connection to cascade the feature information of contraction path of the model with that of deconvolution operation in expansion path, so as to obtain multi-scale feature information to improve network performance. In order to recover the structural features of the original image, U-Net uses 4 skip connection to connect the low-level and high-level feature maps.
## Autoencoder
Overview Autoencoder network is a type of unsupervised learning method. The main objective of autoencoder is to automatically learn and represent the unmarked input data, typically for data dimensionality reduction, compression, fusion. Autoencoder includes two processes: encode and decode. Input images are processed by encoding to get code, and then decode to get output, by using the back-propagation algorithm to train the network so that the output equals the input. High-dimensional data is always difficult to understand; if we reduce the input data to a lower dimension, it will be much more intuitive and easy to handle.
The followingshows the conceptual diagram of an autoencoder. It is a kind of neural network aiming at reconstructing input information and it can give a better description of features than the original data. That is to say, there is the following relationship between the output layer and the input layer of the autoencoder network:
[formula] x i ≈x i [57]. [/formula]
proposed to solve this problem. SegNetand U-Netare the most prevalent semantic segmentation networks based on encoder-decoder, which are also used widely in dense scenes.
SegNet is modified by the VGG-16 network based on FCN. The encoder used in SegNet the first 13 layers of the convolution network of VGG-16. Each encoder layer corresponds to a decoder layer. In the encoder process and decoder process, the convolution method utilized by SegNet is the same convolution. The difference is that in the decoder process, convolution is used to make information lost in the pooling process available through the decoder.
U-Net is an improvement on FCN, consisting of a contraction path and an expansion path. The contraction path is used to get context information, and an expansion path is used to locate accurately. Different from FCN, U-Net uses skip connection to cascade the feature information of contraction path of the model with that of deconvolution operation in expansion path, so as to obtain multi-scale feature information to improve network performance. In order to recover the structural features of the original image, U-Net uses 4 skip connection to connect the low-level and high-level feature maps.
## Autoencoder
Overview Autoencoder network is a type of unsupervised learning method. The main objective of autoencoder is to automatically learn and represent the unmarked input data, typically for data dimensionality reduction, compression, fusion. Autoencoder includes two processes: encode and decode. Input images are processed by encoding to get code, and then decode to get output, by using the back-propagation algorithm to train the network so that the output equals the input. Highdimensional data is always difficult to understand; if we reduce the input data to a lower dimension, it will be much more intuitive and easy to handle.
The followingshows the conceptual diagram of an autoencoder. It is a kind of neural network aiming at reconstructing input information and it can give a better description of features than the original data. That is to say, there is the following relationship between the output layer and the input layer of the autoencoder network:
≈. Generally, different variations of the autoencoder have been proposed to ensure robust features representation for machine learning applications. These include parse autoencoder, denoising autoencoder, stacked autoencoder, and contractive autoencoder.
# Conclusion
These deep learning networks are used to solve various challenges in agricultural intensive scenes due to their own characteristics. To deal with the complex problems encountered in dense scenes, it is inevitable to deepen the depth of the network. Therefore, VGGNet has a broad application in dense scenes analysis. Convolution and reassembly of multiple sizes in GoogLeNet is an effective way to solve the problem of scale inconsistency in dense scenes. The training data of DetectNet is a large image containing multiple objects, which is suitable for dense object detection. The main advantage of ResNet is to destroy the flow of information in the network, but this means that some features can be reused, and the ResNet architecture is easier to optimize than other deep networks. Inception-residual network combines the inception idea, which links the results of convolutional layers with different filter sizes through inception module to capture the characteristics of multiple sizes. In addition, YOLO can recognize the background correctly, and can be used to detect a certain degree of occlusion and overlapping objects. The structure of FCN reduces the repeated calculation, reduces the complexity of the model with filling up the missing detail data during upsampling, and the input image can be any size. It is used for accurate crowd count estimation in crowded scenes. As a semantic segmentation network based on encoder decoder, while the segmentation result of SegNet is not good, so in the dense agricultural image, U-Net is more used. The outstanding advantage of U-Net is that it can get very accurate segmentation results with very few training images. However, it still considers more high-level features and makes little use of low-level features. In general, analysis of agricultural dense scenes is one of computer vision tasks, and CNNs are the main deep learning method. At present, an autoencoder is rarely used in agricultural scenes.
## Applications
Deep learning has been widely used in agriculture, such as disease diagnosis, crop recognition and classification, leaf counting, etc. To a great extent, these methods have solved the agricultural tasks. In fact, they cannot be applied directly to dense scenes. In most agricultural scenes, due to the existence of dense objects, these methods of DL cannot work well. Therefore, this section introduces the application of DL in dense scenes in agriculture.shows the applications of deep learning networks in dense images in different agricultural tasks. We introduce the models used in various applications of agricultural dense scenes and the results. Among them, computing time indicates whether it is real-time. We found that in the dense agricultural scenes, the vast majority of datasets are created by the author. So, in, we describe the datasets used by the methods in.
## Recognition and classification
Object recognition is a process in which a particular object or a type of object is distinguished from other objects or other types of objects. Classification is the further processing of data based on recognition, that is, to recognize multiple classes. Recognition of agricultural dense scenes is mainly used in the recognition of pests and diseases and the recognition of objects as the basis of other tasks, such as fruit picking and object tracking. Classification tasks in dense scenes mainly include crop and weed classification and fine-grained classification in agriculture. In general, the recognition ability of the model increases as the depth deepens. However, it has been shown that simply increasing the depth can cause network degradation. Especially in dense scenes, it still faces the problem of complex background. Therefore, it is necessary to find a DL method to increase the depth of the model while restraining the degradation for dense scenes analysis in agriculture. Images in recognition tasks are quantity-dense scenes, while the images in classification tasks tend to be inner-dense scenes. Therefore, the problem of classification in dense scenes in agriculture is more complex. Usually, multiple DL methods are combined to deal with the overlapping and classification of different classes, i.e., classification-based network and semantic segmentation based network.
The main task of recognition in agriculture is to recognize diseases and insect pests. Timely and accurate diagnosis of plant diseases is one of the main objectives of precision agriculture. Cheng et al.used the 50-layer and 101-layer fine-tuning deep residual learning pre-trained on ImageNet to identify pests. Because of the end-to-end training on the pest dataset with a complex background, their system has strong robustness. The recognition accuracy of 98.67% for 10 classes of crop pest images with complex farmland background was achieved. Besides, recognition is the first step in picking in the orchard. The general system can only recognize the fruit, but in the field environment, it is also very important to recognize obstacles such as leaves and branches. Liu et al.designed a fruit-picking robot working in a natural environment. Yolov3 and two Mark R-CNNs with backbone networks ResNet50 and ResNet152 were used. ResNet50 and ResNet152 contain 16 and 50 residual modules, respectively. The experimental results showed that the Mark R-CNN152 model has the highest comprehensive recognition accuracy of 85.12% and can meet the real-time requirements.
To solve a series of problems such as random growth posture of trees and complex occlusion of branches and leaves to fruits, they divided the types of recognition into Normal Branch, Occlusion, Leaf Occlusion, Slight Occlusion, Overlapping and Main Branch according to the occlusion rate. Then they recognize, classify and plan the path according to the occlusion. Accurate individual recognition is an important step towards automated dense object tracking. In the hive, there are hundreds of dense, often occluded and constantly moving individuals. In order to identify all honeybee individuals in the dense natural environment, and judge the orientation of the individuals, Bozek et al.used the method of combining U-Net and a recurrent component to encode the additional information of the segmented object. In their work, they reduced the size of U-Net to reduce overfitting. The proposed method can effectively identify individuals in iteration.
Crop and weeds are similar in shape, size, and color, so the classification of crop and weeds faces the challenges: the overlap between crop and weeds. For general classification tasks, most of the methods use the fully connected layer of CNNs architecture for final classification. However, in dense scenes, the use of a fully connected layer directly for classification is not effective. Because the pooling layer compresses the original two-dimensional matrix into one-dimensional, thus, spatial information is lost. Therefore, classification tasks are usually divided into two steps in dense scenes. Firstly, object and background are segmented by semantic segmentation methods, and then foreground objects are classified. A new method combining robust pixel-wise with a coarse-to-fine classifier based on CNN was proposed in. Specifically, the segmentation network was based on a modified version of U-Net architecture. The segmented network was trained by datasets from different contexts, plant types, fields and environmental conditions, which was used to generate the binary mask. Then the vegetation clumps in the binary mask were extracted and sent to the fine-tuned CNN classifier based on VGG-16 for crop/weed classification. The proposed method achieved good classification results on challenging data, i.e., serious occlusion of weeds and crops. It is the overlapping of weeds and crops that affects the detection results. Besides traditional fruit or crop classification, some tasks can also be transformed into classification tasks in analyzing dense scenes in agriculture, such as quality assessment, fine-grained classification and so on. In the study of, the evaluation of paddy field quality was transformed into the classification of paddy field density, including sparse density and normal density. Different from traditional disease classification, there are larger intra-class similarities and smaller inter-class variance. So fine-grained classification in dense scenes is more of a challenge. At present, the research in this field is too little and not mature.
## Detection
Detection is based on classification, which requires the network to output not only the category of the object but also the location parameters of the object. Accurate detection is usually the basis of other tasks. In agricultural applications, weed control, fruit picking and so on need to use the detection results as input. In the analysis of agricultural dense scenes, the main challenges of detection are: (1) for the case of quantity-dense scenes, serious occlusion between objects, as well as small objects; (2) for the inner-dense scenes, due to the narrow line spacing, overlapping similar objects often appear. In this context, DL used in detection includes the most advanced object detection networks and semantic image segmentation models.
Weeds detection and control is an important application of precision agriculture, and also an important application of dense scenes in agriculture. Weeds generally have the characteristics of wide distribution, high density and low nutritional value, which do not make use of the growth and development of crops. So, it is of great significance to accurately detect weeds in farmland for controlling weeds, spraying pesticides, machine cutting and so on. Dyrmann et al.used DetectNet-based structure consisting of a fully convolution neural network to detect weeds. They used a model trained on the ImageNet to initialize the weight, and by evaluating the error of bounding boxes and predicting the coverage map, the FCN is trained to determine the position of weeds in the image. The network could detect a single weed in a grain field under the condition of serious leaf occlusion. The results showed that the accuracy was 86.6%, although most of the weeds overlapped with wheat plants. The average IOU was 0.64, according to the IOU standard, the algorithm encountered problems in the detection of very small weeds, grasses and weeds exposed to serious overlap, which led to the performance degradation of the system. Due to the selection of the area with the largest occlusion degree, they found a big challenge of automatic weed detection, that is, small target and serious occlusion, which must be solved for the precise control of wheat field weeds. They also pointed out that not all weeds were annotated in the training data used in the study, so if the training data is fully annotated, the accuracy of the test will be further increased. So, some studies have done weed detection on more challenging datasets. Yu et al.reported three deep convolution neural networks (DCNN) including VGGNet, GoogLeNet and DetectNet for detecting weeds in bermudagrass [Cynodon dactylon (L.) Pers.]. The detection results are very accurate, among them, DetectNet is the most successful because it exhibits high performance in multiple images with very high weed density (more than 100 weeds per image), with F1 scores > 0.99. The networks of all of them have all achieved good detection results under severe blocking conditions, and cannot play a role in a few high-density weed detection images.
The precise location of fruits is also an important application in agricultural dense scenes, detecting fruits helps to reduce one of the most labor-intensive tasks in orchards. Different from the challenge of weed detection task, which mainly depends on the deep object detection network to deal with inter-object occlusion, fruit detection in orchard often still faces the problem of occlusion of other objects in the background, small objects and different appearance changes. So, in a real outdoor farm environment, fruit detection is more challenging than weed detection, and the DL algorithms based on segmentation network is widely used and dense fruit image detection to separate the object from the background.
Experimental results show that the region-based object detection networks can achieve accurate object detection even in the complex background of light changes. In view of the fact that a single sensor mode cannot fully provide the information needed to detect the target fruit, the fruit detection system based on the multi-modal region has begun to be used in dense agricultural scenes. A high-performance fruit detection system was developed by, i.e., a novel multi-modal Faster R-CNN based on VGG-16. By using early and late fusion, combined multiple modes (RGB and NIR) images, they combined the classification decisions of the two modes and changed the structure of the VGGNet input layer. The end-to-end network can significantly reduce the training and prediction time, and further improvements have been made in the detection results. However, the detection results are relatively poor on the fruits with most areas blocked. Bargoti and Underwoodproposed a trained model called Tiled Faster R-CNN to implement fruit detection over orchard data containing between 100-1000 fruits per image. The experiment was carried out on the ZF network and VGG-16 net. By using the attention mechanism, the individual proposals are propagated through the fully connected layers and then become the two sibling layers again, and a more refined output of region classification and associated object boundaries were obtained. In addition, in order to perform more advanced tasks on the original image, they used smaller sliding windows to perform detection or "tiling" on the larger image to obtain tiled Fast R-CNN. Recently, a new detection method has been provided by utilizing additional features and machine learning strategies. Dias et al.studied the removal of apple blossoms at a specific growth stage. In view of the fact that the detection of apple flowers is hampered by variable lighting and blocking of leaves, stems or other flowers, they further fine-tuned the Clarifai model. It had been significantly fine-tuned, and effectively combined the features extracted by CNN with color and morphological information, which made their methods have certain applicability in various occlusion level scenes. Even in datasets that are significantly different from training data, the optimal recall rate and accuracy rate are close to 80%. Different from the above two-stage model, Bresilla et al.proposed an end-to-end fast and accurate fruit detection model based on YOLO. By modifying the standard model, the network scale was expanded, some layers were deleted, and then two other blocks are added. The "splitter" block at the entrance divided the image into four separate images, and the "joiner" block at the end splices the four pieces together, taking into account speed and accuracy. However, their models cannot detect two objects of the same category. Accurate fruit detection is also the key to counting based on detection, because any undiscovered fruit will not be counted. Häni et al.compared detection methods using GMM and image segmentation, deep image segmentation network, U-Net and object-based detection network. Among them, their deep object detection network is based on the work of, but the backbone network is switched to the ResNet50. In order to make up for the poor performance of the network in detecting small objects, they used a feature pyramid network (FPN) with lateral connections proposed into extract more low-level features. The counting results also showed that the DL method using 50-layer ResNet is more accurate. Of course, UNET did quite well. Grimm et al.developed a learning semantic segmentation framework for grape detection, localization, and counting. The serial VGG-16 is used as the encoder and the combination of U-Net, SegNet, and FCN-8 is used as the decoder to effectively segment the object and background. Specifically, the decoder used upsampling and downsampling steps with factors of 2 and used concatenations of complete feature maps of the encoder part with feature maps of the decoder part having the same size. Then, the decoder ended with a final upsampling layer with a factor of 8. It can learn more complex models while having good detection performance, and reduce the memory and running time of image segmentation. In addition to the method based on semantic segmentation, case segmentation is also used in agricultural dense object detection. Gonzalez et al.proposed a new instance segmentation algorithm of blueberry quantization based on Mask R-CNN. ResNet101, ResNet50 and MobilNetV1 were selected as backbone networks and several experiments were carried out in each network. Different from the usual method, their algorithm was trained from the beginning on a challenging new dataset captured in the field. In order to detect small objects in dense images, they improved the standard feature extraction pyramid Sensors 2020, 20, 1520 20 of 33 by adding a second pyramid to access lower and higher-level features. Good detection results were obtained in dense images.
## Counting and yield estimation
Since Seguí et al.began to study the number of interest counted by CNN, the counting method based on DL has become popular. The automatic counting of agricultural products is very helpful for agricultural enterprises to optimize and simplify their harvest. Yield is the criterion for evaluating the production situation and production efficiency, estimation of yield is very important for agricultural producers, and it will be the basis for producers to forecast storage demand, profit and production capacity. In the dense scene of agriculture, counting and yield estimation are similar, because the yield estimation based on DL mainly uses the counting based method. The meaning of counting is to calculate the number of all objects in each frame, while the yield estimation needs to integrate these counts in the whole dataset, and also needs to eliminate duplicate counts. In other words, yield estimation is based on previously obtained data. In general, counting is to estimate part of the yield. It is not hard to imagine that counting and estimating the yield is challenging because it deals with a large number of clustered objects, and the use of artificial methods is time-consuming and laborious, which is not suitable for large-scale cultivation of crops. In dense scenes of the agricultural sector, there are two difficulties: (1) Objects often grow in arbitrarily large clusters.Objects are often occluded by branches, leaves or other objects. Methods of object counting tend to fall within four categories: segmentation-based, detection-based, regression-based, and finally density functions-based.
Segmentation-based methods are used most frequently. French et al.explored a method to segment the scene and counting fish with CNN, which exploited the N 4 -Fieldsas a foreground segmentation method. The N 4 -Fields generated a real-valued output image giving the probability that each pixel is a foreground pixel. After foreground segmentation, the N 4 -Fields algorithm was used to predict edge maps again, and then small convolution was used to segment and count objects in surveillance video. Chen et al.used a blob detector based on FCN to extract candidate regions, and then used a counting algorithm based on second convolution network to estimate the number of each region. Finally, a linear regression model was used to get the final fruit counting. Since FCN does not use any fully connected layer, the network can perform the partitioning task well, and by initializing weights as the VGG network weights, the network can be trained quickly. Their method can realize good counting results even in very dense conditions. It is very difficult to use only one network to deal with the counting problem of large fruit images which differ greatly in appearance and quantity. Further, Liu et al.combined deep segmentation, frame-to-frame tracking and 3D location technology, proposed a new counting method: using FCN model to segment targets and non-targets, using Hungarian algorithm to track inter-frame results, and finally using 3D location to eliminate the trajectories counted repeatedly and correct the count by rejecting false positives.
YOLO network is the most effective counting method based on the detection. In the research of Zhong et al., they designed and implemented a vision-based flying insect counting system. To detect flying insects, the phenomenon of camera out of focus and impurity interference is easy to occur in the process of image acquisition, so a YOLO network with strong anti-interference ability and reliability is selected. For the convenience of counting, they supposed flying insects as a class and used YOLO for detection and rough counting, which was pre-trained on ImageNet and added a convolutional layer to transform classification model into detection model. Then, the boundary box of detection results of the fine-tuned model is provided to SVM for identification to alleviate the problem of insufficient samples.
Through density regression, we can also train a deep model which can solve the counting task. To solve the problem of crowding and occlusion between targets, Arteta et al.proposed a method to define the target density map by foreground segmentation. Compared with the general density regression, this method is easier to learn, and its core is to use the robustness of fcn8s to noise annotation. Specifically, three tasks, foreground background segmentation, explicit local uncertainty estimation, and density estimation, are solved through a single deep structure joint training. The algorithm performs well in very crowded conditions, but there are still many problems to be solved.
Since all previous fruit counting methods based on DL rely on counting the detected instances. In fact, under the conditions of the large variance of illumination, leaf occlusion or some degree of overlap amongst fruits, in it is difficult to count directly even using DL methods. Rahnemoonfarfirst proposed fruit counting based on deep simulation learning. In their method, objects were counted without detecting objects. They modified the Inception-ResNet-A layer of Inception-ResNet to capture the characteristics of multiple scales and generated synthetic images for network training, and good counting results were achieved. The experimental results showed that the average test accuracy of their network in the real image and the synthetic image is 91% and 93%, respectively. However, it is still a long way from promoting this method. In their subsequent further research, on this basis, they reduced the image size and the number of expansion filters, realizing a real-time estimation of tomato yield.
In addition to the above counting methods, there is also region or area-based method for yield estimation. Based on the fact that the panicle number is directly related to the final yield, yield can be predicted by detecting the leaves and panicles of rice.first used UAV remote sensing data to estimate crop yields using CNNs; they proposed a DL network consisting of two independent branches CNN structures for processing RGB and multispectral images respectively. The first branch consists of five convolutional layers and three max-pooling layers, similar to the reduced version of AlexNet, used to learn spatial features in RGB images. The second branch uses the structure of a pool layer and three convolutional layers to extract features from multispectral images. Due to the difficulty in obtaining high-resolution remote sensing images and the corresponding crop yield distribution, the research on the application of dl to the estimation of production at home and abroad is very limited. It is worth mentioning that Hasan et al.established a wheat spikes dataset, which consists of hundreds of high-quality wheat images with 25,000 annotated peaks, in total. Four region-based convolutional neural networks (R-CNNs) were designed to accurately detect and count spike regions in each map in the study and to better classify and improve the overall detection performance, they made some modifications to the implementation method and optimized the hyper-parameters.
After the objects are detected and counted, to solve the problem that fruit parts are occluded or invisible in the cluttered environment, measures should be taken to improve the robustness of the counting method. Clustering and merging single prediction through cross-frame tracking is a routine operation to avoid repeated counting. Some work combined GPS positioning information to accurately track objects. It was later discovered that the accuracy of counting was greatly improved by using multiple perspectives. Das et al.utilized multi-sensor platforms and optical lines, and Hung et al.ensured that the images do not overlap at all by selecting the sampling frequency, thus avoiding tracking. Besides, Structure from Motion (SfM)technology can be used to perform 3D reconstruction, and cluster views can be tracked by estimating camera motion. In short, counting is still a thorny problem in dense agricultural scenes. Generally, the accurate counting results cannot be obtained by using the DL method alone, all of these often need to be combined with other methods to further improve the counting performance.
# Discussion
In the previous section, we reviewed the DL methods and their applications in dense agricultural scenes. In this section, we first present an analysis of the applications of the surveyed DL methods. Then, the limitations and future research directions of DL applications that we think are meaningful in this field are emphasized.
## Analysis of the surveyed dl applications
(1) Data Preprocessing One of the major features of DL is the need for a large amount of data. One of the most prominent factors in promoting DL is the emergence of large, high-quality, publicly accessible annotated datasets. The quality and quantity of training data often determine the depth of the network and the effects of the model. To get better training data, many studies have adopted the method of image generation and data augmentation.shows the methods of data preprocessing. At this stage, most of the datasets studied by DL come from some large technology companies, it is difficult for general researchers to establish high-quality public datasets, but the emergence of data synthesis technology breaks the shackles. Synthesis data is data generated by a computer to simulate real data. There is no cost to generate new data by using synthetic images of computergenerated models to enhance the dataset of images, and simply training synthetic datasets can be successfully applied to real data. In addition, it has been proved that the network trained on synthetic data can be generalized into two separate real image datasets, each of which has a different distribution. Generally, the DL model is trained on synthetic data, then validated on actual data, and finally applied to real data. The experimental results showed that the effect of synthetic data training is fine, and only when there is a big deviation between synthetic data and real data, the accuracy will be reduced. Experiments inwere implemented using lpfg, a plant simulator based on L-system included in the Virtual Laboratory plant modeling environment. Virtual environments allow free placement of any contained element in a scene and generate its semantic annotations without additional work. In addition to generating synthetic data in a virtual environment, there is a simpler and more widely used synthetic data generation method. Rahnemoonfarfilled the entire blank image with green and brown circles to simulate the background and tomato plants and used Gaussian filtering for blurring; then, random circles were drawn at any position to simulate tomatoes with variable sizes. The images overlap with the change of size, proportion, and illumination to fuse complex real scenes.
Data augmentation is commonly used in the use of DL methods to artificially expand training data. There are two purposes of data augmentation: one is that the sample is scarce, which is not enough to support the training of DL networks, so it needs data augmentation to increase the number of the sample. The other one is to prevent over-fitting, over-fitting means that the DL model performs too well in training datasets, resulting in poor performance in testing datasets. Data augmentation usually supplements the original data in quantity and diversity, and we can divide data At this stage, most of the datasets studied by DL come from some large technology companies, it is difficult for general researchers to establish high-quality public datasets, but the emergence of data synthesis technology breaks the shackles. Synthesis data is data generated by a computer to simulate real data. There is no cost to generate new data by using synthetic images of computer-generated models to enhance the dataset of images, and simply training synthetic datasets can be successfully applied to real data. In addition, it has been proved that the network trained on synthetic data can be generalized into two separate real image datasets, each of which has a different distribution. Generally, the DL model is trained on synthetic data, then validated on actual data, and finally applied to real data. The experimental results showed that the effect of synthetic data training is fine, and only when there is a big deviation between synthetic data and real data, the accuracy will be reduced. Experiments inwere implemented using lpfg, a plant simulator based on L-system included in the Virtual Laboratory plant modeling environment. Virtual environments allow free placement of any contained element in a scene and generate its semantic annotations without additional work. In addition to generating synthetic data in a virtual environment, there is a simpler and more widely used synthetic data generation method. Rahnemoonfarfilled the entire blank image with green and brown circles to simulate the background and tomato plants and used Gaussian filtering for blurring; then, random circles were drawn at any position to simulate tomatoes with variable sizes. The images overlap with the change of size, proportion, and illumination to fuse complex real scenes.
Data augmentation is commonly used in the use of DL methods to artificially expand training data. There are two purposes of data augmentation: one is that the sample is scarce, which is not enough to support the training of DL networks, so it needs data augmentation to increase the number of the sample. The other one is to prevent over-fitting, over-fitting means that the DL model performs too well in training datasets, resulting in poor performance in testing datasets. Data augmentation usually supplements the original data in quantity and diversity, and we can divide data augmentation methods into traditional methods and DL methods. Traditional augmentation technology can be divided into geometric transformation and intensity transformation. The geometric transformation includes size adjustment, sample tailoring, rotation, inversion and so on. The intensity transformation includes the transformation of contrast, brightness, color, and noise. Besides, mixup and PCA enhancement technologies are also traditional data augmentation means. Recently, methods based on DL for data augmentation have become popular. The CycleGANis essentially two mirror-symmetrical GAN, it can learn the characteristics of a class of data and generate similar data. The research showed that compared with traditional image enhancement methods, the CycleGAN method greatly enriched the diversity of training data sets.
Although data augmentation technology and synthetic image technology expand some data sets, greatly alleviating the problem of lack of data sets, in fact, at least hundreds of images are needed, which depends on the complexity of the problem studied.
(2) Surveyed models and model optimization Although DL has made substantial progress in many computer vision problems, up to now, the research on dense scenes in agriculture is not mature. However, the review shows that the DL methods are better than the traditional machine learning methods based on the performance index adopted by the authors. Our analysis shows that in most of the related work, the main DL method to deal with dense scenes in agriculture is CNN, among which VGGNet and ResNet are the most commonly used models. Since the closer the network is to the bottom, the more effective the extracted features are, it can be said that the network architecture of VGGNet is superior to other most advanced networks. So, in order to maintain this characteristic, many convolutional neural network models use VGGNet as the basic model for image detection and other tasks. ResNet can not only prevent the degradation of deep neural network, but also improve the training quality of the model, so it is widely used in dense scenes analysis. As reviewed above, data annotation in dense images is very difficult. As a result, due to the character that it needs less data and can use available data accurately, semantic segmentation network including FCN and U-Net is also widely used to solve recognition, detection and counting tasks by segmenting foreground objects.
CNNs are the most important deep learning methods in the dense agricultural scene. At the same time, transfer learning and autoencoder are also used in the dense agricultural scene. The biggest use of transfer learning is to alleviate the lack of data and pre-training the networks. Supervised learning needs a lot of annotation data, which is also one of difficulties of DL in dense scenes. On the one hand, many machine learning methods work only when training data and test data come from the same feature space and have the same distribution. If the distribution of data changes, most models need to be reconstructed. On the other hand, labeling data is a tedious and expensive task. Reducing the amount of data means reducing the overhead, but there is a risk of reducing network performance. Therefore, due to the low data requirements, transfer learning is more and more used in the applications of DL. Transfer learning allows different domains, tasks, and distributions of training and testing data while saving a lot of labeling work. It is also helpful in adapting a training model for one period of time or equipment so that it can be applied to another period of time and transferring knowledge to a new field. In computer vision tasks, it has become a standard operation to transfer the trained features (i.e., fine-tuning) from the trained CNN to the new target tasks. In general, the pre-training model and initial weights are obtained by pre-training on the ImageNet dataset or Pascal VOC dataset. That is to fine-tune the network. Fine-tuning usually requires only a small number of parameters to optimize, and the whole network has not been trained from the beginning. This greatly reduces data requirements for specific tasks. Initialization of convolutional neural networks using models trained on different datasets will result in faster convergence of the networks. After training on thousands or even more annotated images, the DL networks can be better trained in dense agricultural scenes. An autoencoder can also be used to generate data, but it has not yet been used in dense agriculture scenes.
Standard DL models cannot perform dense scenes analysis well, so it is necessary to enhance the basic network architecture. The review shows that there are many methods to optimize the structure of DL network and make it more suitable for dense scenes analysis. One is to integrate different CNNs networks and combine the advantages of each network to deal with occlusion, complex background, and other issues. For example, Grimm et al.designed a segmentation network that combines U-net, SegNet and FCN-8 to segment objects more effectively. The other is to modify the advanced networks, such as changing the size of the convolutional filter, adding or deleting layers of network, updating or adjusting model parameters, etc.. At the same time, there is another network improvement method that has a good performance in the dense agricultural scenes analysis, i.e., CNNs models combined with other classical machine learning methods. Bozek et al.combined U-Net and recurrent components. Zhong et al.used YOLO to count roughly, and then used SVM to classify and recognize accurately. There is also work to divide the objects into clusters, and then detect or count each cluster.
There are many commonly used evaluation indexes for DL. Our review shows that accuracy, recall, and IOU are the most frequently used. Among them, the accuracy rate is best understood, which is an index presenting the proportion of correctly classified samples to the total number of samples. Generally, the accuracy rate is used to evaluate the global accuracy of the model, but it cannot fully evaluate the performance of a model. A similar indicator is the recall, which is the proportion of all positive samples in the test set that are correctly identified as positive samples. IOU is an object detection task evaluation index, which is the percentage of overlapping area between detected and ground truth boundary box in the joint area.
## (3) hyperparameters and optimization technology
Hyperparameters are user-defined and tailored to a specific application task. Hyperparameters have a great influence on DL networks, and the performance of networks is likely to depend on subtle differences of hyperparameters. The setting of hyperparameters is an important work in DL, learning rate, batch size, momentum and weight decay are the most common hyperparameters. The deployment of this work often depends on experience. From the relative literature review, the learning rate is discussed more in the hyperparameter optimization technology, and the setting and optimization of other hyperparameters are less involved. Specifically, there are four ways to set the learning rate: (1) fixed learning rate; (2) linear decay policy, when the number of iterations reaches the upper limit, the training stops; (3) exponential decay; (4) step strategy, we set an iteration number in advance, when the number is reached, the learning rate will be reduced to a certain value. Gradient descent is a common optimization technique in machine learning algorithm. Batch gradient descent and stochastic gradient descent (SGD) are used most in the application of dense agricultural scenes. In addition, Ada-Delta and mini batch gradient descent are also applied. Among them, mini batch gradient descent adds extra hyperparameter to the learning algorithm, the batch size. The batch size and weight decay remain unchanged during training. In the literature we investigated, the momentum is generally the same, and the momentum is generally set to 0.9. In order to make the deep learning model provide the best results, it is necessary and not easy to find the best value of these hyperparameters. Therefore, better hyperparameters optimization technologies need to be proposed.
(4) Solutions to challenges As mentioned in this review, occlusion is the biggest challenge in dense scenes. There are two main methods to alleviate the problem of occlusion. One is to label the object into different categories according to the occlusion situation, and then recognize and classify it. This method avoids the huge error caused by the same feature extraction for all objects. The other is to train DL networks on thousands of large annotated data sets with different sizes, illumination, and occlusion levels, which is the main solution. However it also means collecting large annotated datasets. In the experiment of, it took more than three years to collect Penguin datasets and more than 35000 volunteers to label the images.
Small object is also one of the key challenges in dense scenes. Such a challenge is relieved by optimizing the network. Because the pooling layer in CNNs reduces the feature mapping by two times every time, it does a poor job in detecting small objects. Because the FPN with lateral connections makes up for the difference of spatial resolution in the higher level, it is very suitable for small target detection. In addition, for networks with anchor scale or mini-mask shape, the settings of these two parameters can be changed to improve the detection effect of small objects.
For the case of high background noise, one solution is to select a high-performance network; the other is to add a variety of scenes with high background noise in the test dataset. Some CNNs used for saliency detection are fine-tunedand others choose networks with strong anti-interference ability and reliability, such as YOLO, R-CNN, etc. Although DL has made great progress in dense image processing, there is still room for further improvement in serious occlusion, real-time detection and other aspects. Only relying on the network often cannot perform well in the agricultural complex real scene. It is also necessary to train and test representative visual changes and various high noise datasets to improve the robustness of the network.
Besides, in dense agricultural scenes, for background processing with high similarity to the object, using semantic segmentation to extract the region of interest can achieve good results.
## Limitations to dl applications
(1) Influential Architectures In general, by considering partially occluded objects in training data or detecting partially occluded objects through design models can improve performance. However, by selecting the densest area, it is possible to reduce the performance of the model. It has been proved that the DL methods cannot be extended well when there are clutter and occlusion in the image. The performance of the network will decrease with the increase of the complexity of the dataset. Moreover, in the dense scenes of agriculture, there are objects that are hard to see even with human eyes. Therefore, the DL algorithms cannot solve the problem of serious overlap up to now.
## (2) datasets
In addition to the fact that DL algorithms are still difficult to deal with serious occlusion problems, the collection, and labeling of dense objects datasets is a major problem. The costs of generating new data are high, it takes a lot of time and effort to acquire and annotate datasets. We are familiar with common datasets for DL architecture, but in the field of agricultural applications, there are few public datasets available. In the complex background of dense agricultural scenes, the construction of datasets is even more difficult. Besides, the datasets used by one research institute are often not suitable for other studies. From, we can see that all the experiments except one study that used other people's datasets were self-generated datasets. At present, there are no public datasets for dense agriculture scenes, and few studies can download data directly from the Internet, and most require researchers to build their datasets. For example, to collect images of pests and diseases of tomato plants according to the time, the researchers collected 9230 pictures of wheat crops with different wheat diseases under various conditions.
## Future work
(1) DL optimization Facing the challenges of dense scenes analysis in agriculture, given the current solution to the use of auxiliary means to solve occlusion, scale inconsistency, and other issues, the effect is unlikely to have a qualitative leap, the research focus should also turn to the DL algorithm itself. First, we need to add more layers to improve the performance of the DL network. Plain networks do not perform well in dense scenes, so it is necessary to build a deeper DL structure to deal with more complex problems. However, blindly pursuing the layers of the network may lead to too many model parameters, increased computational complexity, and low training efficiency. Therefore, on the premise of not reducing the training efficiency of the network, increasing the number of network layers will be an important research direction to improve the ability of DL in dense scenes analysis. Therefore, it is a trend to explore lightweight CNNs, which can obtain higher recognition accuracy with less trainable parameters, such as MobileNet. It is possible to propose some new network structures to deal with dense objects. The new compression layer and optimization technology should be considered in the new DL architecture. Next, it is recommended that an in-depth study of the mechanism of DL. In the existing DL applications, the design and optimization of the algorithm mainly focus on the accuracy of the algorithm, rather than studying the black box. Needless to say, DL has a powerful ability, but the process is too abstract. We do not know how it works at present. As a result, Geoffrey Hinton, the father of neural network, said that DL entered a bottleneck period. Therefore, further research on DL theory needs more. Multi-feature fusion is also an effective way to improve DL performance. One is the fusion of multiple features, for example, the combination of spatial features and semantic features can maintain the positioning accuracy and more semantic information. Second, multi-layer features fusion can effectively use spatial information by giving different weights to different feature layers. The third is to combine CNN features with artificial features. In fact, this has been applied in agricultural intensive scenes, but there are more feature combinations, which may achieve the effect of improving network performance. In addition, it is suggested that CNNs combine with other DL algorithms, such as RNN, LSTM. Each DL method has its own application areas. Other DL methods may not be as powerful as CNNs in scenes processing, but they can use their advantages to complement CNNs. Now there are examples of CNNs and RNNs combinationand CNNs and LSTM combination. The next step is to improve them to effectively apply them to dense scenes in agriculture. Besides, how to ensure the accuracy of deep learning in agricultural tasks is also a problem to be considered. Although there may be errors inevitably, it is worth studying how to reduce the errors by using the real values for calibration.
(2) To meet the challenges of big data processing As mentioned in the review, there is no public dense image data set for agriculture. To maximize the potential of DL, it is very important to create a well-annotated large-scale intensive agricultural image database. It is also helpful for researchers to use the same data set for training, verification, and testing, and algorithm comparison. However, it is difficult due to that labeling data is costly and laborious.
Several methods can be used to alleviate the difficulty of datasets generation. One is to generate virtual training data by simulation. There are two issues to consider when applying data generation in an agricultural scene. One is that the generated image is highly similar to the real image, ensuring that it can be tested on the real data set; the other is that the generated image is rich in type, covering all real scenes. Its advantage is that it can express multi-level details and has high modeling efficiency, i.e., it can generate a large amount of data in a very simple way. However, it ignores the complex background in reality, so the difficulty now is to extend the model of synthetic data training to actual data. Therefore, it is a promising method to use the computer to generate a large number of available synthetic data and use these weakly annotated training data to promote the solution of the visual task. Another one is to transform supervised DL into unsupervised or semi-supervised DL. At present, in the field of DL, the use of annotated data for supervised learning is in a leading position. However, with the increase of dense scenes, it is unrealistic to mark all objects. Therefore, there is a lot of research space for semi-supervised and unsupervised DL methods using only part of annotated data. In addition, more and more attention will be paid to the research of automatic annotation. On the other hand, low resolution is another obstacle to the application of DL in dense agricultural scenes, so it is necessary to make a new attempt on low-resolution data based on the DL model.
(3) Apply DL to agriculture through transfer learning DL is applied to agriculture after achieving excellent performance in other fields. So the application level of DL in other dense scenes is far higher than that in agriculture. For example, in the crowd analysis, many improved DL algorithms or DL combined with traditional machine learning methods are proposed. A lot of open datasets have been generated. Based on transfer learning, transferring knowledge from one application field to another is a learning method that is conducive to the application of DL in agriculture. In the long run, this is a very promising application.
(4) Robot system
The research on robots for weeding, feeding, pruning, picking, and harvesting is a hotspot in agricultural application, which greatly improves labor efficiency, saves production costs, and promotes the development of agricultural machinery. Agricultural robots have developed into one of the most promising industries in precision agriculture. Developing an accurate, real-time and reliable computer system for pest identification or yield estimate is an important part of agricultural robot system. It requires complex algorithms to overcome realistic challenges, such as lighting, shape, occlusion and natural changes in viewpoints. The positioning algorithms and modules corresponding to the robot system are also necessary. Combining robot path planning, mechanical design, and DL algorithm to solve practical needs is also the next research focus. As a visual task, tracking is more and more needed in the dense scene analysis of agriculture. After the object in the dense scene is detected, the research of quantitative trajectory reconstruction algorithms will be an important direction in the future. At the same time, it is necessary to combine DL algorithms with modern technologies such as meteorological information and geographic information to form a comprehensive service platform. By considering the portability of mobile platforms such as Android or ios, it can be applied in smart mobile terminals to provide higher practical and promotional value.
# Conclusions
As discussed in this review, the greatest challenge for dense scenes analysis is that objects are too dense and small. This paper performed a review of DL in dense agricultural scenes, mostly from an application perspective. In this review, we review the recent research on DL for application in dense agricultural scenes. Based on these surveys, readers can understand which dense scenes exist in current agricultural applications, and which DL algorithms are used in these dense scenes, as well as the application of DL in general agricultural tasks. By comparing the DL methods with other methods in the survey paper, it shows that DL provides better performance in dense scenes and is superior to other popular image processing techniques. In short, research on the application of DL in dense agricultural scenes is a very meaningful and urgent matter. However, what we also need to know is that research on dense agricultural scenes is still in its infancy and many obstacles need to be overcome, such as datasets. The future development of DL technology still has different opportunities and challenges and has a high development prospect. |
Non-disclosure of tuberculosis diagnosis by patients to their household members in south western Uganda
BackgroundTuberculosis (TB) non-disclosure by adult patients to all household members is a setback to TB control efforts. It reduces the likelihood that household contacts will seek early TB screening, initiation on preventive or curative treatment, but also hinders the implementation of infection controls and home-based directly observed treatment. Therefore, the purpose of this study was to determine the level of TB non-disclosure, its predictors and the effects of disclosure among adult TB patients in Uganda.MethodsWe conducted a cross-sectional study at a large regional referral hospital in Mbarara, southwestern Uganda. Questionnaires were administered to collect patients' sociodemographic and their TB disclosure data. Non-disclosure was considered if a patient did not reveal their TB diagnosis to all household members within 2 weeks post-treatment initiation. Univariate and multivariate logistic regression models were fitted for predictors of non-disclosure.ResultsWe enrolled 62 patients, 74% males, mean age of 32 years, and median of five people per household. Non-disclosure rate was 30.6%. Post-disclosure experiences were positive in 98.3% of patients, while negative experiences suggestive of severe stigma occurred in 12.3% of patients. Being female (OR 6.5, 95% CI: 1.4-29.3) and belonging to Muslim faith (OR 12.4, 95% CI: 1.42-109.1) were significantly associated with TB non-disclosure to household members.
disclosure at household level while minimizing negative effects of stigma should be developed and prioritized.
# Background
Tuberculosis (TB) remains one of the ten causes of death worldwide and the leading cause from a single infectious agent, ranking above HIV/AIDS. The World Health Organization (WHO) estimated that 10 million people fell ill with TB in 2018 and that treatment success rate among new cases remained at only 85% by 2017 globally. Uganda is among the top 30 TB-HIV high burden countries, with an overall TB incidence of 200 per 100,000 people and treatment success rate of 69% by 2018.
In addition to the HIV epidemic, uncontrolled transmission is fueling the global tuberculosis epidemic [bib_ref] Modeling the joint epidemics of TB and HIV in a South African..., Bacaë [/bib_ref]. All indoor environments especially homes and congregate settings are potential sites of transmission since air dilution is limited and occupants are crowded [bib_ref] Rates of tuberculosis transmission to children and adolescents in a community with..., Middelkoop [/bib_ref] [bib_ref] Tuberculosis in a South African prison-a transmission modelling analysis, Johnstone-Robertson [/bib_ref]. Adherence to effective treatment not only ensures a faster reduction in risk of TB transmission from adult patients to their household members [bib_ref] Aerial dissemination of pulmonary tuberculosis. A two-year study of contagion in a..., Riley [/bib_ref] , but also limits the emergence of drug resistance. WHO recommends both social and psychological support as supplements to community-or home-based directly observed treatment. This ensures several successful outcomes namely: completion of treatment, prevention of relapse, death from active TB or its late effects, reduction in transmission to others, and prevention of the development and transmission of drug resistance [bib_ref] Serum drug concentrations predictive of pulmonary tuberculosis outcomes, Pasipanodya [/bib_ref] [bib_ref] Concentration-dependent Mycobacterium tuberculosis killing and prevention of resistance by rifampin, Gumbo [/bib_ref]. Household members are well positioned to provide such needed social and psychological support so as to ensure successful treatment outcomes and infection control, but only if adequately engaged in care of their TB patient, a step that is dependent on patients' disclosure of their TB diagnosis.
Disclosure of TB diagnosis to household members could influence patient adherence [bib_ref] Barriers and facilitators of adherence to TB treatment in patients on concomitant..., Gebremariam [/bib_ref] , but may also promote early initiation of appropriate infection control measures, and access to diagnostic, curative and preventive TB services by their household contacts. However, TB disclosure may be hindered due to TB-related stigma, the association between TB and Human immunodeficiency virus (HIV) infection, perceived incurability of TB, and myths about TB etiology [bib_ref] Assessing the Consequences of Stigma for Tuberculosis Patients in Urban Zambia, Cremers [/bib_ref]. Negative consequences of disclosure like patient isolation, neglect, withdrawal of social support, and divorce have been previously reported [bib_ref] Reemergence And Amplification Of Tuberculosis In The Canadian Arctic, Lee [/bib_ref].
There is a paucity of information on the frequency of disclosure of TB disease status at household level and consequences of disclosure among patients within TB-HIV high-burden countries. This information is useful in designing contextualized interventions to promote disclosure, as part of the efforts to achieve the 2035 targets of the End-TB strategy. Overall, knowledge on TB disclosure dynamics at household level, could guide TB control interventions. Therefore, we conducted a study at a large regional referral hospital in south-western Uganda, aimed at measuring the level of TB non-disclosure to household members by adult patients, the associated factors and patients' experiences post-disclosure.
# Methods
## Design and setting
We conducted a cross sectional study among adult patients started on antituberculosis treatment for pulmonary TB at the TB clinic at Mbarara Regional Referral Hospital. The hospital is a large tertiary health facility in south-western Uganda with a catchment area of nearly 2 million people. Data were collected at the clinic between May and July 2017. In agreement with the National TB Control Program guidelines, each patient is required to identify someone close to them prior to initiation of TB treatment, to serve as a treatment supporter where possible. Trained nurses then counsel patients on TB disease, treatment adherence and infection control before administering treatment. Patients return for drug refills every 2 weeks during the first 2 months of treatment (intensive phase), and thereafter, monthly for the next 4 months (continuation phase).
## Study population
Patients were eligible for enrollment if they were 18 years or older, were either new or retreatment cases with bacteriologically confirmed or unconfirmed pulmonary TB, completed at least 2 weeks of TB treatment, living with at least one household member, and provided written consent. The unit was considered to be a household if there was more than one person living in the same dwelling; sharing meals and living accommodation.
Patients were excluded if they were too ill, or had psychiatric disorders deterring them from responding to interview questions. The informed consent process was administered by trained research assistant study nurses who were not participating in patient care within the TB clinic.
## Study procedures
We administered a structured questionnaire with both closed and open questions in English or translated into the local language of Runyankole based on the patient's preference (S1 Questionnaire and S2 Questionnaire). We collected data on socio-demographic characteristics, household, patients' behavioral factors such as alcohol intake, smoking, medical history, and patients' TB disclosure experiences. We focused at assessing disclosure that occurs early during treatment and we defined this as TB disclosure to household members within the first 2 weeks following treatment initiation. The choice of the first 2 weeks was premised on being the estimated time for a patient on appropriate TB treatment to become non-infectious [bib_ref] Transmission of tubercle bacilli: The effects of chemotherapy, Rouillon [/bib_ref] , and therefore, could serve as the best time to optimize adherence support but also infection control measures so as to minimize TB transmission at household level.
In addition, post-disclosure experiences were elicited by asking patients an open-ended question "Can you tell us what your experience was two weeks after disclosing your TB status to your household members?" Based on the response, the research assistant checked "Yes" or "No" on a list of responses in the listing or would fill in the space provided for "specifying other experiences" that might not have been listed. The open question allowed the patient to respond freely without bias.
The predefined positive post-disclosure experiences were: 1) encouragement from family members, 2) support in taking their medications, 3) support in feeding, 4) financial support to attend clinic days, 5) Other positive experiences, in which case, the researcher would specify the additional positive experiences reported by the patient that were outside the predefined list. The negative post-disclosure experiences of patients elicited by the researcher focused on demonstrating any effects on the patient that is due to potential TB-related stigma among household members. These included experiences of 1) criticism or blame, 2) isolation by household members, 3) withdrawal of support by household, 4) marital separation, and 5) other negative experiences, in which case, the researcher would specify the additional negative experiences reported by the patient that are outside the predefined list.
Patients were then asked to respond with a "Yes" or "No" on whether they perceived that the above positive or negative post-disclosure experiences had impacted on: their treatment intake and adherence to scheduled TB clinic visits, but also on some of their psychosocial aspects, that is; self-esteem, hope of recovery, work, household harmony, and marital relationships in the case of those who were married or co-habiting. Also, patients were asked on whether they would agree to the training of their treatment supporters by healthcare workers so as to support in patients' TB disclosure to household members.
The primary dependent variable for this study was "disclosure", which was derived from a question on whether they had "disclosed their current TB status to their household members", with three responses namely "yes, all = 1", "yes, some = 2" and "no = 3". A patient was considered to have complete disclosure if he or she responded with "yes, all", meaning that they had revealed their TB diagnosis to all their household members. A patient was considered to have partial disclosure if he or she responded with "yes, some", meaning that he or she had revealed their TB diagnosis to only a selected number of their household members and left out others.
For the purpose of the analysis of study outcomes, we generated a binary dependent variable coded 0 = complete disclosure, and 1 = partial or non-disclosure. The patients were considered to have disclosed only if they had complete disclosure, and everyone else was considered as "no disclosure'. The decision was made because both total non-disclosure and partial disclosure positions household members at risk of TB infection, given that TB is an infectious disease. The independent variables included all patients' characteristics of sociodemographics, household characteristics, patients' behavioral factors alcohol intake and smoking and medical characteristics.
## Sample size
A sample size of 62 patients was calculated using a formula for single population proportion with correction for finite population. In the calculation, sample size n = N � X / (X + N-1), where, X = Z α/2 2 � p � (1-p) / MOE 2 , and Z α/2 is the critical value of the Normal distribution at α/2 for α of 0.05, MOE is the margin of error considered at 5%, p is the proportion of non-disclosure which was taken as 50%, an arbitrary value given the lack of comparable data, and N is the population size, estimated at 70 new patients with TB undergoing treatment at the clinic for two weeks to two month period. To cater for attrition, an additional 5% was added to the sample size.
# Data analysis
Data from completed questionnaires were entered into a database designed using Epi Info™ software (V7.2, 1600 Clifton Road Atlanta, GA 30329-4027 USA) and analysis was performed with Stata software (v.13, College Station, Texas, USA). Descriptive statistics were generated for participants' characteristics. The proportion of patients reporting not to have disclosed their TB status (that is, partially or completely) to household members was calculated out of all study participants and presented as a percentage. Percentages of TB patients with positive or negative post-disclosure experiences, and those reporting different aspects of their lives impacted by these post-disclosure experiences at household level were calculated as proportions. The dependent variable was a binary variable of TB status non-disclosure to household members coded 0 = No and 1 = Yes. All patient factors were used as independent variables in this analysis. In Univariate analysis, based on both Chi-square test and Logistic regression, analysis comparing each independent variable with non-disclosure was performed. Unadjusted odds ratios with their corresponding 95% CI were reported. A significance level of 5% was used. All factors with pvalue <0.1 in univariate analysis and those with biological plausibility (e.g. age categories) were considered in the multivariate analysis which was performed to control potential confounding. Assumptions for use of multiple logistic regression, e.g. the absence of multicollinearity among the independent variables, were checked. A manual back-ward stepwise selection method was used in fitting the final multivariate analysis model. In this method, we excluded variables that lost their meaningful association with TB non-disclosure after controlling for the effect of other variables in the model. The goodness-of-fit test was performed on the final model to assess its quality. The factors in the final multivariate model were reported together with their adjusted odds ratios and 95% confidence intervals. A variable was considered significant in this analysis if it had a p<0.05.Percentages were calculated to describe the patients' perception on training treatment supporters so as to enhance their role in supporting patients' disclosure at household level.
## Human subjects
The study was approved by the Mbarara University Faculty of Medicine Research Committee, and Mbarara University of Science and Technology Research Ethics Committee. Study approval no: 16/02-17. Written informed consent was received from all study participants prior to enrollment in the study. We also used separate research assistants (nurses) who were not involved in the routine care of these TB patients to perform the interviews. The patients' treatment was a priority to research and they were only interviewed after they had received their routine care, so as to prevent coercion and therapeutic misconception. No incentives were given for patients' participation in the study. All patients' information was kept confidential.
# Results
We enrolled 62 patients, predominantly males (74%), mean age of 32 years, with half having at least secondary education. Majority (80%) of patients' households had atleast four people. About 95% of the patients had the current episode as their first TB episode, 82% had a treatment supporter, 49% were HIV infected, and two-thirds had been on TB treatment for at least 8 weeks at the time of interview. The details of the sociodemographic characteristics are shown in [fig_ref] Table 1: Participants' characteristics [/fig_ref].
## Tb non-disclosure to household members
Overall, of the 62 patients, 19 (30.6%) had either not disclosed to some or all the members of their household. Significantly higher TB non-disclosure rates were noted among patients who were; female (62.5%, p = 0.001), single (44.1%, p = 0.005), and moslems (71.4%, p = 0.013), as compared to their counterparts. No significant differences in non-disclosure rates were observed across patients' age, occupation, education and residence type, p>0.05. The detail of these results are shown in [fig_ref] Table 2: TB Non-disclosure rates across patients' sociodemographic characteristics [/fig_ref].
Overall, the median duration on TB treatment for patients that had not disclosed to household members by time of interview was 13.3 (IQR 2.0-13.2) weeks.
Of the five patients that had not disclosed to any one, four had been on treatment for more than 4 weeks at the time of interview.
## Motivation for tb disclosure
Of the 57 patients with either complete or partial disclosure, 50 responded to the open-ended question on what had motivated their TB disclosure to household members. The responses were, 1) The desire to receive support/care from household members (40%); 2) need to avoid transmitting the disease (30%); 3) Need for support to collect and remind them to take their medicines (16%); and then, 4) the need to inform their family what they were suffering from (14%).
## Factors associated with tb non-disclosure
In univariate analysis, the factors significantly associated with adult TB non-disclosure to household members are; female gender, being a Moslem, being single or unmarried and having dependents, as shown in [fig_ref] Table 3: Results of univariate analysis for factors associated with non-disclosure of adult TB... [/fig_ref].
However in multivariate analysis, the only factors with independent association with TB nondisclosure to household members were; female gender and being a Moslem, after controlling for [fig_ref] Table 4: Results of multivariate analysis for factors independently associated with non-disclosure of adult... [/fig_ref].
## Tb patients' positive post-disclosure experiences
Of the 57 patients with either complete or partial disclosure, 56 (98.3%) reported positive experiences post-disclosure. These were in form of support from household members in areas of: medication intake; encouragement; in feeding; and financial support to enable attendance of clinic days. The areas with the least reported support offered by household members were; infection control (2.4%) and counselling about TB as a curable disease (4.9%) [fig_ref] Fig 1: Positive post-disclosure experiences in adult TB patients, N = 57 [/fig_ref]. Majority of patients felt that these positive experiences post disclosure, had impacted greatly on their: treatment intake, attendance of clinic visits, self-esteem, hope of recovery, household harmony and work [fig_ref] Fig 2: Areas of TB patients that are impacted by early positive post-disclosure experiences,... [/fig_ref].
## Tb patients' negative post-disclosure experiences
Of the 57 patients with either complete or partial disclosure, 7 (12.3%) had experienced atleast one negative experience following disclosure. The most common negative post-disclosure experiences among them were: marital separation (n = 3), blame and negative criticism (n = 2) and isolation (n = 2) All these 7 patients also reported to have had at least one positive post-disclosure experience.
## Perceptions towards enhancing treatment supporter role in ensuring complete patient disclosure
Of the 62 study participants, 38 (61.3%) expressed agreement to the involvement of a treatment supporter to assist in patients' TB disclosure to household members through training and provision of suitable TB information aides (like charts) by healthcare workers.
# Discussion
Our study is the first, to the best of our knowledge, to explore the dynamics of TB disclosure at a household level in Uganda. Disclosure of TB status is a key strategy in promoting engagement of household members in TB control and efforts to achieve the targets of the End TB strategy. We report a high rate of non-disclosure to household members by adult TB patients (30.6%). The need to attract care was the main motivational factor for patients to disclose to household members, besides other factors like: the desire to avoid transmitting the disease to them and attracting support in picking or reminding them to take their medicines. TB disclosure to household members yielded positive benefits almost to all patients, especially with regard to treatment adherence and psychosocial support, but with little gains on counselling pertaining to infection control and hope of complete cure from TB. Negative experiences occurred to about one out of every eight TB patients following disclosure to household members. Being female or a Muslim was significantly associated with non-disclosure of TB diagnosis to household members. A potential intervention based on training and provision of suitable TB information aides to treatment supporters by healthcare workers, so as to enhance their confidence in supporting the TB patient in the process of disclosure to household members so as, was acceptable by two-thirds of TB patients. The observed high TB non-disclosure rate among adult patients to their household members, is of great public health concern, and can result in uncontrolled TB transmission at household level. Some of the patient factors and their household characteristics in our study seem to be in alignment with this concern, and indicative of a conducive environment for continued TB exposure and potential transmission. The point in case is the fact that; 1) majority (80%) of TB patients were staying with at least 4 household members, who would therefore qualify as TB contacts and at-risk of TB infection; 2) most of the patients (95%) were newly diagnosed with TB, and though not explored in this study, but could perhaps have had insufficient experience and knowledge on TB infection control measures so as to be able to minimize transmission to their household members; and 3) as part of the study eligibility criteria, all patients had been at least 2 weeks on treatment, and we noted a median time of 13.3 weeks since treatment initiation among patients reporting non-disclosure to household members. This suggests a prolonged period of potential uncontrolled TB transmission to household members. Such non-disclosure deprives the patient with TB of the care and support they would have received from their household members. In addition, it deprives their household members especially the children or HIV infected contacts, who are more likely to develop disease and to present with severe forms of TB once infected, of their right and early opportunity to self-protection and access to TB services. Such TB services include; contact screening with initiation on preventive treatment, prompt TB diagnosis, and early initiation on appropriate TB treatment. This is an important reason why the "ethical and human right approach" has been proposed to guide the pursuit of END-TB targets.
The finding in our study that disclosure was motivated by the patient's desire to attract care from household members further emphasize the culturally perceived role of family in Africa in caring for their patients. Such a perception though good, positions household members at risk of contracting TB as they care for their patients if appropriate TB infection control measures are not promoted and enhanced at this level. On the other-hand, it could be expected that anything that could deny a patient this care and support from household/family members could potentially impact on their psychosocial status, treatment adherence, and outcomes of treatment. Therefore, our findings further support the need for social and psychological support interventions within TB patient care. Our study showed that being a female adult TB patient was associated with 6.5 times higher odds of partial or total non-disclosure as compared to males. In as much as the reasons for this gender disparity were not explored in this study, one potential explanation would be on the underlying fear of TB-related stigma and discrimination that could result from disclosure of TB status, just as it has been reported in studies conducted in Zambia and Ghana, with great effect on women than men [bib_ref] Assessing the Consequences of Stigma for Tuberculosis Patients in Urban Zambia, Cremers [/bib_ref] [bib_ref] Individual, household and community level factors associated with keeping tuberculosis status secret..., Amo-Adjei [/bib_ref]. Other fears for disclosure among female TB patients could be on the risk of breaking marital harmony [bib_ref] Reemergence And Amplification Of Tuberculosis In The Canadian Arctic, Lee [/bib_ref]. Importantly, this disparity raises important concerns with a need to explore gender issues within the efforts tailored towards TB control or elimination. Given that the study participants were from a setting in which women rely on their husbands for financial support to access healthcare, non-disclosure could deny women of this financial support for drug collection and could result in poor adherence as already reported in the context of patients on ART in Africa [bib_ref] Hunger, waiting time and transport costs: Time to confront challenges to ART..., Hardon [/bib_ref]. Also, given the culturally perceived role of women, that is, to care for people within their households, positions them as important agents of TB transmission especially to those dependent on their care like under-five children, Adult TB disclosure to household members their spouses and others in vulnerable states e.g. the HIV positive, pregnant women and elderly. More to this, given that the women in rural Africa, bare the greatest burden pertaining to household food production, children feeding, food preparation and child healthcare [bib_ref] Effect of Male Involvement on the Nutritional Status of Children Less Than..., Kansiime [/bib_ref] , issues related to their health could impact greatly on their children's health, especially whether they could be enrolled on preventive treatment or evaluated for TB in the timely manner.
Belonging to the Muslim faith had 12.4 times higher odds of non-disclosure of TB status to household members as compared to patients belonging to the Catholic faith (p = 0.023). This is a new finding especially in the context of TB disclosure. However, shame-related HIV stigma has been reported to be strongly associated with religious beliefs such as the belief that HIV is a punishment from God or that people living with HIV/AIDS have not followed the Word of God [bib_ref] Religion and HIV in Tanzania: influence of religious beliefs on HIV stigma,..., Zou [/bib_ref]. It is however not clear whether such shame-related stigma, motivated by religious beliefs do exist in the context of TB, and if it could explain the disparity in religions with respect to disclosure to household members.
Although statistically non-significant, being single was associated with 4.6 times higher odds of non-disclosure of TB status to household members as compared to those who were married. This is a surprising finding that though may not be explainable based on the data available in this study, but will necessitate further exploration in not only a larger analytical study but also using qualitative methods. Adult TB disclosure to household members As expected, our study reports positive rewards for majority of patients (>98%) following TB disclosure to their household members. The commonest reported positive rewards include; household members supporting their medication intake, offering encouragement, supporting their feeding, and financial support to enable attendance of clinic days. As expected, such positive rewards of disclosure, were considered by patients to have impacted on their treatment adherence and psychosocial aspects. This finding affirms the benefit of a successful engagement of family/household members in TB patient care through patients' TB disclosure, a known foundational strategy in ensuring good treatment outcomes and infection control.
Surprisingly, all patients that had negative experiences following TB disclosure to household members also had atleast one positive experience concurrently. The reported negative experiences especially; marital separation, but also blame and negative criticism, and isolation, seem to indicate the existence of severe effects of potential underlying TB-related stigma among household members and communities at large. These findings are in agreement with those reported in previous studies on patients' TB disclosure to household members, that reported isolation, divorce, among other outcomes [bib_ref] Reemergence And Amplification Of Tuberculosis In The Canadian Arctic, Lee [/bib_ref]. Such negative post-disclosure experiences could be explained by the low awareness about TB disease that could be existing among patients and household members, something that could complicate TB control [bib_ref] Challenges in tuberculosis care in Western Uganda: Health care worker and patient..., Wynne [/bib_ref]. Such deficit in knowledge on TB by household members of TB patients, could be represented indirectly in this study by the fact that counseling on infection control and hope for total cure from TB, where the least performed aspects by household members following patients' TB disclosure. Such deficit in TB knowledge, could offer alternative explanation on why negative experiences erupted concurrently with positive experiences of patients post disclosure. These findings, not only highlight the need for robust interventions aimed at enhancing social and psychological support to TB patientsthrough TB awareness, but also interventions that could directly promote TB disclosure to household members while minimizing on such negative outcomes.
Enhancing the involvement of treatment supporters in facilitating successful stigma-free disclosure of TB by patients to household members through training and provision of suitable TB information aides by healthcare workers, is one potential intervention that was perceived acceptable to two-thirds of TB patients in the study. Since the treatment supporter system is already in existence to support home-based DOTS, and that all TB patients are required to have one where necessary prior to treatment initiation, such disclosure enhancement interventions centered on treatment supporters could easily be adopted. However, the feasibility and efficacy of such models would need to be evaluated in future randomized trials, and its success would depend largely on the healthcare workers' level of knowledge of TB, an argument supported by the already known positive correlation in knowledge on infection control practices between healthcare workers and household members [bib_ref] Health care workers' knowledge, attitudes and practices on tuberculosis infection control, Nepal, Shrestha [/bib_ref]. Efficacy would also depend on the availability of acceptable TB information aides for use by both patient and treatment supporter, so as to facilitate the disclosure process to household members.
The study has several limitations: i) The small sample size may have reduced the power to detect significance of some predictor variables of non-disclosure. This may also limit the conclusions on the predictors of TB non-disclosure to household members ii) The study did not quantify the level of stigma among patients, in as much as it is known to influence disclosure. iii) Patients' disclosure or not of TB to the household members was not assessed in regards to some clinical or treatment outcomes such as completion, cure, adherence or household contact screening, something that should be evaluated in future studies utilizing stronger analytical study designs. iv) The fact that the study never utilized qualitative assessment methods could not allow complete understanding of the major hindrances of disclosure especially among the female and moslem patients.
On the other-hand, our study unfolds the less addressed problem of TB patient non-disclosure to household members that could potentially impend the success of TB control strategies and attainment of the End TB targets of reducing TB incidence and TB-related death by 95% by 2035. Some of such TB control strategies hinged on successful disclosure to household members are: 1) TB Contact screening, which is impossible unless an index case voluntarily reports to have household contacts and willing to disclose to household members, 2) Initiation of TB contacts on preventive treatment, especially children, HIV infected adults, and other vulnerable household members, 3) Early TB diagnosis among symptomatic household members, which may be limited by delayed TB suspicion, hence delayed health seeking process by household members of TB patients, 4) Implementation of appropriate infection control measures at household level, 5) Home-based DOTS, and 6) Involvement of household members to offer social and psychological support to their patients, something that would impact on their treatment adherence and outcomes.
## Recommendations
Based on the study findings, we recommend that: 1) future larger studies should be conducted to conclusively establish the predictors of TB non-disclosure to household members, post-disclosure experiences, and their impact on treatment outcomes within both rural and urban settings; 2) Development of TB information aides to facilitate in process of patient disclosure to household members should be prioritized; 3) Interventions for enhancing disclosure at household level that are gender and religion-sensitive should be developed and evaluated in welldesigned community-based randomized trials so as to ascertain their efficacy and impact on ensuring successful stigma-free TB disclosure, early access to TB diagnostic services, preventive therapy uptake, treatment adherence and extent of family support to the TB patients; 4) Psychosocial counselling should be integrated within the healthcare package for TB patients among TB burden countries; 5) Qualitative studies are needed to explore the reasons for nondisclosure among female and Muslim TB patients but also TB disclosure dynamics and experiences in general at a household level.
# Conclusion
The rate of non-or partial TB disclosure to household members is very high, with important differences across gender and religion. This poses a risk of missing an opportunity to control TB transmission at household level, access to available TB diagnostic, treatment and preventive services by household members, and attracting family support towards their patients' care. Negative post-disclosure experiences reflecting severe forms of TB-related stigma, co-exist amidst the enormous patient-felt benefits of disclosure. There is need to enhance the knowledge of patients on TB infection control measures so as to minimize TB transmission at household level. Gender and religion sensitive interventions to enhance TB disclosure at household level while minimizing effects of stigma are urgently needed so as to increase engagement of household members in TB control efforts and patient care.
[fig] Fig 1: Positive post-disclosure experiences in adult TB patients, N = 57. https://doi.org/10.1371/journal.pone.0216689.g001 [/fig]
[fig] Fig 2: Areas of TB patients that are impacted by early positive post-disclosure experiences, N = 56. https://doi.org/10.1371/journal.pone.0216689.g002 [/fig]
[fig] S1: Dataset. Minimal dataset. (CSV) Questionnaire. English questionnaire. (DOCX) S2 Questionnaire. Runyankole questionnaire. (DOCX) [/fig]
[table] Table 1: Participants' characteristics. [/table]
[table] Table 2: TB Non-disclosure rates across patients' sociodemographic characteristics. [/table]
[table] Table 3: Results of univariate analysis for factors associated with non-disclosure of adult TB status to household members. [/table]
[table] Table 4: Results of multivariate analysis for factors independently associated with non-disclosure of adult TB status to household members. [/table]
|
An Introduction to Bacterial Biofilms and Their Proteases, and Their Roles in Host Infection and Immune Evasion
## The complexity of bacterial biofilms
Bacteria are single-celled microorganisms that are commonly associated with being independent, free-living organisms. However, they frequently adhere to surfaces to create multi-cellular aggregates across species that interact with one another [bib_ref] Biofilms as Complex Differentiated Communities, Stoodley [/bib_ref]. Such bacterial communities are called biofilms, and they are embedded in a self-produced extracellular polymeric substance (EPS), which not only confers increased tolerance to environmental stresses, but, as in the case of many pathogenic bacteria, also resistance against antibiotics and the host immune system. Furthermore, biofilms enhance nutrient capitalization, metabolite exchange, horizontal gene transfer, and cellular communication [bib_ref] What Are the Advantages of Living in a Community?, Santos [/bib_ref]. Importantly, because of their increased pathogenicity and resistance to therapy compared to free-living bacteria, the tendency and capacity to establish biofilms are considered as key virulence factors for a wide range of microorganisms, and the formation of biofilms on medical devices and implants, along with antibiotic resistance, poses one of the major challenges in medicine [bib_ref] Biofilms: An Emergent Form of Bacterial Life, Flemming [/bib_ref] [bib_ref] The Clinical Impact of Bacterial Biofilms, Høiby [/bib_ref].
Depending on the respective environment, three distinct types of biofilm growth can be identified: while pellicles (floating biofilms) are observed at air-water interfaces, colonies or submerged biofilms form on solid surfaces when interfacing with air or water, respectively . Despite variances in the adhesive fibers, proteins, nucleic acids, and exopolysaccharides that embed the respective biofilm, the general stages of biofilm development are remarkably similar across most bacteria, and follow four general steps: (i) reversible attachment to a surface or inaugural aggregation in solution; (ii) microcolony formation within a common EPS; (iii) biofilm growth and maturation; and (iv) dispersion and release of planktonic cells to colonize new sites [bib_ref] Biofilms as Complex Differentiated Communities, Stoodley [/bib_ref].
## Figure 1.
Stages of biofilm development. While the initial adhesion or aggregation planktonic bacterial species is reversible, the secretion of adhesins and different matr stimulates microcolony development and results in cell proliferation within the EPS. maturation, heterogenous zones with diverse transcriptomic, proteomic, and me emerge, including areas with dormant bacteria and subpopulations that are lysed to cellular DNA (eDNA) for the EPS. Finally, biofilms are dispersed either passively forces or actively by the secretion of hydrolytic enzymes that target the EPS compon in the release of planktonic bacteria or smaller cell aggregates and restarting the biof new colonization sites. Exemplary bacteria are depicted as polar flagellated rods and lated cocci in blue and in dark yellow, respectively, with their dormant counterpa The EPS of the pellicle (top) and the surface-associated biofilm (bottom) are shown and light brown, respectively, to indicate differences in chemical composition, and r ing dormant bacteria are highlighted in a darker shade of the same hue. Purple P represent various secreted proteases involved in biofilm remodeling and dispersion. fromand [bib_ref] Emerging Concepts of Biofilms in Infectious Diseases, Hung [/bib_ref].
In the first step, mainly driven by flagella, pili, and other adhesion pr tonic bacteria arrive and reversibly bind to a surface, or, as for pellicles, form cell aggregates in solution [bib_ref] Bacillus Subtilis Biofilm Formation and Social Interactions, Arnaouteli [/bib_ref]. Notably, proteolytically active flagella were covered in several Gram-negative and Gram-positive bacteria [bib_ref] Discovery of a Proteolytic Flagellin Family in Diverse Bacterial Phyla That Assembles..., Eckhard [/bib_ref] [bib_ref] Identification and Characterization of the Proteolytic Flagellin from the Common Freshwater Bacterium..., Eckhard [/bib_ref] , and are play a critical role in biofilm remodeling in these species. The next step is m secretion of high-molecular weight biopolymers, the EPS, thereby establish tional and structural integrity of the biofilm through a shared extracellula allowing irreversible attachment in surface-associated biofilms [bib_ref] Biofilms as Complex Differentiated Communities, Stoodley [/bib_ref]. This t hand in hand with a loss of motility, but crucially not necessarily with a lo [bib_ref] Temporal Gene-Expression in Escherichia Coli K-12 Biofilms, Domka [/bib_ref]. At the same time, biofilm matrix polysaccharides such as Pel and Ps nas aeruginosa are increasingly produced and secreted [bib_ref] Role of Exopolysaccharides in Pseudomonas Aeruginosa Biofilm Formation and Architecture, Ghafoor [/bib_ref] , strengthening th cell-surface interactions. Such formed microcolonies display considerable s . Stages of biofilm development. While the initial adhesion or aggregation of one or more planktonic bacterial species is reversible, the secretion of adhesins and different matrix biopolymers stimulates microcolony development and results in cell proliferation within the EPS. During biofilm maturation, heterogenous zones with diverse transcriptomic, proteomic, and metabolic profiles emerge, including areas with dormant bacteria and subpopulations that are lysed to provide extracellular DNA (eDNA) for the EPS. Finally, biofilms are dispersed either passively due to external forces or actively by the secretion of hydrolytic enzymes that target the EPS components, resulting in the release of planktonic bacteria or smaller cell aggregates and restarting the biofilm life cycle at new colonization sites. Exemplary bacteria are depicted as polar flagellated rods and as non-flagellated cocci in blue and in dark yellow, respectively, with their dormant counterparts in dark red. The EPS of the pellicle (top) and the surface-associated biofilm (bottom) are shown in light green and light brown, respectively, to indicate differences in chemical composition, and regions containing dormant bacteria are highlighted in a darker shade of the same hue. Purple Pacman shapes represent various secreted proteases involved in biofilm remodeling and dispersion. Figure adapted from [bib_ref] Emerging Concepts of Biofilms in Infectious Diseases, Hung [/bib_ref].
In the first step, mainly driven by flagella, pili, and other adhesion proteins, planktonic bacteria arrive and reversibly bind to a surface, or, as for pellicles, form initial small cell aggregates in solution [bib_ref] Bacillus Subtilis Biofilm Formation and Social Interactions, Arnaouteli [/bib_ref]. Notably, proteolytically active flagella were recently discovered in several Gram-negative and Gram-positive bacteria [bib_ref] Discovery of a Proteolytic Flagellin Family in Diverse Bacterial Phyla That Assembles..., Eckhard [/bib_ref] [bib_ref] Identification and Characterization of the Proteolytic Flagellin from the Common Freshwater Bacterium..., Eckhard [/bib_ref] , and are suggested to play a critical role in biofilm remodeling in these species. The next step is marked by the secretion of high-molecular weight biopolymers, the EPS, thereby establishing the functional and structural integrity of the biofilm through a shared extracellular matrix and allowing irreversible attachment in surface-associated biofilms [bib_ref] Biofilms as Complex Differentiated Communities, Stoodley [/bib_ref]. This typically goes hand in hand with a loss of motility, but crucially not necessarily with a loss of flagella [bib_ref] Temporal Gene-Expression in Escherichia Coli K-12 Biofilms, Domka [/bib_ref]. At the same time, biofilm matrix polysaccharides such as Pel and Psl in Pseudomonas aeruginosa are increasingly produced and secreted [bib_ref] Role of Exopolysaccharides in Pseudomonas Aeruginosa Biofilm Formation and Architecture, Ghafoor [/bib_ref] , strengthening the cell-cell and cell-surface interactions. Such formed microcolonies display considerable sessile growth and cell-cell communication such as quorum sensing, allowing the individual bacteria to coordinate and carry out cooperative activities such as biofilm maturation, dispersion, and virulence [bib_ref] Biofilms: Architecture, Resistance, Quorum Sensing and Control Mechanisms, Saxena [/bib_ref] [bib_ref] Potential Emergence of Multi-Quorum Sensing Inhibitor Resistant (MQSIR) Bacteria, Koul [/bib_ref]. As the biofilm matures and differentiates, phenotypically heterogeneous micro-environments and an organized 3D architecture emerge, characterized by complex internal water channels and pores that allow the distribution and transport of oxygen, nutrients, and signaling molecules necessary for concerted growth while also removing waste products and dead cells. Thus, mature biofilms exhibit multicellular-like behavior. The final stage of biofilm development is distinguished by the active (i.e., initiated by the bacteria) or passive (i.e., mediated by external forces) dispersal of the mature biofilm and the release of single planktonic cells, clumps, or pieces of biofilm into the environment to colonize new sites and begin a new cycle of biofilm formation. Importantly, numerous extracellular hydrolases, including saccharolytic enzymes and especially proteases, are produced during the active cell release from the EPS to destabilize and break down the EPS matrix [bib_ref] Characterization of Nutrient-Induced Dispersion in Pseudomonas Aeruginosa PAO1 Biofilm, Sauer [/bib_ref] [bib_ref] Genes Involved in the Synthesis and Degradation of Matrix Polysaccharide in Actinobacillus..., Kaplan [/bib_ref] [bib_ref] Biofilm Dispersal, Kaplan [/bib_ref]. Importantly, each of the biofilm stages is associated with specific transcriptomic, proteomic, and metabolomic profiles, reflecting bacterial adaptation to the respective situation. Likewise, distinct profiles are observed in the differentiated regions of the mature biofilm [bib_ref] Motility Influences Biofilm Architecture in Escherichia Coli, Wood [/bib_ref].
The key components of any EPS are highly hydrated biosynthesized polymers such as polysaccharides, proteins, and DNA at various ratios depending on the bacterial species or even strain, but also strongly depending on the actual type of biofilm formed, the biofilm morphotype [bib_ref] Genes Involved in the Synthesis and Degradation of Matrix Polysaccharide in Actinobacillus..., Kaplan [/bib_ref] [bib_ref] Biofilm Morphotypes and Population Structure among Staphylococcus Epidermidis from Commensal and Clinical..., Harris [/bib_ref]. Thus, the EPS is characterized by remarkable diversity that also allows the biofilm to adapt to different environments depending on environmental signals and quorum sensing. The EPS matrix provides several key advantages that compensate for its energetically expensive synthesis: (i) it can retain water, protecting the embedded microbes against drought; (ii) it can entrap and accumulate nutrients from the environment thus allowing survival in oligotrophic surroundings, and (iii) it can shield the bacteria from toxic metals but also antimicrobial compounds [bib_ref] The Biofilm Matrix, Flemming [/bib_ref] [bib_ref] Microbial Extracellular Polymeric Substances: Ecological Function and Impact on Soil Aggregation, Costa [/bib_ref].
Exopolysaccharides, the carbohydrate-based component of the EPS, often represent the major portion of the matrix. They can be made up of one or more types of monosaccharides (homo-versus heteropolysaccharides), and can be either neutral or charged. In particular, polycationic or polyanionic (i.e., highly positively or negatively charged, respectively) carbohydrates aid the absorption of water molecules and ions, and the formation of a stable network by acting as a glue between individual cells while also providing a scaffold that allows the stratification of the bacterial community [bib_ref] Genes Involved in the Synthesis and Degradation of Matrix Polysaccharide in Actinobacillus..., Kaplan [/bib_ref]. Another key component of microbial biofilms is extracellular DNA (eDNA), which may originate from concerted cell lysis of a small biofilm subpopulation. Its function has remained elusive for many years. However, DNase I treatment of early-stage biofilms causes destabilization and leads to premature dispersion, underlining the structural role of DNA within the 3D matrix. Additional roles have also been proposed, such as a source of nutrition for other colonies, and as a DNA pool for horizontal gene transfer within the biofilm community [bib_ref] Extracellular DNA Required for Bacterial Biofilm Formation, Whitchurch [/bib_ref].
Finally, the EPS contains extracellular proteins, including both glyco-and lipoproteins. In addition to their roles in scaffolding, cell-to-cell interaction, and adhesion, proteins play important enzymatic functions in various metabolic processes inside the bio-film matrix, and they facilitate the controlled breakdown of the various EPS components and thus enable dynamic biofilm remodeling [bib_ref] Bap: A Family of Surface Proteins Involved in Biofilm Formation, Lasa [/bib_ref]. Among the most prevalent proteinaceous structural components of the biofilm matrix are the functional amyloid fibers called curli, produced by numerous Enterobacteriaceae, and the various curli-like adhesins from other species [bib_ref] Amyloid Structures as Biofilm Matrix Scaffolds, Taglialegna [/bib_ref]. These fibrils considerably contribute to the mechanical properties of biofilms. Furthermore, carbohydrate-binding proteins, such as lectins, are frequently found and assumed to provide structural integrity by functioning as adapters between the different EPS components [bib_ref] Curli Biogenesis and Function, Barnhart [/bib_ref] [bib_ref] The Pseudomonas Aeruginosa Lectin LecB Binds to the Exopolysaccharide Psl and Stabilizes..., Passos Da Silva [/bib_ref]. Noteworthily, for each polymer component found in the EPS, there are also enzymes that facilitate their turnover, such as saccharolytic enzymes, proteases, DNases and lipases [bib_ref] The Biofilm Matrix, Flemming [/bib_ref]. This can occur for nutritional purposes, during biofilm remodeling and maturation, or to aid in biofilm dispersion and the release of planktonic bacteria. Furthermore, the same enzymes may play central roles during host colonization and infection, for example by degrading or neutralizing host molecules, allowing the bacteria to evade the immune system and propagate more easily. This takes us to proteases, which not only have a direct influence on the biofilms but also play crucial roles during host infection. Hence, many of them are regarded as major virulence factors. Moreover, in addition to cleaving proteins in the host's extracellular matrix to ease host invasion, they frequently target and inactivate host proteases or immunoglobulins from the immune system, as well as components of the complement system, allowing the bacteria to remain undetected for an extended period of time. As a result, these extracellular proteolytic activities are often directly linked to pathologies, thus rendering the associated proteases critical targets for medical treatments [bib_ref] Protease-Dependent Mechanisms of Complement Evasion by Bacterial Pathogens, Potempa [/bib_ref] [bib_ref] Staphylococcus Aureus Metalloprotease Aureolysin Cleaves Complement C3 to Mediate Immune Evasion, Laarman [/bib_ref] [bib_ref] IdeS, a Novel Streptococcal Cysteine Proteinase with Unique Specificity for Immunoglobulin G, Von Pawel-Rammingen [/bib_ref]. In the next sections, we will give a brief introduction to proteolytic enzymes, to protein secretion, and present several instances of secreted bacterial proteases that act as key virulence factors that directly impact the course of the infection.
## Proteolytic enzymes-how they work and what they do
Proteases, also called peptidases, are enzymes that are able to hydrolyze (i.e., break a chemical bond using a water molecule) peptide bonds within proteins and thus break down their substrates into shorter fragments such as peptides, and eventually into amino acids. While exopeptidases use terminal anchoring interactions and thus cleave only the terminal, penultimate, or antepenultimate peptide bond (i.e., removing up to three amino acids from the respective end of the polypeptide chain), endopeptidases target internal peptide bonds by means of so-called context-driven recognition [fig_ref] Figure 2: Classification of proteolytic enzymes [/fig_ref] [bib_ref] Molecular Basis of Exopeptidase Activity in the C-Terminal Domain of Human Angiotensin..., Naqvi [/bib_ref] [bib_ref] Proteases: A Primer, Hooper [/bib_ref]. In general, proteases utilize the active site water in one of two ways: (i) the protease activates the water molecule (i.e., increasing its nucleophilicity) and thus triggers its attack on the scissile peptide bond (general acid-base catalysis); or, (ii) an activated protein side chain of amino acids such as serine or cysteine acts as the nucleophile, thereby passing through a covalent acyl-enzyme intermediate, which is then resolved by a water molecule to release the bound N-terminal substrate fragment (covalent catalysis). Six distinct classes of proteases have been identified based on the active site and the amino acid or metal ion catalyzing the peptide-bond cleavage, with aspartate, glutamate, and metallopeptidases following acid-base catalysis, while serine, cysteine, and threonine peptidases use covalent catalysis [fig_ref] Figure 2: Classification of proteolytic enzymes [/fig_ref] [bib_ref] Proteases: A Primer, Hooper [/bib_ref] [bib_ref] Proteolytic Cleavage-Mechanisms, Function, and "Omic" Approaches for a Near-Ubiquitous Posttranslational Modification, Klein [/bib_ref] [bib_ref] The MEROPS Database of Proteolytic Enzymes, Their Substrates and Inhibitors in 2017..., Rawlings [/bib_ref].
Metallopeptidases are a large family of peptidases distinguished by an active site metal, typically zinc, that is tetrahedrally coordinated by three proteinaceous ligands and the catalytic water molecule [fig_ref] Figure 2: Classification of proteolytic enzymes [/fig_ref]. Two of the zinc ligands are typically histidines (His; H) derived from a HExxH motif on the so-called active-site helix, which also includes the general acid-base catalyst glutamate (Glu; E). The active site water is sandwiched between the zinc ion and the glutamate, with the latter subtracting a proton and therefore priming the water molecule for the nucleophilic attack on the amide bond carbonyl of the scissile peptide bond upon substrate binding [bib_ref] Families of Zinc Metalloproteases, Hooper [/bib_ref] [bib_ref] Architecture and Function of Metallopeptidase Catalytic Domains, Cerdà-Costa [/bib_ref]. Aspartic and glutamic peptidases belong to the carboxyl peptidase family, which is characterized by a low pH optimum and the respective name-giving acidic active site residue that activates the water molecule. The catalytic dyad of aspartic peptidases is typically formed by a pair of aspartic acid residues, one serving as the general base (i.e., abstracting the proton from the active site water) and the other providing electrophilic support. The situation is similar in glutamic peptidases, where a glutamate residue is typically supported by a glutamine [bib_ref] Catalytic Residues and Substrate Specificity of Scytalidoglutamic Peptidase, the First Member of..., Kataoka [/bib_ref] [bib_ref] The Aspartic Proteinases. An Historical Overview, Blundell [/bib_ref] [bib_ref] The Structure and Function of the Aspartic Proteinases, Davies [/bib_ref] [bib_ref] Identification and Characterization of a Bacterial Glutamic Peptidase, Jensen [/bib_ref].
Serine peptidases represent the most prevalent protease family in bacteria, and they perform so-called covalent peptide hydrolysis. The nucleophilic attack is carried out by a serine side chain inside a catalytic triad, which normally consists of a histidine, aspartate, and serine residue in a highly conserved three-dimensional configuration [fig_ref] Figure 2: Classification of proteolytic enzymes [/fig_ref]. This results in the generation of a covalent acyl-enzyme intermediate, which is then resolved by a water molecule. Every side chain in this acid-base-nucleophile triad has a unique role in catalysis: the acidic aspartate aligns and polarizes the histidine to act as a general base for the serine, enabling it to deprotonate for the nucleophilic attack while also acting as a proton donor for the first cleavage product [bib_ref] Serine Protease Mechanism and Specificity, Hedstrom [/bib_ref] [bib_ref] The Catalytic Triad of Serine Peptidases, Polgár [/bib_ref]. The process of peptide bond hydrolysis in cysteine peptidases is highly similar, and an equivalent cysteine-histidineaspartate/asparagine triad is frequently found [fig_ref] Figure 2: Classification of proteolytic enzymes [/fig_ref]. However, because the sulfhydryl group is more acidic than the hydroxyl group of serine, the importance of the aspartate is reduced and thus dyads containing only the nucleophile (cysteine) and base (histidine) are also found [bib_ref] Modes of Inhibition of Cysteine Proteases, Rzychon [/bib_ref] [bib_ref] Inhibitors of Cysteine Proteases, Vicik [/bib_ref]. Finally, while the overall cleavage mechanism of threonine proteases is similar to serine and cysteine peptidases, a single N-terminal threonine takes over the role of the catalytic dyad. While the secondary alcohol group of the side chain acts as the nucleophile, the N-terminal amino group acts as the general base that either directly or indirectly via a bridged water molecule deprotonates the side chain hydroxyl group to increase its nucleophilicity [bib_ref] A Protein Catalytic Framework with an N-Terminal Nucleophile Is Capable of Self-Activation, Brannigan [/bib_ref] [bib_ref] Catalytic Triads and Their Relatives, Dodson [/bib_ref].
acids from the respective end of the polypeptide chain), endopeptidases tar peptide bonds by means of so-called context-driven recognition [fig_ref] Figure 2: Classification of proteolytic enzymes [/fig_ref] , proteases utilize the active site water in one of two ways: (i) the protea the water molecule (i.e., increasing its nucleophilicity) and thus triggers its at scissile peptide bond (general acid-base catalysis); or, (ii) an activated protein of amino acids such as serine or cysteine acts as the nucleophile, thereby passi a covalent acyl-enzyme intermediate, which is then resolved by a water mol lease the bound N-terminal substrate fragment (covalent catalysis). Six distinc proteases have been identified based on the active site and the amino acid o catalyzing the peptide-bond cleavage, with aspartate, glutamate, and metallo following acid-base catalysis, while serine, cysteine, and threonine peptidase lent catalysis [fig_ref] Figure 2: Classification of proteolytic enzymes [/fig_ref] [bib_ref] Proteases: A Primer, Hooper [/bib_ref] [bib_ref] Proteolytic Cleavage-Mechanisms, Function, and "Omic" Approaches for a Near-Ubiquitous Posttranslational Modification, Klein [/bib_ref] [bib_ref] The MEROPS Database of Proteolytic Enzymes, Their Substrates and Inhibitors in 2017..., Rawlings [/bib_ref]. [bib_ref] Extracellular Proteases of Staphylococcus Epidermidis: Roles as Virulence Factors and Their Participation..., Martínez-García [/bib_ref]. Left: the serine peptidase Esp (PDB entry 4jcn [bib_ref] Secreted Proteases Control Autolysin-Mediated Biofilm Growth of Staphylococcus Aureus, Chen [/bib_ref] , with a catalytic triad (from left to right) formed by Asp159/His117/Ser235. Middle: the cysteine peptidase EcpA with a Asn362/His341/Cys245 triad (based on a AlphaFold2 [bib_ref] Highly Accurate Protein Structure Prediction with AlphaFold, Jumper [/bib_ref] model of UniProt ID Q5HKF6). Right: the gluzincin metallopeptidase SepA, with His355, Glu375, and His351 as the proteinaceous zinc-binding ligands, and Glu352 as the general base (HE 352 xxH(x) 19 E). The catalytic zinc ion is highlighted in yellow, while two structural calcium ions are shown in red. The depicted SepA active site is based on a AlphaFold2 [bib_ref] Highly Accurate Protein Structure Prediction with AlphaFold, Jumper [/bib_ref] model of UniProt ID P0C0Q4, while the metal ions and the active site water were inferred from PDB entry 1bqb [bib_ref] Amino-Acid Sequence and Three-Dimensional Structure of the Staphylococcus Aureus Metalloproteinase at 1.72..., Banbula [/bib_ref]. All molecular graphics representation were created using an open-source build of PyMOL version 2.5 (The PyMOL Molecular Graphics System, Schrödinger, LLC, New York, NY, USA) [bib_ref] The Case for Open-Source Software in Drug Discovery, Delano [/bib_ref].
Beyond doubt, proteases serve critical roles in many, if not all, fundamental biological functions, including protein maturation, signal peptide removal during protein secretion, and as signaling scissors by modulating biochemical pathways [bib_ref] The ADAMs: Signalling Scissors in the Tumour Microenvironment, Murphy [/bib_ref]. Importantly, more than 95% of the secreted proteases in bacteria are annotated as serine (~48%), metallo (~31%), and cysteine (~17%) peptidases [bib_ref] Distributions of Extracellular Peptidases Across Prokaryotic Genomes Reflect Phylogeny and Habitat, Nguyen [/bib_ref] , and thus not surprisingly, the proteases involved in biofilm remodeling (e.g., by cleaving EPS proteins or participating in quorum sensing) and pathogenicity (e.g., by cleaving host proteins such as collagen, complement factors, and antibodies), also belong nearly exclusively to these three protease families.
## Protein secretion to the extracellular space
Bacterial proteases can be intracellular, extracellular, or, as in the case of Gram-negative bacteria, periplasmic. While intracellular proteases are commonly associated with protein degradation, they also function as essential signaling scissors, precisely activating and inactivating proteins, thereby altering biochemical pathways and controlling cellular activity [bib_ref] Role of Intracellular Proteases in the Antibiotic Resistance, Motility, and Biofilm Formation..., Fernández [/bib_ref]. Periplasmic proteases, on the other hand, such as the serine peptidase DegP and the metallopeptidases BepA and YcaL from E. coli, are commonly involved in outer membrane protein biogenesis and assembly, and act as quality control factors that clear proteins with folding defects [bib_ref] Distinctive Roles for Periplasmic Proteases in the Maintenance of Essential Outer Membrane..., Soltes [/bib_ref]. Finally, the archetypal role of extracellular proteases is likely to scavenge resources, for example by degrading proteins in the surrounding environment and supplying peptides as easily accessible nutrients for the cell. However, they are also involved in biofilm formation and modulation, as already briefly outlined above, as well as in the breakdown of host cell matrices during pathogenesis. Moreover, they have been shown in several instances to increase virulence by targeting components of the host's immune system, thus allowing bacterial immune evasion, and directly linking secreted proteases to pathogenesis and disease progression [bib_ref] Bacterial Proteases and Virulence, Frees [/bib_ref].
Due to the critical roles of extracellular proteins and proteolysis for biofilm formation and bacterial fitness, bacteria have developed several mechanisms for protein secretion. The most commonly used and preserved mechanisms across all kingdoms of life are the "general secretion" (Sec) and "twin arginine translocation" (Tat) pathways [fig_ref] Figure 3: Schematic overview of known secretion systems in Gram-negative bacteria [/fig_ref] [bib_ref] Bacterial Secretion Systems: An Overview, Green [/bib_ref]. Both systems rely on signal peptides located at the N-termini of the proteins to be transported. While the Sec pathway primarily conveys unfolded proteins that fold once they reach the periplasm, the Tat pathway transports proteins already folded in the cytosol, distinct signal peptides ensure correct targeting to the respective secretion system [bib_ref] Sec-and Tat-Mediated Protein Secretion across the Bacterial Cytoplasmic Membrane-Distinct Translocases and Mechanisms, Natale [/bib_ref] [bib_ref] Twin-Arginine-Specific Protein Export in Escherichia Coli, Müller [/bib_ref]. Importantly, Sec and Tat both only bridge the cytoplasmic membrane. Thus, they are sufficient for Gram-positive bacteria, where secreted proteins arrive at the peptidoglycan layer and are either covalently linked to the cell wall and thus immobilized, or they passively diffuse to the exterior. Furthermore, some Gram-positive bacteria harness type VII and type VII-like secretion systems (T7SS) to transport proteins across the cell wall. The best studied T7SS are from Mycobacterium tuberculosis, which possesses five sub-types (ESX1 to ESX5), each with distinct secretomes (i.e., set of proteins transported). Importantly, current structural knowledge suggests that the T7SS core complex is positioned solely in the plasma membrane and does not reach the outer mycomembrane, which is located distally to the peptidoglycan layer and represents a unique feature of the mycobacterial envelope. Although specific details on the secretion process are still unavailable, members of the PE-PPE protein family are proposed to form substrate secretion pores in the mycomembrane [bib_ref] The Tat System of Gram-Positive Bacteria, Goosens [/bib_ref] [bib_ref] Protein Secretion in Gram-Positive Bacteria: From Multiple Pathways to Biotechnology, Anné [/bib_ref] [bib_ref] Type VII Secretion Systems in Gram-Positive Bacteria, Bottai [/bib_ref].
In Gram-negative bacteria, on the other hand, additional components are required to continue protein transport from the periplasm through the outer membrane to the extracellular space, such as the various type II (T2SS) and type V secretion systems (T5SS) [fig_ref] Figure 3: Schematic overview of known secretion systems in Gram-negative bacteria [/fig_ref]. While type V systems are structurally quite simple and are typically made up of a single polypeptide chain that spans the outer membrane and allows for autonomous transport ("autotransporters"), type II secretion systems are a bit more complex since they span both membranes. However, they only have a translocation unit for the outer membrane, and thus, as with the type V system, rely on the prior transport of the cargo proteins into the periplasm by the SEC/TAT pathways. Furthermore, Gram-negative bacteria possess type I secretion systems (T1SS) that directly transport proteins in one step from the cytosol to the extracellular space and highly evolved mechanisms for the direct delivery of bacterial effector proteins and virulence factors into the cytosol of host target cells, such as the type IV and type VI secretion systems, as well as the type III injectosome [bib_ref] Bacterial Secretion Systems: An Overview, Green [/bib_ref] [bib_ref] Secretion Systems in Gram-Negative Bacteria: Structural and Mechanistic Insights, Costa [/bib_ref]. and type VII-like secretion systems (T7SS) to transport proteins acro best studied T7SS are from Mycobacterium tuberculosis, which poss (ESX1 to ESX5), each with distinct secretomes (i.e., set of protein portantly, current structural knowledge suggests that the T7SS core co solely in the plasma membrane and does not reach the outer mycom located distally to the peptidoglycan layer and represents a unique f bacterial envelope. Although specific details on the secretion process a members of the PE-PPE protein family are proposed to form substrat the mycomembrane [bib_ref] The Tat System of Gram-Positive Bacteria, Goosens [/bib_ref] [bib_ref] Protein Secretion in Gram-Positive Bacteria: From Multiple Pathways to Biotechnology, Anné [/bib_ref] [bib_ref] Type VII Secretion Systems in Gram-Positive Bacteria, Bottai [/bib_ref]. , and type VI (T6SS) secretion systems are depicted, with the latter three being capable of directly injecting effector proteins into host cells. Sec (general secretion route) and Tat (twin-arginine translocation pathway) transfer unfolded and folded substrates of T2SS and T5SS across the inner membrane, respectively. The other secretion systems are one-step mechanisms that transport their cargo directly from the bacterial cytosol to the extracellular space or into eukaryotic host cells. Importantly, Sec and Tat are sufficient to cross the single bacterial membrane and reach the surface-exposed peptidoglycan layer in Gram-positive bacteria. ].
## The immune system and how pathogenic bacteria evade to establish infection
The immune system is a network of specialized cells and molecules that cooperate to protect the body from infections. Antibodies belong to the immunoglobulin superfamily and represent a vital detection system of the adaptive immune system in vertebrates, capable of detecting and neutralizing foreign substances such as pathogenic bacteria and viruses by binding to surface antigens or secreted effector proteins. Immunoglobulins are composed of two identical heavy chains that designate the antibody class (e.g., IgG, IgM, IgA, IgD, and IgE) and two identical light chains. While the heavy chain alone forms the Fc region, the two antigen-binding Fab regions are each made up of the variable regions of one light and one heavy chain. Communication with the cells of the immune system occurs through the Fc region, which is recognized by surface receptors (FcRs) that are primarily expressed on innate immune cells. Upon the binding of antigen-antibody complexes, the immune cells initiate cell-based and broad-ranging immune responses such as the neutralization and clearance of targeted compounds (e.g., via uptake by macrophages, neutrophils, or natural killer cells) or programming and training of the adaptive immunity (e.g., after internalization by dendritic cells, the presentation of processed antigenic peptides on major histocompatibility complex (MHC) proteins to naïve T cells leads to their expansion and maturation into antigen-specific effector and memory T cell populations) [bib_ref] The Immune System, Nicholson [/bib_ref] [bib_ref] Human Antibody-Fc Receptor Interactions Illuminated by Crystal Structures, Woof [/bib_ref].
Another pathway initiated by the recognition of the Fc region of an antigen-antibody complex is the classical complement cascade, which is part of the innate immune system and constitutes a powerful barrier against pathogens by mediating bacterial lysis, triggering phagocytosis, and recruiting immune cells [bib_ref] The Complement System, Sarma [/bib_ref] [bib_ref] Complement and Its Role in Innate and Adaptive Immune Responses, Dunkelberger [/bib_ref]. When IgM or IgG binds to a bacterial surface antigen, complement complex C1 recognizes the respective Fc region and initiates a series of proteolytic cleavage events that culminate in the formation of C3 convertase (C4bC2b) and the cleavage of C3 into C3a and C3b. While the anaphylatoxin C3a recruits leukocytes, C3b continues the cascade by forming C5 convertase (C4b2b3b) and interacting with additional components to generate the membrane attack complex (C5b-9), which then causes bacterial lysis by forming a pore in the bacterial cell membrane. Alternatively, the lectin pathway can activate the complement system, in which mannose-binding lectin (MBL) takes over pathogen recognition by binding to carbohydrates or glycoproteins on the bacterial surface. MBL oligomers are then recognized by one of three mannose-binding lectin-associated serine proteases (MASPs) in their inactive zymogen form. Binding leads to MASP activation, allowing for the cleavage of C4 and C2, and thus to the formation of C3 convertase and downstream processing as in the classical cascade. Additionally, there is the alternative pathway that involves factors such as B, D, H, and I that are exclusive to this complement pathway. Together with C3b, they produce an alternative C3 convertase (C3bBb) which can activate more C3 and thus enhance the cascade. Notably, while bacterial and fungal cell wall components enhance C3bBb formation, the presence of mammalian cells inhibits it [bib_ref] The Complement System, Sarma [/bib_ref] [bib_ref] Complement and Its Role in Innate and Adaptive Immune Responses, Dunkelberger [/bib_ref]. Due to the pivotal role of both immunoglobulins and the complement system for host defense, it is not surprising that bacterial proteases capable of targeting these molecules represent major virulence factors that facilitate infection and bacterial survival by disarming the immune system.
Antimicrobial peptides (AMPs) are another effective first line of defense against pathogenic microorganisms by the host innate immune system. They exhibit a typical length of 10 to 50 amino acids and can cause irreversible cell membrane damage. They are characterized as being either α-helical, β-sheet, or extended coil based on their secondary structure, with the latter being disordered on their own. Because of their amphipathic properties (i.e., exposing both hydrophilic and hydrophobic regions), these host defense peptides can interact with bacterial membranes through their hydrophobic interfaces while interacting with polar components via their hydrophilic regions, resulting in a destabilization of the membrane [bib_ref] Bacterial Evasion of Host Antimicrobial Peptide Defenses, Cole [/bib_ref]. AMPs comprise a diverse collection of peptides with limited sequence conservation that are primarily stored in the lysosomes of macrophages and polymorphonuclear leukocytes, with cathelicidins and defensins being the most prominent representatives. Importantly, cathelicidins require proteolytic activation; for example, the only cathelicidin described in humans, the α-helical AMP LL-37, requires maturation from its precursor CAP-18 by the serine peptidase proteinase 3. Defensins likewise need proteolytic processing for activity; however, many defensins have been described in man. They are classified as αand β-defensins in vertebrates based on their amino acid sequence and their synthesis by primarily neutrophils or epithelial cells, respectively. While they share a conserved triple-stranded antiparallel β-sheet structure constrained by intramolecular disulfide bonds, the molecular differences between individual molecules can be significant. However, defensins are always cysteine-rich and cationic at neutral pH, and they exhibit subclass-specific disulfide connectivity (Cys1-Cys6, Cys2-Cys4 and Cys3-Cys5 in α-defensins and Cys1-Cys5, Cys2-Cys4 and Cys3-Cys6 in β-defensins) [bib_ref] Bacterial Evasion of Host Antimicrobial Peptide Defenses, Cole [/bib_ref] [bib_ref] The Expanding Scope of Antimicrobial Peptide Structures and Their Modes of Action, Nguyen [/bib_ref] [bib_ref] Mode of Action of Linear Amphipathic Alpha-Helical Antimicrobial Peptides, Oren [/bib_ref].
Several modes of AMP action have been proposed. In the barrel-stave model, the peptides interact with the membrane via their hydrophobic faces while engaging with each other through their hydrophilic regions, resulting in a perpendicular arrangement of the AMPs to the membrane surface and the formation of a hydrophilic channel. The toroidal model is quite similar, except that lipid molecules intercalate between the individual peptides, and thus the formed channel is created by both the antimicrobial peptides and the hydrophilic head groups of the phospholipids. In the carpet model, however, the peptides bind parallel to the membrane surface and disrupt the membrane in a detergent-like fashion by creating micelle-like units, thus dissolving the cell membrane [bib_ref] The Expanding Scope of Antimicrobial Peptide Structures and Their Modes of Action, Nguyen [/bib_ref] [bib_ref] Mode of Action of Linear Amphipathic Alpha-Helical Antimicrobial Peptides, Oren [/bib_ref].
olecules 2022, 12, x FOR PEER REVIEW. Models of lipid membrane disruption by antimicrobial peptides. A are first adsorbed to the membrane surface. Once a certain threshold is reache tion occurs. In the barrel-stave model, the peptides are aligned parallel to the bacterial membrane, generating a continuous barrel-like pore. In the toroidal hydrophilic transmembrane channel is created, but with the lipid moieties fold hydrophilic phospholipid heads participate in channel formation. In the carpe peptides break down the lipid bilayer in a detergent-like manner into micell rounded by the peptides. Importantly, secreted bacterial proteases, indicat shapes, can cleave and hence inactivate antimicrobial peptides. . Models of lipid membrane disruption by antimicrobial peptides. Antibacterial peptides are first adsorbed to the membrane surface. Once a certain threshold is reached, membrane disruption occurs. In the barrel-stave model, the peptides are aligned parallel to the phospholipids of the bacterial membrane, generating a continuous barrel-like pore. In the toroidal pore model, a similar hydrophilic transmembrane channel is created, but with the lipid moieties folding inwards, and the hydrophilic phospholipid heads participate in channel formation. In the carpet model, accumulated peptides break down the lipid bilayer in a detergent-like manner into micelle-like structures surrounded by the peptides. Importantly, secreted bacterial proteases, indicated as green Pacman shapes, can cleave and hence inactivate antimicrobial peptides. [bib_ref] The Expanding Scope of Antimicrobial Peptide Structures and Their Modes of Action, Nguyen [/bib_ref] [bib_ref] The Role of Antimicrobial Peptides in Preventing Multidrug-Resistant Bacterial Infections and Biofilm..., Park [/bib_ref].
However, just as vertebrates have developed various defense mechanisms, pathogens have evolved multifaceted strategies to avoid or inactivate both the adaptive and the innate immune system to ensure their survival and spread to different sites within a host, such as by changing their proteins on the surface, or by secreting proteases that can target antibodies, complement factors, or AMPs. For example, by cleaving antibodies between the Fab and Fc regions, the recognition and effector domains of the antibody are separated, inhibiting downstream signaling and thus allowing bacteria to stay undetected, leading to enhanced bacterial survival. Moreover, numerous bacterial proteases have been shown to cleave and neutralize antimicrobial peptides, top right) or to inactivate various components of the complement cascade, hence preventing not only the assembly of the membrane attack complex, but also deterring the recruitment of immune cells to the site of infection [bib_ref] Bacterial Evasion of Host Antimicrobial Peptide Defenses, Cole [/bib_ref] [bib_ref] Staphylococci Evade the Innate Immune Response by Disarming Neutrophils and Forming Biofilms, De Vor [/bib_ref] [bib_ref] IdeS and SpeB: Immunoglobulin-Degrading Cysteine Proteinases of Streptococcus Pyogenes, Von Pawel-Rammingen [/bib_ref].
Among the various proteases capable of cleaving immunoglobulins, the cysteine peptidases IdeS (streptopain) and SpeB (streptococcal erythrogenic toxin B) from the Grampositive pathogenic bacterium Streptococcus pyogenes are the best studied [bib_ref] IdeS and SpeB: Immunoglobulin-Degrading Cysteine Proteinases of Streptococcus Pyogenes, Von Pawel-Rammingen [/bib_ref]. IdeS is a highly selective protease that separates the Fc and Fab regions of IgG antibodies by sequentially cleaving the two heavy chains of the lower hinge region. While two conserved glycine residues (Gly236-Gly237) form the preferred cleavage site, the Gly-Ala motif observed in IgG2 can also be processed, but with decreased efficiency [bib_ref] Rapid IgG Heavy Chain Cleavage by the Streptococcal IgG Endopeptidase IdeS Is..., Vindebro [/bib_ref] [bib_ref] Structure of the Streptococcal Endopeptidase IdeS, a Cysteine Proteinase with Strict Specificity..., Wenig [/bib_ref]. Importantly, cleavage of just one of the Fc heavy chains already results in an approximately 20-fold reduction in affinity to the corresponding Fc receptor, leading to reduced or eliminated bacterial cell clearance [bib_ref] Tumor-Associated and Microbial Proteases Compromise Host IgG Effector Functions by a Single..., Brezski [/bib_ref]. Unlike IdeS, which has a high degree of specificity, its structural homolog SpeB exhibits a significantly broader substrate profile [bib_ref] Proteolytic Profiling of Streptococcal Pyrogenic Exotoxin B (SpeB) by Complementary HPLC-MS Approaches, Blöchl [/bib_ref]. In addition to multiple proteins from the host matrix, SpeB is capable of cleaving complement factor C3 as well as immunoglobulins IgG, IgM, IgA, and IgD, but only IgG and IgD cleavage occurs in the hinge region [bib_ref] Effect of SpeB and EndoS from Streptococcus Pyogenes on Human Immunoglobulins, Collin [/bib_ref]. Furthermore, a recent study emphasized the importance of SpeB during infection, discovering that its specific inhibition resulted in a 28% decrease in bacterial survival in the presence of human neutrophils, confirming the notion that SpeB-mediated proteolysis impairs complement-mediated host defense [bib_ref] An Irreversible Inhibitor to Probe the Role of Streptococcus Pyogenes Cysteine Protease..., Woehl [/bib_ref].
Another intriguing set of proteases secreted by Streptococci are the subtilisin-like cell-envelope serine peptidases (CEPs), which are covalently attached to the bacterial cell wall due to their C-terminal LPXTG motif for sortase-mediated processing [bib_ref] The Role of Streptococcal Cell-Envelope Proteases in Bacterial Evasion of the Innate..., Mckenna [/bib_ref]. While ScpC (streptococcal chemokine protease C; also known as SpyCEP) inactivates interleukin 8 (IL-8) by removing its C-terminal region required for signaling and neutrophil recruitment [bib_ref] Specific C-Terminal Cleavage and Inactivation of Interleukin-8 by Invasive Disease Isolates of..., Edwards [/bib_ref] , the C5a peptidase ScpA destroys the chemotactic function of this anaphylatoxin [bib_ref] C5a Peptidase Alters Clearance and Trafficking of Group A Streptococci by Infected..., Ji [/bib_ref]. As a result, both proteases aid the pathogen in evading the host's immune system, while in the absence of either ScpA or ScpC, significantly increased bacterial clearance is observed as the neutrophil clearance mechanism via so-called neutrophil extracellular traps (NETs) is not impaired by these two proteases [bib_ref] Nizet, V. The IL-8 Protease SpyCEP/ScpC of Group A Streptococcus Promotes Resistance..., Zinkernagel [/bib_ref].
A wide variety of secreted proteases that interfere with the host immune system are also described in Gram-negative bacteria. One of them is the serine peptidase Pic secreted by enteroaggregative E. coli (EAEC) strains. Pic shows a broad substrate specificity profile capable of degrading multiple complement system components such as C2, C3, C3b, C4 and C4b, and thereby aiding immune evasion [bib_ref] The Serine Protease Pic From Enteroaggregative Escherichia Coli Mediates Immune Evasion by..., Abreu [/bib_ref]. Enterohemorrhagic E. coli (EHEC) produces EspP, a serine peptidase and Pic homolog that likewise inactivates C3/C3b and C5, interfering with the complement cascade and thus preventing bacterial opsonization and complement-mediated killing and phagocytosis [bib_ref] Serine Protease of Enterohemorrhagic Escherichia Coli, Impairs Complement Activation by Cleaving Complement..., Orth [/bib_ref]. E. coli also possesses a β-barrelshaped outer membrane aspartyl endopeptidase of the omptin protease family named OmpT, a housekeeping protease that degrades foreign peptide material encountered by the bacterium [bib_ref] Crystal Structure of the Outer Membrane Protease OmpT from Escherichia Coli Suggests..., Vandeputte-Rutten [/bib_ref]. Unsurprisingly, the same protease is also active against certain antimicrobial peptides (AMPs). This was discovered by comparing the effects of protamine on wild-type versus OmpT knock-out strains, and observing that only wild-type could clear the AMP and protect the cells from damage. On the other hand, deletions in the serine endopeptidase DegP or the metallopeptidase pitrilysin had no effect on protamine susceptibility [bib_ref] Identification of OmpT as the Protease That Hydrolyzes the Antimicrobial Peptide Protamine..., Stumpe [/bib_ref]. However, E. coli also possesses other mechanisms for AMP inactivation and clearance.
For example, even though increased OmpT activity correlates with symptomatic uropathogenic E. coli (UPEC) infections, the minimum inhibitory concentration of the human antimicrobial peptide LL-37 remained unchanged in a deletion mutant, despite OmpT being capable of partially cleaving LL-37 [bib_ref] Role of Uropathogenic Escherichia Coli OmpT in the Resistance against Human Cathelicidin..., Brannon [/bib_ref]. However, the most virulent UPEC strains encode not only for OmpT but also for an OmpT-like protease called ArlC. Additionally, in combination, these strains can degrade and inactivate a much wider selection of AMPs. In Salmonella enterica, the outer membrane aspartyl peptidase PgtE, which exhibits a 46% sequence identity with E. coli OmpT, confers resistance against the antimicrobial peptides C18G and LL-37 [bib_ref] A PhoP-Regulated Outer Membrane Protease of Salmonella Enterica Serovar Typhimurium Promotes Resistance..., Guina [/bib_ref]. Moreover, plasmid-derived PgtE complementation in Salmonella knock-outs and recombinant expression in E. coli demonstrated activity against human complement factors C3b, C4b, and C5, highlighting its capacity to interfere with several lines of host defense. Additionally, PgtE mutants showed reduced growth in the presence of normal human serum but not with heat-inactivated serum, indicating that PgtE confers a certain level of resistance against certain proteinaceous human serum components [bib_ref] The Surface Protease PgtE of Salmonella Enterica Affects Complement Activity by Proteolytically..., Ramu [/bib_ref].
In Pseudomonas aeruginosa, on the other hand, LL-37 has been shown to be capable of disrupting initial surface attachment and biofilm maturation [bib_ref] Human Host Defense Peptide LL-37 Prevents Bacterial Biofilm Formation, Overhage [/bib_ref]. However, the secreted metallopeptidase LasB, an elastase, can dampen the impact and improve cell survival by cleaving the AMP at multiple sites [bib_ref] Proteinases of Common Pathogenic Bacteria Degrade and Inactivate the Antibacterial Peptide LL-37, Schmidtchen [/bib_ref]. Consequently, even though LasB does not provide complete protection, the inferred increased AMP tolerance may allow for biofilm establishment and development, and is thus crucial for pathogenicity. Furthermore, due to its broad substrate specificity, LasB is also reported to degrade various extracellular matrix proteins, causing severe tissue damage while aiding host invasion [bib_ref] Clinical Strains of Pseudomonas Aeruginosa Secrete LasB Elastase to Induce Hemorrhagic Diffuse..., Zhu [/bib_ref] , and cleaves both complement components [bib_ref] Effect of Pseudomonas Aeruginosa Elastase and Alkaline Protease on Serum Complement and..., Hong [/bib_ref] and immunoglobulins [bib_ref] Experimental Studies of the Pathogenesis of Infections Owing to Pseudomonas Aeruginosa: Elastase,..., Holder [/bib_ref] [bib_ref] Degradation of IgA Proteins by Pseudomonas Aeruginosa Elastase, Heck [/bib_ref] , thus facilitating immune evasion.
Another Gram-negative bacterium capable of modulating the immunoglobulin response is Neisseria meningitidis, a bacterial colonizer of the human nasopharynx in approximately 10% of human adults. However, when leaving its commensal environment, Neisseria harnesses its IgA1-specific serine peptidase IgA1P, which is also capable of degrading IgG3 [bib_ref] Neisseria Meningitidis IgA1-Specific Serine Protease Exhibits Novel Cleavage Activity against IgG3, Spoerry [/bib_ref] , and an IdeS-like peptidase capable of cleaving the IgA hinge region, to evade the immune system, thereby causing blood sepsis and bacterial meningitis.
## Biofilm remodeling and the role of proteases in virulence
Extracellular proteases execute crucial roles throughout the biofilm life cycle, and they frequently represent virulence factors. Likewise, bacterial biofilm formation and modulation capacities are increasingly recognized as so-called passive virulence factors [bib_ref] Biofilms by Bacterial Human Pathogens: Clinical Relevance-Development, Composition and Regulation-Therapeutical Strategies. Microb, Schulze [/bib_ref]. Proteolytic activity against proteins of the host extracellular matrix (ECM) often represents a first crucial cornerstone during colonization, as seen for example in clostridial infections and the rapid collagen degradation mediated by their secreted collagenases [bib_ref] Clostridial Hydrolytic Enzymes Degrading Extracellular Components, Matsushita [/bib_ref] [bib_ref] Structure of Collagenase G Reveals a Chew-and-Digest Mechanism of Bacterial Collagenolysis, Eckhard [/bib_ref] [bib_ref] Structural Basis for Activity Regulation and Substrate Preference of Clostridial Collagenases G,..., Eckhard [/bib_ref] , or for enteroaggregative E. coli (EAEC) and its serine peptidase Pic, which readily cleaves mucins, heavily O-glycosylated linear glycoproteins that are secreted by higher organisms to protect the surfaces of epithelial cells [bib_ref] The Pic Protease of Enteroaggregative Escherichia Coli Promotes Intestinal Colonization and Growth..., Harrington [/bib_ref]. Moreover, as shown in studies using colonoids (i.e., an organoid derived from human colon crypts), the increased colonization rate of EAEC goes hand in hand with improved biofilm formation capabilities, and both are linked with proteolytic activity [bib_ref] Mucus Layer Modeling of Human Colonoids during, Liu [/bib_ref].
Another major proteolytic virulence factor is fragilysin, a secreted 20 kDa zincdependent metallopeptidase that is also known as Bacteroides fragilis enterotoxin (BFT), from enterotoxigenic Bacteroides fragilis (ETBF). This bacterium is not only linked with acute diarrhea and inflammatory bowel disease, but is also associated with colorectal cancer [bib_ref] Function and Latency Regulation of a Bacterial Enterotoxin Potentially Derived from a..., Goulas [/bib_ref] [bib_ref] Purification and Characterization of an Enterotoxin from Bacteroides Fragilis, Van Tassell [/bib_ref]. Fragilysin can cleave a wide variety of ECM substrates, including E cadherin, collagen type IV, actin, fibrinogen, and myosin, thus disrupting the cell-to-cell junctions and disrupting the integrity of the intestinal epithelial barrier and allowing the bacterium to penetrate and invade the host [bib_ref] The Bacteroides Fragilis Toxin Fragilysin Disrupts the Paracellular Barrier of Epithelial Cells, Obiso [/bib_ref] [bib_ref] The E-Cadherin Cleavage Associated to Pathogenic Bacteria Infections Can Favor Bacterial Invasion..., Devaux [/bib_ref]. Notably, enterotoxigenic ETBF strains (i.e., strains that express fragilysin) have a greater tendency to form biofilms [bib_ref] Genomic Diversity of Enterotoxigenic Strains of Bacteroides Fragilis, Pierce [/bib_ref] [bib_ref] Anti-Biofilm and Antimicrobial Effects of Zerumbone against Bacteroides Fragilis, Kim [/bib_ref] , and mouse studies using ETBF strains resulted in mucosal thickening, infiltration of inflammatory cells, crypt abscesses, epithelial cell exfoliation, erosion, and ulcerations in all mouse strains used, and led to even lethal colitis in germfree mice [bib_ref] Induction of Persistent Colitis by a Human Commensal, Enterotoxigenic Bacteroides Fragilis, in..., Rhee [/bib_ref]. Furthermore, fragilysin has been associated with colorectal cancer development resulting from the oxidative stress provoked by the upregulation of spermine oxidase due to E-cadherin degradation [bib_ref] The Mechanism of Bacteroides Fragilis Toxin Contributes to Colon Cancer Formation. Malays, Cheng [/bib_ref].
Another prominent protease capable of cleaving the adherens junction protein Ecadherin is the homooligomeric serine peptidase high temperature requirement A (HtrA) from the group 1 carcinogen Helicobacter pylori that colonizes the hostile environment of the stomach [bib_ref] Helicobacter Pylori HtrA Is a New Secreted Virulence Factor That Cleaves E-Cadherin..., Hoy [/bib_ref]. HtrA removes the ectodomain of this tumor suppressor on gastric epithelial cells, thus disrupting the epithelial integrity and propelling the infection. The relationship between HtrA and virulence was further confirmed using a HtrA overexpression system, which resulted in increased (i) rates of E-cadherin cleavage, (ii) bacterial transmigration, and (iii) delivery of the effector protein CagA into host cells [bib_ref] Overexpression of Serine Protease HtrA Enhances Disruption of Adherens Junctions, Paracellular Transmigration..., Harrer [/bib_ref]. HtrA also plays a key role in bacterial protein quality control, both protease-independent and protease-dependent, by acting as a molecular chaperone for certain nonnative polypeptides while degrading other misfolded proteins, which may explain why its knock-out leads to increased sensitivity to thermal, osmotic, and acidic stress, as well as amplified susceptibility to the amino nucleoside antibiotic puromycin [bib_ref] Properties of the HtrA Protease From Bacterium Helicobacter Pylori Whose Activity Is..., Zarzecka [/bib_ref] [bib_ref] Chaperone Activity of Serine Protease HtrA of Helicobacter Pylori as a Crucial..., Zarzecka [/bib_ref]. Notably, in Streptococcus mutans biofilms, a lack of HtrA results in aberrant biofilm structures with granular patches instead of a smooth confluent layer. This biofilm phenotype may be linked to the role of HtrA in the processing and maturation of various extracellular proteins, such as glucan-binding protein B but also surface-associated glycolytic enzymes such as glucosyltransferases and fructosyltransferase, thereby connecting HtrA activity with biofilm formation and maturation, alongside its role in host invasion [bib_ref] Role of HtrA in Surface Protein Expression and Biofilm Formation by Streptococcus..., Biswas [/bib_ref].
One key group of adhesion proteins involved in Staphylococcus aureus biofilm formation is the biofilm-associated protein (Bap) family, which has also been shown to play an important role in adhesion to the surface of epithelial cells and thus in pathogenic biofilm formation [bib_ref] Bap: A Family of Surface Proteins Involved in Biofilm Formation, Lasa [/bib_ref]. This Bap-dependent biofilm formation process is regulated by several proteases such as the metallopeptidase aureolysin and the staphylococcus serine peptidase A (SspA) [bib_ref] The Role and Regulation of the Extracellular Proteases of Staphylococcus Aureus, Shaw [/bib_ref]. However, the exact molecular mechanism is still unclear, as the various known substrates and the two proteases are strongly interconnected (e.g., aureolysin is required for activation of the SspA zymogen), and as it is likely that additional factors are involved, a phenomenon frequently described as the protease web [bib_ref] Network Analyses Reveal Pervasive Functional Regulation between Proteases in the Human Protease..., Fortelny [/bib_ref]. In a related study, it was demonstrated that the biofilm matrix interferes with the entry of S. aureus into epithelial cells [bib_ref] Bap, a Biofilm Matrix Protein of Staphylococcus Aureus Prevents Cellular Internalization through..., Valle [/bib_ref] , emphasizing the critical role of proteases in biofilm dispersal, for the controlled release of the pathogen from the EPS, and thus for virulence and pathogenicity [bib_ref] Staphylococcus Aureus Manipulates Innate Immunity through Own and Host-Expressed Proteases, Pietrocola [/bib_ref] [bib_ref] An Engineered Protein-Based Submicromolar Competitive Inhibitor of the Staphylococcus Aureus Virulence Factor..., Mendes [/bib_ref] Consequently, aureolysin represents a key virulence factor by cleaving not only proteins from the host's extracellular matrix and innate immune system, but also of the bacterial cell wall and the biofilm EPS, and by proteolytically activating other secreted toxins. This greatly emphasizes the complex roles of proteases in biofilm biology and bacterial infection on a variety of levels and through a number of different mechanisms that go far beyond cleaving a single substrate.
Furthermore, overexpression of SasG (surface protein G) in S. mutans has been found to result in polysacharride-independent biofilm formation, highlighting its role in surface adherence [bib_ref] The Role of Staphylococcus Aureus Surface Protein SasG in Adherence and Biofilm..., Corrigan [/bib_ref]. Sortases likewise play a key role during early surface attachment, especially in Gram-positive bacteria, as they covalently bond adhesion proteins such as the major surface protein P1 to the bacterial cell wall, allowing for protein-mediated cell-cell or cell-surface interactions [bib_ref] Mutational Analysis of the C-Terminal Anchoring Domains of Streptococcus Mutans P1 Antigen:..., Lee [/bib_ref]. In turn, and equally important, protease-mediated adhesin degradation represents a key mechanism during enzyme-driven biofilm dispersal [bib_ref] Staphylococcus Aureus Biofilms: Recent Developments in Biofilm Dispersal, Lister [/bib_ref] , and certain cell-associated proteases may be sortase-dependent. Furthermore, S. mutans deletion mutants of sortase A (StrA) were unable to form cell aggregates, likely due to the lower quantity of proteins being anchored to the peptidoglycan surface layer of the cell wall [bib_ref] Roles of Sortase in Surface Expression of the Major Protein Adhesin P1,..., Lee [/bib_ref] , and thus sortases are commonly recognized as critical virulence factors in Gram-positive bacteria [bib_ref] The Biology of Gram-Positive Sortase Enzymes, Paterson [/bib_ref]. This is further supported by an in vivo nasopharyngeal colonization model in chinchillas, where StrA-positive bacteria were detectable in the nasopharynx even 21 days after application, while sortase mutants were not [bib_ref] Sortase A Contributes to Pneumococcal Nasopharyngeal Colonization in the Chinchilla Model, Chen [/bib_ref]. Finally, the autolysin-mediated cell lysis of a biofilm subpopulation to provide eDNA for the EPS is also regulated by proteases, as shown in Staphylococcus epidermidis biofilms [bib_ref] Secreted Proteases Control Autolysin-Mediated Biofilm Growth of Staphylococcus Aureus, Chen [/bib_ref] [bib_ref] Role of Autolysin-Mediated DNA Release in Biofilm Formation of Staphylococcus Epidermidis, Qin [/bib_ref].
All of these examples demonstrate the intricacy of the numerous mechanisms that drive biofilm development and bacterial pathogenicity, as well as the critical involvement of proteases in these processes, making them not only essential virulence factors but also important therapeutic targets, as they could allow one to inhibit biofilm maturation and dispersal [bib_ref] Bacterial Proteolytic Complexes as Therapeutic Targets, Raju [/bib_ref]. Furthermore, biofilm-remodeling enzymes may be used to specifically target bacterial biofilms, thereby disrupting biofilm integrity and thus increasing its susceptibility to antibiotics. Additionally, because of reciprocal host-pathogen coevolution, many of these proteolytic virulence factors exhibit high substrate specificity and processivity, rendering them highly attractive for both biomedical and biotechnological applications.
## Biomedical and biotechnological applications of bacterial proteases
Various bacterial proteases have been explored for medicinal purposes, and some have even been approved as therapeutics. Two of the probably best-known protease therapeutics are the botulinum toxins A and B, two metallopeptidases secreted by Clostridium botulinum and commercialized as Botox and Myobloc, respectively [bib_ref] Proteases as Therapeutics, Craik [/bib_ref]. Notably, despite different substrate specificities, they both cleave components in the synaptic fusion complex, thus abolishing the release of the neurotransmitter acetylcholine into the synaptic cleft. Besides their use for cosmetic purposes, Botox is approved for the treatment of urinary incontinence associated with primary bladder muscle overactivity, abnormal muscle tightness due to prolonged muscle contraction (spasticity), and axillary hyperhidrosis (excessive underarm sweating), among other medical indications [bib_ref] Proteases as Therapeutics, Craik [/bib_ref] [bib_ref] Treatment of Hyperhidrosis with Botulinum Toxin, Doft [/bib_ref] , while neck dystonia (severe neck muscle spasms) and chronic sialorrhea (drooling or excessive salivation) may be treated with Myobloc [bib_ref] Clinical Use of Non-A Botulinum Toxins: Botulinum Toxin Type B, Dressler [/bib_ref]. Furthermore, the large secreted collagenolytic zinc-metallopeptidases ColG and ColH from Clostridium histolyticum are used for the treatment of Dupuytren's contracture and Peyronie's disease [bib_ref] CORD I Study Group Injectable Collagenase Clostridium Histolyticum for Dupuytren's Contracture, Hurst [/bib_ref]. However, despite good therapeutic performance, the drug was recently pulled from a number of markets including the European Union based on commercial grounds, while still being available in the USA [bib_ref] The End of an Era: Withdrawal of Xiapex (Clostridium Histolyticum Collagenase) from..., Cocci [/bib_ref]. Recently, IdeS from Streptococcus pyogenes (Imlifidase ® ) was used to desensitize patients with strong positive crossmatches for liver transplantation, enabling the procedure [bib_ref] IdeS (Imlifidase): A Novel Agent That Cleaves Human IgG and Permits Successful..., Lonze [/bib_ref]. Similarly, the IgA peptidase from Haemophilus influenzae was able to deplete the IgA complexes deposited in mouse glomeruli, and thus could represent a treatment option for IgA nephropathy [bib_ref] Microbial IgA Protease Removes IgA Immune Complexes from Mouse Glomeruli in Vivo:..., Lamm [/bib_ref].
In addition to these biomedical applications, various bacterial proteases are also used in biotechnology. For example, the glycylglycine endopeptidase lysostaphin from Staphylococcus simulans is a potent antimicrobial against S. aureus, and its transgenic expression protected cows from bacterial mastitis [bib_ref] Genetically Enhanced Cows Resist Intramammary Staphylococcus Aureus Infection, Wall [/bib_ref]. Furthermore, treatment of roasted peanut extracts with the Bacillus licheniformis serine peptidase subtilisin (Alcalase ® ) has been shown to significantly reduce the IgE reactivity and thus the allergenicity of the derived products, and thus may render common foods containing trace amounts of peanuts from production less dangerous for people with a peanut allergy [bib_ref] Influence of Enzymatic Hydrolysis on the Allergenicity of Roasted Peanut Protein Extract, Cabanillas [/bib_ref]. Bacterial proteases are also used in the laundry detergent industry. For example, alkaline proteases from Bacillus spp. have been used since the 1960s as additives due to their robustness and capacity to remove protein stains [bib_ref] Laundry Detergent Compatibility of the Alkaline Protease from Bacillus Cereus, Banik [/bib_ref].
Finally, immunoglobulin-cleaving peptidases from Streptococcus pyogenes, such as IdeS and SpeB, are increasingly used in the biopharmaceutical industry to specifically digest antibodies and render them suitable for middle-up proteomic analysis and especially glycoprofiling by mass spectrometry [bib_ref] Proteolytic Profiling of Streptococcal Pyrogenic Exotoxin B (SpeB) by Complementary HPLC-MS Approaches, Blöchl [/bib_ref] [bib_ref] Quantitative N-Glycan Profiling of Therapeutic Monoclonal Antibodies Performed by Middle-Up Level HILIC-HRMS..., Duivelshof [/bib_ref] , for example, to compare originator and biosimilar therapeutic monoclonal antibodies. Additionally, since the global market for therapeutic antibodies is predicted to be over USD 300 billion by 2025 [bib_ref] Development of Therapeutic Antibodies for the Treatment of Diseases, Lu [/bib_ref] , demand for these and similar enzymes is expected to surge over the next few years.
All these examples greatly highlight how bacterial virulence factors and proteases involved in biofilm remodeling can be transformed from "foes" during infection and pathogenesis into "friends" for biotechnological and biomedical applications. Surprisingly, this context-dependent potential of beneficial versus detrimental proteolysis is frequently overlooked, while it could enable new research avenues and therapeutic strategies.
# Conclusions
Proteolytic enzymes play intricate roles in many biological systems, including biofilm development and pathogenicity. With this review, we tried to provide a comprehensive reference point for early-career scientists working on microbial proteases and interested in how these enzymes impact bacterial life. We have highlighted several secreted proteolytic enzymes with well-defined roles in biofilm formation, maturation, and dispersion, and showcased proteases that act as virulence factors in host colonization and immune evasion by degrading components of the host extracellular matrix and the immune system. We then conclude with documented biotechnological and biomedical applications of some of secreted bacterial proteases, including their use as laundry detergents, in mass spectrometry, and to desensitize donor organs for patients with strong positive crossmatches.
[fig] Figure 2: Classification of proteolytic enzymes. (a) Proteases can be classified by the site cleavage. While endopeptidases cleave within proteins, exopeptidases remove N-or Cidues from the target protein. (b) Based on the active site responsible for cleavage o peptide bond, six classes of peptidases can be differentiated, of which threonine, serine peptidases act through a covalent acyl-enzyme intermediate, while metallopeptidases a [/fig]
[fig] Figure 3: Schematic overview of known secretion systems in Gram-negative bacteria. The basic structural features of the type I (T1SS), type II (T2SS), type V (T5SS), type III (T3SS), type IV (T4SS) [/fig]
[fig] Figure 4: Models of lipid membrane disruption by antimicrobial peptides. Antibacterial peptides are first adsorbed to the membrane surface. Once a certain threshold is reached, membrane disruption occurs. In the barrel-stave model, the peptides are aligned parallel to the phospholipids of the bacterial membrane, generating a continuous barrel-like pore. In the toroidal pore model, a similar hydrophilic transmembrane channel is created, but with the lipid moieties folding inwards, and the hydrophilic phospholipid heads participate in channel formation. In the carpet model, accumulated peptides break down the lipid bilayer in a detergent-like manner into micelle-like structures surrounded by the peptides. Importantly, secreted bacterial proteases, indicated as green Pacman shapes, can cleave and hence inactivate antimicrobial peptides. Figure adapted from [77,79]. [/fig]
[fig] Author: Contributions: Conceptualization, J.S.R.-L. and U.E.; writing-original draft preparation, J.S.R.-L.; writing-review and editing, U.E.; visualization, U.E. All authors have read and agreed to the published version of the manuscript. Funding: The authors gratefully acknowledge a FPI fellowship from the Spanish Ministry of Science and Innovation (MICINN) to J.S.R.L (PRE2020-096731), and a Beatriu de Pinós COFUND fellowhip from the Secretary of Universities and Research (Government of Catalonia) and the European Union Horizon 2020 program to U.E. (2018BP00163). [/fig]
|
HER2-Altered Non-Small Cell Lung Cancer: Biology, Clinicopathologic Features, and Emerging Therapies
Multiple oncogenic molecular alterations have been discovered that serve as potential drug targets in non-small cell lung cancer (NSCLC). While the pathogenic and pharmacological features of common targets in NSCLC have been widely investigated, those of uncommon targets are still needed to be clarified. Human epidermal growth factor receptor 2 (HER2, ERBB2)-altered tumors represent a highly heterogeneous group of diseases, which consists of three distinct situations including mutation, amplification and overexpression. Compared with breast and gastric cancer, previous studies have shown modest and variable results of anti-HER2 treatments in lung cancers with HER2 aberrations, thus effective therapies in these patients represent an unmet medical need. By far, encouraging efforts towards novel treatment strategies have been made to improve the clinical outcomes of these patients. In this review, we describe the biological and clinicopathological characteristics of HER2 alterations and systematically sum up recent studies on emerging therapies for this subset of patients.
## Introduction: the history of her2 at a glance
In the 1980s, researchers in Robert Weinberg's laboratory isolated a cDNA clone (which was termed as Neu) from carcinogen-induced tumors and found it displayed a protein structure that highly resembled epidermal growth factor receptor (EGFR) as earlier identified. From that point, successful efforts by three independent laboratories to identify EGFR-related cDNA sequences yielded the human orthologue of Neu (now called HER2 or ERBB2) [bib_ref] The ERBB Network: At Last, Cancer Therapy Meets Systems Biology, Yarden [/bib_ref]. However, unlike the mutation in the sequence of rodent Neu reported by Weinberg, human HER2 is usually amplified in tumors. Based on preclinical studies of breast cancer cell lines, HER2 amplification was found in a subset of patients with breast cancers and emerged as an important predictor of resistance to hormonal and chemotherapy regimens, time to relapse and overall survival (OS) [bib_ref] Evolving Standards of Care and New Challenges in the Management of HER2-Positive..., Choong [/bib_ref]. This significant finding was continuously confirmed by later studies and extended to gastric cancer [bib_ref] Trastuzumab in Combination With Chemotherapy Versus Chemotherapy Alone for Treatment of HER2-Positive..., Bang [/bib_ref].
HER2 amplification and protein overexpression lead to the dimerization of the receptor and activation of several signaling pathways that drive tumorigenesis. Consequently, targeting HER2 has been investigated as a promising therapeutic strategy. In 1998, trastuzumab, a monoclonal antibody (mAb) against HER2, was the first HER2-targeted agent approved by the FDA for treating metastatic breast cancer, which represented the beginning of a turnaround for the poor clinical outcomes of HER2-positive disease. Later on, considerable progress has been made as several categories of HER2targeting agents, including additional mAbs, signal transduction inhibitors, novel tyrosine kinase inhibitors (TKIs) which showed excellent efficacy in both early-stage and metastatic HER2-positive breast cancer. The second malignancy suitable for trastuzumab-based therapy as a standard of care was gastric cancer. Additionally, HER2 amplification/overexpression has also been observed in other solid tumors including the biliary tract, ovarian, endometrial, bladder, colon and NSCLC [bib_ref] HER2 Expression Status in Diverse Cancers: Review of Results From 37,992 Patients, Yan [/bib_ref]. Disappointingly, targeting HER2 with traditional anti-HER2 agents which were proved to be effective in breast and gastric cancer has failed in other tumor types, indicating the histological and biological diversity of HER2 alternations in distinct malignancies [bib_ref] Oncogenic Alterations in ERBB2/HER2 Represent Potential Therapeutic Targets Across Tumors From Diverse..., Chmielecki [/bib_ref].
In the subsequent sections, we will review available data and describe the biology of the HER2 pathway in normal and tumorigenesis processes, trying to provide a comprehensive overview of its dysregulation, clinical implications, as well as recent studies of emerging therapies for NSCLC patients with HER2 alterations.
## The biology of her2 and its dysregulation in nsclc
The ERBB family is comprised of four members that belong to the transmembrane tyrosine kinase receptors (TKR), including EGFR (also known as HER1), HER2, HER3 and HER4. HER2 encodes a transmembrane TKR which consists of three domains: an extracellular domain (ECD), a transmembrane domain (TMD) and an intracellular tyrosine kinase domain (TKD). Ligand binding results in heterodimerization or homodimerization between the ERBB receptors, and it sequentially stimulates the transactivation of the intracellular tyrosine kinase domain and activates downstream signaling pathways concerning cellular proliferation, differentiation, migration and apoptosis. However, HER2 lacks specific endogenous ligands and retains in the active conformation, making it continuously available for dimerization and to be the preferred heterodimerization partner. In contrast, HER3 has several ligands but lacks intrinsic tyrosine kinase activity. Interestingly, HER2-HER3 pairing displays the highest potency regarding the interaction strength and downstream signaling cascade, suggesting a complementary action between them [bib_ref] Structural Analysis of the Catalytically Inactive Kinase Domain of the Human EGF..., Jura [/bib_ref] [bib_ref] NSCLC and HER2: Between Lights and Shadows, Ricciardi [/bib_ref].
HER2 is a common oncogene identified in various cancer types and dysregulation of HER2 signaling can be caused by mutation, amplification and overexpression. All three types of HER2 dysregulation could appear in NSCLC with almost no overlap between mutation and amplification [bib_ref] HER2 Amplification and HER2 Mutation Are Distinct Molecular Targets in Lung Cancers, Li [/bib_ref]. Thus, associations of the three types of HER2 alterations are much more complex. This phenomenon partially explains the observed poor outcomes of classic HER2-targeted therapies in the setting of NSCLC than in breast cancer, where its oncogenesis predominantly relies on HER2 overexpression attributed to gene amplification. HER2 mutation is identified in 2%-4% of NSCLC and encompasses heterogeneous alterations distributed in the ECD, TKD and TMD. Exon 20 insertions that occurred within the kinase domain are the dominant forms of all the HER2 mutations and these insertions might account for about 1.5% of NSCLC. The most common variant is a 12-base-pair (encoding YVMA) in-frame duplicated insertion at codon 775 of exon 20, which affects the aC-b4 loop of the kinase domain and is identified as an early event in lung adenocarcinoma (LUAD) tumorigenesis [bib_ref] Clinicopathologic Associations, and Molecular Spectrum of ERBB2 (HER2) Tyrosine Kinase Mutations in..., Arcila [/bib_ref]. The HER2 YVMA subtype accounts for 34%-83% of HER2-mutated NSCLC, followed by G778_P780dup and G776delinsVC [bib_ref] Mutation Variants and Co-Mutations as Genomic Modifiers of Response to Afatinib in..., Fang [/bib_ref] [bib_ref] Pan-Cancer Landscape and Analysis of ERBB2 Mutations Identifies Poziotinib as a Clinically..., Robichaux [/bib_ref] [bib_ref] Conformational Landscapes of HER2 Exon 20 Insertions Explain Their Sensitivity to Kinase..., Zhao [/bib_ref]. In addition, there are more types of point mutations affecting the TKD but with a lower prevalence. Mutations in the TKD lead to conformational changes of ATP-binding pocket, which enhances kinase activity and downstream signaling. Other rarer mutations could also affect the ECD (mostly S310 in exon 8) and TMD (mostly V659 and G660 in exon 17) [bib_ref] HER2 Transmembrane Domain (TMD) Mutations (V659/G660) That Stabilize Homo-and Heterodimerization Are Rare..., Ou [/bib_ref] [bib_ref] Functional Analysis of Receptor Tyrosine Kinase Mutations in Lung Cancer Identifies Oncogenic..., Greulich [/bib_ref]. Generally, different mutation variants have heterogeneous behaviors, which could affect inhibitor binding affinity and sensitivity [bib_ref] Conformational Landscapes of HER2 Exon 20 Insertions Explain Their Sensitivity to Kinase..., Zhao [/bib_ref]. In rare cases, HER2 mutations could be found in germline and cause hereditary and sporadic LUADs [bib_ref] Novel Germline Mutation in the Transmembrane Domain of HER2 in Familial Lung..., Yamamoto [/bib_ref]. Concomitant EGFR somatic mutations are often detected in EGFR/ERBB2 germline mutations, suggesting that patients carrying EGFR/ ERBB2 germline mutations could also acquire somatic mutations in EGFR that eventually drive tumorigenesis [bib_ref] EGFR and ERBB2 Germline Mutations in Chinese Lung Cancer Patients and Their..., Lu [/bib_ref].
HER2 amplification accounts for 2%-4% in NSCLC, which is far less common compared with breast cancer [bib_ref] Lung Cancer That Harbors an HER2 Mutation: Epidemiologic Characteristics and Therapeutic Perspectives, Mazieres [/bib_ref]. Studies have shown that HER2 amplification and HER2 mutations were distinct molecular targets that may have different therapeutic and prognostic values. But they could co-exist in very few cases [bib_ref] NGS Sequencing Based Liquid/Tissue Biopsy Identified Coexistence of HER2 Amplification and Mutation..., Chen [/bib_ref]. Though an official consensus is not available, generally HER2 amplification is defined as HER2/CEP17 ≥ 2.0 by fluorescent in situ hybridization (FISH) testing. Notably, it should be distinguished from HER2 copy number gain (CNG), which happens even more frequently as a consequence of chromosome 17 polysomy and is not supposed to drive tumorigenesis [bib_ref] Clinical Significance of Epidermal Growth Factor Receptors in Non-Small Cell Lung Cancer..., Al-Saad [/bib_ref] [bib_ref] Evaluation of HER-2/Neu Gene Amplification and Protein Expression in Non-Small Cell Lung..., Hirsch [/bib_ref]. Chromosome 17 polysomy leads to the increased copy number of HER-2 per cell, so as to the activity of upstream promoter, resulting in the larger expression of HER2 gene. HER2 CNG is usually defined by HER2/CEP17 < 2.0 and HER2 gene copy number ≥ 6 per cell. By far, the predictive or prognostic role of HER2 CNG in NSCLC remains unclear. Apart from de novo tumorigenesis, HER2 amplification is one of the most frequent acquired resistance mechanisms following the EGFR T790M mutation in EGFR-mutant NSCLC treated with firstor second-generation EGFR-TKIs, which accounts for about 10% of cases [bib_ref] Second-Line Treatment of EGFR T790M-Negative Non-Small Cell Lung Cancer Patients, Liao [/bib_ref]. It can also confer resistance to osimertinib therapy, with a lower incidence rate of 2%-5% [bib_ref] Resistance Mechanisms to Osimertinib in EGFR-Mutated Non-Small Cell Lung Cancer, Leonetti [/bib_ref]. A combination of osimertinib and anti-HER2 agent trastuzumab emtansine was reported to overcome osimertinib resistance in T790M-positive EGFR-mutated NSCLC cell lines which gained HER2 amplification.
The incidence rate of HER2 overexpression is reported with a wide range of 2.5%-34% in NSCLC, probably due to the inconsistency among different methods for positivity assessment and low concordance assessment by pathologists [bib_ref] HER2-Targeted Therapies -A Role Beyond Breast Cancer, Oh [/bib_ref] [bib_ref] HER2 Status in Lung Adenocarcinoma: A Comparison of Immunohistochemistry, Yoshizawa [/bib_ref] [bib_ref] Impact of HER2 Gene and Protein Status on the Treatment Outcome of..., Kuyama [/bib_ref]. Currently, there is no consensus on how to define HER2 overexpression using IHC in NSCLC. Immunohistochemistry (IHC) scoring system and H-score are both applied for the assessment, with the former used more widely. Contrary to breast cancer, the co-occurrence of HER2 overexpression and amplification has not been well confirmed in lung cancer. However, an overlap between IHC 3+ staining and HER2 amplification was reported in NSCLC, albeit IHC low/negative has been also shown to be FISH positive. In the IHC 2+ cases, perhaps mechanisms other than gene amplification (HER2 polysomy, mutation and unknown reasons) cause the immuno-positive results [bib_ref] HER2 Status in Lung Adenocarcinoma: A Comparison of Immunohistochemistry, Yoshizawa [/bib_ref]. Therefore, HER2 overexpression in NSCLC represents a heterogeneous group with distinct molecular features, making it a less accurate indicator of inhibitor sensitivities and patient outcomes. Further analyses need to be done to discriminate among these possibilities. The associations of HER2 mutation, amplification and overexpression are depicted in .
## Clinical characteristics of her2-altered nsclc
Mutations in HER2 are more frequent in females, never smokers and lung adenocarcinoma, similar to those observed in patients with EGFR mutations. However, HER2-mutated patients have a worse prognosis than their EGFR and ALK counterparts, partially due to the lack of highly-selective targeted agents [bib_ref] Mutational Landscape and Characteristics of ERBB2 in Non-Small Cell Lung Cancer, Wei [/bib_ref] [bib_ref] HER2 Mutations in Lung Adenocarcinomas: A Report From the Lung Cancer Mutation..., Pillai [/bib_ref]. HER2 mutations also exhibit a tendency of brain metastases on treatment [bib_ref] Frequency and Outcomes of Brain Metastases in Patients With HER2-Mutant Lung Cancers, Offin [/bib_ref]. Similarly, another study found that the HER2 YVMA subtype was associated with a higher estimated 12month brain metastasis incidence compared with the non-YVMA group (40.2% vs. 3.6%, P=0.002) [bib_ref] Exon 20 YVMA Insertion Is Associated With High Incidence of Brain Metastasis..., Yang [/bib_ref]. De novo HER2 mutations in NSCLC are supposed to be mutually exclusive with other driver genes. But the oncogenic role of HER2 varied among different domains, with most of them occurring in the kinase domain. A study demonstrated that the frequency of EGFR or KRAS co-mutation was significantly higher in the non-TKD mutation compared to the TKD mutation, but OS was comparable between the two groups [bib_ref] Mutational Landscape and Characteristics of ERBB2 in Non-Small Cell Lung Cancer, Wei [/bib_ref]. Another retrospective database study revealed that patients harboring TMD mutations were diagnosed at more advanced stages (P<0.001) and had poorer OS (median OS 10.0m vs. 61.6m, P<0.001) than non-TMD mutations [bib_ref] HER2 Transmembrane Domain Mutation: Comprehensive Characteristics and Real-World Evidence of Treatment Response..., Jia [/bib_ref]. TP53 aberrations were the most prevalent co-mutations in HER2-mutated patients, followed by aberrations in the PI3K/AKT/mTOR pathway. Both two co-mutation variants were correlated with shorter progression-free survival of afatinib treatment [bib_ref] Mutation Variants and Co-Mutations as Genomic Modifiers of Response to Afatinib in..., Fang [/bib_ref]. This trend was also observed in another study which reported an impaired OS in patients with co-mutations in the cell-cycle pathway especially TP53 [bib_ref] Co-Occurring Alterations of ERBB2 Exon 20 Insertion in Non-Small Cell Lung Cancer..., Yuan [/bib_ref]. These observations indicated different clinical implications of HER2 mutation variants and their co-mutations.
De novo HER2 amplification and HER2 overexpression were detected more often in smokers and male patients, implying their inconsistent origins of tumorigenesis with HER2 mutations [bib_ref] The Efficacy of Ado-Trastuzumab Emtansine in Patients With ERBB2-Aberrant Non-Small Cell Lung..., Huang [/bib_ref]. HER2 amplification seemed to have a controversial role on the prognosis of NSCLC, while HER2 overexpression was a marker of poor prognosis, especially for adenocarcinoma, early-stage NSCLC and small cell lung cancer (SCLC) as suggested by a meta-analysis [bib_ref] The Role of Human Epidermal Growth Factor Receptor 2 As a Prognostic..., Liu [/bib_ref]. Compared with their mutation variants, co-mutations of other driver genes were more frequently seen in HER2 amplifi cation and overexpression tumors but still remained at a relatively low incidence rate. Concomitant HER2 alterations were proved to have an impact on the efficacy of targeted therapies. A retrospective study demonstrated that patients with concurrent EGFR mutation and HER2 amplification had a longer median time on treatment with EGFR-TKIs than those with EGFR Trastuzumab is a humanized IgG monoclonal antibody that could selectively inhibit the proliferation and survival of HER2addictive tumors by targeting the extracellular domain of HER2 [bib_ref] Trastuzumab-Mechanism of Action and Use in Clinical Practice, Hudis [/bib_ref]. A phase II study launched by the Eastern Cooperative Oncology Group (ECOG) evaluated the efficacy of combining carboplatin, paclitaxel and trastuzumab in advanced NSCLC patients with HER-2/neu positivity (1+ to 3+). The reported ORR, median PFS and median OS were 24.5%, 3.3 months and 10.1months, respectively. Notably, patients with 3+ HER-2/neu expression experienced a survival exceeding that of historical data, suggesting potential application for trastuzumab in this rare subgroup [bib_ref] Eastern Cooperative Oncology Group S. Trastuzumab in the Treatment of Advanced Non-Small-Cell..., Langer [/bib_ref]. Similarly, in a randomized phase II trial, the addition of trastuzumab to gemcitabine-cisplatin chemotherapy was beneficial for HER2 3+ or FISH-positive patients, but this subgroup is too small to provide definitive information [bib_ref] Randomized Phase II Trial of Gemcitabine-Cisplatin With or Without Trastuzumab in HER2-Positive..., Gatzemeier [/bib_ref]. A phase IIa multiple basket study (Mypathway) also demonstrated a moderate efficacy of dual blockade with trastuzumab and pertuzumab in HER2-altered NSCLC patients [bib_ref] Targeted Therapy for Advanced Solid Tumors on the Basis of Molecular Profiles:..., Hainsworth [/bib_ref]. HOT1303-B trial was a multicenter, single-arm phase II study of trastuzumab for pretreated HER2-altered NSCLC patients, which were defined as HER2 mutations, IHC 3+ or IHC 2+/dual color in situ hybridization [DISH]+. Disappointingly, trastuzumab monotherapy did not produce any response (ORR was 0%) in this cohort, although DCR was 70.0% and the median PFS reached 5.2 months [bib_ref] A Phase II Study of Trastuzumab Monotherapy in Pretreated Patients With Non-Small..., Kinoshita [/bib_ref]. A multicenter, phase II study (IFCT 1703-R2D2 trial) enrolled HER2-mutated advanced NSCLC patients progressing after platinum-based treatment and evaluated the efficacy and safety of triple therapy with trastuzumab, pertuzumab, and docetaxel [bib_ref] Combination of Trastuzumab, Pertuzumab, and Docetaxel in Patients With Advanced Non-Small-Cell Lung..., Mazieres [/bib_ref]. The ORR, median PFS and median OS were 29%, 6.8 months and 17.6 months, with tolerable toxicity, suggesting this triple therapy regimen becoming an option for pretreated patients.
In general, the efficacy of traditional chemotherapy for the HER2-altered population is far from satisfactory. The successes of trastuzumab observed in breast and gastric cancer might not be replicated in NSCLC. The reasons for differences in efficacy are complicated, but could possibly be explained by a different spectrum of HER2 alterations and other aspects of disease biology among cancer types. A combination of trastuzumab, pertuzumab, and docetaxel seemed to compete favorably with single mAb or chemotherapy, but this regimen remained to be defined. Other targeting strategies and agents for NSCLC patients with HER2 alterations are urgently needed.
## Non-selective tkis
Second-generation irreversible TKIs developed for the treatment of EGFR mutations represented early attempts to target HER2 in NSCLC. De Grève and colleagues reported the clinical benefits of afatinib in HER2 ex20ins LUAD patients [bib_ref] Clinical Activity of Afatinib (BIBW 2992) in Patients With Lung Adenocarcinoma With..., De Greve [/bib_ref]. Among the five identified patients, three patients evaluable for the response all showed an objective response. Later on, the efficacy of afatinib was evaluated in a global named patient use program that enrolled HER2-mutant patients who had exhausted other treatments. Median TTF was 2.9 months, ORR and DCR were 19% and 69%, respectively. Notably, for the group of HER2 YVMA subtype, median TTF was 9.6 months and 40% continued treatment for more than one year [bib_ref] Activity of Afatinib in Heavily Pretreated Patients With ERBB2 Mutation-Positive Advanced NSCLC:..., Peters [/bib_ref]. Conversely, other studies presented different perspectives, showing that the clonality status of HER2 ex20ins and the HER2 YVMA subtype were potential indicators for poor response to afatinib, while G778_P780dup and G776delinsVC subtypes derived favorable outcomes from afatinib, suggesting a further investigation into the clinical implications of mutation variants to help optimize outcomes with HER2-targeted therapies [bib_ref] Mutation Variants and Co-Mutations as Genomic Modifiers of Response to Afatinib in..., Fang [/bib_ref] [bib_ref] Co-Occurring Alterations of ERBB2 Exon 20 Insertion in Non-Small Cell Lung Cancer..., Yuan [/bib_ref]. A prospective phase II NICHE trial exploring the efficacy of afatinib in pretreated advanced HER2-mutant NSCLC patients reported that ORR and DCR were 7.7% and 53.8%. Thus, the accrual into the trial was terminated with 13 patients enrolled altogether. Median PFS and OS were 15.9 and 56.0 weeks, respectively [bib_ref] Afatinib in NSCLC With HER2 Mutations: Results of the Prospective, Open-Label Phase..., Dziadziuszko [/bib_ref]. Additionally, another phase II trial aimed to investigate the efficacy of afatinib among the Asian population with HER2-mutant NSCLC and consisted of two parts. In total, 18 patients were recruited and received afatinib in Part A. None of them achieved PR, 11 patients achieved SD and 6 patients had progressive disease as their best response. Median PFS and OS were 2.76 and 10.02 months, respectively. No patients met the Part B inclusion criteria, resulting in the termination of this study [bib_ref] Afatinib in Patients With Advanced Non-Small Cell Lung Cancer Harboring HER2 Mutations,..., Fan [/bib_ref]. The study of the EUHER2 cohort assessed the efficacy of chemotherapy and/or HER2-targeted agents in 101 advanced NSCLC patients with HER2 ex20ins. Sixty-five patients received HER2-targeted therapies.
Eleven patients were treated with afatinib, with an ORR and median PFS of 18.2% and 3.9 months, respectively [bib_ref] Lung Cancer Patients With HER2 Mutations Treated With Chemotherapy and HER2-Targeted Drugs:..., Mazières [/bib_ref]. Modest clinical activity of afatinib was also observed in a retrospective multicenter study, in which 27 patients included showed an ORR and median duration of response (DOR) of 13% and 6 months [bib_ref] Afatinib in Patients With Metastatic or Recurrent HER2-Mutant Lung Cancers: A Retrospective..., Lai [/bib_ref]. A Chinese retrospective study included patients harboring HER2 mutation and amplification treated with afatinib and reported an ORR of 24%. Median PFS and OS were 3.3 and 13.9 months, respectively [bib_ref] Efficacy and Resistance of Afatinib in Chinese Non-Small Cell Lung Cancer Patients..., Song [/bib_ref]. Collectively, studies above showed inconsistent efficacy of afatinib for HER2-mutant patients. A meta-analysis study integrated and reanalyzed the existing data regarding afatinib treating HER2mutant lung cancers. Pooled ORR and DCR were 21% and 66%, respectively, with the HER YVMA subtype deriving greater clinical benefit [bib_ref] The Efficacy of Afatinib in Patients With HER2 Mutant Non-Small Cell Lung..., Zhou [/bib_ref]. Thus, afatinib was not recommended as the regular application for treating NSCLC patients with HER2 mutation.
Dacomitinib is a pan-HER2 TKI that could bind to EGFR, HER2 and HER4 tyrosine kinases. A prespecified cohort from a phase II study enrolled NSCLC patients with HER2 mutations (n=26) or amplification (n=4). Three of 26 HER2-mutant patients had partial responses, but no partial responses were observed in four patients with HER2-amplified tumors. Median PFS and OS for HER2-mutant patients were 3 and 9 months, respectively. Interestingly, two patients harboring p. P780_Y781insGSP changes showed the longest responses, and the remaining patient with a partial response was identified with a p. M774delinsWLV mutation. No responses occurred in the most common HER2 YVMA subtype [bib_ref] Targeting HER2 Aberrations As Actionable Drivers in Lung Cancers: Phase II Trial..., Kris [/bib_ref]. Subsequently, a preclinical study demonstrated that the IC50 values of the HER2 ex20ins Ba/F3 cells harboring the dacomitinib-sensitive mutations (InsGSP, InsWLV, and InsCPG) were significantly lower than the other HER2 mutants or Wildtype HER2 (P=0.031), consistent with the previous clinical data [bib_ref] Response Heterogeneity of EGFR and HER2 Exon 20 Insertions to Covalent EGFR..., Kosaka [/bib_ref].
Neratinib is another type of TKI which was applied and evaluated in HER2-mutant NSCLC. A preclinical in vitro study assessed the activity of drugs in HER2 mutation variants and found neratinib and afatinib more effective than other inhibitors for the HER2 YVMA subtype [bib_ref] High-Throughput Functional Evaluation of Variants of Unknown Significance in ERBB2, Nagano [/bib_ref]. A randomized 2-stage phase II study compared neratinib monotherapy and the combination of neratinib with temsirolimus in stage IIIB/IV HER2-mutant NSCLC patients. Twenty-seven patients were enrolled in stage 1, and 3 of 14 patients (21%) in the combination group had a response, resulting in a median PFS of 4 months (60). In the subsequent expansion cohort, the dual inhibition group obtained an ORR of 19%, a median PFS of 4.1 months and a median OS of 15.8 months. The incidence of grade 3 diarrhea was 12% and could be managed with loperamide prophylaxis [bib_ref] Neratinib ± Temsirolimus in HER2-Mutant Lung Cancers: An International, Randomized Phase II..., Gandhi [/bib_ref]. In the SUMMIT phase II basket trial, neratinib was evaluated across multiple cancer types. In patients with lung cancer (n=26), only one patient harboring L775S kinase missense mutation was observed with objective response (ORR=3.8%), suggesting its limited efficacy in HER2-mutated lung cancer. The median PFS in recurrent NSCLC was 5.5 months with 6 patients continuing therapy for more than one year [bib_ref] HER Kinase Inhibition in Patients With HER2-and HER3-Mutant Cancers, Hyman [/bib_ref].
The activities of non-selective TKIs in HER2-mutant NSCLC patients yield moderate or even disappointing results, though sporadic responses have been reported as HER2 mutation location could affect the drug binding affinity. This may be explained because HER2 ex20ins seem to tighten the drugbinding pocket, restricting the binding of large-sized inhibitors. Therefore, structurally novel pan-HER2 TKIs have been developed to achieve better outcomes in NSCLC with HER2 alterations.
## New-generation tkis
As a covalent and irreversible EGFR/HER2 inhibitor, poziotinib has a smaller size and flexible structure. A preclinical study compared the activity of different TKIs in Ba/F3 cells with HER2 exon 20 mutations, poziotinib showed the most potent activity. Also, the secondary C805S mutation was identified as a potential mechanism of acquired resistance to poziotinib [bib_ref] Activity of a Novel HER2 Inhibitor, Poziotinib, For HER2 Exon 20 Mutations..., Koga [/bib_ref]. Another study indicated that HER2 20 exon insertions tightened the size of the drug-binding pockets and restricted the binding of large, rigid inhibitors using 3D modeling. Poziotinib can avoid these spatial changes owing to its small size and flexibility, thus becoming a potent inhibitor of the most common HER2 variant. Its efficacy was further confirmed in vitro and in vivo studies [bib_ref] Mechanisms and Clinical Activity of an EGFR and HER2 Exon 20-Selective Kinase..., Robichaux [/bib_ref]. Similar trends were also reported in the pan-cancer landscape and functional analysis of HER2 mutation by Robichaux and colleagues [bib_ref] Pan-Cancer Landscape and Analysis of ERBB2 Mutations Identifies Poziotinib as a Clinically..., Robichaux [/bib_ref]. In addition, in their preclinical models, poziotinib upregulated HER2 expression at the cell surface and potentiated the T-DM1 activity. Early results from a phase I study showed an encouraging activity of poziotinib and further justified its application in patients with EGFR/HER2 alterations [bib_ref] Phase 1 Studies of Poziotinib, an Irreversible Pan-HER Tyrosine Kinase Inhibitor in..., Kim [/bib_ref]. Based on this study and preclinical evidence, a phase II clinical trial of poziotinib in NSCLC patients with EGFR and HER2 exon 20 mutations was initiated. All HER2-mutant participants enrolled harbored the Y772dupYVMA or G778dupGSP insertions. Response was confirmed in 5 of 12 HER2-mutated patients (confirmed ORR=42%). The DCR and median PFS were 83% and 5.6 months. No patients discontinued treatment due to poziotinib-related toxicity [bib_ref] Pan-Cancer Landscape and Analysis of ERBB2 Mutations Identifies Poziotinib as a Clinically..., Robichaux [/bib_ref]. A poziotinib expanded access program enrolling NSCLC patients with EGFR or HER2 ex20ins showed a median PFS of 5.6 months and a median OS of 9.5 months. The ORR was higher in HER2 subgroup (50% vs. 23%). Grade 3 AEs were reported in 66% of the patients, and the toxicity rate was high leading to frequent dose interruption and reduction [bib_ref] Poziotinib for EGFR and HER2 Exon 20 Insertion Mutation in Advanced NSCLC:..., Prelaj [/bib_ref]. A single-arm, open-label, phase II study assessed the efficacy and safety profiles of poziotinib in HER2mutant advanced NSCLC. The confirmed ORR was 27% with responses observed across mutation subtypes. Median PFS and OS were 5.5 and 15.0 months, respectively. One possible treated-related death due to pneumonitis was reported [bib_ref] Poziotinib for Patients With HER2 Exon 20 Mutant Non-Small-Cell Lung Cancer: Results..., Elamin [/bib_ref]. ZENITH20 study evaluated poziotinib in previously treated NSCLC patients with HER2 exon 20 insertions. In cohort 2, patients received poziotinib once daily, the ORR and DCR were 27.8% and 70.0%, respectively. Most patients (74%) had tumor reduction, with a median PFS was 5.5 months. Clinical benefits were seen regardless of types and lines of previous treatment, presence of brain metastasis and HER2 mutation variants. Severe treatment-related AEs (grade≥3) included rash (48.9%) diarrhea (25.6%), and stomatitis (24.4%) (68). Updating results from cohort 4 of ZENITH20 presented a promising efficacy in treatment-naive patients, with an ORR of 44% and a median PFS of 5.6 months, the safety profile was similar to cohort 2 of this study.
Pyrotinib is an oral, irreversible pan-HER TKI. The enhanced antitumor activity of another irreversible pan-HER TKI pyrotinib was observed in organoids as well as patients-derived xenograft (PDX) models relative to afatinib and trastuzumab-emtansine (T-DM1). In a phase II cohort of 15 HER2-mutant NSCLC patients, pyrotinib resulted in an ORR and a median PFS of 53.3% and 6.4 months. AEs were all grade 1-2 and no dose reduction or treatment discontinuation occurred [bib_ref] HER2 Exon 20 Insertions in Non-Small-Cell Lung Cancer Are Sensitive to the..., Wang [/bib_ref]. In another single-arm prospective study, pyrotinib exhibited promising efficacy and acceptable safety in 27 advanced HER2-amplified NSCLC patients, reaching a confirmed ORR of 22.2%, a median PFS of 6.3 months and a median OS of 12.5 months. Of note, Patients who received pyrotinib as a first-line treatment achieved a median PFS of 12.4 months. Treatment-related AEs occurred in all patients, but no grade 4 or higher AEs were documented [bib_ref] Pyrotinib in Patients With HER2-Amplified Advanced Non-Small Cell Lung Cancer: A Prospective,..., Song [/bib_ref]. A larger phase II study evaluated the efficacy and safety profiles of pyrotinib in stage IIIB-IV LUAD patients harboring HER2 mutations after the failure of platinum-based chemotherapy. Independent reading committeeassessed ORR was 30.0%, with a favorable ORR observed across all HER2 subtypes and between patients with and without brain metastases (25.0% vs. 31.3%). The median PFS and OS were 6.9 and 14.4 months, respectively. Grade 3-4 treatment-related AEs occurred in 28.3% of patients and diarrhea (20.0%) appeared to be the most common type [bib_ref] Pyrotinib in HER2-Mutant Advanced Lung Adenocarcinoma After Platinum-Based Chemotherapy: A Multicenter, Open-Label,..., Zhou [/bib_ref]. In addition, the encouraging efficacy of pyrotinib was also reported in a study in which patients with EGFR mutation and HER2 amplification could obtain clinical benefits from combining EGFR-TKIs and pyrotinib, suggesting the potential of pyrotinib in targeting the HER2 pathway in patients after progressing on EGFR-TKIs [bib_ref] HER2 Amplification in Advanced NSCLC Patients After Progression on EGFR-TKI and Clinical..., Gan [/bib_ref].
The hypoxia-activated prodrug (HAP) tarloxotinib is a potential treatment for NSCLC with HER2 alterations. Prior work demonstrates that the tarloxotinib can be converted into tarloxotinib-E as its active form in a hypoxic tumor microenvironment. Preclinical studies showed that tarloxotinib-E could interfere with cell signaling and proliferation by inhibiting phosphorylation and activation of ERBB heterodimers in PDX models. In vivo, tarloxotinib inhibited tumor growth and progression. The pharmacokinetic analysis also confirmed the accumulations of tarloxotinib-E in tumor sites than plasma or skin [bib_ref] Tarloxotinib Is a Hypoxia-Activated Pan-HER Kinase Inhibitor Active Against a Broad Range..., Estrada-Bernal [/bib_ref]. Another in vitro study demonstrated that the IC50 of tarloxotinib for wildtype HER2 was 180 times higher than that of tarloxotinib-E, suggesting a wide therapeutic index of tarloxotinib [bib_ref] Activity and Mechanism of Acquired Resistance to Tarloxotinib in HER2 Mutant Lung..., Koga [/bib_ref]. The phase I RAIN-701 trial (NCT03805841) enrolled advanced patients with HER2 activating mutations (cohort B). First results showed that 2 of 9 patients experienced confirmed PR (22%) and 4 patients had SD. Grade 3 TEAEs included prolonged QTc (34.8%), increased ALT (4.3%), diarrhea (4.3%) and rash (4.3%) [bib_ref] First Analysis of RAIN-701: Study of Tarloxotinib in Patients With Non-Small Cell..., Liu [/bib_ref].
Mobocertinib (TAK-788) is a research-based oral EGFR/ HER2 inhibitor designed against ex20ins. The IC50 of mobocertinib was higher than poziotinib and comparable with pyrotinib, neratinib and afatinib in HER2 ex20ins cell lines. Mobocertinib exhibited the lowest HER2 ex20ins IC50/wildtype EGFR IC50 ratio, implying its excellent selectivity profile. Additionally, lung cancers with HER2 G776delinsVC subtypes reported a superior response to mobocertinib than the YVMA subtypes. The combination of ado-trastuzumab emtansine (T-DM1) and mobocertinib had a synergistic function in HER2 YVMA tumors [bib_ref] Targeting HER2 Exon 20 Insertion-Mutant Lung Adenocarcinoma With a Novel Tyrosine Kinase..., Han [/bib_ref]. A phase I/II trial (NCT02716116) showed promising antitumor activities of mobocertinib in advanced NSCLC patients harboring EGFR ex20ins, with a similar safety profile compared to other EGFR-TKIs [bib_ref] Updated Results From a Phase I/II Study of Mobocertinib (TAK-788) in NSCLC..., Riely [/bib_ref]. Results from the expansion cohort 2 of this study which enrolled NSCLC patients with HER2 exon 20 alterations are still awaited.
Compared with non-selective TKIs originally developed for EGFR mutations, these novel TKIs have shown greater activities and broader anti-tumor effects across exons in HER2-mutant NSCLC. For patients who had previously received platinumbased chemotherapy where there are limited therapeutic drugs, new generation TKIs could be an option to consider.
## Antibody-drug conjugates
ADCs are characterized by the covalent coupling of drugs to mAbs as an alternative to naked antibody-targeted therapy. The development of ADCs represents groundbreaking progress in the treatment of malignancies with actionable targets. After their successes in breast and gastric cancer, these agents are gradually attracting much attention in NSCLC with HER2 alterations.
Trastuzumab emtansine (T-DM1) is a second-generation anti-HER2 ADC composed of trastuzumab and emtansine (DM1), which is an inhibitor of microtubule aggregation. This complex enters into HER2-positive cells via receptor-mediated endocytosis. Proteolytic degradation of the antibody moiety in lysosomes leads to the release of conjugated agents [bib_ref] Targeting HER2-Positive Breast Cancer With Trastuzumab-DM1, an Antibody-Cytotoxic Drug Conjugate, Phillips [/bib_ref]. In a phase II study of T-DM1 in relapsed HER2-positive NSCLC (IHC 3+, IHC 2+/FISH+, or exon 20 mutations), among fifteen assessable patients, only one patient achieved a PR (ORR=6.7%). The median PFS and OS were 2.0 and 10.9 months, respectively. Grade 3-4 AEs included thrombocytopenia (40%) and hepatotoxicity (20%), with no treatment-related deaths [bib_ref] A Phase II Study of Trastuzumab Emtansine in HER2-Positive Non-Small Cell Lung..., Hotta [/bib_ref]. However, another phase II basket trial enrolled 18 advanced LUAD patients harboring HER2 mutations. The treatment with T-DM1 might achieve an ORR as high as 44% and a median PFS of 5 months. Responses to T-DM1 could be seen across all the subtypes. The toxicities included grade 1-2 elevated hepatic transaminases, thrombocytopenia and infusion reactions [bib_ref] Ado-Trastuzumab Emtansine for Patients With HER2-Mutant Lung Cancers: Results From a Phase..., Li [/bib_ref]. Peters et al. evaluated the efficacy and safety of T-DM1 in 49 advanced HER2-overexpressing NSCLC patients (29 IHC 2+ and 20 IHC 3+) who were previously treated. Although there were no treatment responses presented in the IHC 2+ cohort, four PR were observed in the IHC 3+ cohort (ORR=20%). Median PFS and OS were similar between the two cohorts (median PFS: 2.6 and 2.7 months; median OS: 12.2 and 15.3 months). Forty-five patients (92%) reported an AE of any grade and ten patients (20%) reported grade 3 AEs. One patient with a history of brain metastases reported grade 4 seizures while receiving seizure therapy. There were no deaths due to AEs [bib_ref] Trastuzumab Emtansine (T-DM1) in Patients With Previously Treated HER2-Overexpressing Metastatic Non-Small Cell..., Peters [/bib_ref].
Trastuzumab deruxtecan (T-Dxd, also known as DS-8201a) is a novel HER2-ADC composed of trastuzumab and a novel topoisomerase I inhibitor (MAAA-1181) linked by an enzymatically cleavable peptide. The drug moiety of T-Dxd could bind to topoisomerase I-DNA complexes and induce DNA doublestrand breaks. T-Dxd has a drug-to-antibody ratio (DAR) of 8, which is almost two-fold higher than T-DM1 (DAR of [bib_ref] Trastuzumab in Combination With Chemotherapy Versus Chemotherapy Alone for Treatment of HER2-Positive..., Bang [/bib_ref] [bib_ref] HER2 Expression Status in Diverse Cancers: Review of Results From 37,992 Patients, Yan [/bib_ref]. Preclinical results suggested that T-Dxd had antitumor activities towards a broad range of HER2-positive models and acceptable safety profiles. Featured by a highly membrane-permeable payload, it can exert the by-stander effect and is favorable in treating tumors that are insensitive to T-DM1. Thus, T-Dxd is expected to be a promising therapy to cope with HER2-positive or HER2-low-expressing tumors that do not respond to T-DM1 [bib_ref] DS-8201a, A Novel HER2-Targeting ADC With a Novel DNA Topoisomerase I Inhibitor, Ogitani [/bib_ref] [bib_ref] Bystander Killing Effect of DS-8201a, a Novel Anti-Human Epidermal Growth Factor Receptor..., Ogitani [/bib_ref]. The first evidence of T-DM1 in NSCLC from a dose-expansion phase I study suggested that T-Dxd had great potential for HER2-expressing/mutant solid tumors. In the HER2-mutant/HER2-expressing NSCLC subgroup, 10 of 18 patients (55.6%) had a confirmed objective response, with a median PFS of 11.3 months. Among the subset of HER2-mutant NSCLC patients, the confirmed ORR even reached 72.7% (8/11) and the median PFS was 11.3 months. Notably, NSCLC patients with documented HER2 mutations had more pronounced tumor shrinkage than those without mutations, regardless of IHC status. All patients experienced at least 1 AE, and 2 of 18 patients (11.1%) had serious AEs. Three patients were adjudicated as interstitial lung disease (ILD) related to the study drug, and one patient experienced an AE of respiratory failure which was associated with a fatal outcome [bib_ref] Targeting HER2 With Trastuzumab Deruxtecan: A Dose-Expansion, Phase I Study in Multiple..., Tsurutani [/bib_ref]. The DESTINY-Lung01 phase II trial investigated the efficacy of T-Dxd in NSCLC patients and consisted of two cohorts. Cohort 1 contained patients with HER2 overexpression (IHC 2+ or IHC 3+), and cohort 2 contained patients with HER2 mutation. Results from cohort 1 showed an ORR and a median PFS of 24.5% and 5.4 months. Response rates were comparable according to HER2 IHC expression levels (ORR 25.6% vs. 20.0% in IHC2+ and IHC3+ patients, respectively). All patients had at least one treatment-emergent AEs, and grade 3 AEs were reported in 73.5% of patients. There were 8 cases of drug-related ILD as adjudicated by an independent committee. Treatment-emergent AEs were associated with dose interruption in 26 patients (53.1%), dose reduction in 17 patients (34.7%), and treatment discontinuation in 11 patients (22.4%) [bib_ref] Trastuzumab Deruxtecan in HER2-Overexpressing Metastatic Non-Small Cell Lung Cancer: Interim Results of..., Nakagawa [/bib_ref]. Recently published results from cohort 2 were promising and more spectacular in comparison with cohort 1. Among 91 patients enrolled, centrally confirmed ORR was 55%, the median PFS was 8.2 months and the median OS was 17.8 months [bib_ref] Trastuzumab Deruxtecan in HER2-Mutant Non-Small-Cell Lung Cancer, Li [/bib_ref]. Preclinical findings revealed that HER2-activating mutations facilitated receptor-mediated endocytosis of the HER2-ADC complex [bib_ref] HER2-Mediated Internalization of Cytotoxic Agents in ERBB2 Amplified or Mutant Lung Cancers, Li [/bib_ref]. This may provide the mechanistic foundation for its higher efficacy in HER2-mutant NSCLC patients in contrast to the lower response rates among HER2-overexpressing patients. Regarding T-Dxd toxicity, common events included gastrointestinal and hematologic events, decreased appetite, and alopecia. Grade 3 or higher drug-related AEs occurred in 42 patients (46%). Twenty-three patients (25%) discontinued treatment because of investigator-reported, drugrelated AEs, including pneumonitis in 12 patients and ILD in 5 patients [bib_ref] Trastuzumab Deruxtecan in HER2-Mutant Non-Small-Cell Lung Cancer, Li [/bib_ref]. Given the exciting results from DESTINY-Lung01, a randomized, open-label, phase 3 trial (DESTINY-Lung04; NCT05048797) is underway to further evaluate the efficacy and safety of T-Dxd compared to standard of care (pembrolizumab combined with chemotherapy) in non-squamous NSCLC patients harboring a HER2 exon 19 or 20 mutations.
By far, ADC-based therapies seem to provide the highest response rates and best clinical outcomes among the anti-HER2 agents, both in HER2-mutant and HER2-overexpressed NSCLC patients. However, drug-related ILD is observed during the treatment and further studies should be carried out to determine which patients are at great risks and how to manage this potentially fatal AE.
## Immunotherapy
Immune checkpoint inhibitors (ICIs) demonstrated significant improvements in overall response and survival for driver-negative NSCLC patients, while their application in oncogene-addicted tumors remains to be elucidated. The presence of HER2 mutations in NSCLC is correlated with a "cold" tumor microenvironment, characterized by a relatively lower PD-L1 positive expression rate and tumor mutation burden (TMB), similar to those of EGFR-dependent tumors [bib_ref] Oncogene-Specific Differences in Tumor Mutational Burden, PD-L1 Expression, and Outcomes From Immunotherapy..., Negrao [/bib_ref] [bib_ref] Immune Microenvironment Features and Efficacy of PD-1/PD-L1 Blockade in Non-Small Cell Lung..., Chen [/bib_ref]. The efficacy of ICIs monotherapy was evaluated in several studies. A retrospective study identified 26 patients who were treated with ICIs. ORR was 12% (3/26), the median PFS and OS were 1.9 and 10.4 months, respectively. Of the three responders, none had a HER2 YVMA mutation, two had PD-L1 ≥50%, and two had TMB ≥median [bib_ref] PD-L1 Expression, Tumor Mutation Burden and Response to Immune Checkpoint Blockade in..., Lai [/bib_ref]. In the IMMUNOTARGET registry, 29 NSCLC patients with HER2 mutation received ICIs and the ORR was 7%. The median PFS was 2.5 months and was significantly associated with positive smoking status (3.4 months in smokers vs. 2 months in nonsmokers, P=0.04) [bib_ref] Immune Checkpoint Inhibitors for Patients With Advanced Lung Cancer and Oncogenic Driver..., Mazieres [/bib_ref]. The French Lung Cancer Group (GFPC) reported 6 out of 23 relapsed HER2-mutant NSCLC patients had objective responses to ICIs, with a median DOR of 15.2 months. Survival data were close to previous reports, with a median PFS of 2.2 months and a median OS of 20.4 months [bib_ref] Efficacy and Safety of Anti-PD-1 Immunotherapy in Patients With Advanced NSCLC With..., Guisier [/bib_ref]. Other studies also showed moderate or even dismal responses and survival outcomes of ICIs monotherapy against HER2 mutations [bib_ref] Oncogene-Specific Differences in Tumor Mutational Burden, PD-L1 Expression, and Outcomes From Immunotherapy..., Negrao [/bib_ref] [bib_ref] Immune Microenvironment Features and Efficacy of PD-1/PD-L1 Blockade in Non-Small Cell Lung..., Chen [/bib_ref] [bib_ref] Clinicopathologic Characteristics, Treatment Outcomes, and Acquired Resistance Patterns of Atypical EGFR Mutations..., Patil [/bib_ref].
Considering the synergistic antitumor effects of combining ICIs and chemotherapy, ICIs-based therapies were explored and been evaluated. A multicenter retrospective study enrolled 26 HER2-mutant NSCLC patients, most of which received immunochemotherapy combination regimens. The ORR, DCR and median PFS were 38.5%, 84.6% and 7.4 months, respectively [bib_ref] Treatment Efficacy of HER2-Mutant Lung Adenocarcinoma by Immune Checkpoint Inhibitors: A Multicenter..., Chu [/bib_ref]. Additionally, results from another study demonstrated that the ORR, median PFS, and one-year OS rate of ICIs combined with chemotherapy for treatment-naive HER2-mutant NSCLC were 52%, 6 months and 88%, respectively (97). Tian et al. reported similar clinical outcomes of chemo-immunotherapy in 13 stage IV HER2 ex20ins LUAD patients, with an ORR of 31% and a median PFS of 8.0 months. They also found a higher TMB, a trend toward lower clonality of tumors and a trend toward lower TCR diversity of peripheral blood in responders compared with non-responders (P=0.0067, 0.071 and 0.085, respectively). Patients with baseline TMB-high combined with mutations in DNA damage repair-related pathways or SWI/SNF complex was associated with favorable outcomes of chemo-immunotherapy combinations [bib_ref] Lung Adenocarcinoma With ERBB2 Exon 20 Insertions: Comutations and Immunogenomic Features Related..., Tian [/bib_ref].
Evidence on the efficacy of ICIs in HER2-mutant NSCLC patients remains limited and is heavily derived from retrospective results. These studies do not encourage the use of ICIs monotherapy as a therapeutic strategy in HER2-mutant NSCLC patients. Immunotherapy-based approaches could be a potential treatment option for this patient group and further clinical trials are required to confirm these results. Clinical trials of anti-HER2 agents in NSCLC patients mentioned in this review are summarized in .
# Discussion
HER2 alterations, including mutation, amplification and overexpression, have emerged as novel potential targets for anti-HER2 agents in NSCLC. Following new drug designations rapidly changing the treatment landscape of this particular subset of NSCLC patients, several matters still need to be discussed and addressed. First of all, the inconsistency in the clinical efficacy of anti-HER2 drugs among HER2 variants implies their distinct molecular entities and features. Currently, most studies evaluating anti-HER2 agents in NSCLC patients are relatively small-sampled and do not distinguish the three types of HER2 alterations. Thus, it leaves us the question of how we define the term 'HER2-positive' and which types of 'HER2-positive' lung cancers are suitable for receiving anti-HER2 targeted therapy. As mentioned above, apart from breast and gastric cancers, the patterns of HER2 staining in other cancer types have not been thoroughly investigated. This probably explains the variations of IHC overexpression rates reported in different studies. Consequently, IHC expression is not a definitive biomarker for anti-HER2 activity in NSCLC. In order to define HER2 alterations and determine their response patterns, it is crucial to advocate for standardized methods in the future to establish the definitive formulation of HER2 activating mutation, amplification and overexpression. Patient cohorts enrolled for evaluating HER2targeted drugs should also be distinguished by the particular HER2 alteration present.
Secondly, given the substantial variant efficacy and safety profiles of anti-HER2 drugs, it is imperative to investigate the pharmacodynamic and pharmacokinetic properties of these agents. Studies also reported that HER2-mutated NSCLCs are associated with central nervous system (CNS) metastases during treatment in approximately half of the patients. Therefore, frequent monitoring for the early identification of CNS involvement along with the attentive assessment of the intracranial activity of both existing and forthcoming anti-HER2 drugs are of great importance.
Thirdly, most of the current studies focused on the efficacy of single anti-HER2 agents. But preclinical studies have demonstrated the additional or synergistic effects of combining ADCs with irreversible TKIs or immunotherapy. Li et al. reported that cotreatment with irreversible pan-HER inhibitors promoted receptor ubiquitination and consequent internalization of ADC complex, resulting in improved efficacy. Switching from T-DM1 to T-Dxd, which exhibited a different cytotoxic payload, could achieve durable responses after developing resistance to T-DM1 [bib_ref] HER2-Mediated Internalization of Cytotoxic Agents in ERBB2 Amplified or Mutant Lung Cancers, Li [/bib_ref]. Moreover, T-Dxd increased tumor-infiltrating CD8+ T cells and enhanced the expression of PD-L1 and MHC-I on tumor cells. Combining T-Dxd and PD-1 inhibitors is more effective than single-agent monotherapy in mouse models. These findings provide evidence for the rationale of combination therapy in patients with HER2altered NSCLC [bib_ref] A HER2-Targeting Antibody-Drug Conjugate, Trastuzumab Deruxtecan (DS-8201a), Enhances Antitumor Immunity in a..., Iwata [/bib_ref]. Therefore, a paradigm shift from monotherapies towards combination therapies will probably change the current treatment landscape for HER2-addicted NSCLC. ADC-based therapies seem to provide the highest response rates and the best survival outcomes in HER2-altered NSCLC patients by far. Several clinical trials (NCT03334617, NCT04686305) have been launched to address this issue. Also, with more agents in the pipeline are waited to be approved, the sequencing of these novel therapies is attracting more attention.
Another issue is the not-infrequent situation of concomitant mutations in HER2-altered NSCLC patients (for example, TP53), which can impair the efficacy of anti-HER2 agents and affect prognosis. Furthermore, acquired HER2 mutation and amplification are considered to be the main mechanisms of bypass pathways driving EGFR-TKIs resistance, suggesting the expansion of TKIs-induced selection of HER2-driven cell clones. Under these circumstances, concomitant treatment of EGFR-TKIs and HER2-targeted agents could become a promising therapeutic strategy. Further investigations are needed to provide the rationale for targeting HER2 in these clinical settings.
Not long ago, HER2 alterations were considered poor targets in NSCLC and were deemed to have inferior clinical outcomes than those with other genetic alterations. Nowadays, owing to advances in preclinical biology-based drug development, the wide array of the investigating HER2-targeted agents is a glimmer of hope shooting through the past prejudice. With emerging therapies requiring further clinical validation, exploration of targeting HER2 in NSCLC is underway.
[table] TABLE 1 |: Clinical trials of anti-HER2 agents in NSCLC patients. N, number; ORR, objective response rate; PFS, progression-free survival; OS, overall survival; CT, chemotherapy; IHC, immunohistochemistry; ECD, extracellular domain; DISH, dual color in situ hybridization; FISH, fluorescence in situ hybridization; NA, not available; NE, not estimable; NR, not reached. a Tumor response data were available for 16 patients. b This phase II study enrolled 30 patients, 22 had EGFR 20 exon mutations and 8 had HER2 mutations. c PFS and OS data were evaluated based on the whole cohort (n=30). [/table]
[table] Table 1: Promising strategies to cope with HER2-altered NSCLC and their underlying mechanisms are shown in [/table]
|
Effects of prenatal nutrient supplementation and early life exposures on neurodevelopment at age 10: a randomised controlled trial - the COPSYCH study protocol
# Introduction
Mounting evidence suggests that health and disease are programmed in early life, and that a wide range of prenatal and early life exposures affect brain development. [bib_ref] The influence of maternal prenatal and early childhood nutrition and maternal prenatal..., Marques [/bib_ref] Early
## Strengths and limitations of this study
► The Copenhagen Prospective Studies on Asthma in Childhood 2010 cohort consists of 700 children and is a randomised, placebo-controlled trial on effects of prenatal nutrient supplementation on prenatal n-3 long-chain polyunsaturated fatty acids and high-dose vitamin D supplementation on child development. ► The COpenhagen Prospective Study on Neuro-PSYCHiatric Development (COPSYCH) study comprises an extensive 2-day 10-year clinical visit and focuses on brain development, neurocognition and psychopathology. ► Deep phenotyping is available from the previous 14 clinical visits, and a limited selection of neurodevelopmental measures have been continuously assessed. ► The randomised controlled trial design allows for unique, causal inferences on effects of prenatal nutrient supplementation on neurodevelopment at age 10. ► The planned COPSYCH study would be strengthened by longitudinal neurodevelopmental data, thus future similar evaluations of the cohort will be applied for.
Open access life may therefore represent a window of opportunity for promoting a healthy neurodevelopment. Nevertheless, the evidence regarding early life exposures and their effects on neurodevelopment are limited by observational study designs. Neurodevelopmental disorders, for example, autism spectrum disorder, attention-deficit/ hyperactivity disorder (ADHD) and intellectual disability are all highly heritable with complex, polygenetic contributions [bib_ref] Group of the Psychiatric Genomics Consortium. Identification of risk loci with shared..., Cross-Disorder [/bib_ref] [bib_ref] A review of the evidence from family, twin and adoption studies for..., Shih [/bib_ref] [bib_ref] Neurodevelopmental hypothesis of schizophrenia, Owen [/bib_ref] [bib_ref] The Kraepelinian dichotomy -going, going… but still not gone, Craddock [/bib_ref] and have all been associated with brain structural and functional aberrations and neurocognitive impairment. Prenatal exposures in the form of supply of micronutrients, like folic acid, are shown to be pivotal to decrease risk of congenital neural tube defects, including anencephaly, spina bifida and encephalocele [bib_ref] Effects and safety of periconceptional folate supplementation for preventing birth defects, De-Regil [/bib_ref]. Our previous study on the Copenhagen Prospective Studies on Asthma in Childhood 2010(COPSAC 2010 ) cohort showed that boys born to mother receiving n-3 long-chain polyunsaturated fatty acids (n-3 LCPUFA) supplementation had improved motor development, improved global neurocognitive function at age 2½ years of age as assessed with the neurocognitive part of Bayley Scales of Infant and Toddler development, third edition [bib_ref] Validity and crosscultural differences of the Bayley scales of infant and toddler..., Hoskens [/bib_ref] and reduced functional impact of emotional and behavioural problems by age 6 years according to parents' replies to the Strengths and Difficulties Questionnaire (SDQ). [bib_ref] Psychometric properties of the strengths and difficulties questionnaire, Goodman [/bib_ref] [bib_ref] The extended version of the strengths and difficulties questionnaire as a guide..., Goodman [/bib_ref] [bib_ref] Strengths and difficulties questionnaire (SDQ-Dan). A new instrument for psychopathologic screening of..., Obel [/bib_ref] [bib_ref] Fish oil supplementation in pregnancy and neurodevelopment in Childhood-A randomized clinical trial, Sass [/bib_ref] No effect on neurodevelopment in the first 6 years of life was observed in the offspring born to highdose vitamin D3 supplemented mothers. [bib_ref] High-Dose vitamin D supplementation in pregnancy and neurodevelopment in childhood: a Prespecified..., Sass [/bib_ref] The risk for manifestation of psychiatric disorders later in life is associated with environmental insults, like maternal infections during pregnancy, infections during childhood and adolescence [bib_ref] Hospital contacts with infection and risk of schizophrenia: a population-based cohort study..., Nielsen [/bib_ref] [bib_ref] Infections and exposure to antiinfective agents and the risk of severe mental..., Köhler [/bib_ref] [bib_ref] Autoimmune diseases and severe infections as risk factors for schizophrenia: a 30-year..., Benros [/bib_ref] [bib_ref] A nationwide study in Denmark of the association between treated infections and..., Köhler-Forsberg [/bib_ref] or ongoing inflammatory processes like asthma. [bib_ref] Association of atopic diseases and attention-deficit/hyperactivity disorder: a systematic review and meta-analyses, Jvander [/bib_ref] [bib_ref] Schizophrenia in patients with atopic disorders with particular emphasis on asthma: a..., Pedersen [/bib_ref] [bib_ref] Inflammatory processes in schizophrenia: a promising neuroimmunological target for the treatment of..., Meyer [/bib_ref] [bib_ref] Association between childhood asthma and attention deficit hyperactivity or autism spectrum disorders:..., Kaas [/bib_ref] Furthermore, emerging evidence has indicated that n-3 LCPUFA and vitamin D are important for brain development in particular maturation of cerebral white matter myelin and cognition. [bib_ref] Polyunsaturated fatty acid biostatus, phospholipase A2 activity and brain white matter microstructure..., Mcnamara [/bib_ref] [bib_ref] Vitamin D3-implications for brain development, Mcgrath [/bib_ref] [bib_ref] Maternal omega 3 fatty acid supplementation during pregnancy to a micronutrient-imbalanced diet..., Sable [/bib_ref] [bib_ref] White matter and cognition: making the connection, Filley [/bib_ref] The beneficial effects of maternal n-3 LCPUFA supplementation on neurocognitive development are not consistent in previous randomised controlled trials (RCTs), and the interpretation is complicated by methodological issues. [bib_ref] The effect of maternal omega-3 (n-3) LCPUFA supplementation during pregnancy on early..., Gould [/bib_ref] Still, there is some evidence that n-3 LCPUFA supplement may reduce the prevalence rate of children with lower neurocognitive scores indicative of delayed neurocognitive development at 18 months of age. [bib_ref] Effect of DHA supplementation during pregnancy on maternal depression and neurodevelopment of..., Makrides [/bib_ref] This may be hypothesised to be due to downregulation of the pro-inflammatory n-6 LCPUFA-derived eicosanoids rather than the more anti-inflammatory eicosanoids since n-3 LCPUFA and n-6 LCPUFA competes for the same enzymes. [bib_ref] Evidence regarding an effect of marine n-3 fatty acids on preterm birth:..., Salvig [/bib_ref] However, to the best of our knowledge, no RCT of prenatal supplementation with n-3 LCPUFA and high-dose vitamin D in pregnancy has examined the neurocognitive outcome in late childhood.
The composition of the gut microbiome matures within the first years of life and the microbiome may have the ability to affect host inflammatory status at a time, when significant brain development takes place. Perturbation of the microbial and viral homeostasis during this critical period of development may affect the signalling pathways between gut and brain (immune, neural and endocrine pathways, the so-called 'gut-brain axis'). [bib_ref] The microbiota-gut-brain axis, Cryan [/bib_ref] Thus, the microbiome may be an intermediary player in the interaction between the host and its environment in the mechanisms that determine the transition from health to disease. In the human gut viruses are at least as abundant as bacteria. The viral community (virome) mostly target bacteria (bacteriophages) while only a fraction directly affect humans. Their role in disease development is poorly understood, but may be detrimental in the transition from health to disease in early life, either through manipulation of the bacterial community or through direct virome-host effects. [bib_ref] Ménage à trois in the human gut: interactions between host, bacteria and..., Mirzaei [/bib_ref] [bib_ref] Intestinal virome changes precede autoimmunity in type I diabetes-susceptible children, Zhao [/bib_ref] [bib_ref] Dynamic modulation of the gut microbiota and metabolome by bacteriophages in a..., Hsu [/bib_ref] Most of the hypotheses regarding early risk factors are based on studies reporting associations between early life exposures and a later outcome, which could be confounded by other exposures and cannot infer causality. [bib_ref] Applying the Bradford Hill criteria in the 21st century: how data integration..., Fedak [/bib_ref] The factorial, randomised controlled design of the current study provide the unique opportunity to unravel causal effects of prenatal n-3 LCPUFA and vitamin D supplementation on neuropsychiatric and neurocognitive status as well as neuroimaging, all reflecting neurodevelopment at age 10. Furthermore, we have extensive information on a broad range of environmental exposures, which were collected through personal interviews with the families and when possible validated against register data, and assessed objectively and longitudinally, including repeated assessment of the airway and gut microbiome from birth throughout childhood.
The main objective of COpenhagen Prospective Study on Neuro-PSYCHiatric Development (COPSYCH) is characterising the neurocognitive, psychopathological and brain structural and functional outcomes, and relating these to the n-3 LCPUFA and vitamin D interventions and early life exposures. Data are collected at a comprehensive, ongoing 2-day 10-year follow-up visit of the COPSAC 2010 cohort. Comprehensive data on early exposures have been meticulously collected at previous follow-up visits of the COPSAC 2010 cohort. The ambitious goal is to provide the basis for preventive interventions in prenatal or early life in order to improve neurocognitive development and prevent development of psychiatric disorders.
# Methods and analysis
COPSYCH is a 5-year translational research alliance between COPSAC and Center for Clinical Intervention and Neuropsychiatric Schizophrenia Research (CINS) initiated in January 2019.
First clinical examination took place on 11 February 2019. Expected end of clinical examinations is December 1st 2021, followed by data processing and analysis. The COPSAC 2010 cohort and previous assessments are briefly summarised below and detailed previously. [bib_ref] Deep phenotyping of the unselected COPSAC2010 birth cohort study, Bisgaard [/bib_ref] Open access
## Patient and public involvement
Patients or the public were not involved in the design, or conduct, or reporting, or dissemination plans of our research.
## The copsych 10-year visit
The ongoing 10-year COPSYCH visit extends over 2 days. Day 1 is a 6-8 hour visit at the COPSAC research clinic. The first part consists of a neurocognitive assessment (table 1) and the second part is a neuropsychiatric assessment, where categorical and dimensional psychopathology is assessed using a diagnostic interview and parental and teacher questionnaire ratings [fig_ref] Table 2: Psychopathological interviews and questionnaires [/fig_ref].
## Neurocognition
The neurocognitive test battery is based on selected subtests from several neuropsychological tests. The Cambridge Neuropsychological Test Automated Battery (CANTAB) [bib_ref] Computerized assessment in neuropsychiatry using Cantab: discussion paper, Sahakian [/bib_ref] is a neuropsychological test battery developed from a neuroscience approach. The CANTAB battery has been validated for use between the ages 4-80 years [bib_ref] Normative data from the CANTAB. I: development of executive function over the..., De Luca [/bib_ref] and has been extensively used in, for example, lesion studies as well as neuropsychiatric and neurological populations. The CANTAB tests have been used in clinical cohorts of both children [bib_ref] Assessment of neurocognitive functions in 7-year-old children at familial high risk for..., Hemager [/bib_ref] and adults with neurodevelopmental disorders including schizophrenia and ADHD, and healthy controls [bib_ref] Assessment of neuropsychological function through use of the Cambridge neuropsychological testing automated..., Luciana [/bib_ref]. To assess specific neurocognitive ¶The Behaviour Rating Inventory of Executive Function-second edition. **A prime measures target sensitivity. † † Extra-dimensional stage errors denotes the number of errors made in the extra-dimensional stage of the task, where the child is required to make an extradimensional shift. ‡ ‡Span length denotes the longest sequence successfully recalled by the child (three attempts at each level). § § Total errors comprise the number of times a box is selected that do not have hidden target that must be found and therefore should not have been visited by the child. BRIEF-2, Behaviour Rating Inventory of Executive Function, Second Edition; CANTAB, Cambridge Neuropsychological Test Automated Battery; D-KEFS, Delis-Kaplan Executive Function System; hs-CRP, High-sensitivity C-Reactive Protein; TNFα, Tumor Necrosis Factor α; TOMAL-2, Test of Memory and Learning, Second Edition; WISC-IV, Wechsler Intelligence Scale for Children.
## Open access
functions across the selected neurocognitive domains from CANTAB, we have included the following: rapid visual information processing to assess sustained attention; reaction time task to assess both simple and fivechoice reaction time and movement time; and paired associates learning to assess visual memory. Executive functions are assessed using the intra-extra dimensional set shifting task examining mental flexibility; the stockings of cambridge examining planning, and the spatial span and spatial working memory tasks, respectively, testing working memory span, and spatial working memory manipulation and strategy formation.
Two subtests (vocabulary and matrices) from the Wechsler Intelligence Scale for Children, fourth edition (WISC-IV)are used to estimate intelligence; additionally other subtests are used to characterise verbal working memory (digit span and letter-number sequencing) and processing speed (coding and symbol search). From the Delis-Kaplan Executive Function System, trail making 2 is applied as an additional measure of processing speed, and the trail making 4 to measure flexibility/set shift. Moreover, selected subtests (Word Selective Reminding and Object Recall) from Test of Memory and Learning-Second Edition are used for evaluating verbal memory (2011). Fine motor dexterity is assessed with Grooved Pegboard and Reaction Time in CANTAB. [bib_ref] Updated developmental norms for fine motor functions as measured by finger tapping..., Skogan [/bib_ref] The parent and teacher fill out Behaviour Rating Inventory of Executive Function, Second Edition (see table 1) 47 measuring daily life executive functions in the home environment and school environment, respectively.
All the neurocognitive tests are well validated, reliable and appropriate for this age group. [bib_ref] Normative data from the CANTAB. I: development of executive function over the..., De Luca [/bib_ref] The time for completing the full test battery is approximately 2 hours, but the duration is individualised with breaks when needed. Tests are administered in a fixed order.
## Categorical psychopathology
The diagnostic interview Kiddie-Schedule for Affective Disorders and Schizophrenia for School-Age Children-Present and Lifetime Version (K-SADS-PL) [bib_ref] Schedule for affective disorders and schizophrenia for school-age Children-Present and lifetime version..., Kaufman [/bib_ref] is used to provide information on current and lifetime psychopathology. K-SADS-PL is a semi-structured clinical diagnostic interview containing the same questions for parents and children (except in the Autism Spectrum Supplement that contains extra questions for the parents only). According to the K-SADS-PL guidelines the interview is first completed with a parent/caregiver and afterwards with the child younger than 11 years of age. The K-SADS-PL consists of a general screening interview. If the threshold of a symptom has been reached in the general screening interview, K-SADS-PL contains additional questions in supplements, which target specific disorders. K-SADS-PL Schizophrenia Spectrum and Other Psychotic Disorders Supplement is applied whether or not any psychotic symptoms appear in the screening interview to detect possible psychotic-like experiences (PLE). [bib_ref] The CCC2000 birth cohort study of register-based family history of mental disorders..., Jeppesen [/bib_ref] To obtain information on potential current and earlier use of psychotropic medicine, questions from the Schedules for Clinical Assessment in Neuropsychiatry 52 are reworded to interview about prescribed psychotropic medication to the child (antidepressants, mood stabilisers, anxiolytics, hypnotics, antipsychotics, central stimulants).
Children's global assessment of functioning 53 (included in K-SADS-PL), is used for assessment of the child's global level of functioning.
Research diagnoses are made according to both International Classification of Diseases 10th Revision(ICD-10)and Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5).Diagnoses are based on information from all available sources (clinical observations during the neuropsychological testing and psychopathological assessments of each participating child). All K-SADS-PL interviews are video recorded. The videos are used for supervision and to reach consensus on diagnostics. Neurocognitive assessments and clinical interviews are supervised by a senior researcher, psychologist with specialty in child and adolescent psychiatry (JRMJ) and consensus diagnoses are made weekly (between JRMJ and Open access at least two examiners). All diagnoses are furthermore re-evaluated at a monthly conference with a professor of child and adolescent psychiatry (NB). Additionally, cases with an inconclusive diagnosis are supervised by NB. Ratings of The Test Observation Form (TOF), the verbal subtest from WISC-IV (vocabulary) and Children's Global Assessment Scale are also based on consensus rating with JRMJ. Inter-rater reliability will be estimated based on 10 K-SADS-PL screening interview videos recorded (by JRMJ) during the first half data collection period. Each 20th child that is not assigned a diagnosis will undergo re-evaluation with NB. The four interviewers have supervision on a weekly basis.
## Dimensional psychopathology
In order to also quantify psychopathology, which may or may not meet the threshold for a psychiatric diagnosis, dimensional assessments of psychiatric symptoms and behavioural problems are undertaken on all children. For a detailed registration of inattentive and impulsive-hyperactive behaviours in different situations, the ADHD-Rating Scale (ADHD-RS) questionnaire is administered (filled out by parents and teachers, respectively). [bib_ref] Psychometric properties of the parent and teacher ADHD rating scale (ADHD-RS): measurement..., Makransky [/bib_ref] Similarly, the Social Responsiveness Scale-2 is used for parental ratings of the severity of autism spectrum disorder traits. In terms of dimensions of child and adolescent psychopathology, the parents report their evaluation in the Child Behaviour Checklist (CBCL) ; The teacher's report form (TRF) is completed by the child's teacher. In most aspects the TRF corresponds to the CBCL as most of its items have counterparts in the CBCL. SDQ is a multiinformant screening tool for emotional and behavioural problems in children and adolescents, answered by parents and teachers. [bib_ref] Psychometric properties of the Danish strength and difficulties questionnaire: the SDQ assessed..., Niclasen [/bib_ref] TOF will be used to score severity of behavioural and emotional problems observed during the neurocognitive test session. It consists of 125 items, scored on a 4-point Likert scale. TOF is completed by the examiner immediately after the neurocognitive testing. 62
## Mri (day 2)
Day 2 is a 1-hour visit [fig_ref] Table 2: Psychopathological interviews and questionnaires [/fig_ref]. MRI is a non-invasive medical scanning technique, which can acquire images of the brain without application of radioactive ligands or X-rays. MRI scans are acquired on a 3T Philips Achieva scanner (Philips Healthcare, Best, the Netherlands). Both brain structural and functional sequences are acquired [fig_ref] Table 3: MRI sequences [/fig_ref]. The total duration of MRI acquisition is 35 min, which including breaks effectively corresponds to 40-45 min of scan time.
All children undergo a virtual MRI 'practice scanning' using a pair of three-dimensional (3D) goggles on examination day 1. In order to minimise movement and impatience, all children are given the opportunity to watch an animation movie while inside the MRI scanner, using a mirror and video projection. We extract head motion parameters using the Tracoline system, which is an advanced 3D stereo vision system to monitor and capture head movements using 3D surfaces from structured invisible infrared light, enabling markerless tracking of the subject's head in a clinical MRI setup (https:// tracinnovations. com/ markerless-technology/). These motion parameters are used to retrospectively correct for scanner head motion.
COPSAC 2010 cohort COPSYCH is based on the COPSAC 2010 cohort consisting of 700 mother-child pairs followed prospectively from pregnancy week 24. A population-based sample of 736 pregnant women were recruited during 2008-2010 and participated in a double blinded, randomised and placebo controlled study of fish oil, n-3 LCPUFA (2.4 g/ day n-3 LCPUFA) and/or high-dose vitamin D (2400 IU/ day) supplementation during the third trimester with the primary outcome of persistent wheeze/asthma until age 3, where the trials were unblinded ( ClinicalTrials. gov: NCT00798226 and NCT00856947). Data validation and quality control followed the guidelines of Good Clinical Practice. Exclusion criteria were chronic cardiac, endocrinological, nephrological or lung disease other than asthma or a daily vitamin D intake above 600 IU.
Neurodevelopmental outcomes were prespecified secondary outcomes specified in the primary study protocol (including milestones, language and The Bayley Scales of Infant and Toddler Development) The parents were informed that they participated in a cohort study lasting at least 3 years, and that we planned to follow the children forward if we succeeded in financing. The clinical visits after 3 years have been planned accordingly for the purpose of gaining as much knowledge in the prespecified fields as possible, and to replicate and continue further research of the previous findings
The mothers and their children have been closely followed prospectively at the COPSAC research unit with 14 completed clinical visits scheduled from pregnancy week 24 until 8 years and with day-to-day diaries from birth capturing all symptoms related to airways, skin and infectious episodes the first 3 years between visits Open access (figure 1). Retention rate for the previous visit at age 8 years was 92%. The ongoing neurocognitive and psychopathological 10-year examinations are undertaken by trained medical doctors and experienced nurses, who will remain blinded to the intervention status of the children throughout the data collection and statistical analyses.
## Previous visits and examinations
At all previous 14 clinical visits, the parents were interviewed about a variety of socioeconomic factors such as smoking during pregnancy, family size, household pets and so on. Furthermore, information has been collected on delivery mode, Apgar scores, maternal diet during pregnancy 4 weeks prior to randomisation, breastfeeding length, fatty acid composition of breast milk and introduction of solid foods. The child's daily activity level and sleep patterns were also measured at 2 years and 10 years of age by accelerometry (wristband that measures daily physical activity, omnidirectional Actical accelerometer, Philips Respironics, Murrysville, Pennsylvania, USA). [bib_ref] Objective assessment of levels and patterns of physical activity in preschool children, Brasholt [/bib_ref] Metabolomics Global metabolomics profiling of blood was done in the mothers at 24 weeks gestation and 1-week post partum and in the children at age 0.5, 1.5 and 6 years. [bib_ref] Fish-Oil supplementation in pregnancy, child metabolomics and asthma risk, Rago [/bib_ref] Further, dry blood spot metabolomics profiling was done in the newborn child and urine metabolomics by age 1 month. [bib_ref] Neonatal urine metabolic profiling and development of childhood asthma, Chawes [/bib_ref] These metabolomic assessments allow discovery of yet unknown mechanisms affecting aspects of neurodevelopment.
## Genetics
Genomewide scan has been conducted on the cohort using the Illumina HumanOmniExpressExome-8 V1.2 BeadChip, including genetic variants of importance for n-3 LCPUFA and vitamin D levels and metabolism, which will be included in the analyses of the intervention. This data will allow us to analyse to which extent the effect of early exposures are modified by host genetics.
## Microbiome
Longitudinal gut microbiome composition profiling has been performed in the cohort using 16S rRNA gene amplicon sequencing 66 at 1 week, 1 month, 1 year, 4 and 6 years and metagenome at 1 year. These early samples together with samples collected at 8 and 10 years will allow us to characterise the microbial development until age 10 and how this may impact components of neurodevelopment.
## Inflammation in mothers and children
We measured systemic inflammatory surrogate markers including high-sensitivity C-Reactive Protein (hs-CRP), interleukin (IL)-6, Tumor Necrosis Factor alpha (TNFα), IL-1β and CXC Chemokine Ligand 8 (CXCL-8). Furthermore we assessed a panel of immune mediators in the upper airways related to specific immune signalling pathways in unstimulated upper airway mucosal lining fluid at 1 month, 2 years and 6 years of age and at episodes with acute respiratory symptoms. Finally, at 18 months a deep immunological phenotyping was done with an extensive profiling of immune cell distribution and function including detailed phenotyping of 25 immune cell populations and 8 ligand simulations in freshly collected whole blood (figure 1). [bib_ref] Distinct immune phenotypes in infants developing asthma during childhood, Thysen [/bib_ref] Infections and prenatal information Mothers were interviewed regarding infectious symptoms and use of medication including all courses of antibiotics in pregnancy at the scheduled visits preintervention and post-intervention. The first 3 years of the child's life the families kept diary cards monitoring signs of infection categorised as common cold, pneumonia, pharyngitis, otitis, fever, gastrointestinal infection and absence from daycare because of illness. The COPSAC physician validated the diary cards entries with the parents at each visit and interviewed them regarding other types of infections. Information was checked against hospital and medication prescription registries. After age 3, data on hospital admissions and information about prescribed medication were gathered at the scheduled visits. Diagnosed preeclampsia, gestational diabetes and pre-pregnancy weight were registered. These data allow Diagnosis of asthma, allergy and eczema Asthma/persistent wheeze was diagnosed according to a validated symptom-based quantitative algorithm (five episodes of troublesome lung symptoms, typical asthma symptoms, symptom relief when using inhaled shortacting B 2 -agonists and response to a 3-month trial of inhaled corticosteroid). Lung function was measured with whole-body plethysmography and spirometry, and Open access bronchial responsiveness was evaluated by methacholine challenge. Allergen specific IgE levels along with skin prick tests have been measured for the most common inhalant and food allergens and allergic rhinitis was diagnosed prospectively. Eczema diagnosis was based on the Hanifin-Rajka criteria [bib_ref] Diagnostic features of atopic dermatitis, Hanifin [/bib_ref] and the SCORing Atopic Dermatitis (SCORAD) index was used to measure severity during flares (severity scoring of atopic dermatitis) (figure 2).
## Previous neurodevelopmental and cognitive assessments
The children's neurodevelopmental status [bib_ref] Determinants of neurodevelopment in early childhood -results from the Copenhagen prospective studies..., Bjarnadóttir [/bib_ref] has been examined throughout the years. Motor milestone achievements were continuously evaluated during the first years of life by the parents using a registration form based on the Denver Developmental Index, [bib_ref] The Denver II: a major revision and restandardization of the Denver developmental..., Frankenburg [/bib_ref] and WHO milestones. Language development was evaluated with MacArthur Bates Communicative Developmental Inventory at age 1 and 2 years. [bib_ref] The Danish communicative developmental inventories: validity and main developmental trends, Bleses [/bib_ref] By the age of 2½ years global neurocognitive functioning was examined using the neurocognitive part of Bayley Scales of Infant and Toddler development, third edition. At age 3 years general neurodevelopment was assessed by the Ages and Stages Questionnaire.At ages 6 and 8 years, parents were asked to fill out a SDQ 8-10 focusing on emotional and behavioural problems. Moreover, at age 8 years, parents were also asked to fill out the ADHD-RS which is a questionnaire merely focused on ADHD symptoms and oppositional defiant disorder symptoms (figure 3).
## Power calculation
At the last visit at 8 years, the retention rate was 92%, and we anticipate a similar participation rate at the 10-year visit. With a total sample size of N=644 and an alpha level of 0.05, t and a SD of 1.00, this study has a power of 80% to detect an effect size of 0.22 SD. Based on the SD of the psychopathological and global functioning raw scores reported in a large, recent Danish sample of population-based control children 7 years of age, 40 the effect size of 0.22 SD is equivalent to 3.3 raw scores on the child behaviour checklist total problem scale, and 3.0 raw scores on the CGAS, respectively. We assume that the raw score distributions observed in these children with a mean age of 7.8 years are equivalent for 10-year-old children. Finally, this effect size is equivalent to 3.3 scores on the Processing Speed Index and the Working Memory Index using the published Danish WISC-IV norms.Risks, side effects and potential benefits There are no known risks of simple MRI scanning. Discomfort during the actual scanning will be minimised with earplugs and hearing protectors.
Children with psychiatric symptoms who have not been evaluated prior to the visit may likely benefit from the thorough COPSYCH evaluation. Early identification of mental challenges or detection of a psychiatric diagnosis is crucial in order to optimise, for example, the child's schooling performance. [bib_ref] Association of mental disorder in childhood and adolescence with subsequent educational achievement, Dalsgaard [/bib_ref] Children who have psychiatric disorders are referred directly to child and adolescent mental health services (after the parents give their permission and consent).
## Main outcomes neurocognition
Neurocognitive functioning reflecting intelligence and the following seven neurocognitive domains: speed of processing, attention/vigilance, psychomotor speed, verbal memory, visual memory, executive functions and motor functioning [fig_ref] Table 1: Neurocognitive test battery [/fig_ref].
## Categorical psychopathology
Categorical psychopathology, that is, child and adolescent disorders according to ICD 10 54 and DSM 5.Dimensional psychopathology Dimensional psychopathology scores will be based on ratings in all psychopathological questionnaires (total scores and subscales scores (table 2) and on the number of mental and behavioural symptoms reported in the Open access K-SADS-PL. [bib_ref] Schedule for affective disorders and schizophrenia for school-age Children-Present and lifetime version..., Kaufman [/bib_ref] Severity is thus reflected in a quantitative composite index of psychiatric symptoms which is the number of symptoms endorsed by a severity score of 'three' in the complete K-SADS-PL interview, that is, total number of symptoms at/over the threshold. Present symptoms will be counted separately from previous symptoms. Additionally, the number of symptoms within different subdomains will be calculated, for example, affective and anxiety disorders; neurodevelopmental and behavioural disorders. Counting symptoms does not interfere with the standard administration of the K-SADS-PL interview.
Using this method we will quantify the symptom load of each child since a dimensional approach may be more sensitive than the categorical diagnostic approach in our context. A similar approach has been applied here where they use symptom composite measure from the K-SADS. [bib_ref] Mediating links between maternal depression and offspring psychopathology: the importance of independent..., Burt [/bib_ref] The symptom count method will include all children irrespective of their diagnostic status.
## Ple
We will examine if the children have experienced PLE. PLE will be measured by semi structured interviews using the K-SADS-PL-items on psychotic symptoms (Schizophrenia Spectrum and Other Psychotic Disorders Supplement), each symptom scored as 'not present' (1) 'likely present' [bib_ref] Group of the Psychiatric Genomics Consortium. Identification of risk loci with shared..., Cross-Disorder [/bib_ref] or 'definitely present' (3). A similar method was used in a previous Danish Copenhagen Child Cohort study (CCC2000). [bib_ref] Psychotic experiences from preadolescence to adolescence: when should we be worried about..., Rimvall [/bib_ref] A score of 2 is reflective of a PLE. PLE will be consensus rated with a medical doctor (MR) with extensive experience from a prior similar study. [bib_ref] Psychotic experiences from preadolescence to adolescence: when should we be worried about..., Rimvall [/bib_ref] Brain structure and function Structural MRI measures of grey and white matter will be used to disentangle the putative link between early inflammation, psychosocial functioning, neurocognition and psychopathology. We will use conventional software packages to obtain measures of cortical thickness and surface area, curvature and volumes of specific grey matter brain structures (eg, FreeSurfer (http://surfer.nmr.mgh. harvard.edu/): Diffusion data we will used to assess white matter integrity using, for example, Tract-based Spatial Statistics (TBSS) (https://fsl.fmrib.ox.ac.uk/fsl/fslwiki/ TBSS). Moreover we will use the functional data to assess cerebral blood flow and the spectroscopy data will provide novel data on cerebral lactate metabolism. [bib_ref] Higher physiological vulnerability to hypoxic exposure with advancing age in the human..., Vestergaard [/bib_ref]
## Ethics and dissemination
This study is conducted in accordance with the guiding principles of the Declaration of Helsinki and was approved by the Local Ethics Committee: H-B-2008-093 and the Danish Data Protection Agency 2015-41-3696. All families will receive written information about the study prior to participation and both parents will give written informed consent before enrolment.
Acquired data are stored in the secure Research Electronic Data Capture system provided by the Capital Region of Copenhagen. Direct electronic typing of as much data as possible is enforced, and double data entry by independent staff ensures validity of data.
## Perspectives
This is the largest study of its kind, with a broad range of exposures, which were validated by personal interviews with the families. Identifying the possible linkage between neurocognitive impairments, psychopathology and early life exposures may become a game changer in future prevention of psychiatric disorders and neurocognitive impairments, for example, in terms of change of recommended guidelines of micronutrient and vitamin supplementation or other exposures during pregnancy. Knowledge delineating the mechanisms involved in linking early life exposure to aberrant neurodevelopment will potentially enable a more targeted therapeutic approach for the prevention or amelioration of structural and functional brain and neurocognitive developmental abnormalities, and the onset of psychopathology. Thereby, the results of this research may have direct, easily accessible and essential impact on children's brain health.
Positive, negative and inconclusive results will be presented at national and international conferences. Papers will be submitted to peer-reviewed journals.
Author affiliations 1
[fig] Figure 1: Overview of inflammation and infection markers in COPSAC2010 (mothers) -Modified from: http://copsac.com/ home/copsac-cohorts/copsac2010-clinic/) [/fig]
[fig] Figure 2, Figure 3: Overview of inflammation and infection markers in COPSAC2010 (children) -Modified from: http://copsac.com/ home/copsac-cohorts/copsac2010-clinic/) Neurodevelopmental outcomes in COPSAC 2010 . ADHD-RS, Attention Deficit/Hyperactivity Disorder-Rating Scale; ASQ, Ages and Stages Questionnaire; COPSAC 2010 , Copenhagen Prospective Studies on Asthma in Childhood 2010; COPSYCH, COpenhagen Prospective Study on Neuro-PSYCHiatric Development; SDQ, Strengths and Difficulties Questionnaire. [/fig]
[table] Table 1: Neurocognitive test battery [/table]
[table] Table 2: Psychopathological interviews and questionnaires [/table]
[table] Table 3: MRI sequences [/table]
|
Individual and partnership characteristics associated with consistent condom use in a cohort of cisgender men who have sex with men and transgender women in Nigeria
Background: This study reports on the individual and partnership characteristics that influence consistent condom use in cisgender men who have sex with men (MSM) and transgender women (TGW) attending trusted community centers that provide HIV prevention and treatment services in Nigeria. Methods: Adults assigned male at birth who reported anal sex with male partners who enrolled between March 2013-2019 and had information about at least one male sexual partner were included in these analyses. At enrollment and follow-up visits every 3 months for up to 18 months, participants were administered detailed questionnaires that collected information about demographics, sexual practices, HIV risk behaviors, and characteristics and behaviors of their partners in the previous year (at enrollment) or the preceding 3 to 6-months (at follow-up visits). Logistic regression models with generalized estimating equations were used to assess the odds ratio (OR) and 95% confidence intervals (CI) of individual, partner, and partnership characteristics associated with consistent condom use (CCU). A participant was defined as consistently using condom if they reported always using condoms all the time they had insertive, receptive or both types of anal sex with a male partner. Results: At the individual level, CCU was positively associated with higher education, disclosure of key population status to a healthcare worker and negatively associated with poor access to condoms. At the partner and partnership level, CCU was associated with partners with higher education (aOR: 1.36; 95% CI: 1.07-1.72), casual relationships (aOR: 1.22; 95% CI: 1.11-1.34) and relationships in which partners encouraged the participant to use condoms with other partners (aOR: 1.14; 95% CI: 1.02-1.28). Relationships in which the partner was married to a woman and/or the partner's HIV status positive or unknown were negatively associated with CCU.
Conclusions: These findings suggest that individuals in relationships where partners were more open and encouraged safer sex were more likely to consistently use condoms. HIV prevention programs should consider leveraging communication to sexual partners to encourage condom use as this may support condom use with other sexual partners. Given sustained and growing HIV and STI epidemics among MSM and TGW, even with preexposure prophylaxis scale-up, it is crucial to continue to study optimal implementation strategies to increase condom use.
Keywords: Men who have sex with men, Nigeria, condoms, Networks, HIV, Epidemiology, Sexual and gender minorities Background Even in the context of a mixed HIV epidemic in Nigeria, cisgender men who have sex with men (MSM) and transgender women (TGW) bear a disproportionate burden of infection with estimated HIV prevalence 18-66% compared to about 4% among reproductive aged adults. This increased burden is due to the high HIV transmission risks associated with condomless anal intercourse with viremic partners. It has been estimated that the 5-year cumulative incidence of HIV among MSM would be reduced by 80-98% if the efficiency of HIV transmission in anal sex was similar to the efficiency of transmission in condomless vaginal sex. While the use of pre-exposure prophylaxis (PrEP) is effective in reducing transmission of HIV, it is not routinely available in many sub-Saharan African (SSA) countries. Thus, consistent condom use (CCU) remains a cost-effective HIV prevention strategy writ large.
Individual-level determinants of CCU among MSM in Nigeria have been shown to include knowledge of HIV transmission, frequency of sex, and alcohol or drug use before or during sex. However, factors influencing condom use may not be fully captured by a single measure and understanding the dynamics between the attributes of the sexual relationship, partner characteristics, and individual level factors is necessary to fully characterize the determinants of condom usage and opportunities for intervention.
Among MSM, research has consistently shown unprotected anal intercourse is more likely in committed or more serious relationships compared to relationships that are more casual. This dynamic has been attributed to higher levels of trust and familiarity with committed partners and the perception that condom use may affect intimacy. Perceived and actual partner characteristics have also been shown to influence individual decisions to use condom. For example, MSM at risk for HIV may be more likely to use condom with partners they perceive to be more likely to be living with HIV and serosorting is common among MSM when both partners have disclosed their HIV statuses. Other factors such as age difference between partners, sexual partner concurrency, and partner norms around condoms use may also impact CCU.
Although TGW have been shown to have higher odds of reporting condomless receptive anal sex with male partners compared to MSM, few studies have characterized the factors that drive this association in TGW specifically. A study among TGW in China found that subjective norms, condom use attitude, condom use skills, and self-efficacy were associated with frequent condom use and knowledge of HIV was not related to condom use . Other factors that have been shown to be associated with condomless sex among TGW include feminizing medical intervention, factors related to sex-work .
While various studies have provided evidence for the partner and relationship characteristics that may influence individual decisions to consistently use condoms, there is limited research on the subject matter among MSM and TGW in sub-Saharan Africa, where cultural factors, resource limitations and other factors may influence the dynamics of condom negotiation and decisionmaking in ways that differ from those in westernized settings. Given limited condom uptake among cisgender MSM and TGW in Nigeriaand the continued importance of characterizing optimal strategies to encourage CCU, we assessed the individual, relationship and partner-level factors associated with CCU among a cohort of cisgender MSM and TGW enrolled in two community-engaged clinics in Abuja and Lagos, Nigeria, thus identifying characteristics that may aid in developing targeted strategies.
# Methods
## Study design and population
TRUST/RV368 study is a prospective cohort study in Lagos and Abuja, Nigeria that assesses the effectiveness of network-based recruitment of MSM and TGW into an HIV prevention and treatment program. As described elsewhere, participants are recruited into the study using respondent-driven sampling; RDS. Briefly, participants are eligible to participate if they were assigned male at birth, are aged ≥16 years (Abuja) or ≥ 18 years (Lagos), and reported receptive or insertive anal intercourse with a male partner in the past year. Initial seed participants were recruited through community-based sampling of individuals who were identified to be wellconnected within the population of interest. Each seed was given three RDS coupons to distribute to eligible peers and each individual subsequently recruited was also given three coupons to distribute until equilibrium was achieved.
At enrollment and follow-up visits every 3 months for up to 18 months, participants were administered a detailed questionnaire that collected information about demographics, sexual practices, and HIV risk behaviors in the previous year (at enrollment) or the preceding 3 to 6-months (at follow-up visits). Participants also provided information related to demographics and risk behaviors of up to five of their sexual partners from the past year. At each visit, participants were counselled and screened for HIV and other sexually transmitted infections (STIs). Blood samples were tested for HIV using a parallel testing algorithm of two rapid tests according to national guidelines in Nigeria. The questionnaires used (supplemental file) were developed for the study incorporating modifications of previously validated scales such as the PHQ9 depression module and the World Bank's Social Capital Tool.
Baseline data for participants enrolled in the study between April 2013 and April 2019 who provided information on at least one sexual partner at enrollment were included in these analyses. All participants provided written informed consent in English or Hausa. The study was approved by the University of Maryland Baltimore Institutional Review Board (IRB); the Federal Capital Territory Health Research Ethics Committee, Abuja; and Walter Reed Army Institute of Research IRB.
## Outcomes of interest
The three outcomes of interest were CCU with insertive anal sex, CCU with receptive anal sex, and CCU with both types of anal sex in the past year. For each anal sex type, participants were asked "In the past year, when you had anal sex with [partner name], how often was a condom used?". A participant was defined as consistently using condom if they reported always using condoms all the time they had insertive, receptive or both types of anal sex with a male partner. Each outcome was examined in a separate model.
## Predictors of interest participant-level factors
Participant-level factors of interest included demographic characteristics such as age, education status, marital status, sexual orientation and gender identity.
Other participant-level variables of interest were disclosure of sexual or gender identity to a healthcare worker, drinking habits in the past 30 days, access to condoms, knowledge of HIV risks, history of STI and HIV testing, previous STI diagnosis, self-reported HIV status, and HIV testing results at baseline.
## Partner and relationship-level factors
Partner and relationship-level factors were all reported from the perspective of the participant. Partner-level variables included partner's age, education status, sexual orientation, marital status, sexual concurrency, condom use with their other partners, alcohol or drug use, STI history, and HIV status.
Relationship-level variables included relationship type, age difference between participant and partner socioeconomic difference between participant and partner, frequency of sexual intercourse with partner, how often HIV was discussed in relationship, and HIV serostatus concordance based on self-reported HIV status. Sexual network density-based on the answer to the question "Does sexual partner one know sexual partner two?" and so on up to the fifth sexual partner as necessary-was also assessed. Sexual network density was defined as degree of connection between the sexual partners of a participant and was calculated as the total number of actual ties divided by the total number of potential ties. Network density scores were dichotomized with < 50% considered "smaller" and ≥ 50% considered "larger" network density.
## Statistical analyses
Demographic characteristics and sexual behaviors of study participants and at most five of their partners were examined at enrollment using descriptive statistics.
Odds ratios (ORs) and adjusted odds ratios (aORs) with corresponding 95% confidence intervals (CI) of participant and partnership covariates associated with CCU during insertive, receptive, and both types of anal sex were calculated using logistic regression with generalized estimating equations to account for clustering within participants. Variables significant at p ≤ 0.10 and prespecified variables including age, marital status, gender identity, and education status previously known to be associated with CCU were included in the multiple regression analyses regardless of statistical significance. 1 0.05 was considered significant in the multiple regression analyses. Analyses were carried out as complete case analyses and refusal/don't know responses were kept as categories the final model where applicable. Sub-group analyses assessing the associated factors in the two KP groups separately i.e., MSM and TGW were also carried. All analyses were performed using Stata (version 15; Sta-taCorp LP, College Station, Texas USA).
# Results
Between April 2013 and April 2019, 2591 eligible MSM and TGW participants were recruited into the study. Of those, 1786 reported on a total of 6432 sexual partners at enrollment and were included in these analyses. The average sex network size was 3.6.
## Participant characteristics
An overview of the demographic and behavioral characteristics of the participants in this study is presented in. Seventy-four percent (n = 1324) of the participants were recruited from Abuja and the median age of the participants was 24 years (interquartile range [IQR]:. Eighty percent (n = 1424) self-identified as male and the median number of sexual partners reported in the past year was 4 (IQR: 2-8). Five hundred and thirty (30%) participants reported being gay or homosexual while 70% reported being bisexual (n = 1238). Most of the participants had a secondary education or higher and majority of them reported being single or never married. Ninety-six percent (n = 1715) of participants reported either buying condoms or getting them for free. A majority of the participants reported that all types of sex were associated with equal risk of HIV transmission, while 22% (n = 384) reported anal sex being the type of sex associated with the highest risk of transmission. Seventy-nine percent (n = 1411) of the participants had ever been previously tested for HIV and a quarter of the participants reported being very worried about HIV in the past 12 months Of the 1786 participants included in these analyses, 47% (n = 842) were living with HIV.
## Partner and relationship characteristics
The participant reported characteristics of the partners and relationships are presented in. The median age of partners reported on was 27 (IQR:and 65% (n = 4106) of the partners were reported to be casual partners. Most of the partners (n = 5285; 82%) were single or never married and 53% (n = 3418) were reported to be bisexual. A majority of the partners had a senior secondary education or higher and more than half of the partners reported on had a socio-economic status that was higher than the participants'.
Sixty-eight percent of the partners reported having multiple regular sexual partners at the same time, while only 15% of the partners were believed to always consistently use condoms when they had had sex with their other sexual partners. Condom use was encouraged by 52% of the partners and 18% of the partners discussed HIV with the participant at least once a week.
The HIV status of 54% of the partners were unknown by the participants while 6% of the partners were reported to be HIV positive. Using the self-reported HIV status of the participants and the participant reported HIV status of the partners, 4% of the partnerships were HIV positive seroconcordant and 75% of the partnership were HIV-serodiscordant or had unknown seroconcordance status.
## Individual characteristics associated with consistent condom use
The bivariate and multivariable analyses of the participant characteristics associated with CCU with insertive, receptive, both types of anal sex are presented in. The results of the analysis of CCU with both types of anal sex are reported here.
In the bivariate analysis, higher education status disclosure of MSM status to a healthcare worker, previous STI and HIV tests, self-reported positive HIV status, and knowledge that anal sex had the highest HIV transmission risk were all positively associated with CCU.
Age between 20 and 24, participants who identified as women, buying condoms or neither buying nor getting condoms for free, being very worried about HIV in the previous 12 months, and a self -reported positive HIV status were all negatively associated with CCU.
In the multivariable analysis, higher education status (aOR: 1.80; 95% CI: 1.10-2.94) and disclosure of MSM status to a healthcare worker (aOR: 1.22 1.00-1.48) remained positively associated with CCU. All age categories except 35 years or older were negatively associated with CCU and so was neither buying nor getting condoms for free (aOR: 0.26; 95% CI: 0.11-0.59), and being very worried about HIV in the previous 12 months (aOR: 0.66; 95% CI: 0.53-0.82). Results for the sub-group analyses are presented as supplemental tables in the appendix. In the multivariable analysis among TGW (Supplemental, participants who reported being previously married or widowed and participants who reported that they had previously tested for HIV more than once were more likely to consistently use condoms with both types of anal sex. Conversely, transgender participants who reported their sexual orientation as other were less likely to consistently use condoms with both types of anal sex.
Findings for MSM only (Supplementalwere similar to results in for the combined analysis. In the multivariable analysis among MSM only, higher education and participants who reported that oral sex had the highest HIV transmission risk were more likely to consistently use condom while younger age group, neither buying or getting condoms for free, and MSM participants who reported being very worried about HIV in the previous 12 months were less likely to consistently use condoms with both types of anal sex.
## Partner and relationship characteristics associated with consistent condom use
In the multivariable analysis, relationships in which the partners had a higher education, casual relationships, relationships where the partner encouraged the participant to use condoms and where HIV was discussed at least once a week and larger network sizes were found to be positively associated with CCU. Older partners (35 and older), relationships where the partner was reported to be married to a woman, all frequencies of sexual acts, and relationships where the partner's HIV status were reported to be positive, or unknown were all negatively associated with CCU.
In the sub-group analyses for TGW (Supplemental, results for the multivariable analysis showed that CCU with both types of anal sex was significantly more likely where the participant's partner had an education at the secondary level or higher, or where the relationship was described as casual, while CCU was significantly less likely with any frequency of sexual acts. In the multivariable analysis among MSM (Supplemental, CCU was significantly more likely where the relationship was described as casual, where the participant reported that their partner had encouraged the participant to use condoms with their other sexual partners, and a larger network size. Consistent condom use was significantly less likely if the participant's partner was 45 or older, with any frequency of sexual acts, and if the participants reported that their partners' HIV status was positive or unknown.
# Discussion
In these analyses among MSM and TGW in Nigeria, the data demonstrated on the individual level, CCU during both insertive and receptive anal sex were positively associated with higher education and disclosure of sexual behavior to a healthcare worker, and on the partnership level was associated with relationship type, partners encouraging condom use, and frequent discussion of HIV by partners. Given that consistent condom use was reported in only 44% of the partnerships reported here and the prevalence of HIV remains high among MSM and TGW in Nigeria, our results highlight possible correlates of condom use which may be leveraged for HIV prevention messaging. The findings that condom use was more likely in relationships where condom use was encouraged and in relationships where HIV was discussed frequently is consistent with previous studies which highlight that relationship and partner norms influence health behaviors, including condom use. In a study among MSM in the US, it was found that having at least one person in a sexual dyad who did not disapprove of condomless anal sex was associated with an increased likelihood of condomless sex compared to those who did not have such a person in their sexual network. Relationships present opportunities for information exchange and involved parties often influence each other leading to sustained behavior change. This dynamic could be leveraged to create interventions that promote open discussion of safer sex practices. Such interventions may require training for effective communication, as openly discussing sexual behaviors may be uncommon in some communities, particularly ones where HIV and same-sex sexual practices are heavily stigmatized.
Results showing that CCU was more likely in casual relationships as compared to regular relationships is also consistent with findings from other studies among MSM where higher rates of unprotected anal intercourse (UAI) have been observed in relationships considered more serious. In these relationships, couples may engage in UAI to increase intimacy and trust in a relationship. While risk perception may decrease with more regular partners, there may be sustained risks for both STIs and HIV in non-exclusive relationships, therefore the importance of prevention modalities should be emphasized for MSM or TGW in non-exclusive relationships. Moreover, previous studies have associated steady partnership as a source of HIV infection among gay men. These findings highlight both the importance and the potential of couple-oriented testing and counseling practices. Couple-based approaches have been shown to be highly effective in several settings and evaluation in the Nigerian context for MSM and TGW may be valuable. Our results also show that participants in relationships where the partners were married to women were less likely to consistently use condoms. The majority of participants in this study reported that they were bisexual and marriage to women was common among their partners. Given the presence of punitive laws prohibiting same sexrelationships in Nigeria, this is unsurprising. Many individuals with gay or bisexual identities may choose to marry women due to fear of stigma, societal pressures, or to avoid being outed. Previous surveys of men who have sex with men and women (MSMW) have shown them to have riskier HIV-related behaviors that potentiate HIV infection, including a higher probability of having an unknown HIV status and a lower probability of encountering HIV prevention activities and materials due to presence of "biphobia". Thus, MSMW could play a potential role in HIV transmission to women by bridging the epidemics in two different populations. This dynamic could also apply in TGW who have sex with cisgender women. These results suggest that addressing the prevention needs of MSMW and TGW and all their sex partners is needed for HIV reduction among both KP groups and women. Tailored prevention strategies requires context and understanding of behaviors and sexual networks of TGW who have sex with women and MSMW.
Unlike findings from other studies that have shown age-disparate relationships to be associated with more high-risk HIV-related behaviors including unprotected anal intercourse, our results showed increased odds of condom use during receptive anal sex in relationships where there was a ten year or more difference in age between partners. The risks associated with age-disparate partnerships are often higher in relationships where adolescents have partners who are much older than them. This is due to the increased likelihood for older partners to be living with HIV and a high propensity for condomless sex and alcohol use during sex with older partners. There is also evidence that it may be linked to power dynamics in the relationships. This could lead to a limited ability on the part of younger MSM to adopt or advocate for HIV risk reduction behaviors like condom use even when protected sex is desired. Thus, a ten-year age difference between two middle-aged persons may not be associated with the same risk as a tenyear age difference between an adolescent and a middleaged person; a nuance which is not sufficiently captured in this analysis and may therefore be the reason for the difference in the results observed.
On the individual level, disclosure of sexual behavior to a healthcare practitioner was also positively associated with CCU. Disclosure of key-population status may indicate more engagement with healthcare services and therefore reduced likelihood for risky behaviors due to more access to targeted HIV prevention messaging. Disclosure of MSM status in Nigeria may be challenged by the presence of laws criminalizing same sex behaviors and the high levels of stigma faced by MSM in the country. Studies have demonstrated that structural elements, where the risks of disclosure of sexual practices for a meaningful risk assessment outweigh the potential benefits of better engagement in HIV prevention services, have affected the uptake of HIV prevention services including condom uptake among MSM at risk for HIV acquisition. This is particularly important as our results showed that participants who reported neither buying condoms nor getting condoms for free were significantly less likely to consistently use condoms, a prevention modality commonly available for free in healthcare centers. Given the importance of the healthcare sector in managing and reducing the HIV epidemic among MSM in the country, training and sensitizing healthcare workers about the impacts of stigmatizing behavior is needed. De-centralization of HIV prevention services by leveraging KP community-engaged facilities such as the trusted community health centers involved in this study may also increase healthcare engagement and alter HIV-related risk behaviors. Longitudinal analysis of data from this cohort showed an improvement in the uptake of condoms and water-based lubricants by MSM and TGW, suggesting effectiveness of these facilities in improving engagement in care.
The findings from this study should be interpreted in the context of several limitations. Firstly, all behavioral data collected, including partner and relationship characteristics, were self-reported by the participant in the study. This makes our findings susceptible to multiple biases, including recall and social desirability. The presence of missing data may have biased our results however, missingness was around 1% for many of the variables and is unlikely to significantly impact our interpretation. Additionally, "refusal" and "don't know" categories were included in models where applicable as these responses may be informative about behavioral risks. Although the study was longitudinal, we carried out cross-sectional analyses using the baseline data. Thus, causality cannot be inferred from our results. While our results may capture the individual and relationship characteristics of MSM and TGW in Lagos and Abuja, Nigeria, our findings may not be generalizable to the broader key population in Nigeria or other countries in sub-Saharan Africa.
# Conclusions
HIV prevention programs may benefit from encouraging open communication of safer sex practices as a strategy for HIV risk reduction given that MSM and TGW were more likely to consistently use condoms when their partners encouraged and openly discussed safer sex. Even with the introduction and scale up of other biomedical HIV prevention modalities such as PrEP, condom use remains an important mode for reducing onward transmission of STIs as HIV and STI infection rates remains sustained among MSM. Implementation science research could support the evaluation of optimal strategies that integrate open discussions about safe sex into existing behavioral interventions. However, while research is helpful, leveraging existing strategies to increase condom use including the provision of condoms in safe spaces, ensuring access to better condoms and condom-compatible lubricants, and teaching condom negotiation skills may support increased consistent condom use among men who have sex with men and transgender women across Nigeria.
## Supplementary information
The online version contains supplementary material available at https://doi. org/10.1186/s12889-021-11275-w.
Additional file 1. Trust Questionnaire.
Additional file 2: Supplemental. Participant characteristics associated with consistent condom use with anal sex among TGW. Supplemental. Partner and relationship characteristics associated with consistent condom use with anal sex among TGW. Supplemental. Participant characteristics associated with consistent condom use with anal sex among MSM. Supplemental. Partner and relationship characteristics associated with consistent condom use with anal sex among MSM. |
Optical excitation and external photoluminescence quantum efficiency of Eu3+ in GaN
We investigate photoluminescence of Eu-related emission in a GaN host consisting of thin layers grown by organometallic vapor-phase epitaxy. By comparing it with a reference sample of Eu-doped Y 2 O 3 , we find that the fraction of Eu 31 ions that can emit light upon optical excitation is of the order of 1%. We also measure the quantum yield of the Eu-related photoluminescence and find this to reach (,10%) and (,3%) under continuous wave and pulsed excitation, respectively. W ide-gap semiconductors doped with rare-earth ions are of great interest for applications in lightemitting diodes (LEDs), because of their temperature insensitive, sharp and stable emission and an ease of current injection 1,2 . Blue and green LEDs based on galliumnitrides are already successfully commercialized. However, red emission from these materials is still lacking and monolithic full-color displays are at this moment impossible until this problem is solved. Large effort is thus spent on research of GaN materials that emit in the red. Among them, Eu-doped GaN (GaN:Eu) attracts special attention for its intense red emission around 622 nm, originating from the 5 D 0 R 7 F 2 transition of the Eu 31 ions 3-7 . In combination with Er 31 and Tm 31 doped GaN, GaN:Eu could be applied in monolithic full-color displays 8 . In fact, for displays with the full color gamut, a phosphor with an emission line around 610 nm is needed, and it is clear that only the Eu 31 ion can satisfy this requirement 9 . However, the light output power of the presently available GaN:Eu LEDs is still too low to compete with 'conventional' red-light LEDs based on III-V materials. To overcome this, material characteristics of GaN:Eu layers need to be optimized. For that, we have to determine the limiting factor in the optical performance.The main concerns are the optical accessibility -the percentage of Eu 31 ions which contribute to the emissionand the external quantum efficiency (of both electro-and PL). Past research has revealed that a variety of Eurelated emitting centers, as well as a variety of energy transfer routes exist 10-14 . These findings imply that the potential for improvement of emission is mainly to be found in material engineering towards optimization of the energy transfer between the GaN-host and Eu 31 ions 3-6 . We present here in this work the careful analysis of Eu luminescence in GaN and come to the conclusion that the problem lies in the low percentage of Eu that is optically active. We were able to draw this conclusion by comparing our GaN:Eu sample to a well-characterized standard reference sample of Y 2 O 3 :Eu. The europium in this reference sample, by being embedded in a non-active and insulating host matrix without any defects that can act as quenching centers, has an internal quantum efficiency of over 90% with, moreover, basically all Eu is participating in emission; the reference sample can thus be used as a way to calibrate the optical accessibility of europium in other matrices.GaN:Eu material can be prepared in various ways. In particular, samples with high crystalline quality can be produced by organometallic vapor phase epitaxy (OMVPE, the de facto standard for GaN growth). For this technique, it has been shown that samples grown under atmospheric pressure feature significant enhancement of PL compared to those grown under low pressure, while the absolute Eu 31 concentration is smaller. Since the effective PL lifetime is not affected, this implies that the enhancement is due to a larger number of Eu 31 ions contributing to emission 3,4 . In this study we quantify this notion; we investigated the level of optical accessibility of Eu 31 and its temperature dependence in GaN layers grown by OMVPE and determined the fraction of Eu 31 dopants participating in emission to be in the order of only some percent. We also determined the external quantum yield (QY) of the Eu-related PL under different excitation conditions.
## Experimental
In this study, the GaN:Eu layers were grown by the OMVPE technique on a sapphire (0001) substrate. Specific details on the sample preparation procedure can be found in Ref. [bib_ref] Improved luminescence properties of Eu-doped GaN light-emitting diodes grown by atmospheric-pressure organometallic..., Nishikawa [/bib_ref]. The particular sample used in this study has the optically active GaN:Eu layer with a thickness of approximately 400 nm and nominal concentration of Eu 31 ions N Eu 5 3 3 10 19 cm 23 , as measured by secondary-ion massspectroscopy. The reference sample is a conventional industrial Y 2 O 3 powder with a high concentration of Eu 31 ions, N Eu 5 1. [bib_ref] Plasma synthesis and liquid-phase surface passivation of brightly luminescent Si nanocrystals, Mangolini [/bib_ref] 3 10 21 cm 23 , compressed to a solid pellet to facilitate handling in optical experiments. Y 2 O 3 is an insulating host typically used in rareearth-based efficient phosphors; the wide band-gap character of this host makes that the Eu 31 ions can only be excited resonantly and can attain near 100% optical accessibility. Therefore, the chosen material can be suitably used as a reference, allowing for the determination of the percentage of Eu 31 ions contributing to the photo-luminescence, i.e., being optically accessible.
The PL experiments were conducted under pulsed and continuous wave (CW) excitations. For the former, a Nd:YAG-pumped optical parametric oscillator with a repetition rate of f 5 100 Hz and a pulse duration of Dt^7 ns was used to obtain resonant excitation of the Eu in Y 2 O 3 . For GaN-host-mediated excitation, the 3 rd harmonic of the same laser pump was used, l exc 5 355 nm. The CW GaN-mediated excitation was achieved with an Ar-ion laser operating at a wavelength of l exc 5 364 nm. The spectral information was resolved by a multi-grating monochromator and detected by a CCD camera; timeresolved PL signals were registered with a water-cooled photo-multiplier tube with 1 ms resolution in the applied settings. The QY measurements were performed in both pulsed (resonant) and CW (host-mediated) excitation conditions, where for the latter a lowintensity xenon lamp was used in combination with a monochromator selecting the appropriate excitation wavelength, and recorded with a CCD camera. The sample was placed inside an integrating sphere for homogeneous distribution of the emitted and excitation light; The evaluation of the PL QY was done following the procedure of Refs. For the evaluation of the optical accessibility of Eu 31 in GaN, the integrated PL intensity was measured as a function of the applied photon fluence per pulse. The methodology is based on a comparison of the measured PL intensity with that of Eu 31 ions in the reference sample. In that way, the experimentally measured PL intensity can be directly related to a specific concentration of Eu 31 contributing to emission [bib_ref] Concentration of Er 31 ions contributing to 1.5-mm emission in Si/Si:Er nanolayers, Vinh [/bib_ref] [bib_ref] Sensitization of Er luminescence by Si nanoclusters, Wojdak [/bib_ref]. In this experimental approach, identical excitation and detection configurations for both samples were maintained in order to ensure constant alignment, independent of the excitation settings. A pinhole placed in front of the investigated sample defines the dimensions of the excitation spot, which, in combination with the exact measurement of the energy per laser pulse after the pinhole, determines the photon fluence value (the number of photons per cm 2 per pulse, W).
# Results
For the initial optical characterization, the sample was measured upon CW band-to-band excitation, as done in previous studies on similar materials 3-6 . [fig_ref] Figure 1 |: PL spectra of GaN [/fig_ref] shows the normalized PL spectra of GaN:Eu measured at room temperature and at T 5 10 K in red and blue, respectively. The peaks associated with the intra-4f shell transitions are visible 9 , with the main contribution arising from the 5 D 0 R 7 F 2 transition at 621.5 nm, as labeled in the figure. In addition to the Eu 31 -related PL peaks, the spectrum measured at room temperature (red) shows a broad structure centered around l 5 575 nm and some 100s of nm wide, which decreases in intensity for lower temperatures. This feature appeared independently of the excitation photon flux -it has a constant relative amplitude -and can possibly be assigned to a recombination path involving defects and is not further discussed here. With the exception of this specific band, the PL spectral line positions of the investigated material are independent of temperature. The inset of [fig_ref] Figure 1 |: PL spectra of GaN [/fig_ref] shows the temperature dependence of the PL intensity, obtained under a photon flux of j exc 5 9.4 3 10 18 cm 22 s 21 . Saturation at low temperatures was seen, in agreement with previous reports [bib_ref] Improved luminescence properties of Eu-doped GaN light-emitting diodes grown by atmospheric-pressure organometallic..., Nishikawa [/bib_ref].
The experiments for the evaluation of the percentage optical accessibility have been performed under pulsed excitation. The reference Y 2 O 3 :Eu sample was excited resonantly (l exc 5 470 nm, coinciding with the 7 F 0 R 5 D 2 transition 9,20,21 ), while the GaN:Eu sample was pumped with over-bandgap-energy photons. shows the normalized PL spectra of both GaN:Eu and Y 2 O 3 :Eu at room temperature in black and blue, respectively, with the corresponding excitation mechanisms schematically illustrated in the insets. In both samples, multiple peaks due to the intra-4f shell transitions are observed (labeled in the figure), with the whole emission band being clearly blue-shifted for Y 2 O 3 :Eu in comparison to GaN:Eu. The dominant PL feature in both cases can be associated with the 5 D 0 R 7 F 2 transition 9 , and is observed at l max 5 612 nm for Y 2 O 3 :Eu and l max 5 621.5 nm for GaN:Eu. Apart from the blue-shift of the main PL line, the influence of the host matrix can also be observed in the effective PL lifetime of the dominant peak, shown in . While in both materials a mono-exponential decay is measuredas indicated by dashed lines -the time constants differ by almost an order of magnitude, being t eff < 250 ms (GaN, upper panel) and t eff < 1.2 ms (Y 2 O 3 , lower panel).
The PL intensity measurements were conducted from room temperature cooling down. shows the temperature dependence of the integrated PL intensity at a photon fluence of W 5 4.6 3 10 16 cm 22 . As can be seen, upon cooling, the PL intensity initially increases 2 down to T 5 150 K -and then quenches somewhat. A similar dependence has been reported before for a Mg 21 co-doped GaN:Eu sample upon CW over-bandgap excitation [bib_ref] Effect of thermal annealing on luminescence properties of Eu,Mgcodoped GaN grown by..., Lee [/bib_ref] [bib_ref] Current status for lightemitting diode with Eu-doped GaN active layer grown by..., Wakahara [/bib_ref]. Such behavior might be due to thermal activation of specific defect centers participating in excitation and de-excitation paths, which can then either enhance or quench the PL. At the same time, the lowering of temperature results in a moderate increase of PL effective lifetime from 220 ms to 265 ms, as shown in Figs. 3(b) and 3(c). This lack of correlation between PL intensity and lifetime temperature dependencies indicates that the effect of temperature is stronger on the excitation than on the recombination, as argued by Lee et al.. The difference between the temperature dependence of the integrated PL intensity of GaN:Eu observed in our research and the research reported by Fujiwara et al. is most likely due to the difference between pulsed and CW excitation 3-6 . The differences are consistent with the concept of a variety of different Eu 31 centers, and site-selective excitation paths with individual thermal characteristics, as concluded in the past on basis of high-resolution PL and PL excitation investigations [bib_ref] Site and sample depemdent electron-phonon coupling of Eu ions in epitaxial-grown GaG..., Woodward [/bib_ref] [bib_ref] Excitation pathways and efficiency of Eu ions in GaN by siteselective spectroscopy, Fleischman [/bib_ref] [bib_ref] Site-specific excitation of Eu ions in GaN, Penn [/bib_ref] [bib_ref] The photoluminescence/excitation (PL/E) spectroscopy of Eu-implanted GaN, O'donnell [/bib_ref] [bib_ref] Lattice site location of optical centers in GaN:Eu light emitting diode material..., Lorenz [/bib_ref]. Figure 3(d) shows the integrated PL spectra of Y 2 O 3 :Eu and GaN:Eu as a function of the applied photon fluence. The red, purple and blue squares correspond to the integrated PL intensity of GaN:Eu at T 5 290 K, T 5 150 K and T 5 20 K, respectively. Initially, in the low photon fluence regime, a linear dependence can be observed for the sample, after which saturation sets in; a stretched exponential was fit to the data and shown as solid lines in the figure. It has to be noted that the PL spectrum remained unaltered within the investigated excitation flux range. The black circles depict the integrated PL intensity of the Y 2 O 3 :Eu reference sample measured at room temperature, and the solid black line is a linear fit to the data.
# Discussion
Based on this, an estimation of the optical accessibility of Eu 31 in GaN can be made by comparing the time-integrated PL intensity of the investigated sample with that of the Y 2 O 3 :Eu reference. The instantaneous PL intensities of both samples are proportional to N Ã Eu t ð Þ t rad , where N Ã Eu and t rad correspond to the density of excited Eu 31 -ions and their radiative lifetimes, respectively. From here on, omitting the time (t) signifies the number at t 5 0. Since the PL signals are integrated over time, the result will be proportional to N Ã Eu |t eff t rad , where t eff is the effective PL lifetime, as measured in the experiment, and N Ã Eu the initial number of excited Eu ions. (Note:
The ratio of effective and radiative lifetimes is the internal quantum efficiency -number of photons created divided by the number of photons absorbed 2 IQE 5 t eff /t rad ). To find the number of photons coming out of the sample and the external quantum efficiency (EQE), this value has to be multiplied by the extraction efficiency g that is determined by the refractive index of the material for that specific wavelength; (EQE 5 g 3 IQE; We ignore in this definition other effects such as photon reabsorption, as well as light-insertion efficiencies, the latter not entering the calculation anyway). Therefore, the total equation for the ratio of the number of photons emitted by GaN:Eu and Y 2 O 3 :Eu is given by
[formula] I GaN I Y2O3~N Ã Eu GaN ð Þ : EQE GaN N Ã Eu Y2O3 ð Þ : EQE Y2O3 N Ã Eu GaN ð Þ N Ã Eu Y2O3 ð Þ : g GaN g Y2O3 : t Eu GaN ð Þ eff . t Eu GaN ð Þ rad t Eu Y 2 O 3 ð Þ eff . t Eu Y 2 O 3 ð Þ rad :ð1Þ [/formula]
On basis of this, we can calculate the number of excited europium ions in our GaN sample, N Ã Eu GaN ð Þ , from the measured PL intensities I GaN and I Y2O3 . This goes as follows.
In the first step, we calculate the extraction efficiencies, knowing that these are determined by the refractive index of that particular [fig_ref] Table 1 |: Parameter values found by fitting www [/fig_ref]. material and wavelength. By Snell's law, light will not manage to come out at an angle greater than the critical angle that depends on the ratio of refractive indexes. For isotropic emission into air, it can be shown that the (double sided) extraction efficiency of a planar device is equal to
[formula] g l ð Þ~1{cos sin {1 n air n l ð Þ ! :ð2Þ [/formula]
In our case, for GaN at l 5 621.5 nm, n 5 2.39, and for Y 2 O 3 at l 5 612 nm, n 5 1.93. With n air assumed unity, the extraction coefficients can be calculated as: g GaN 5 0.091 and g Y2O3~0 :144.
The measured EQE of the Y 2 O 3 :Eu sample was 29.5% and this is much larger than the above implied maximum possible EQE of 14.4%. This shows that the above calculation for g Y2O3 is incorrect, possibly due to a different geometry (a pellet rather than a planar device) and surface roughness. However, it does indicate that the internal quantum efficiency must be close to 100%. We continue, therefore, with a best guess of the extraction efficiency of g Y2O3~0 :295 and an internal quantum efficiency of
[formula] IQE Y2O3~t Eu Y2O3 ð Þ eff . t Eu Y2O3 ð Þ rad~1 :ð3Þ [/formula]
This introduces a substantial uncertainty in the following calculation. Continuing, the ratio of extraction efficiencies is then
[formula] g GaN g Y 2 O 3~0 :31:ð4Þ [/formula]
These values were used in our calculations. Next, the effective lifetime of GaN:Eu PL, t Eu GaN ð Þ eff , is measured via PL transients, as in . The radiative lifetime, t Eu GaN ð Þ rad , can be found from the radiative lifetime of Y 2 O 3 :Eu using Fermi's golden rule for spontaneous emissions that states that radiative lifetimes scale with the refractive index of the host material, t rad / 1/n. With for Y 2 O 3 the radiative and effective lifetimes equal (as discussed above) and the effective lifetime Y 2 O 3 , again, measured in transients, the radiative lifetime in GaN can be found as
[formula] t Eu GaN ð Þ rad~n Y 2 O 3 n GaN |t Eu Y2O3 ð Þ eff :ð5Þ [/formula]
Finally, for this comparison we assume that all Eu 31 -ions in Y 2 O 3 :Eu are equivalent and all contribute to the PL (in contrast to those in GaN:Eu). We can then calculate the number of excited europium atoms in Y 2 O 3 :Eu. Because the pulse duration in this experiment is much shorter than the effective lifetime t eff of Eu 31 , it can be assumed that recombination does not take place during an excitation pulse and that N Ã
[formula] Eu Y2O3 ð Þ follows the equation 26 N Ã Eu Y2O3 ð Þ~1 00%|N Eu Y 2 O 3 ð Þ 1{exp {s abs Y 2 O 3 ð Þ W À Á Â Ã ,ð6Þ [/formula]
where N Eu Y2O3 ð Þ~1 :16|10 21 cm {3 is the density of Eu-atoms, s abs Y2O3 ð Þ~3 :9|10 {20 cm 2 is the absorption cross-section of Eu 31 for the excitation wavelength l exc 5 470 nm. It is based on a value given by Wakamatsu et al. [bib_ref] Luminescence properties of Eu-doped GaN under resonant excitation and quantitative evaluation of..., Wakamatsu [/bib_ref] for GaN scaled to Y 2 O 3 . As shown in For the PL of GaN:Eu, however, a stretched exponential behavior is observed, and the excitation of Eu atoms in GaN through absorption therefore follows the equation
[formula] N Ã Eu GaN ð Þ~a N Eu GaN ð Þ 1{exp { s abs GaN ð Þ W À Á b n o h i ,ð7Þ [/formula]
with N Eu(GaN) 5 3 3 10 19 cm 23 , and a the fraction (between 0 and 1) of ions that do take part in optical processes, i.e., the parameter we are trying to establish in this work. The PL is proportional to this number and follows a stretched exponential too. shows that it fits the data quite well at all temperatures. The parameters of the fits are given in [fig_ref] Table 1 |: Parameter values found by fitting www [/fig_ref] , with A GaN the amplitude of the curves normalized to saturation PL at 290 K. With this, the ratio of excited ions thus becomes
[formula] N Ã Eu GaN ð Þ N Ã Eu Y2O3 ð Þ~a N Eu GaN ð Þ 1{exp { s abs GaN ð Þ W À Á b n o h i N Eu Y2O3 ð Þ s abs Y2O3 ð Þ W ,ð8Þ [/formula]
where s abs(GaN) and b are found from fitting a curve to the data. Note that, to get a rough idea, s abs(GaN) is basically the reciprocal fluence at which the PL intensity reaches 63% (1 2 e) of its saturation value.
With the PL behavior known (I GaN and I Y2O3 , see ) -the geometric photon collection efficiencies are equal and thus cancel in the ratio, i.e., no integrating sphere needs to be used -the only remaining unknown in the starting equation (Eq. 1) is the fraction of active europium centers in GaN:Eu, a, which can thus be derived. As an example, for a fluence of W 5 10 16 cm 22 , a ratio of I GaN =I Y2O3~0 :0665 can be found from or [fig_ref] Table 1 |: Parameter values found by fitting www [/fig_ref] for T 5 290 K. When we substitute this into Eq. 1, with the parameters of we get a value of 2.6%. In the same way, we find values of 2.9% and 3.3% for 20 K and 150 K, respectively.
While the particular experimental measurements leading to these numbers are quite accurate (PL intensity vs. photon fluence), the overall accuracy of the optical accessibility determination depends on the method itself and the assumptions which have been made. The errorbar on the final value is estimated to be about a factor 2, which is mainly due to the poor knowledge of the extraction efficiency as discussed before. Other factors that might influence it are the concentration on Eu 31 ions in the reference sample and the assumption that all of them contribute to PL. Finally, (multiple) internal reflections might play a role. Using Fresnel's equation (for perpendicular rays), light is reflected off the interfaces of air-GaN, GaN-sapphire, and sapphire-air. Using the refractive indexes of these materials (2.4, 1.77), we estimate that effectively 83% of the incident light is in the GaN layer under study. The absorption cross section found by us is then 17% under-estimated. This further increases the error bar. Yet, undoubtedly, a lot of europium ions are in the GaN host that never participate in any way in luminescence and this is the main conclusion of this work.
In addition, the PL QY of GaN:Eu was measured upon low flux CW (Xe lamp) and pulsed indirect excitation at l 5 355 nm. This method provides information on the efficiency of the energy transfer, as it gives directly the ratio of the number of emitted and absorbed photons, without consideration of the concentration of the available optically active centers. Under CW excitation, the PL QY was <3.2%, while under pulsed excitation it was found to be about a factor 2.5 higher, QY < 8%. This difference is in agreement with the results observed in the PL spectra, since the defect-related emission shoulder, as revealed for CW excitation, is likely to lower the overall QY.
# Conclusions
The current work showed that the problem of low yield in GaN:Eu LEDs lies in the relatively low levels of optical accessibility of Eu dopants, in the order of a few percent. Here thus lies the opportunity for improvement, basically in material preparation. The temperature-dependent percentage of optically active Eu 31 ions -the 'optical accessibility' -has been determined for state-ofthe-art GaN:Eu layers grown by OMVPE. The values between 2.6% and 3.3% have been found depending on the temperature (T 5 20-300 K). This has been achieved by comparison of the integrated PL intensity of Eu-related emission in the investigated materials and a reference sample. Also the external quantum yield of the hostmediated Eu PL in GaN has been determined as QY < 10% and QY < 3%, for pulsed and CW band-to-band excitation modes, respectively.
[fig] Figure 1 |: PL spectra of GaN:Eu measured under CW excitation at T 5 10 K (blue line) and T 5 290 K (red line). The inset shows the temperature dependence of the wavelength-integrated PL intensity. [/fig]
[fig] Figure 2 |Figure 3 |: (a) PL spectra of GaN:Eu and Y 2 O 3 :Eu at room temperature under pulsed excitation, with the peaks labeled according to convention 9 and consistent with measurements by Tallant et al. 21 . The insets show the respective excitation paths: the Eu 31 -ions in GaN are excited by means of an indirect excitation mechanism, where photons are initially absorbed by the host, with the subsequent energy transfer to Eu 31 ions. In case of Y 2 O 3 :Eu, excitation is accomplished via direct resonant pumping of the 7 F 0 R 5 D 2 transition. (b) Decay profiles of PL from GaN:Eu (upper panel) and Y 2 O 3 :Eu (lower panel), measured at a wavelength of the maximum PL intensity, corresponding to recombination through the 5 D 0 R 7 F 2 transition. (a) Temperature dependence of the integrated PL of GaN:Eu for the temperature range of T 5 6-300 K upon pulsed excitation with the photon fluence W 5 4.6 3 10 16 cm 22 . (b) Decay profile of PL in GaN:Eu at T 5 6 and T 5 300 K. (c) Temperature dependence of the PL decay time of GaN:Eu. (d) Time-integrated PL intensity of GaN:Eu at T 5 20 K, T 5 150 K and T 5 290 K and Y 2 O 3 :Eu at T 5 290 K as a function of the applied photon fluence. The solid lines show linear (Y 2 O 3 ) and stretched exponential (GaN) fits to the data. Parameters of the fits are given in [/fig]
[table] Table 1 |: Parameter values found by fitting www.nature.com/scientificreports SCIENTIFIC REPORTS | 4 : 5235 | DOI: 10.1038/srep05235 [/table]
|
SNPAnalyzer 2.0: A web-based integrated workbench for linkage disequilibrium analysis and association analysis
Background:Since the completion of the HapMap project, huge numbers of individual genotypes have been generated from many kinds of laboratories. The efforts of finding or interpreting genetic association between disease and SNPs/haplotypes have been on-going widely. So, the necessity of the capability to analyze huge data and diverse interpretation of the results are growing rapidly.Results:We have developed an advanced tool to perform linkage disequilibrium analysis, and genetic association analysis between disease and SNPs/haplotypes in an integrated web interface. It comprises of four main analysis modules: (i) data import and preprocessing, (ii) haplotype estimation, (iii) LD blocking and (iv) association analysis. Hardy-Weinberg Equilibrium test is implemented for each SNPs in the data preprocessing. Haplotypes are reconstructed from unphased diploid genotype data, and linkage disequilibrium between pairwise SNPs is computed and represented by D', r 2 and LOD score. Tagging SNPs are determined by using the square of Pearson's correlation coefficient (r 2 ). If genotypes from two different sample groups are available, diverse genetic association analyses are implemented using additive, codominant, dominant and recessive models. Multiple verified algorithms and statistics are implemented in parallel for the reliability of the analysis.Conclusion: SNPAnalyzer 2.0 performs linkage disequilibrium analysis and genetic association analysis in an integrated web interface using multiple verified algorithms and statistics. Diverse analysis methods, capability of handling huge data and visual comparison of analysis results are very comprehensive and easy-to-use.
# Background
Since the completion of the HapMap project, huge numbers of individual genotypes have been generated from many kinds of laboratories. The efforts of finding or inter-preting genetic association between disease and SNPs/ haplotypes have been on-going widely, and the necessity of the capability to analyze huge data and diverse interpretation of the result are growing rapidly. Recently devel-oped software programs are well suited for constructing linkage disequilibrium blocks, estimating haplotypes or detecting genetic association between disease and SNPs [bib_ref] SNPStats: a web tool for the analysis of association studies, Sole [/bib_ref] [bib_ref] SNPAnalyzer: a web-based integrated workbench for single-nucleotide polymorphism analysis, Yoo [/bib_ref] [bib_ref] Efficient multilocus association mapping for whole genome association studies using localized haplotype..., Browning [/bib_ref] [bib_ref] PLINK: a toolset for whole-genome association and population-based linkage analyses, Purcell [/bib_ref] [bib_ref] HapBlock: haplotype block partitioning and tag SNP selection software using a set..., Zhang [/bib_ref] [bib_ref] Haploview: analysis and visualization of LD and haplotype maps, Barrett [/bib_ref]. However, some software programs have drawbacks such as long computation time for the association analysis [bib_ref] SNPStats: a web tool for the analysis of association studies, Sole [/bib_ref] , limited size of dataset [bib_ref] SNPStats: a web tool for the analysis of association studies, Sole [/bib_ref] [bib_ref] SNPAnalyzer: a web-based integrated workbench for single-nucleotide polymorphism analysis, Yoo [/bib_ref] , inconvenient user interface [bib_ref] Efficient multilocus association mapping for whole genome association studies using localized haplotype..., Browning [/bib_ref] [bib_ref] PLINK: a toolset for whole-genome association and population-based linkage analyses, Purcell [/bib_ref] [bib_ref] HapBlock: haplotype block partitioning and tag SNP selection software using a set..., Zhang [/bib_ref] and limited number of genetic models or statistics for the association analysis [bib_ref] Haploview: analysis and visualization of LD and haplotype maps, Barrett [/bib_ref]. We have developed an advanced analysis software program, SNPAnalyzer 2.0, which performs sample-specific linkage disequilibrium analysis and implements genetic association analysis using multiple genetic models in an integrated web interface. It can handle hundreds of thousands of SNPs and thousands of samples in a rather manageable time as compared with other software programs.
## Implementation
The analysis engine was developed by C and interface by JAVA, and the operation of the software program is executed using JAVA applet after accessing through a web browser. Although the implementation of the software program is triggered by a web browser, any information about the user's data is not transmitted anywhere because all the analysis are performed locally using JAVA applet. Raw data and all the analyzed results are stored to the user's computer only. If genotypes from two different samples are available, sample-specific analysis and sample-merged analysis are simultaneously implemented in data preprocessing, haplotype estimation and LD blocking. For diverse interpretation of the genetic effects, one allelic or haplotype association test and three genotypic or diplotype association tests are possible. The free implementation of SNPAnalyzer 2.0 and free download of test dataset are available [7].
# Results
SNPAnalyzer 2.0 comprises of four main analysis modules. All the processes are sequentially implemented and results are displayed in comprehensive tables and graphs. The main features and functions are as follows.
## Data import
Genotypes of biallelic SNPs should be coded in a tab delimited text file. From the first to the fourth column separately represent marker name, chromosome number, chromosome position and dbSNP rs number of each SNP. Subsequent columns represent individual genotypes of each SNP. First row describes headers and individual identifications. Second row describes sample types, i.e. case sample and control sample that are represented as "0" or "1". Sample type should be in dichotomous number and subsequent rows represent SNPs. Individual genotypes should be coded as allele1, slash and allele2 (e.g. "A/A", "A/G", "G/G"). If input data contains missing genotype, it is coded as "N/N". Detailed information on input data format can be checked in the supplementary information Individual genotype data import display Individual genotype data import display. Missing genotypes are represented in red. Allele frequencies, genotype frequencies and p-value of the Hardy-Weinberg Equilibrium test are shown in the table. Triggering tabs for Data Import, Haplotype Estimation, LD Blocking and Case-Control Study are shown on the left panel.
[7]. If data format is correct, data preprocessing is automatically implemented and the results are displayed in the data import interface. shows the result of data importing and data preprocessing.
## Data preprocessing
Once the data is input, data quality check and preprocessing is automatically implemented to drop out erroneous SNPs such as monomorphic SNP. SNPs of which minor allele frequencies and missing genotype frequencies are below the specified threshold are also dropped out. Missing genotype can be replaced by heterozygous genotype. Hardy-Weinberg Equilibrium (HWE) test is sequentially implemented to each SNPs, and Bonferroni correction can be applied in the HWE test to prevent excluding SNPs by chance. Red colors in show missing genotypes. Allele frequencies, genotype frequencies, and the result of the HWE test are displayed in tables.
## Haplotype estimation
A haplotype is a particular pattern of alleles at sequential loci on a single chromosome. In order to reconstruct haplotypes from the unphased diploid genotype data, we have used EM-based algorithm [8] and PL-EM algorithm [bib_ref] Bayesian haplotype inference for multiple linked single-nucleotide polymorphisms, Niu [/bib_ref]. For the performance of reconstruction, 25 or less SNPs are recommended for the EM-based algorithm. PL- EM algorithm can analyze more than 25 SNPs. Reconstructed haplotypes are displayed in an integrated interface [fig_ref] Figure 2: Haplotype estimation display [/fig_ref]. The most likely haplotypes and their frequencies in a given sample are displayed in histogram and table. Reconstructed individual haplotypes and accuracies of the reconstruction are displayed in a separate table. The sample-specific analysis result can be saved as a tab delimited text file.
## Haplotype estimation display
## Linkage disequilibrium (ld) blocking
The degree of genetic linkage between two different SNPs can be estimated by several linkage disequilibrium indices like D', r 2 , LOD score, or by four gamete test [bib_ref] A comparison of linkage disequilibrium measures for fine-scale mapping, Devlin [/bib_ref]. Representative SNP that has strong correlation (r 2 > 0.8) with other SNPs is designated as pairwise tagging SNP. The entire pattern of linkage disequilibrium and tagging SNPs are displayed in a reverse triangle. Several SNPs that are in strong linkage disequilibrium can be bound into one LD block and we construct LD blocks using Gabriel's method. Crossover percentages between haplotypes that are reconstructed within adjacent LD blocks and multi-allelic D' are simultaneously calculated. [fig_ref] Figure 3: LD blocking display [/fig_ref] shows the linkage disequilibrium pattern, LD blocks and reconstructed haplotypes. Red color in the linkage disequilibrium pattern graph means that there exists strong pairwise linkage disequilibrium between adjacent SNPs, and the area enclosed by a thick black line shows LD block. Haplotypes, haplotype frequencies and multi-allelic D' are displayed in the bottom of the linkage disequilibrium pattern graph. shows the different linkage disequilibrium patterns of two samples and merged sample.
# Association analysis
Genetic association between disease and SNP is analyzed using Pearson's chi-square test if the input data contains two different samples such as case sample and control LD blocking display sample. We applied goodness of fit test and likelihood ratio test simultaneously to avoid biased results acquired by applying only a single statistics. False positive control is implemented by both Bonferroni correction and false discovery rate [bib_ref] Statistical significance for genomewide studies, Storey [/bib_ref]. Odds ratios (OR) and 95% confidence interval of odds ratios are calculated simultaneously with chi-square test. If there are haplotypes reconstructed from haplotype estimation or LD blocking, genetic association analysis between disease and haplotypes is performed using the same statistics as SNPs. For analyzing different genetic effects conveniently, four genetic models are available. Additive model deals with allelic or haplotype association, and genotypic or diplotype association can be analyzed using codominant model, dominant model or recessive model. [fig_ref] Figure 5: Case-control study display [/fig_ref] shows the result of association analysis with SNPs and haplotypes. Bar chart displays the log transformed p-values that are sorted by descending order.
In the association analysis with haplotypes, we applied a haplotype-specific test with one degree-of-freedom. Estimation of haplotype effects was not implemented because the current version handles only the haplotype frequencies previously reconstructed in the LD blocking analysis. Several algorithms for estimating haplotype effects have been developed by many researchers [bib_ref] Inference on haplotype effects in casecontrol studies using unphased genotype data, Epstein [/bib_ref] [bib_ref] WHAP: haplotype-based association analysis, Purcell [/bib_ref] [bib_ref] Score tests for association between traits and haplotypes when linkage phase is..., Schaid [/bib_ref] [bib_ref] A new algorithm for haplotype-based association analysis: the Stochastic-EM algorithm, Tregouet [/bib_ref]. Software programs like THESIASand Haplo Statsare freely available and widely used for the analysis of haplotype effects.
## Data export
All the analyzed results can be saved as tab delimited text files for user's convenience. shows the results of association analysis, false discovery rate and reconstructed haplotypes.
## Accuracy measure
For the measurement of the accuracy of the haplotype estimation, we applied two methods [bib_ref] SNPAnalyzer: a web-based integrated workbench for single-nucleotide polymorphism analysis, Yoo [/bib_ref] [bib_ref] A new statistical method for haplotype reconstruction from population data, Stephens [/bib_ref] : (i) the accuracy measured with the average error rate and (ii) the discrepancy between the true haplotype frequencies and the estimated haplotype frequencies. The average error rate is the ratio of the number of incorrectly reconstructed samples to the total number of samples. The discrepancy was calculated using index D, given as , where is the estimated haplotype frequency and f j is the true haplotype frequency of the jth sample. The true haplotype datasets were obtained from the dbSNP database at NCBI, and the detailed description of the data isshows the accuracies of the haplotype estimation for the EM-based algorithm and PL-EM algorithm employed by SNPAnalyzer 2.0.
For the reliability of LD blocking, we compared the results produced by SNPAnalyzer 2.0 with the results by Haploview program [bib_ref] Haploview: analysis and visualization of LD and haplotype maps, Barrett [/bib_ref]. [fig_ref] Figure 7: Comparison of LD blocking [/fig_ref] shows the results of LD block- ing. For the African American group, all the LD blocks produced by two software programs were the same, except the third block. For the Asian American group, the structure of the second LD block was mismatched. The overall structures of the LD blocks are similar by both programs when using the simulated genotype dataset consisting of 100 SNPs and 135 samples. All the test data are downloadable [7].
## Performance
The performance of the software program was tested by computation time in seconds according to the numbers of SNPs and samples used for association analysis. For the performance test, we simulated several genotype datasets having different number of SNPs. All the simulated datasets contained 1,000 control samples and 1,000 case samples. Two other publicly available software programs, BEAGLE [bib_ref] Efficient multilocus association mapping for whole genome association studies using localized haplotype..., Browning [/bib_ref] and PLINK [bib_ref] PLINK: a toolset for whole-genome association and population-based linkage analyses, Purcell [/bib_ref] , were used for comparison. shows the results of computation time. The computation time increased linearly with the increasing number of SNPs. SNPAnalyzer 2.0 was slightly faster than two other software programs in spite of the fact that it created graphic results as well as statistical results. PLINK and Data export display Data export display. All the analyzed results can be saved as tab delimited text files for user's convenience: (A) association analysis, (B) false discovery rate, (C) reconstructed haplotypes. BEAGLE programs created text files only containing statistical results.
## Case-control study display
## Comparison of ld blocking
The limit of the analyzable dataset size depends on the random access memory (RAM) of user's computer. We checked that the association analysis using genotype data with over 100,000 SNPs and 2,000 samples was possible.
All the test datasets are downloadable [7].
# Discussion
In the past work, we have developed a software program that calculates linkage disequilibrium between SNPs, reconstructs haplotypes and performs quantitative trait analysis [bib_ref] SNPAnalyzer: a web-based integrated workbench for single-nucleotide polymorphism analysis, Yoo [/bib_ref]. To meet the increasing demand for wholegenome association study, we have developed SNPAnalyzer 2.0 that can handle the genetic linkage disequilibrium analysis and the genetic association analysis between disease and SNPs/haplotypes in an integrated web interface. For the accuracy of the analysis, it implements several verified algorithms and statistics. The accuracy of the haplotype estimation was very high and the results of LD blocking were similar both by SNPAnalyzer 2.0 and Haploview program [bib_ref] Haploview: analysis and visualization of LD and haplotype maps, Barrett [/bib_ref]. Some mismatched structures of LD blocks are due to the different usage of the detailed parameters or algorithms applied by each software programs. For example, Haploview program used an accelerated EM algorithm. However, SNPAnalyzer 2.0 used both the EM-based algorithm and PL-EM algorithm for haplotype estimation. Comparison among control, case and merged samples is possible for linkage disequilibrium analysis using many LD indices. False positive control is implemented by multiple test correction and false discovery rate (FDR) in the association analysis. All the results are provided as tab delimited text files for user's convenience. We plan to implement more statistical analysis in future versions: stratification analysis, interaction analysis using multiple SNPs, haplotype effects analysis, and classification analysis for multiple samples.
# Conclusion
SNPAnalyzer 2.0 performs linkage disequilibrium analysis and genetic association analysis in an integrated web interface. It implements multiple verified algorithms and statistics for the enhanced reliability of the analysis. Visual comparison and interpretation of the analysis result between two different sample groups are very comprehensive. The allelic or haplotype association and genotypic or diplotype association can be analyzed using multiple genetic models. Hundreds of thousands of SNPs and thousands of samples are analyzable in moderate time, and the analysis results are displayed in figures and tables for user's convenience.
## Availability and requirements
[fig] Figure 2: Haplotype estimation display. Haplotype estimation can be implemented in parallel using two different algorithms like EMbased algorithm and PL-EM algorithm. The upper panel shows the control options for PL-EM algorithm. Middle panel shows the observed alleles and allele frequencies. The histogram and tables in the bottom panel shows the most likely haplotypes and their frequencies in a given sample. Individual haplotypes and estimation accuracies are shown on the right part of the bottom panel. [/fig]
[fig] Figure 3: LD blocking display. The upper panel shows the control options for LD blocking. The following panel shows the linkage disequilibrium pattern. The part in deep red means that there exists strong pairwise linkage disequilibrium between adjacent SNPs and the area enclosed by a thick black line designates LD block. Tagging SNPs are represented as light blue bars up in the linkage disequilibrium pattern graph. Haplotypes, haplotype frequencies and multi-allelic D' estimated in the adjacent LD blocks are displayed in the bottom of the linkage disequilibrium pattern graph. [/fig]
[fig] Figure 5: Case-control study display. The p-values from association analysis (A) with 15 SNPs and (B) with four haplotypes in the third LD block are separately shown in the bar charts. P-values are log transformed and sorted by descending order. Odds ratios (OR) and 95% confidence interval of odds ratios are displayed simultaneously with p-value of the chi-square test in the upper table. Green horizontal line represents significance level. [/fig]
[fig] Figure 7: Comparison of LD blocking. The results of LD blocking of (A) African American ethnic group and (B) Asian American ethnic group are produced by SNPAnalyzer 2.0, and (C) African American ethnic group and (D) Asian American ethnic group by Haploview program. The structures of LD blocks consisting of 100 SNPs are produced by (E) SNPAnalyzer 2.0 and by (F) Haploview program. [/fig]
[table] Table 1: The numbers of individuals of each ethnic group and the numbers of SNPs used for redefining haplotypes [/table]
|
Use of Antihypertensive Agents and Association With Risk of Adverse Outcomes in Chronic Kidney Disease: Focus on Angiotensin‐Converting Enzyme Inhibitors and Angiotensin Receptor Blockers
Background--Our objective was to determine patterns of antihypertensive agent use by stage of chronic kidney disease (CKD) and to evaluate the association between different classes of antihypertensive agents with nonrenal outcomes, especially in advanced CKD.Methods and Results--We studied 3939 participants of the CRIC (Chronic Renal Insufficiency Cohort) study. Predictors were timedependent angiotensin-converting enzyme inhibitor or angiotensin receptor blocker , b-blocker, and calcium channel blocker use (versus nonuse of agents in each class). Outcomes were adjudicated heart failure events or death. Adjusted Cox models were used to determine the association between predictors and outcomes. We also examined whether the associations differed based on the severity of CKD (early [stage 2-3 CKD] versus advanced disease [stage 4-5 CKD]). During median follow-up of 7.5 years, reninangiotensin-aldosterone system inhibitor use plateaued during CKD stage 3 (75%) and declined to 37% by stage 5, while b-blocker, calcium channel blocker, and diuretic use increased steadily with advancing CKD. Renin-angiotensin-aldosterone system inhibitor use was associated with lower risk of heart failure (hazard ratio, 0.79; 95% confidence interval, 0.67-0.97) and death (hazard ratio, 0.78; 95% confidence interval, 0.67-0.90), regardless of severity of CKD. Calcium channel blocker use was not associated with risk of heart failure or death, regardless of the severity of CKD. b-Blocker use was associated with higher risk of heart failure (hazard ratio, 1.62; 95% confidence interval, 1.29-2.04) and death (hazard ratio, 1.22; 95% confidence interval, 1.03-1.43), especially during early CKD (P<0.05 for interaction).Conclusions--Angiotensin-converting enzyme inhibitor and angiotensin receptor blocker use decreased, while use of other agents increased with advancing CKD. Use of agents besides angiotensin-converting enzyme inhibitors or angiotensin receptor blockers may be associated with suboptimal outcomes in patients with CKD. ( J Am Heart Assoc. 2018;7:e009992.
H ypertension affects over 85 million Americans, and >15% of patients with hypertension also have chronic kidney disease . [bib_ref] NMA/PCNA guideline for the prevention, detection, evaluation, and management of high blood..., Whelton [/bib_ref] Yet the class or classes of antihypertensive medications that should be used to optimize outcomes in the CKD population remain unclear. [bib_ref] NMA/PCNA guideline for the prevention, detection, evaluation, and management of high blood..., Whelton [/bib_ref] Use of renin-angiotensin-aldosterone system inhibitors such as angiotensin-converting enzyme inhibitors (ACEIs) or angiotensin receptor blockers (ARBs) has been shown to delay the progression of chronic kidney disease (CKD). Thus, ACEIs and ARBs are currently the preferred agents in the CKD population. [bib_ref] Effects of losartan on renal and cardiovascular outcomes in patients with type..., Brenner [/bib_ref] [bib_ref] Renoprotective effect of the angiotensin-receptor antagonist irbesartan in patients with nephropathy due..., Lewis [/bib_ref] [bib_ref] Proteinuria predicting outcome in renal disease: nondiabetic nephropathies (REIN), Remuzzi [/bib_ref] [bib_ref] Use of renin-angiotensin system blockade in advanced ckd: An NKF-KDOQI controversies report, Weir [/bib_ref] While there are data to support the benefit of ACEI and ARB use during the early stages of CKD (stage 3) for both renal and cardiovascular benefit, [bib_ref] Renin-angiotensin system inhibitors and kidney and cardiovascular outcomes in patients with CKD:..., Xie [/bib_ref] especially among patients with diabetes mellitus, [bib_ref] Comparative effectiveness of renin-angiotensin system blockers and other antihypertensive drugs in patients..., Wu [/bib_ref] less data are available to support the benefit of ACEI and ARB use on outcomes in more advanced CKD (stage 4 or 5). There are also sparse data about how calcium channel blockers (CCBs) or b-blockers (BBs) may be associated with risk of cardiovascular outcomes or death in the CKD population. Few head-to-head comparisons of the effect of different classes of antihypertensive medications on renal or nonrenal outcomes have been conducted, and many of the trials of the use of different antihypertensive classes have been placebo controlled. [bib_ref] Renin-angiotensin system inhibitors and kidney and cardiovascular outcomes in patients with CKD:..., Xie [/bib_ref] [bib_ref] Comparative effectiveness of renin-angiotensin system blockers and other antihypertensive drugs in patients..., Wu [/bib_ref] The objectives of this study were to examine patterns and predictors of the class of antihypertensive medications used across CKD stages 2 to 5 in a well-characterized cohort of participants with CKD followed longitudinally in the CRIC (Chronic Renal Insufficiency Cohort) study, with a focus on ACEI and ARB use. We also examined the association of ACEI and ARB, CCB, or BB use with risk of adjudicated heart failure (HF) events and all-cause mortality, and determined whether these associations varied based on the severity of CKD (early versus advanced stages).
# Methods
## Study population
The CRIC Study is a national multicenter observational cohort that enrolled participants with an estimated glomerular filtration rate (eGFR) between 20 and 70 mL/min per 1.73 m 2 based on the Modification of Diet in Renal Disease equation between June 2003 and September 2008. Inclusion and exclusion criteria were published previously, and the study is ongoing. [bib_ref] Chronic Renal Insufficiency Cohort Study I. The Chronic Renal Insufficiency Cohort (CRIC)..., Feldman [/bib_ref] [bib_ref] Chronic Renal Insufficiency Cohort Study G. Chronic Renal Insufficiency Cohort (CRIC) study:..., Lash [/bib_ref] Informed consent was obtained from CRIC participants at all local sites. We used data from the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) Central Repository in this analysis, and follow-up time was censored as of . The University of California, San Francisco Institutional Review Board considers this study exempt human subjects research.
The data and study materials will not be made available to other researchers for purposes of reproducing the results, as these data are publicly available at no cost through the NIDDK Central Repository, and the NIDDK Central Repository prohibits re-release of these data.
Predictors of the Use of Antihypertensive Agents by Stage of CKD Medication use was reported by CRIC participants at annual visits. We were primarily interested in the use of ACEIs or ARBs, CCBs, BBs, and diuretic use (especially loop and thiazide diuretics) across the different stages of CKD and the changes in therapy that occurred with the progression of CKD. We used person-specific trajectories of renal function decline from mixed models to identify the time points when transitions to each subsequent CKD stage occurred as previously described. [bib_ref] Time-centered approach to understanding risk factors for the progression of CKD, Ku [/bib_ref] In analyses of use of antihypertensive agents across the different stages of CKD, participants were classified as "users" if they reported use of the medication at the first visit after their identified transition to the subsequent stage of CKD. The stages of CKD were defined based on eGFR falling below 90, 60, 45, 30, and 15 mL/min per 1.73 m 2 for CKD stages 2, 3a, 3b, 4, and 5, respectively. 11 Because use of ACEIs and ARBs may be more common among patients with substantial proteinuria, we also examined the prevalence of ACEI or ARB use according to the degree of proteinuria at the time of entry into each stage of CKD (≥1 versus <1 g/g of proteinuria).
If data on use of antihypertensive agents were missing, then data from the prior visit were carried forward. We also examined the number of total antihypertensive medications used in each stage of CKD (including diuretics).
## Clinical predictors of acei or arb use by each ckd stage
Next, we examined predictors of ACEI or ARB use at entry into each stage of CKD in (separate) multivariable logistic regression models as specified a priori given the solid indication for the use of ACEIs or ARBs (over other classes of antihypertensive agents) according to guideline recommendations in CKD. [bib_ref] KDOQI US commentary on the 2012 KDIGO clinical practice guideline for management..., Taler [/bib_ref] The outcome of interest in these logistic models was ACEI or ARB use (yes/no) as a binary outcome. Factors of interest included age at enrollment (≥60 versus <60 years), sex, race, annual household income, proteinuria (≥1 or <1 g/g), HF, myocardial infarction or revascularization, stroke, diabetes mellitus, obesity (body mass index ≥30 kg/m 2 ), uncontrolled systolic blood pressure
## Clinical perspective
What Is New?
- We examined patterns of antihypertensive use with advancing chronic kidney disease, and how the class of antihypertensive agent used associate differentially with adverse outcomes.
What Are the Clinical Implications?
- Use of b-blockers, calcium channel blockers, and diuretics steadily increased, whereas use of angiotensin-converting enzyme inhibitors or angiotensin receptor blockers decreased with advancing chronic kidney disease. - Use of angiotensin-converting enzyme inhibitors or angiotensin receptor blockers was associated with lower risk of heart failure and death, whereas use of b-blockers was associated with higher risk of both outcomes. - Use of calcium channel blockers was not associated with these adverse outcomes. Nonuse of angiotensin-converting enzyme inhibitors or angiotensin receptor blockers with advancing chronic kidney disease may associate with suboptimal outcomes.
(≥140 mm Hg versus <140 mm Hg), concurrent use of other antihypertensive medications (CCBs and BBs), concurrent diuretic use, and potassium concentration. We also tested for interactions between each of these risk factors and stage of CKD in fully adjusted models.
## Association between time-updated antihypertensive medication use and risk of hf
We examined the association between use of each class of antihypertensive agent as a time-dependent predictor of HF, a primary outcome of interest. Time-dependent use of ACEIs or ARBs was based on self-report and updated annually, with missing values carried forward in Fine-Gray models using a counting process formulation of these models. [bib_ref] Time-varying covariates and coefficients in Cox regression models, Zhang [/bib_ref] We chose to focus on HF as the cardiovascular outcome of interest, as it was the most common event and hence would provide the most power. Two independent reviewers adjudicated HF events in the CRIC study, and we included only events classified as "definite." [bib_ref] Urine neutrophil gelatinase-associated lipocalin and risk of cardiovascular disease and death in..., Liu [/bib_ref] [bib_ref] Hsu CY; CRIC Study Investigators. Change in measured gfr versus egfr and..., Ku [/bib_ref] [bib_ref] Persistent high serum bicarbonate and the risk of heart failure in patients..., Dobre [/bib_ref] Criteria used for the definition of an HF event were based on symptoms, radiographic changes (such as pulmonary edema), or physical examination findings (such as presence of rales and peripheral edema), central venous hemodynamic monitoring data, and review of echocardiogram data. [bib_ref] Persistent high serum bicarbonate and the risk of heart failure in patients..., Dobre [/bib_ref] [bib_ref] Risk factors for heart failure in patients with chronic kidney disease: the..., He [/bib_ref] Our primary models were adjusted Fine-Gray models (treating death as a competing risk) that accounted for age, sex, race, income status, baseline HF, baseline myocardial infarction, baseline peripheral artery disease, baseline stroke, baseline eGFR (by Chronic Kidney Disease Epidemiology Collaboration creatinine-based equation [bib_ref] Estimating glomerular filtration rate from serum creatinine and cystatin C, Inker [/bib_ref] , baseline proteinuria (<1 versus ≥1 g/g), and time-dependent covariates including diabetes mellitus, obesity (body mass index ≥30 kg/m 2 ), systolic blood pressure (BP), statin use, aspirin use, diuretic use, and concurrent use of other antihypertensive agents (CCBs and BBs) for the outcome of HF. Because we were interested in the risks or benefits associated with the use of each class of antihypertensive agent independent of achieved BP control, we chose to adjust for absolute systolic BP levels in these models.
We repeated our primary analysis using time-dependent BB or CCB use at each visit as our primary predictors of interest (and therefore adjusting for ACEI or ARB use) for the outcome of HF. We did not focus on diuretics as a primary predictor in these main analyses given the greater potential for confounding by indication due to volume overload, especially since HF was the outcome of interest, although we did adjust for diuretic use in our primary models. To reduce the likelihood of confounding by indication, we performed a sensitivity analysis for the outcome of HF in which we repeated our analysis among the subgroup without baseline HF (who would not be receiving treatment with specific classes of antihypertensive agents for secondary prevention of HF). In secondary analyses, we examined the association between diuretic use (versus nonuse) and the risk of HF.
We did not adjust for time-updated proteinuria and eGFR in our primary models, given that both BP control and ACEI or ARB use may reduce proteinuria or be associated with changes in eGFR and may be a potential mediator of outcomes of interest. However, in sensitivity analysis, we additionally adjusted for time-updated eGFR and proteinuria (categorized as <1 versus ≥1 g/g).
## Association between time-updated antihypertensive medication use and risk of death
We used unadjusted and adjusted Cox models to examine the association between use of each class of antihypertensive agent and death, adjusting for the same covariates as described above. Deaths were identified through report from next of kin, retrieval of death certificates or obituaries, review of hospital records, and linkage with the Social Security Death Index.
## Association between class of antihypertensive medication use with risk of adverse outcomes by ckd severity
We tested for interaction between use of each class of antihypertensive medication (ACEIs and ARBs, CCBs, or BBs) and CKD severity (early CKD defined as eGFR by Chronic Kidney Disease Epidemiology Collaboration equation ≥30 mL/min per 1.73 m 2 [stage 2-3] versus advanced CKD defined as <30 mL/ min per 1.73 m 2 [stage 4-5 disease]) in unadjusted and adjusted Fine-Gray or Cox models for both the outcomes of HF and death, respectively.
All analyses were conducted using SAS 9.4 (SAS Institute, NC). Informed consent was obtained for participation at all CRIC study sites.
# Results
## Patterns and prevalence of antihypertensive agent use
A total of 3939 CRIC participants (100%) were included for analysis; baseline characteristics of CRIC participants were previously described. [bib_ref] Chronic Renal Insufficiency Cohort Study I. The Chronic Renal Insufficiency Cohort (CRIC)..., Feldman [/bib_ref] Overall, mean age was 58 years, 42% were black, mean body mass index was 32 kg/m 2 , median eGFR was 43 mL/min per 1.73 m 2 , median urine protein/ creatinine ratio was 0.2 g/g, and one half had diabetes mellitus at the time of enrollment [fig_ref] Table 1: Clinical Predictors of ACEI or ARB Use as the Outcome of Interest... [/fig_ref]. Approximately 70% of participants reported ACEI or ARB use at baseline, about one half reported BB use, and 41% reported CCB use at enrollment.
With advancing stages of CKD, the number of participants receiving 3 or more antihypertensive medications steadily increased from 25% in CKD stage 2 to 75% in CKD stage 5 . By CKD stage 5, 98% of participants were receiving at least 1 antihypertensive agent .
Use of antihypertensive agents by stage of CKD is shown in the Figure-Panel A. Only approximately one half of CRIC participants were using ACEIs or ARBs in CKD stage 2. The prevalence of ACEI or ARB use increased to >75% by stage 3b and then declined to 37% in stage 5 . Thiazide diuretic use after CKD stage 3a was also less common. Use of other non-renin-angiotensin-aldosterone system antihypertensive agents steadily increased with advancing CKD . Across stages of CKD, the pattern of ACEI or ARB use was similar among the subgroup with significant proteinuria (≥1 g/g) at the first visit upon entry into each stage of CKD .
## Clinical predictors of acei or arb use by each ckd stage
In multivariable analysis, women were less likely to receive an ACEI or ARB compared with men in CKD stages 2 to 4 [fig_ref] Table 1: Clinical Predictors of ACEI or ARB Use as the Outcome of Interest... [/fig_ref]. Higher household income, black race, diabetes mellitus, obesity, and higher potassium levels were associated with ACEI or ARB use across CKD stages 3 to 5 [fig_ref] Table 1: Clinical Predictors of ACEI or ARB Use as the Outcome of Interest... [/fig_ref]. Diuretic use was associated with ACEI or ARB use in stage 3a and 3b CKD but not in more advanced stages of CKD when diuretic use became more prevalent. A history of HF was not statistically significantly associated with ACEI or ARB use across any of the CKD stages [fig_ref] Table 1: Clinical Predictors of ACEI or ARB Use as the Outcome of Interest... [/fig_ref].
## Association between time-updated antihypertensive use and risk of hf
During median follow-up of 7.0 years, 491 participants had an adjudicated HF event. Among those with HF and echocardiogram data that were available after the onset of HF (N=219), the mean ejection fraction was 44% (standard deviation, 13%).
In multivariable analysis, ACEI or ARB use was associated with lower risk of HF (hazard ratio [HR], 0.79; 95% confidence interval [CI], 0.64-0.97) regardless of CKD severity [fig_ref] Table 2: Risk of Adjudicated Heart Failure* Based on Time-Updated Antihypertensive Medication Use Heart... [/fig_ref]. CCB use was not statistically significantly associated with risk of HF (HR, 0.96; 95% CI, 0.79-1.16), regardless of CKD severity. BB use was associated with higher risk of HF, regardless of CKD severity (HR, 1.62; 95% CI, 1.29-2.04; [fig_ref] Table 2: Risk of Adjudicated Heart Failure* Based on Time-Updated Antihypertensive Medication Use Heart... [/fig_ref].
Among the subgroup with no HF at baseline (N=3557), we observed the same general patterns of risk. For example, ACEI or ARB use was associated with lower risk of HF (HR, 0.74; 95% CI, 0.59-0.95), whereas BB (HR, 1.60; 95% CI, 1.23-2.08) was persistently associated with higher risk of HF among those without baseline HF. CCB use was not associated with risk of HF (HR, 1.16; 95% CI, 0.92-1.47).
In secondary analyses, diuretic use was associated with higher risk of HF in adjusted Fine-Gray models (HR, 1.83; 95% CI, 1.43-2.32).
In sensitivity analysis, when we additionally adjusted models for proteinuria as a time-updated covariate, our overall findings were similar for the outcome of HF [fig_ref] Table 2: Risk of Adjudicated Heart Failure* Based on Time-Updated Antihypertensive Medication Use Heart... [/fig_ref]. However, the protective association between ACEI or ARB use and HF was further attenuated with the addition of timeupdated eGFR and proteinuria as covariates (HR; 0.82; 95% CI, 0.67-1.02).
## Association between time-updated antihypertensive use and risk of death
For the outcome of all-cause mortality, use of ACEIs or ARBs was associated with lower risk of death (HR, 0.78; 95% CI, 0.67-0.90), regardless of CKD severity (P=0.19 for interaction). CCB use was not associated with risk of death (HR, 0.92; 95% CI, 0.79-1.06). BB use was associated with a higher risk of death (HR, 1.22; 95% CI, 1.03-1.43), although the association was stronger during the early versus more advanced stages of CKD (P=0.02 for interaction; [fig_ref] Table 3: Risk of Death Based on Time-Updated Antihypertensive Medication Use [/fig_ref].
In sensitivity analysis, when we additionally adjusted models for proteinuria as a time-updated covariate, our overall findings were similar for the risk of death [fig_ref] Table 2: Risk of Adjudicated Heart Failure* Based on Time-Updated Antihypertensive Medication Use Heart... [/fig_ref].
In secondary analysis, diuretic use was not associated with risk of death (HR, 1.09; 95% CI, 0.94-1.28) in adjusted Cox models.
# Discussion
A number of trials have examined the effect of different classes of BP medications on cardiovascular and kidney outcomes, 3,20-27 but the majority of these trials were conducted in people with normal kidney function or mild renal impairment, and few trials have provided long-term data on mortality outcomes. [bib_ref] Cardiovascular and renal outcomes with telmisartan, ramipril, or both in people at..., Tobe [/bib_ref] [bib_ref] Chronic kidney disease, cardiovascular risk, and response to angiotensin-converting enzyme inhibition after..., Tokmakova [/bib_ref] In this study, we found that ACEI or ARB use began to decline after CKD stage 3b and was not as prevalent as would be expected, especially among participants with substantial proteinuria for whom guidelines clearly recommend ACEIs or ARBs as first-line agents. [bib_ref] KDOQI US commentary on the 2012 KDIGO clinical practice guideline for management..., Taler [/bib_ref] In contrast, use of CCBs, diuretics, and BBs increased steadily with advancing stages of CKD. Even after accounting for systolic BP control, we found that ACEI or ARB use was associated with a substantially lower risk of HF and death.
CCB use was not associated with risk of HF or death. However, BB use was associated with a higher risk of HF and death, especially during the early stages of CKD.
Currently, the American Heart Association guidelines for high BP recommend using ACEIs or ARBs as first-line therapy for patients with CKD stage 3 and higher, or for patients with CKD stage 1 to 2 with albuminuria ≥300 mg/g. [bib_ref] NMA/PCNA guideline for the prevention, detection, evaluation, and management of high blood..., Whelton [/bib_ref] These recommendations are primarily based on the added benefit of proteinuria reduction associated with the use of ACEIs or ARBs, which has been shown to slow CKD progression. [bib_ref] KDOQI US commentary on the 2012 KDIGO clinical practice guideline for management..., Taler [/bib_ref] However, we noted that ACEI or ARB use decreased substantially from stage 3b to stage 5, and 63% of participants were not receiving ACEIs or ARBs by the time they reached CKD stage 5 despite the fact that 98% were receiving antihypertensive agents. Even at their peak use, only three quarters of CRIC participants with proteinuria ≥1 g/g were receiving guideline-recommended ACEI or ARB therapy. Participants who were obese and had diabetes mellitus were more likely to receive ACEI or ARB therapy, whereas women were less likely to receive these agents. Whether this observation is related to concerns about fetal risks during pregnancy with these agents is unclear.
Although we do not have data to address why so many participants were not receiving guideline-recommended treatment of CKD and hypertension, we expect that providers might have been concerned about risk (or occurrence) of acute kidney injury or hyperkalemia with use of ACEIs or ARBs. [bib_ref] The frequency of hyperkalemia and its significance in chronic kidney disease, Einhorn [/bib_ref] Despite their known efficacy, the prevalence of ACEI and ARB use in patients with CKD has been reported to be low, even when there are solid indications for their use (eg, diabetic nephropathy or HF). [bib_ref] Angiotensin-converting enzyme inhibitors and angiotensin receptor blockers in patients with congestive heart..., Berger [/bib_ref] [bib_ref] Underuse of ACE inhibitors and angiotensin II receptor blockers in elderly patients..., Winkelmayer [/bib_ref] [bib_ref] Influence of renal insufficiency on the prescription of evidence-based medicines in patients..., Peng [/bib_ref] [bib_ref] Antihypertensive medication use in older patients transitioning from chronic kidney disease to..., Chang [/bib_ref] However, we believe that our study provides new information that may warrant greater caution in withdrawing ACEIs or ARBs in the short term and highlights the need for interventional trials of alternative strategies in the future. For example, the availability of newer potassium binders that may be better tolerated than sodium polystyrene and safer for long-term use [bib_ref] Effect of patiromer on serum potassium level in patients with hyperkalemia and..., Bakris [/bib_ref] [bib_ref] Patiromer in patients with kidney disease and hyperkalemia receiving RAAS inhibitors, Weir [/bib_ref] [bib_ref] Current and future potassium binders, Bakris [/bib_ref] [bib_ref] Advances in treatment of hyperkalemia in chronic kidney disease, Sarafidis [/bib_ref] presents the opportunity to continue ACEIs or ARBs, even in the advanced stages of CKD. In light of the beneficial association between ACEI or ARB use and risk of HF and death (that may at least be partially mediated by their effect on proteinuria), prospective clinical trials are warranted to examine whether these strategies reduce HF and mortality. [bib_ref] Patiromer in patients with kidney disease and hyperkalemia receiving RAAS inhibitors, Weir [/bib_ref] We believe the finding of a substantially higher risk of HF and mortality risk with b-blocker use in our study to be important, particularly since this association was present even among those without known HF at enrollment. Patients receiving BBs have also been reported to have a higher risk of cardiovascular events (including HF) in cohorts without CKD and in the perioperative setting. [bib_ref] Risk of cardiovascular events in patients with diabetes mellitus on beta-blockers, Tsujimoto [/bib_ref] [bib_ref] Association of beta-blocker therapy with risks of adverse cardiovascular events and deaths..., Andersson [/bib_ref] In addition, a number of large meta-analyses have suggested that BB use was associated with a higher risk of stroke and death in the general population. [bib_ref] Should beta blockers remain first choice in the treatment of primary hypertension?..., Lindholm [/bib_ref] [bib_ref] Beta-blockers for hypertension, Wiysonge [/bib_ref] In contrast, a meta-analysis of trials of patients with CKD suggested that BBs conferred a benefit on the risk of mortality and HF despite a higher risk of hypotension and bradycardia. [bib_ref] Effects of beta-adrenergic antagonists in patients with chronic kidney disease: a systematic..., Badve [/bib_ref] The strong association between BB use and risk of adverse outcomes in our study deserves further investigation to understand the reasons for this observation. However, we do emphasize that there may be other compelling indications for the continuation of b-blockade (such as in the setting of angina or myocardial infarction) outside of the outcomes of interest that we examined.
Of note, we did not find CCB use to be associated with the risk of HF or death. Given that few trials have tested CCBs against alternate classes of antihypertensive agents in patients with CKD (especially those with advanced disease), further studies are needed to confirm our findings. In our secondary analyses, we found a higher risk of HF but not death with diuretic use (versus nonuse), but we recognize that these results may be prone to confounding by indication.
The strengths of our study include the large cohort size, its national representativeness, long duration of follow-up, large number of events, and formal adjudication of HF events. In addition, we believe that the use of research-grade data collected during routine study visits (as opposed to using electronic health records where ascertainment bias may occur in sicker patients who are seen more frequently) is a strength. However, because of the observational nature of our study, we cannot rule out the presence of residual confounding or confounding by indication, and our results do not imply causation. In addition, we did not have granular detail on the specific medications that were used (eg, atenolol versus metoprolol), their dosage, or participant adherence to therapy. Although medication use in CRIC was ascertained on an annual basis, we are unable to capture any changes that occurred between the yearly visits and their associated changes in BP and albuminuria in the short term. We also did not have data on the reasons that patients were not receiving ACEIs or ARBs or other antihypertensive agents, including historical episodes of hyperkalemia or acute kidney injury that may have limited renin-angiotensin-aldosterone system inhibition. We also lacked data on causes of death and hence were unable to examine cardiovascular-specific mortality. Finally, CRIC participants may not be representative of patients who are not managed by nephrologists.
In conclusion, we observed an association between ACEI or ARB use and lower risk of HF and death, but >60% of the CRIC ACEI indicates angiotensin converting enzyme inhibitors; ARB, angiotensin receptor blockers; BMI, body mass index; BP, blood pressure; CI, confidence interval; eGFR, estimated glomerular filtration rate; HF, heart failure; MI, myocardial infarction; PAD, peripheral artery disease. *Adjusted for baseline age, race, sex, income, baseline HF, baseline MI, baseline stroke, baseline PAD, baseline eGFR, baseline proteinuria (≥1 or <1 g/g), and time-updated covariates including diabetes mellitus, obesity (yes/no BMI >30 kg/m 2 ), systolic BP, aspirin use, statin use, diuretic use, and other antihypertensive medication use (calcium channel blocker, ACEIs or ARBs, or b-blockers).
population was not receiving ACEIs or ARBs in some stages of disease. BB use was associated with higher risk of HF and death, especially during the earlier stages of CKD. CCBs were not associated with risk of HF or death. Given the decline in use of ACEIs or ARBs with advancing stages of CKD and increase in use of alternative antihypertensive agents, optimizing use of ACEIs or ARBs throughout all stages of CKD may be associated with improved outcomes, and use of BBs may be associated with adverse outcomes. Future trials are needed to determine the optimal antihypertensive agents in the various stages of CKD, and the inclusion of more patients with more advanced CKD in clinical trials is needed to enhance our understanding of how to optimize outcomes for this high-risk population.
# Acknowledgments
The data from the CRIC study reported here were supplied by the NIDDK Central Repositories. This manuscript was not prepared in collaboration with Investigators of the CRIC study and does not necessarily reflect the opinions or views of the CRIC study, the NIDDK Central Repositories, or the NIDDK.
# Sources of funding
Dr Ku was funded by National Institutes of Health K23 HL131023. This work was also supported by the National Institutes of Health K24 DK85153 to Dr Johansen. The CRIC study was conducted by the CRIC Investigators and supported by the NIDDK.
# Disclosures
None.
[fig] Figure.: A, Use of different classes of antihypertensive agents by stage of CKD. B, Use of different classes of antihypertensive agents by stage of CKD and degree of proteinuria. ACE indicates angiotensin-converting enzyme; ARB, angiotensin receptor blocker; CKD, chronic kidney disease; UPCR, urine protein/creatinine ratio. [/fig]
[table] Table 1: Clinical Predictors of ACEI or ARB Use as the Outcome of Interest by Each CKD Stage [/table]
[table] Table 2: Risk of Adjudicated Heart Failure* Based on Time-Updated Antihypertensive Medication Use Heart Failure (N=3939) Unadjusted Hazard Ratio (95% CI) Adjusted Hazard Ratio † (95% CI) Adjusted P Value of Interaction by eGFR (≥ or <30 mL/min per 1.73 m 2 )ACEI indicates angiotensin converting enzyme inhibitor; ARB, angiotensin receptor blocker; BMI, body mass index; BP, blood pressure; CI, confidence interval; CKD-EPI, Chronic Kidney Disease Epidemiology Collaboration; eGFR, estimated glomerular filtration rate; HF, heart failure; MI, myocardial infarction; PAD, peripheral artery disease. *Defined as HF (definite event). † Adjusted for baseline age, race, sex, income, baseline CHF, baseline MI, baseline stroke, baseline PAD, baseline eGFR (by CKD-EPI), baseline proteinuria (≥1 or <1 g/g), and time-updated covariates including diabetes mellitus, obese (yes/no BMI ≥30 kg/m 2 ), systolic BP, aspirin use, statin use, diuretic use, and other antihypertensive medication use (calcium channel blocker, ACEIs or ARBs, or b-blockers). [/table]
[table] Table 3: Risk of Death Based on Time-Updated Antihypertensive Medication Use [/table]
|
Background: Spinal epidural hematoma is a rare condition usually secondary to trauma and coagulopathy. To the best of our knowledge, we present the first case of a patient with an iatrogenic hypercoaguable state performing self-neck manipulation, which resulted in a spinal epidural hematoma and subsequent quadriparesis. Case presentation: A 63-year-old man presented to the emergency department with worsening interscapular pain radiating to his neck 1 day after performing self-manipulation of his cervical spine. He was found to be coagulopathic upon admission, secondary to chronic warfarin therapy for the management of atrial fibrillation. Approximately 48 h after the manipulation, the patient became acutely quadriparetic and hypotensive. Urgent magnetic resonance imaging revealed a multilevel spinal epidural hematoma from the lower cervical to thoracic spine. Conclusions: Partial C7, complete T1 and T2, and partial T3 bilateral laminectomy was performed for evacuation of the spinal epidural hematoma. Following a 2-week course of acute inpatient rehabilitation, the patient returned to his baseline functional status. This case highlights the risks of self-manipulation of the neck and potentially other activities that significantly stretch or apply torque to the cervical spine. It also presents a clinical scenario in which practitioners of spinal manipulation therapy should be aware of patients undergoing anticoagulation therapy.
# Background
Spinal epidural hematoma (SEH) is an uncommon condition with potential for severe neurological compromise. Differentials for SEH include spontaneous, traumatic, iatrogenic, and coagulopathy related. In rare instances, SEH has followed spinal manipulation therapy (SMT). SMT commonly involves high-velocity, low-amplitude forces applied to articulations of the spinal column. In 2015, Huang et al.presented 12 cases of SEH related to SMT, highlighting the rarity of this clinical phenomenon. Acute (< 24 h) epidural hematoma of the cervical spine following SMT has only been reported 6 times in the English literature. Self-neck manipulation, described as a self-applied force of rapid cervical spine lateral rotation with a "popping" sensation, has been linked to cervical artery dissectionbut just once to SEH. To the best of our knowledge, we present the first case of a patient with an iatrogenic hypercoaguable state performing self-neck manipulation, which resulted in SEH and subsequent quadriparesis. It is critical for health care practitioners to be aware of potential risks when considering SMT as part of a treatment regimen in candidates undergoing anticoagulation therapy.
## Case presentation
A 63-year-old man presented to the emergency department (ED) with unremitting interscapular pain radiating to his neck, which he suspected to be from a myocardial infarction (MI). Onset was 1 day after performing selfmanipulation of his cervical spine. He described the maneuver by grabbing his chin and applying a rapid, rotating motion bilaterally. He reported a long history (> 20 years) of performing this maneuver intermittently with perceived benefit. Initial workup in the ED aimed to identify the potential source of pain and consisted of computed tomography angiogram of the chest and abdomen, which excluded dissecting thoracic aneurysm, MI, and pulmonary embolism. Relevant past medical history included atrial fibrillation (AF) managed with long-term warfarin therapy. His international normalized ratio (INR) was 7.84 upon admission to the ED, and he was admitted to the hospital for closer monitoring of any signs or symptoms of neurovascular compromise. The following day, the patient became acutely quadriparetic and hypotensive. His blood pressure decreased to 75/38 mmHg, and his neurological level was determined to be C7 American Spinal Injury Association Impairment Scale D.
Urgent magnetic resonance imaging (MRI) of the cervical and thoracic spine demonstrated multilevel SEH from the lower cervical to thoracic spine. With identification of an active bleed and an INR remaining close to 7.0, the patient's hypercoaguable state was reversed with fresh frozen plasma and factor VII prior to emergency surgery. The patient underwent partial C7, complete T1 and T2, and partial T3 bilateral laminectomy for evacuation of SEH. Placement of an epidural drainage catheter was also performed with the use of Carm fluoroscopy.
Throughout the first 6 days following surgery, the patient recovered some function of the upper and lower extremities while undergoing hospital-based physical therapy. Post-operative MRI showed resolution of the SEH and cord compression without resultant cord damage . The patient was admitted 7 days postoperatively to an acute inpatient rehabilitation unit with: 1) INR of 1.23; 2) acute kidney injury, which improved with hydration; 3) atrial fibrillation converted to sinus bradycardia on antiarrhythmic medications, not to be anticoagulated; 4) neurogenic bladder secondary to spinal cord injury; 5) diabetes mellitus controlled with medication; 5) and anemia, which improved with iron supplementation. Physical examination revealed 4/5 manual muscle strength testing to the upper and lower extremities. He exhibited moderate dysmetria with finger-to-nose testing and slow initiation of movement. Deep tendon reflexes were 1+/4 to the bilateral lower extremities but 2+/4 to the upper extremities. There were no findings of ankle clonus or pronator drift.
The patient completed a 2-week course of acute interdisciplinary rehabilitation. The epidural drainage catheter remained in place for most of his post-operative stay. His neurogenic bladder with urinary retention resolved. At hospital discharge, the patient's greatest deficits remained his fine motor skills to his right hand along with residual paresthesia in his lower extremities. He also exhibited mild dyspraxia with finger-to-nose testing on the right and heel-to-shin testing bilaterally. The patient regained full manual muscle strength of all 4 extremities and ultimately returned to his baseline functional status, which included ambulating independently with a cane and independence with activities of daily living.
# Discussion and conclusions
SEH is a rare surgical emergency associated with trauma, coagulopathy, spontaneous development, and iatrogenic procedures, such as spine surgery and spinal punctures. The incidence of SEH is estimated between 0.15 and 1.5 per 100,000 people annually. Even more infrequent, is the incidence of SEH following SMT. This is the first reported case of SEH following self-neck manipulation in a patient on chronic anticoagulation therapy. A report of one case should not suggest causality. It should be noted that SEH symptoms manifested approximately 48 h after the neck manipulation, and the patient had an elevated INR for a sustained period, which may also have been a contributing factor. Adjusted-dose warfarin requires routine monitoring in AF patients to maintain a therapeutic INR target of 2.0 to 3.0, as one of the strongest predictors of major bleeding is INR above 3.0. The patient in our case did not adhere to routine monitoring standards, rendering him susceptible to a bleed. Although INR monitoring is intimately related to warfarin dosing, practitioners should also understand INR does not apply to patients on non-vitamin K oral anticoagulants, such as dabigatran, rivaroxaban, apixaban, and edoxaban. Nonetheless, this report highlights a potential complication of cervical spinal manipulation and other maneuvers requiring extreme cervical rotation in patients who are over-anticoagulated. This relationship requires further investigation.
In a review of the literature, there is only one case that described self-neck "twisting" temporally associated with development of a cervical epidural hematoma. Interestingly, this self-applied maneuver occurred 2 h after a spinal puncture but in a patient without impaired coagulation. SMT is generally considered a safe procedure but should only be performed by trained health care professionals, following a complete medical history and a thorough physical examination. The incidence of serious complications from SMT, associated with neurological and vascular injury, is estimated between 1 in 400,000 and 1 in 2,000,000 manipulations. The incidence of adverse events from self-neck manipulation is unknown. Risk factors for serious complications of SMT may include failure to identify patients with coagulation disorders. Manipulation of the spine of an overanticoagulated patient could potentially increase the likelihood of SEH.
The mechanism of injury between SMT and SEH may be related to the source of bleeding. Although arterial bleeding has been debated as the origin of hemorrhage in SEH, a considerable body of evidence supports the theory that the venous system is the primary source. The posterior epidural venous plexus is a delicate network of vessels, largely unprotected from trauma or changes in abdominal or thoracic pressure, potentially making it susceptible to rupture. This may explain why a forceful movement, such as self-neck manipulation or SMT, could be capable of injuring vulnerable epidural venous structures. In our case, the patient's neurovascular deficits clinically manifested as quadriparesis and hypotension approximately 48 h following reported selfneck manipulation. Bleeding from epidural veins without Post-operative sagittal T2-weighted MRI of the cervical spine demonstrates resolution of the epidural hematoma with decompression of the spinal cord (asterisk) at the C7 level. There has been restoration of normal spinal canal volume and there is no evidence of residual cord compression or myelomalacia. Note the absent visualization of the C7, T1, T2, and T3 spinous processes. The patient had partial C7, complete T1 and T2, and partial T3 laminectomy for decompression and hematoma evacuation valves could explain this slow progression of spinal cord compression.
From a clinical perspective, once the epidural space is breached, the source of hemorrhage does not appear to be a major prognosticator. Regardless of etiology, SEH causes compression of the spinal cord, which can result in irreversible neurological deficits if not treated in a timely manner. The strongest predictor of clinical outcome in SEH patients is directly related to the length of time between ictus and surgical intervention. Although spontaneous resolution has been reported, early diagnosis via MRI and surgical decompression remain the standard of care. The patient in our case likely made a full recovery due to the emergency surgical intervention that occurred promptly after his progressive neurological deficits manifested.
The rarity of traumatic SEH makes it difficult to establish cause and effect relationships, though all potential risks of SMT should be communicated and discussed during consenting of patients. Of great clinical importance are the potential reasons why bleeding into the epidural space occurs. Oral anticoagulation use resulting in supratherapeutic INR has become a predominant risk factor. Until the relationship between SEH and manipulation of the spine is better understood, we advise that SMT and exercises or stretches that apply significant torque to the cervical spine should be cautioned in patients with known coagulopathy. Further research is needed to determine whether certain side effects of anticoagulation therapy should be considered contraindications to SMT, necessitating consideration of a more careful clinical course. |
Nothing in Evolution Makes Sense Except in the Light of Genomics: Read–Write Genome Evolution as an Active Biological Process
The 21st century genomics-based analysis of evolutionary variation reveals a number of novel features impossible to predict when Dobzhansky and other evolutionary biologists formulated the neo-Darwinian Modern Synthesis in the middle of the last century. These include three distinct realms of cell evolution; symbiogenetic fusions forming eukaryotic cells with multiple genome compartments; horizontal organelle, virus and DNA transfers; functional organization of proteins as systems of interacting domains subject to rapid evolution by exon shuffling and exonization; distributed genome networks integrated by mobile repetitive regulatory signals; and regulation of multicellular development by non-coding lncRNAs containing repetitive sequence components. Rather than single gene traits, all phenotypes involve coordinated activity by multiple interacting cell molecules. Genomes contain abundant and functional repetitive components in addition to the unique coding sequences envisaged in the early days of molecular biology. Combinatorial coding, plus the biochemical abilities cells possess to rearrange DNA molecules, constitute a powerful toolbox for adaptive genome rewriting. That is, cells possess "Read-Write Genomes" they alter by numerous biochemical processes capable of rapidly restructuring cellular DNA molecules. Rather than viewing genome evolution as a series of accidental modifications, we can now study it as a complex biological process of active self-modification.In this brief review, we will examine some of the evolutionary lessons from genome sequence data. Given the broad extent of the subject matter, only a select range of examples will receive attention. The investigation shows that many evolutionary changes result from cellular processes that produce abrupt changes in genome structure and organismal characters. These processes include symbiogenesis, horizontal DNA transfer, hybrid speciation and natural genetic engineering, especially the action of mobile DNA elements[6,7]. At the molecular level, evolution is often saltational rather than gradual[8], and punctuated equilibrium [9] is the default pattern to expect in the history of phenotypic properties. As we shall see, this view is quite different from Dobzhansky's thinking. In particular, molecular sequence analysis documents many genomic elements and novel forms of hereditary variation that were hardly conceivable in Dobzhansky's lifetime.The Genome as a Highly Formatted Sequence DatabaseGenomes serve as controlled read-write databases for the reliable transmission and regulated expression of DNA sequence information[7,8,10]. Because controls over expression, replication and transmission are essential to survival and reproduction, genomes contain abundant cis-acting DNA formatting elements in addition to coding sequences determining RNA and protein primary structures[11].Many of the formatting DNA elements are generic and present at multiple genome locations. Consequently, repetitive DNA is a major component of all genomes, and repeats frequently represent the majority of an organism's DNA[12]. In the human genome, the repetitive content is about two-thirds of the total DNA, compared to well under 5% for protein-coding sequences[13].For evolution, the significance of genome formatting by repeat DNA elements is that important phenotypic alterations can occur as a result of changes to so-called "non-coding" sequences. Recent discoveries about the functions of "non-coding" ncRNA molecules illustrate this point[14][15][16][17][18][19], as do Evo-Devo findings that many developmental alterations distinguishing related organisms occur in regulatory DNA elements rather than coding sequences [20,21].
# Introduction
The title of this mini-review is a paraphrase of Dobzhansky's famous dictum, "Nothing in biology makes sense except in the light of evolution" [bib_ref] Nothing in biology makes sense except in the light of evolution: Theodosius..., Ayala [/bib_ref] [bib_ref] Nothing in biology makes sense except in the light of evolution, Dobzhansky [/bib_ref]. The reason for this paraphrase is to emphasize how, since Dobzhansky's day, genetics and evolution science have moved into the era of a revolutionary new technology, DNA sequencing. Whenever such a technological revolution occurs, science must always ask itself what impact data from the new methods has on the validity of prevailing concepts.
Our current ideas of heredity center on DNA replication and transmission [bib_ref] Genetical implications of the structure of deoxyribonucleic acid, Watson [/bib_ref]. Although cell and organismal heredity involves transmission of all cellular molecules, we conventionally view RNAs, proteins, lipids, polysaccharides and other biomolecules as derivative products resulting from biochemical activities determined by genomic DNA sequences. According to this DNA-centered perspective, stable inherited changes in organismal properties result primarily from alterations in the genome. New DNA sequences can encode new biochemical capabilities that lead to novel traits. If evolution is the acquisition of new characters over time [bib_ref] Origin of Species, Darwin [/bib_ref] , the most basic causal events should emerge from the processes that generate new DNA sequences. Those events are traceable in genome sequences, which today serve as the ultimate empirical data to test evolutionary hypotheses.
## Molecular phylogenies based on core information-processing systems
The most basic impact of genomics on evolution science has been the use of sequence data to establish relationships among different organisms. In the 1970s, Carl Woese pioneered molecular phylogenetic methodologies . Woese chose to base his initial phylogenies on the sequence of small subunit ribosomal RNA for two reasons: (i) ssrRNA was abundant and amenable to 1970s sequence analysis methods; and (ii) the ribosome is a highly conserved central component of information transfer from the genome to the proteome. Because all cells have similar but still distinctive ribosomes, this organelle provided phylogenetic data to establish connections between very diverse organisms.
Using ssrRNA sequence analysis, Woese and his colleagues unexpectedly discovered that there are in fact two separate kingdoms of prokaryotic organisms (those lacking separate cell nuclei), not one bacterial kingdom as previously believed . The two groups of prokaryotic organisms turned out to exhibit ssrRNA sequence clusters as phylogenetically distant from each other as both clusters were from the ssrRNA sequences of eukaryotic (nucleated) organisms. The evolutionary distance between the two prokaryotic groups was further confirmed by major differences in their cell membranes as well as by differences in the basic processes of genome replication and expression. The newly discovered prokaryotic cell kingdom was labeled Archaea because the first members to be studied were organisms isolated from extreme environments thought to resemble the early Earth . Today, however, there is no reason to assume that Archaea are any more ancient than Bacteria.
The most basic tenet of evolution science-the genetic relatedness of all living organisms-is abundantly supported by molecular biology, in particular the universal features of the triplet code for amino acids and the similarities of core cell structures, like the ribosome, associated translation factors , and DNA and RNA polymerases. Nonetheless, the first phylogenetic application of genomics revealed a previously unknown complexity in the biosphere and, as we shall see in the next section, provided compelling evidence for a large number of evolutionary processes excluded by conventional evolutionary thinking based on population genetics and the Modern Synthesis . As highlighted below, the phylogenies of Bacteria, Archea, eukaryotes and giant viruses continue to pose intriguing challenges in understanding the networked evolutionary connections between all domains of life .
## Eukaryotic origins and major eukaryotic taxonomic originations through symbiogenesis
The identification of an unsuspected bifurcation among prokaryotes, the most abundant living organisms , immediately raised questions about the historical relationships between Archaea and Bacteria without nuclei, on the one hand, and Eukarya, the organisms with nucleated cells, on the other. Eukaryotes formed a coherent separate group based on ribosomal RNA sequences, and molecular phylogenies of different eukaryotic groups confirmed well-established taxonomic classifications, such as fungi, plants and animals. But the generic eukaryotic cell was closer to Bacteria in some features-membrane composition, metabolic pathways-and closer to Archaea in other features-replication, transcription and translation . This phenotypic dichotomy gave support for longstanding but hotly disputed arguments championed by Lynn Margulis and others that symbiogenetic cell fusions served to create complex eukaryotic cells from simpler prokaryotic progenitors [29,32-36].
## Endosymbiotic bacterial origins of eukaryotic organelles
Molecular phylogenetics unequivocally documented the role of symbiogenesis in Eukarya evolution by analysis of the two cell organelles that have their own genomes: (1) the mitochondrion, carrying out oxidative metabolism; and (2) the chloroplast, carrying out oxidative photosynthesis. All eukaryotic cells contain a functional mitochondrion or a non-oxidative derivative organelle [37-39]. Thus, the ancestral eukaryote must have acquired an endosymbiotic mitochondrial precursor, and molecular phylogenetics unambiguously identified the mitochondrion as belonging to the Alpha-proteobacteria group . Similarly, photosynthetic eukaryotes, including red and green algae and plants, have chloroplasts, and molecular phylogenetics identified the chloroplast as a member of the cyanobacteria [43-45]. The chloroplast, and derivative plastid organelles in phylogenetically related non-photosynthetic eukaryotes, therefore must have descended from one or more endosymbiotic cyanobacteria [46,47].
## Dna transfer between endosymbiotic organelle and nuclear genomes
The genomes of mitochondria and chloroplasts do not encode all the proteins inherited from their bacterial ancestors. Many organelle proteins are encoded in the nuclear genome as a result of DNA transfers into the nuclear genome from mitochondria
## Origins of photosynthetic eukaryotic taxa by secondary endosymbiogenesis
A large number of photosynthetic eukaryotes did not evolve directly from algae or plants and are most closely related taxonomically to non-photosynthetic organisms . These organisms originated from "secondary" eukaryote to eukaryote symbiogenetic events where a red or green alga has become an endosymbiont in the initially non-photosynthetic lineage [bib_ref] Phylogenomic evidence for separate acquisition of plastids in cryptophytes, haptophytes, and stramenopiles, Baurain [/bib_ref] [bib_ref] Chromalveolates and the evolution of plastids by secondary endosymbiosis, Keeling [/bib_ref] [bib_ref] Plastid evolution: Remnant algal genes in ciliates, Archibald [/bib_ref] [bib_ref] Macroevolution via secondary endosymbiosis: A Neo-Goldschmidtian view of unicellular hopeful monsters and..., Kutschera [/bib_ref] [bib_ref] Differential gene transfers and gene duplications in primary and secondary endosymbioses, Zauner [/bib_ref] [bib_ref] A common red algal origin of the apicomplexan, dinoflagellate, and heterokont plastids, Janouškovec [/bib_ref]. The resulting photosynthetic cells have four different genome compartments that exchange DNA segments: nucleus, mitochondrion, plastid and nucleomorph (descended from the algal nucleus) [bib_ref] Nucleomorph genomes, Moore [/bib_ref] [bib_ref] Phylogeny and nucleomorph karyotype diversity of chlorarachniophyte algae, Silver [/bib_ref] [bib_ref] Nucleomorph genomes: Structure, function, origin and evolution, Archibald [/bib_ref] [bib_ref] Permuted tRNA genes in the nuclear and nucleomorph genomes of photosynthetic eukaryotes, Maruyama [/bib_ref] [bib_ref] Algal genomes reveal evolutionary mosaicism and the fate of nucleomorphs, Curtis [/bib_ref] [bib_ref] Nucleomorph genome sequence of the cryptophyte alga Chroomonas mesostigmatica CCMP1168 reveals lineage-specific..., Moore [/bib_ref] [bib_ref] Nucleomorphs: Enslaved algal nuclei, Cavalier-Smith [/bib_ref]. As with mitochondria and chloroplasts, the major intracellular DNA transfers occur from the organelles, including the nucleomorph, into the nuclear genome. The photosynthetic taxa arising from secondary endosymbiosis include euglenids and chlorachniophytes from green algal endosymbiosis and the chromalveolates from red algal endosymbiosis . The large chromalveolate phylum includes major photosynthetic organisms responsible for a large fraction of atmospheric oxygen, such as brown algae, dinoflagellates and diatoms.
Photosynthetic endosymbiosis is not restricted to unicellular eukaryotes. There are cases where animals have acquired photosynthetic capabilities by symbiogenetic events [bib_ref] Chloroplasts as functional organelles in animal tissues, Trench [/bib_ref]. The photosynthetic animals include sea slugs [117-119] and molluscs [120,121].
## Formation of a primitive eye-like organ in a unicellular eukaryote by serial endosymbioses
Among the photosynthetic dinoflagellates resulting from algal ensymbiosis, there is a group labeled "ocelloids," which possess a remarkable light-harvesting organ (the ocelloid) that resembles a complex camera-like animal eye . The ocelloid has analogues to the cornea, lens, iris and retina. A noteworthy recent paper reports that genomic analysis reveals that each of these structures resulted from a distinct endosymbiogenetic event [124]:
"Here we show, using a combination of electron microscopy, tomography, isolated-organelle genomics, and single-cell genomics, that ocelloids are built from pre-existing organelles, including a cornea-like layer made of mitochondria and a retinal body made of anastomosing plastids. We find that the retinal body forms the central core of a network of peridinin-type plastids, which in dinoflagellates and their relatives originated through an ancient endosymbiosis with a red alga. As such, the ocelloid is a chimaeric structure, incorporating organelles with different endosymbiotic histories."
The genomics-verified example of ocelloid formation by serial endosymbiogenesis has to be considered in light of Darwin's famous statement: "If it could be demonstrated that any complex organ existed, which could not possibly have been formed by numerous, successive, slight modifications, my theory would absolutely break down." [bib_ref] Origin of Species, Darwin [/bib_ref] (p. 189). While the endosymbiogenetic origin of the ocelloid does not demonstrate the impossibility of camera eyes evolving by Darwinian gradualism, this striking case of adaptive evolution by repeated cell fusion events clearly lies outside the parameters of evolutionary processes envisioned by the author of Origin of Species as well as by the Modern Synthesis.
## Eukaryotic speciation by inter-specific hybridization, whole genome duplications and genome restructuring
When we ask how novel species arise from human intervention, it is significant that there are no cases where selection has led to species formation. Selection only modifies existing characteristics by reducing or amplifying them. Artificial species arise through hybridization, as in the case of the wheat-rye hybrid Triticale [125,126], and involve genome mergers and whole genome duplication (WGD) events . A similar "cataclysmic evolution" process involving hybridization of wild grasses was at the origin of flour wheat (Triticum) several thousand years ago and can be reproduced in real time . Ongoing abrupt hybrid speciation has been observed to occur in wild sunflowers . A recent paper reports laboratory formation of a novel tobacco species with a double genome by fusion of tissue culture cells from two different natural Nicotiana species [132].
Many genomes contain duplicate copies of extended chromosome regions holding homologous genetic loci in a conserved ("syntenic") order . By documenting the prevalence of numerous syntenic duplications, genomic analysis provides compelling evidence that hybrid speciation and WGD have been common events in the history of evolution in all eukaryotic groups, including yeasts and other fungi [bib_ref] Beyond the whole-genome duplication: Phylogenetic evidence for an ancient interspecies hybridization in..., Marcet-Houben [/bib_ref] [bib_ref] Polyploidy in fungi: Evolution after whole-genome duplication, Albertin [/bib_ref] , ciliated protists [bib_ref] Global trends of whole-genome duplications revealed by the ciliate Paramecium tetraurelia, Aury [/bib_ref] , plants [bib_ref] Genome duplication and the origin of angiosperms, De Bodt [/bib_ref] [bib_ref] Widespread genome duplications throughout the history of flowering plants, Cui [/bib_ref] and animals, including two successive WGD events at the origins of vertebrates [bib_ref] Two rounds of whole genome duplication in the ancestral vertebrate, Dehal [/bib_ref] [bib_ref] The medaka draft genome and insights into vertebrate genome evolution, Kasahara [/bib_ref] [bib_ref] The 2R hypothesis: An update, Kasahara [/bib_ref] [bib_ref] Genome duplication, extinction and vertebrate evolution, Donoghue [/bib_ref] [bib_ref] Whole-genome duplications in the ancestral vertebrate are detectable in the distribution of..., Hughes [/bib_ref] , followed by further WGD events in bony fishes [bib_ref] From 2R to 3R: Evidence for a fish-specific genome duplication (FSGD), Meyer [/bib_ref] [bib_ref] Whole-genome duplication in teleost fishes and its evolutionary consequences, Glasauer [/bib_ref].
In addition to WGD, interspecific hybridization leads to episodes of genome instability and restructuring [bib_ref] Genomic instability within centromeres of interspecific marsupial hybrids, Metcalfe [/bib_ref] [bib_ref] Genomic instability in Solanum tuberosumˆSolanum kurtzianum interspecific hybrids, Marfil [/bib_ref] [bib_ref] Rapid genomic changes in interspecific and intergeneric hybrids and allopolyploids of Triticeae, Han [/bib_ref]. Chromosome rearrangements have long been recognized as major features of speciation and taxonomic divergence [bib_ref] Chromosomes of the vertebrates, White [/bib_ref] [bib_ref] Chromosomal rearrangements and karyotype evolution in carnivores revealed by chromosome painting, Nie [/bib_ref]. Repetitive and mobile DNA elements play special roles in chromosome restructuring [bib_ref] Gross chromosome rearrangements mediated by transposable elements in Drosophila melanogaster, Lim [/bib_ref] [bib_ref] Recombination between retrotransposons as a source of chromosome rearrangements in the yeast..., Mieczkowski [/bib_ref] [bib_ref] Chromosome rearrangements and transposable elements, Lonnig [/bib_ref] [bib_ref] Transposable elements as catalysts for chromosome rearrangements, Zhang [/bib_ref] , as exemplified in primate evolution by the recently published gibbon genome [bib_ref] Gibbon genome and the fast karyotype evolution of small apes, Carbone [/bib_ref].
## Adaptations acquired and comingled by horizontal transfers
The conventional view of evolution maintained by Dobzhansky and his colleagues is that traits are transmitted vertically from progenitors to offspring, with evolutionarily important hereditary changes occurring within each particular line of descent. That is the pattern of "descent with modification" illustrated at the end of Origin of Species [bib_ref] Origin of Species, Darwin [/bib_ref]. According to this perspective, each new adaptation has to evolve within a distinct lineage. In contrast to this conventional view, genomic analysis revealed that the molecular phylogenies of genetic loci encoding certain adaptive functions do not always match the taxonomic histories of the basal host genomes . The genomic evidence indicated that organisms could acquire independently evolved DNA providing adaptive benefits from unrelated organisms by "horizontal" DNA transfer [bib_ref] Evolutionary implications of horizontal gene transfer, Syvanen [/bib_ref]. Genomic evidence shows that it often proves more efficient to adapt to a new ecological niche by borrowing functions from distant taxa rather than evolving them internally from the pre-existing genome.
## Horizontal transfer among prokaryotes
The ability of prokaryotic organisms to exchange advantageous DNA segments is evident from studies of antibiotic resistance starting in the early years of molecular genetics (http://shapiro.bsd.uchicago.edu/ ExtraRefs.AntibioticResistanceAndHorizontalTransfer.shtml) [bib_ref] Infectious drug resistance, Watanabe [/bib_ref] [bib_ref] Infective heredity of multiple drug resistance in bacteria, Watanabe [/bib_ref] [bib_ref] Multilevel populations and the evolution of antibiotic resistance through horizontal gene transfer, Andam [/bib_ref] [bib_ref] Antibiotic-induced lateral transfer of antibiotic resistance, Hastings [/bib_ref]. Horizontal DNA transfer can occur between prokaryotic cells by uptake of DNA released by cells to the environment (transformation), direct cell-to-cell contact (conjugation), or viral infection (transduction). Sometimes the horizontally acquired DNA constituted an independently replicating molecule in the prokaryotic cell, and sometimes the horizontally acquired DNA was integrated into the existing genome, often as extended "genomic islands" encoding many different proteins for complex traits like metabolism, defense and pathogenicity [bib_ref] Comparative analysis of mobilizable genomic islands, Daccord [/bib_ref] [bib_ref] Conjugative and mobilizable genomic islands in bacteria: Evolution and diversity, Bellanger [/bib_ref] [bib_ref] Defense islands in bacterial and archaeal genomes and prediction of novel defense..., Makarova [/bib_ref] [bib_ref] Genomic islands and the evolution of catabolic pathways in bacteria, Van Der Meer [/bib_ref]. Special site-specific recombination structures called "integrons" exist in the genomes of some bacteria for the serial integration of antibiotic resistance cassettes [bib_ref] Integrons and gene cassettes: Hotspots of diversity in bacterial genomes, Hall [/bib_ref] [bib_ref] The role of integrons in antibiotic resistance gene capture, Rowe-Magnus [/bib_ref] and, in particular species, "super-integrons" have accumulated cassettes encoding up to hundreds of diverse adaptive functions [bib_ref] Super-integrons, Rowe-Magnus [/bib_ref] [bib_ref] Resistance integrons and super-integrons, Fluit [/bib_ref] [bib_ref] The integron: Adaptation on demand, Escudero [/bib_ref] [bib_ref] The function of integron-associated gene cassettes in Vibrio species: The tip of..., Rapa [/bib_ref]. Horizontal transfer occurs across the Bacteria-Archaea divide and can even lead to the formation of novel taxa [bib_ref] Evidence for massive gene exchange between archaeal and bacterial hyperthermophiles, Sclafani [/bib_ref] [bib_ref] Origins of major archaeal clades correspond to gene acquisitions from bacteria, Nelson-Sathi [/bib_ref] [bib_ref] Inter-domain conjugal transfer of DNA from bacteria to archaea, Dodsworth [/bib_ref] [bib_ref] Horizontal transfer of catalase-peroxidase genes between archaea and pathogenic bacteria, Faguy [/bib_ref] [bib_ref] Genomics of bacteria and archaea: The emerging dynamic view of the prokaryotic..., Koonin [/bib_ref].
The prevalence of prokaryotic horizontal DNA exchange gave rise to the idea of a vast super-cellular pan-genome that prokaryotes can sample facultatively by transformation, conjugation and transduction to assemble novel genomes for cell adaptation to particular ecological niches [bib_ref] Evolution of the genomic systems of prokaryotes and its momentous consequences, Sonea [/bib_ref] [bib_ref] A tentative unifying view of bacteria, Sonea [/bib_ref]. This controversial notion has gained wider credibility in recent years as a result of metagenomic analysis, which has revealed unexpectedly large numbers of known and unknown coding functions present in environmental DNA samples, particularly those encapsidated in virus particles [bib_ref] Comparative metagenomics of microbial traits within oceanic viral communities, Sharon [/bib_ref] [bib_ref] New dimensions of the virus world discovered through metagenomics, Kristensen [/bib_ref] [bib_ref] Estimating the extent of horizontal gene transfer in metagenomic sequences, Tamames [/bib_ref] [bib_ref] Discovery of new protein families and functions: New challenges in functional metagenomics..., Ufarte [/bib_ref] [bib_ref] Counts and sequences, observations that continue to change our understanding of viruses..., Wommack [/bib_ref] [bib_ref] Single cell genomics indicates horizontal gene transfer and viral infections in a..., Labonté [/bib_ref] [bib_ref] Expanding the marine virosphere using metagenomics, Mizuno [/bib_ref]. There has even been the suggestion that environmental metagenome data indicate the possible existence of major new cell types [bib_ref] Highly divergent ancient gene families in metagenomic samples are compatible with additional..., Lopez [/bib_ref].
## Horizontal dna transfer from prokaryotes and fungi to multicellular eukaryotes
Horizontal DNA transfer is not limited to prokaryotes. Genomic analysis provides abundant examples of eukaryotic adaptations with prokaryotic (and fungal) origins. These include:
## ‚
Biochemical pathways [bib_ref] Lateral transfer of genes from fungi underlies carotenoid production in aphids, Moran [/bib_ref] [bib_ref] A horizontal gene transfer supported the evolution of an early metazoan biomineralization..., Jackson [/bib_ref] [bib_ref] Horizontally transferred fungal carotenoid genes in the two-spotted spider mite Tetranychus urticae, Altincicek [/bib_ref] [bib_ref] Energetics and genetics across the prokaryote-eukaryote divide, Lane [/bib_ref] [bib_ref] Identification of horizontally transferred genes in the genus Colletotrichum reveals a steady..., Jaramillo [/bib_ref] ;
‚ Phytopathogenicity in Botrytis fungi [bib_ref] Interkingdom gene transfer may contribute to the evolution of phytopathogenicity in Botrytis..., Zhu [/bib_ref] ;
‚ Ability to live in extreme environments [bib_ref] Gene transfer from bacteria and archaea facilitated evolution of an extremophilic eukaryote, Schönknecht [/bib_ref] ;
‚ Capacity of plant parasitic nematodes to digest cellulose and other phytopolymers [bib_ref] Plant parasitic nematodes: Habitats, hormones, and horizontally-acquired genes, Bird [/bib_ref] [bib_ref] Evolution of plant parasitism among nematodes, Baldwin [/bib_ref] [bib_ref] Role of horizontal gene transfer in the evolution of plant parasitism among..., Mitreva [/bib_ref] [bib_ref] Multiple lateral gene transfers and duplications have promoted plant parasitism ability in..., Danchin [/bib_ref] [bib_ref] Horizontal gene transfer in nematodes: A catalyst for plant parasitism?, Haegeman [/bib_ref] [bib_ref] Horizontal gene transfer of microbial cellulases into nematode genomes is associated with..., Mayer [/bib_ref] [bib_ref] Lateral gene transfers have polished animal genomes: Lessons from nematodes, Danchin [/bib_ref].
We know that this horizontal DNA transfer strategy was used repeatedly because each lineage of plant parasitic nematodes acquired their digestive enzymes from different fungi or bacteria;
‚ Energy metabolism and defense functions subject to purifying selection in a marine shrimp [bib_ref] Horizontally transferred genes in the genome of Pacific white shrimp, Litopenaeus vannamei, Yuan [/bib_ref] ;
‚ Sequences of unknown but selectively conserved function transferred from marine bacteria to fish after the divergence of teleosts from other vertebrates [bib_ref] Horizontal functional gene transfer from bacteria to fishes, Sun [/bib_ref].
Many bacteria live as endosymbionts in animals, and there is abundant evidence that parts or all of endosymbiont genomes have be incorporated into host genomes [bib_ref] Serendipitous discovery of Wolbachia genomes in multiple Drosophila species, Salzberg [/bib_ref] [bib_ref] Widespread lateral gene transfer from intracellular bacteria to multicellular eukaryotes, Hotopp [/bib_ref] [bib_ref] Aphids acquired symbiotic genes via lateral gene transfer, Nikoh [/bib_ref] [bib_ref] Horizontal gene transfer between bacteria and animals, Dunning Hotopp [/bib_ref]. In Drosophila ananassae, for example, more than 2% of the genome comes from Wolbachia endosymbionts.
## Horizontal transfer from eukaryotes to bacteria
Although less widely documented than prokaryote to eukaryote horizontal transfer, the analysis of endosymbiotic and pathogenic bacteria infecting eukaryotic cells has turned up examples where these prokaryotes appear to have integrated eukaryotic host cell sequences into their genomes [bib_ref] Genomic analysis of 38 Legionella species identifies large and diverse effector repertoires, Burstein [/bib_ref] [bib_ref] Evidence for acquisition of Legionella type IV secretion substrates via interdomain horizontal..., De Felipe [/bib_ref] [bib_ref] Horizontal transfer and the evolution of host-pathogen interactions, De La Casa-Esperon [/bib_ref] [bib_ref] Extensive recombination events and horizontal gene transfer shaped the Legionella pneumophila genomes, Gomez-Valero [/bib_ref]. The restricted taxonomic distribution of the eukaryotic domains among the infectious bacteria indicates recent horizontal acquisition rather than shared vertical ancestry with eukaryotes [bib_ref] Genomic analysis of 38 Legionella species identifies large and diverse effector repertoires, Burstein [/bib_ref].
Bacteria use the proteins containing typically eukaryotic domains encoded by horizontally acquired DNA sequences as injected "effector" molecules to modulate host cell metabolism and defenses in order to facilitate the infection process [bib_ref] Legionella pneumophila, armed to the hilt: Justifying the largest arsenal of effectors..., Ensminger [/bib_ref] [bib_ref] Ankyrin domains across the Tree of Life, Jernigan [/bib_ref] [bib_ref] Functional diversity of ankyrin repeats in microbial proteins, Al-Khodor [/bib_ref] [bib_ref] Molecular characterization of the Dot/Icm-translocated AnkH and AnkJ eukaryotic-like effectors of Legionella..., Habyarimana [/bib_ref] [bib_ref] ThANKs for the repeat: Intracellular pathogens exploit a common eukaryotic domain, Voth [/bib_ref] [bib_ref] Bringing host-cell takeover by pathogenic bacteria to center stage, Dubreuil [/bib_ref]. The ability of many infectious bacteria to grow in diverse eukaryotic hosts, such as amoebae and mammals [bib_ref] Genome dynamics in Legionella: The basis of versatility and adaptation to intracellular..., Gomez-Valero [/bib_ref] , apparently plays an important role in the acquisition of eukaryotic domains from one kind of host (e.g., amoebae) and their utilization as invasion functions in another kind of host (e.g., mammals).
## Horizontal dna transfer among eukaryotes
In addition to fungi, other eukaryotes can transfer DNA horizontally across taxonomic boundaries [bib_ref] Evolutionary implications of horizontal gene transfer, Syvanen [/bib_ref] [bib_ref] Horizontal gene transfer in eukaryotic evolution, Keeling [/bib_ref]. This phenomenon is prevalent among unicellular protists [bib_ref] Phylogenetic analyses of diplomonad genes reveal frequent lateral gene transfers affecting eukaryotes, Andersson [/bib_ref] [bib_ref] Horizontal gene transfer in eukaryotic parasites: A case study of Entamoeba histolytica..., Alsmark [/bib_ref] [bib_ref] Patterns of prokaryotic lateral gene transfers affecting parasitic microbial eukaryotes, Alsmark [/bib_ref] , but various classes of DNA transfer have been documented between multicellular lineages:
## ‚
Mitochondrial genomes in flowering plants [bib_ref] Massive mitochondrial gene transfer in a parasitic flowering plant clade, Xi [/bib_ref] [bib_ref] Gorgeous mosaic of mitochondrial genes created by horizontal transfer and gene conversion, Hao [/bib_ref] [bib_ref] Widespread horizontal transfer of mitochondrial genes in flowering plants, Bergthorsson [/bib_ref] ;
‚ Chloroplast genomes in plants [bib_ref] Horizontal transfer of chloroplast genomes between plant species, Stegemann [/bib_ref] ;
‚ Mobile DNA elements [bib_ref] Repeated horizontal transfer of a DNA transposon in mammals and other tetrapods, Pace [/bib_ref] [bib_ref] Horizontal transfer of transposable elements in plants, Fortune [/bib_ref] [bib_ref] Pervasive horizontal transfer of rolling-circle transposons among animals, Thomas [/bib_ref] [bib_ref] Widespread evidence for horizontal transfer of transposable elements across Drosophila genomes, Bartolome [/bib_ref] [bib_ref] Independent and parallel lateral transfer of DNA transposons in tetrapod genomes, Novick [/bib_ref] [bib_ref] Horizontal transposon transfer in eukarya: Detection, bias, and perspectives, Wallau [/bib_ref] [bib_ref] Jumping the fine LINE between species: Horizontal transfer of transposable elements in..., Ivancevic [/bib_ref] ;
‚ Sequences encoding diverse adaptive functions, including glyoxylate cycle enzymes in metazoa [bib_ref] Evolution of glyoxylate cycle enzymes in Metazoa: Evidence of multiple horizontal transfer..., Kondrashov [/bib_ref] , photosynthetic carbon cycles in plants [bib_ref] Lateral transfer and recompartmentalization of Calvin cycle enzymes of plants and algae, Rogers [/bib_ref] [bib_ref] Multiple photosynthetic transitions, polyploidy, and lateral gene transfer in the grass subtribe..., Christin [/bib_ref] , anti-freeze proteins in fish [bib_ref] Lateral transfer of a lectin-like antifreeze protein gene in fishes, Graham [/bib_ref] , and mimicry pattern determinants in butterflies [bib_ref] Butterfly genome reveals promiscuous exchange of mimicry adaptations among species, Heliconius Genome Consortium [/bib_ref] ;
‚ Miscellaneous expressed functions acquired by a parasitic plant from its host [bib_ref] Horizontal transfer of expressed genes in a parasitic flowering plant, Xi [/bib_ref].
Infectious bacterial endosymbionts and pathogens are widely considered as potential vectors for horizontal transfer between multicellular organisms [bib_ref] Possible horizontal transfer of Drosophila genes by the mite Proctolaelaps regalis, Houck [/bib_ref] [bib_ref] A role for host-parasite interactions in the horizontal transfer of transposons across..., Gilbert [/bib_ref] [bib_ref] Microvesicles and intercellular communication in the context of parasitism, Barteneva [/bib_ref] [bib_ref] Algal endosymbionts as vectors of horizontal gene transfer in photosynthetic eukaryotes, Qiu [/bib_ref] [bib_ref] Endosymbiotic bacteria associated with nematodes, ticks and amoebae, Taylor [/bib_ref]. Investigators have documented endosymbiont transfers between different multicellular host species [bib_ref] Independent origins and horizontal transfer of bacterial symbionts of aphids, Sandström [/bib_ref] [bib_ref] Facultative symbionts in aphids and the horizontal transfer of ecologically important traits, Oliver [/bib_ref] , and many bacteria known as vertebrate pathogens also infect lower eukaryotes, especially amoebae [bib_ref] The professional phagocyte Dictyostelium discoideum as a model host for bacterial pathogens, Bozzaro [/bib_ref] [bib_ref] The genomic sequence of the accidental pathogen Legionella pneumophila, Chien [/bib_ref] [bib_ref] Pathogen-host interactions in Dictyostelium, Legionella, Mycobacterium and other pathogens, Steinert [/bib_ref] [bib_ref] Interactions of some common pathogenic bacteria with Acanthamoeba polyphaga, Huws [/bib_ref] [bib_ref] Increased persistence of Salmonella enterica serovar Typhi in the presence of Acanthamoeba..., Douesnard-Malo [/bib_ref] [bib_ref] Genetic and physiological interactions in the amoeba-bacteria symbiosis, Jeon [/bib_ref]. Among these multivalent infectious bacteria, a number are capable of taking up DNA from their immediate environment [bib_ref] Antibiotics and UV radiation induce competence for natural transformation in Legionella pneumophila, Charpentier [/bib_ref] [bib_ref] Competence and natural transformation in vibrios, Sun [/bib_ref] [bib_ref] Ubiquitous late competence genes in Bacillus species indicate the presence of functional..., Kovács [/bib_ref] [bib_ref] Structure-function relationships of the competence lipoprotein ComL and SSB in meningococcal transformation, Benam [/bib_ref].
## Viral integrations into host genomes
Viruses of all kinds (including RNA viruses) insert their genomes into eukaryotic host genomes with surprisingly high frequency [bib_ref] Sequences of flavivirus-related RNA viruses persist in DNA form integrated in the..., Crochu [/bib_ref] [bib_ref] Occurrence of a DNA sequence of a non-retro RNA virus in a..., Tanne [/bib_ref] [bib_ref] Evolutionary capture of viral and plasmid DNA by yeast nuclear chromosomes, Frank [/bib_ref] [bib_ref] Detection of novel insect flavivirus sequences integrated in Aedes albopictus (Diptera: Culicidae)..., Roiz [/bib_ref] [bib_ref] Filoviruses are ancient and integrated into mammalian genomes, Taylor [/bib_ref] [bib_ref] Sequences from ancestral single-stranded DNA viruses in vertebrate genomes: The Parvoviridae and..., Belyi [/bib_ref] [bib_ref] Unexpected inheritance: Multiple integrations of ancient bornavirus and ebolavirus/marburgvirus sequences in vertebrate..., Belyi [/bib_ref] [bib_ref] Endogenous non-retroviral RNA virus elements in mammalian genomes, Horie [/bib_ref] [bib_ref] A four-partner plant-virus interaction: Enemies can also come from within, Iskra-Caruana [/bib_ref] [bib_ref] Discovery and characterization of mammalian endogenous parvoviruses, Kapoor [/bib_ref] [bib_ref] Endogenous viral elements in animal genomes, Katzourakis [/bib_ref] [bib_ref] Widespread horizontal gene transfer from double-stranded RNA viruses to eukaryotic nuclear genomes, Liu [/bib_ref] [bib_ref] Non-retroviral fossils in vertebrate genomes, Horie [/bib_ref] [bib_ref] Widespread endogenization of genome sequences of non-retroviral RNA viruses into plant genomes, Chiba [/bib_ref] [bib_ref] Widespread endogenization of densoviruses and parvoviruses in animal and human genomes, Liu [/bib_ref] [bib_ref] The evolution of endogenous viral elements, Holmes [/bib_ref] [bib_ref] Endogenous viruses: Insights into viral evolution and impact on host biology, Feschotte [/bib_ref] [bib_ref] Endogenous RNA viruses of plants in insect genomes, Cui [/bib_ref]. Integration can occur by retroviral integrase functions, sometimes followed by recombination with other viral sequences [bib_ref] Recombination of retrotransposon and exogenous RNA virus results in nonretroviral cDNA integration, Geuking [/bib_ref] , or by non-homologous end-joining (NHEJ) at DNA breaks [bib_ref] Genomic DNA double-strand breaks are targets for hepadnaviral DNA integration, Bill [/bib_ref] [bib_ref] DNA double-strand breaks, potential targets for HBV integration, Hu [/bib_ref]. Integration events at DNA breaks have the potential to generate novel sequence configurations.
6.6. The Amoeba-Megavirus "Melting Pot" of Sequences from All Three Cell Kingdoms
Viruses have long been considered both as substrates for evolutionary innovation [bib_ref] Virus world as an evolutionary network of viruses and capsidless selfish elements...., Koonin [/bib_ref] [bib_ref] A protein domain-based view of the virosphere-host relationship, Abroi [/bib_ref] and as vectors for horizontal DNA transfer. Particular attention has recently focused on a group of Nucleocytoplasmic Large DNA Viruses (NCLDVs) with genomes comprising hundreds of thousands or millions of base-pairs [bib_ref] The concept of virus in the post-megavirus era, Claverie [/bib_ref] [bib_ref] Giant virus Megavirus chilensis encodes the biosynthetic pathway for uncommon acetamido sugars, Piacente [/bib_ref]. Most significantly, NCLDVs acquire cellular genome fragments and have been found to carry a mixture of DNA sequences from all three major domains of life (http://shapiro.bsd.uchicago.edu/Viral_Composites.html) [bib_ref] Phylogenetic evidence for extensive lateral acquisition of cellular genes by Nucleocytoplasmic large..., Filee [/bib_ref] [bib_ref] Lateral gene transfer, lineage-specific gene expansion and the evolution of Nucleo Cytoplasmic..., Filee [/bib_ref] [bib_ref] Gene exchange and the origin of giant viruses, Filee [/bib_ref] [bib_ref] Route of NCLDV evolution: The genomic accordion, Filee [/bib_ref]. NCLDVs can infect both protists and multicellular hosts and thus transfer the incorporated cellular sequences. Amoebae are common hosts for many of these large DNA viruses, and amoeba have consequently been designated to constitute an evolutionary "melting pot" (http://shapiro.bsd.uchicago.edu/ Amoebal_Viruses.html) [bib_ref] Giant Marseillevirus highlights the role of Amoebae as a melting pot in..., Boyer [/bib_ref]. The designation is especially appropriate for two reasons: (1) amoebae are phagocytic and can acquire DNA sequences from engulfed cells [bib_ref] The giant Cafeteria roenbergensis virus that infects a widespread marine phagocytic protist..., Colson [/bib_ref] and (2) amoebae are hosts to bacteria that both exchange DNA [bib_ref] Evidence of transfer by conjugation of type IV secretion system genes between..., Saisongkorh [/bib_ref] and infect more complex eukaryotes, including both plants and animals [bib_ref] The professional phagocyte Dictyostelium discoideum as a model host for bacterial pathogens, Bozzaro [/bib_ref] [bib_ref] The genomic sequence of the accidental pathogen Legionella pneumophila, Chien [/bib_ref] [bib_ref] Pathogen-host interactions in Dictyostelium, Legionella, Mycobacterium and other pathogens, Steinert [/bib_ref] [bib_ref] Interactions of some common pathogenic bacteria with Acanthamoeba polyphaga, Huws [/bib_ref] [bib_ref] Increased persistence of Salmonella enterica serovar Typhi in the presence of Acanthamoeba..., Douesnard-Malo [/bib_ref] [bib_ref] Genetic and physiological interactions in the amoeba-bacteria symbiosis, Jeon [/bib_ref].
Combining the phenomenology of viral infection, amoebael phagocytosis, multivalent bacterial infectivity and endosymbioses with the documentation of viral and prokaryote to eukaryote DNA transfers, it is more than clear that multiple pathways exist by which cells can acquire and transmit DNA segments from one eukaryotic host to another. Specific biochemical activities are distributed among distantly related domains of life [bib_ref] Evolution of the DUT gene: Horizontal transfer between host and pathogen in..., Mcclure [/bib_ref] , and we can expect further genomically-documented examples of horizontally acquired adaptations to multiply with continued genome sequencing.
## Protein evolution by exon shuffling and exonization from "non-coding" dna
Among the most striking results of genomics was the realization that many proteins and the DNA that encodes them are not continuous unitary structures. Many proteins consist of strings of functionally different but interacting "domains," each one of which may be found iterated in distinct proteins [bib_ref] Evolutionarily mobile modules in proteins, Doolittle [/bib_ref] [bib_ref] Initial sequencing and analysis of the human genome, Lander [/bib_ref] [bib_ref] Exonization of transposed elements: A challenge and opportunity for evolution, Schmitz [/bib_ref]. Correspondingly, the cognate protein-coding regions are often discontinuous and composed of expressed coding segments ("exons") separated by intervening segments ("introns") [bib_ref] The exon theory of genes, Gilbert [/bib_ref] [bib_ref] DNA sequencing and gene structure, Gilbert [/bib_ref]. While not always the case, DNA exons tended to encode functional protein domains or subdomains [bib_ref] Protein domains correlate strongly with exons in multiple eukaryotic genomes-Evidence of exon..., Liu [/bib_ref].
The segmented nature of proteins and protein-coding DNA has major implications for genome functionality and protein evolution:
## ‚
A given genetic locus can encode multiple protein products by joining different combinations of exons into the final mRNA by alternative splicing [bib_ref] Alternative splicing and RNA selection pressure-Evolutionary consequences for eukaryotic genomes, Xing [/bib_ref] [bib_ref] The evolutionary landscape of alternative splicing in vertebrate species, Barbosa-Morais [/bib_ref] [bib_ref] Incorporating alternative splicing and mRNA editing into the genetic analysis of complex..., Hassan [/bib_ref] [bib_ref] Alternative splicing: A pivotal step between eukaryotic transcription and translation, Kornblihtt [/bib_ref] [bib_ref] Functional consequences of developmentally regulated alternative splicing, Kalsotra [/bib_ref] [bib_ref] How did alternative splicing evolve?, Ast [/bib_ref] [bib_ref] Mechanisms of alternative splicing regulation: Insights from molecular and genomics approaches, Chen [/bib_ref] [bib_ref] The origins, evolution, and functional potential of alternative splicing in vertebrates, Mudge [/bib_ref] ;
‚ New protein functionalities can arise rapidly by Lego-like assembly of exons from different sources, known as exon or domain shuffling [bib_ref] Domain shuffling and the evolution of vertebrates, Kawashima [/bib_ref] [bib_ref] Signatures of domain shuffling in the human genome, Kaessmann [/bib_ref] [bib_ref] Molecular mechanisms of exon shuffling: Illegitimate recombination, Van Rijk [/bib_ref] [bib_ref] Evolutionary history of exon shuffling, Franca [/bib_ref] ;
‚ Totally new protein domains can originate rapidly by conversion of non-coding DNA segments into exons ("exonization") and contribute to novel biochemical functions by subsequent duplication and exon shuffling [bib_ref] Exonization of transposed elements: A challenge and opportunity for evolution, Schmitz [/bib_ref] [bib_ref] The birth of new exons: Mechanisms and evolutionary consequences, Sorek [/bib_ref] [bib_ref] Genome-wide survey of ds exonization to enrich transcriptomes and proteomes in plants, Liu [/bib_ref] [bib_ref] Widespread exonization of transposable elements in human coding sequences is associated with..., Huda [/bib_ref].
Both the rapid evolution of new functionalities by exon shuffling and the origination of sequences encoding extended domains by exonization have potential for protein innovation beyond what is possible through codon-by-codon changes to existing proteins. Exon shuffling has an inherently high probability of producing adaptive novelties because it rearranges previously evolved sequences that encode established protein functionalities. This potential has been exploited in biotechnology where domain shuffling has proved an efficient method of protein engineering [bib_ref] Anticipatory evolution and DNA shuffling, Bacher [/bib_ref] [bib_ref] DNA shuffling by random fragmentation and reassembly: In vitro recombination for molecular..., Stemmer [/bib_ref] [bib_ref] Rapid evolution of a protein in vitro by DNA shuffling, Stemmer [/bib_ref].
Mobile DNA elements play major roles in both exon shuffling and exonization. The genomic record indicates transposons and retrotransposons can incorporate and relocate exons [bib_ref] Trans mobilization of genomic DNA as a mechanism for retrotransposon-mediated exon shuffling, Ejima [/bib_ref] [bib_ref] Pack-MULE transposable elements mediate gene evolution in plants, Jiang [/bib_ref] [bib_ref] Gene duplication and exon shuffling by helitron-like transposons generate intraspecies diversity in..., Morgante [/bib_ref] [bib_ref] 5'-Transducing SVA retrotransposon groups spread efficiently throughout the human genome, Damert [/bib_ref] [bib_ref] Exon-trapping mediated by the human retrotransposon SVA, Hancks [/bib_ref] [bib_ref] Bs1, a new chimeric gene formed by retrotransposon-mediated exon shuffling in maize, Elrouby [/bib_ref] [bib_ref] Pack-Mutator-like transposable elements (Pack-MULEs) induce directional modification of genes through biased insertion..., Jiang [/bib_ref] , and mobile DNA-mediated exon shuffling has been studied in real time [bib_ref] Exon shuffling by L1 retrotransposition, Moran [/bib_ref] [bib_ref] Transposition and exon shuffling by group II intron RNA molecules in pieces, Hiller [/bib_ref] [bib_ref] Directed evolution of proteins by exon shuffling, Kolkman [/bib_ref]. Genomics reveals numerous instances of exonization from mobile element insertions in humans and other mammals [bib_ref] Exonization of transposed elements: A challenge and opportunity for evolution, Schmitz [/bib_ref] [bib_ref] Widespread exonization of transposable elements in human coding sequences is associated with..., Huda [/bib_ref] [bib_ref] Exonization of the LTR transposable elements in human genome, Piriyapongsa [/bib_ref] [bib_ref] Alu exonization events reveal features required for precise recognition of exons by..., Schwartz [/bib_ref] [bib_ref] Characteristics of transposable element exonization within human and mouse, Sela [/bib_ref] [bib_ref] Alu-SINE exonization: En route to protein-coding function, Krull [/bib_ref] [bib_ref] Beyond DNA: RNA editing and steps toward Alu exonization in primates, Möller-Krull [/bib_ref] as well as in plants [bib_ref] Genome-wide survey of ds exonization to enrich transcriptomes and proteomes in plants, Liu [/bib_ref].
## Regulatory signal evolution involving mobile dna elements
The involvement of mobile DNA in genome change marks one of the most basic divergences between Dobzhansky's Modern Synthesis perspective and a genomics-based view of the evolutionary process. The Modern Synthesis focused on isolated allelic changes at individual loci. Mobile elements, on the other hand, are distributed at many sites throughout the genome and have the potential to generate coordinated changes rewiring distributed regulatory networks involving many loci (http://shapiro.bsd.uchicago.edu/Table5C-1.MobileElementsFoundtobeExaptedascis-RegulatoryControlSitesinAnimals.html) [bib_ref] Transposable elements donate lineage-specific regulatory sequences to host genomes, Marino-Ramirez [/bib_ref] [bib_ref] Species-specific endogenous retroviruses shape the transcriptional network of the human tumor suppressor..., Wang [/bib_ref] [bib_ref] A c-Myc regulatory subnetwork from human transposable element sequences, Wang [/bib_ref] [bib_ref] Transposable elements have rewired the core regulatory network of human embryonic stem..., Kunarso [/bib_ref] [bib_ref] Rewirable gene regulatory networks in the preimplantation embryonic development of three mammalian..., Xie [/bib_ref] [bib_ref] Transposable elements and the evolution of regulatory networks, Feschotte [/bib_ref] [bib_ref] Inherited adaptation of genome-rewired cells in response to a challenging environment, David [/bib_ref] [bib_ref] Rewiring the transcriptional regulatory circuits of cells, Scannell [/bib_ref] [bib_ref] Measuring the evolutionary rewiring of biological networks, Shou [/bib_ref].
Genomics documents at least three episodes of regulatory innovation involving mobile elements in the course of vertebrate evolution [bib_ref] Three periods of regulatory innovation during vertebrate evolution, Lowe [/bib_ref] [bib_ref] Distinct groups of repetitive families preserved in mammals correspond to different periods..., Jurka [/bib_ref] , and mobile DNA has been a major source of regulatory motifs in human genome evolution [bib_ref] Repetitive DNA elements, nucleosome binding and human gene expression, Huda [/bib_ref] [bib_ref] Origin of a substantial fraction of human regulatory sequences from transposable elements, Jordan [/bib_ref]. Moreover, the fact that mobile elements are often the most taxonomically specific genome components potentially confers distinctive evolutionary trajectories on different lineages [bib_ref] Repetitive sequences in complex genomes: Structure and evolution, Jurka [/bib_ref] [bib_ref] Families of transposable elements, population structure and the origin of species, Jurka [/bib_ref] [bib_ref] How repeated retroelements format genome function, Sternberg [/bib_ref].
## Adaptations and innovations in mammalian reproduction arising by natural genetic engineering processes involving mobile dna and "non-coding" ncrna molecules
Because of their medical relevance, reproductive biology, embryonic development and stem cell biology in mammals have received particular attention from genomicists. The analysis has uncovered major roles for mobile DNA elements and ncRNAs derived from them. Rather than serving as "fossils that litter our genomes" [bib_ref] Human genomics. Sleeping dogs of the genome, Gorbunova [/bib_ref] , as conventional evolutionary thinking would assert, these elements are both essential evolutionary tools and active participants in contemporary genome function.
## Retroviral involvement in placenta evolution
Mammalian reproduction depends upon development of a syncytial placenta from the zygote to nourish the fetus during pregnancy. At repeated stages in mammalian evolution, distinct endogenous retroviral "envelope" (Env) proteins have been exapted as fusionogenic "syncytins" involved in forming placental tissue in different mammalian lineages (http://shapiro.bsd.uchicago.edu/ Retroviral_involvement_in_placenta_evolution.html) [bib_ref] Effects of retroviruses on host genome function, Jern [/bib_ref] [bib_ref] Captured retroviral envelope syncytin gene associated with the unique placental structure of..., Cornelis [/bib_ref] [bib_ref] Paleovirology of "syncytins", retroviral env genes exapted for a role in placentation, Lavialle [/bib_ref] [bib_ref] From ancestral infectious retroviruses to bona fide cellular genes: Role of the..., Dupressoir [/bib_ref] [bib_ref] A placenta-specific receptor for the fusogenic, endogenous retrovirus-derived, human syncytin-2, Esnault [/bib_ref] [bib_ref] Retroviruses facilitate the rapid evolution of the mammalian placenta, Chuong [/bib_ref]. Moreover, endogenous retroviruses (ERVs) provide transcriptional regulatory signals to direct imprinted expression of other proteins, such as insulin-like growth factor, required for placental function [bib_ref] Hypomethylation of functional retrotransposon-derived genes in the human placenta, Macaulay [/bib_ref] [bib_ref] Retrotransposon hypomethylation in melanoma and expression of a placenta-specific gene, Macaulay [/bib_ref] [bib_ref] The origin and evolution of genomic imprinting and viviparity in mammals, Renfree [/bib_ref].
## Mobile dna recruitment of maternal functions
The endometrium is the maternal tissue that nourishes the placenta in mammalian pregnancy. Endometrial development in the uterus involves the hormone-regulated expression of over 1,500 different proteins. The transcriptional regulatory signals coordinating biogenesis of this pregnancy-specific cohort evolved mainly from mobile DNA elements, both transposons and retrotransposons [bib_ref] Ancient transposable elements transformed the uterine regulatory landscape and transcriptome during the..., Lynch [/bib_ref] [bib_ref] Transposable elements are major contributors to the origin, diversification, and regulation of..., Kapusta [/bib_ref] [bib_ref] Transposon-mediated rewiring of gene regulatory networks contributed to the evolution of pregnancy..., Lynch [/bib_ref] [bib_ref] Transposable element recruitments in the mammalian placenta: Impacts and mechanisms, Emera [/bib_ref]. Convergent patterns of regulatory rewiring can be traced in the endometria of distinct mammalian lineages with well-sequenced genomes [bib_ref] Convergent evolution of endometrial prolactin expression in primates, mice, and elephants through..., Emera [/bib_ref]. Thus, evolutionary innovations for both fetal and maternal sides of viviparous reproduction arose, to a large degree, through the ability of mammalian cells to mobilize repetitive components of their genomes to adaptive locations.
## Mobile dna and lncrnas in stem cell programming and early embryogenesis
Transcripts from human endogenous retroviruses (HERVs) have been found to be the most stage-specific RNAs expressed during early human embryonic development [bib_ref] Szczerbinska, I. Dynamic transcription of distinct classes of endogenous retroviral elements marks..., Goke [/bib_ref] , and there is intrinsic retroviral reactivation in human pre-implantation embryos and pluripotent stem cells [bib_ref] Intrinsic retroviral reactivation in human preimplantation embryos and pluripotent cells, Grow [/bib_ref]. Although functional studies are not possible in human embryos, direct functionality has been established for a retroviral RNA in mouse, where MuERV-L transcripts expressed just 8-10 h after fertilization at the 2-cell stage are necessary for developmental competence at the 4-cell stage but not afterwards [bib_ref] MuERV-L is one of the earliest transcribed genes in mouse one-cell embryos, Kigami [/bib_ref].
HERVs and other mobile DNA elements are the major components of long non-coding lncRNAs, which play important roles in (re)programming embryonic and stem cell genomes ("83% of lncRNAs contain at least one TE (transposable element), while of the total number of base pairs that comprise lncRNA sequences, 42% is derived from TEs" [bib_ref] Transposable elements at the center of the crossroads between embryogenesis, embryonic stem..., Hutchins [/bib_ref]. These include the lincRNA ROR ("regulator of reprogramming") beginning with a HERV-H transcript needed for formation of human induced pluripotent stem cells (HiPSCs) [bib_ref] Large intergenic non-coding RNA-RoR modulates reprogramming of human induced pluripotent stem cells, Loewer [/bib_ref] [bib_ref] Transposable elements reveal a stem cell-specific class of long noncoding RNAs, Kelley [/bib_ref] , plus LINC01108 (Linc-ES3) and human L1TD1 lncRNAs required to maintain stem cell pluripotency [bib_ref] Long non-coding RNAs in stem cell pluripotency, Ng [/bib_ref] [bib_ref] RNA-binding protein L1TD1 interacts with LIN28 via RNA and is required for..., Narva [/bib_ref]. A recent genomic census of over 4000 human-specific binding sites for the transcription factors which reprogram HiPSCs remarkably found between 99.8% and 100% of the sites to be located in mobile DNA repeats [bib_ref] Transposable elements and DNA methylation create in embryonic stem cells human-specific regulatory..., Glinsky [/bib_ref].
There is a burgeoning literature relating mobile DNA to the evolution of ncRNA-based circuitry essential for mammalian (and, more specifically, human) stem cell and embryonic development [bib_ref] Transposable elements have rewired the core regulatory network of human embryonic stem..., Kunarso [/bib_ref] [bib_ref] Transposable elements are major contributors to the origin, diversification, and regulation of..., Kapusta [/bib_ref] [bib_ref] Transposable elements reveal a stem cell-specific class of long noncoding RNAs, Kelley [/bib_ref] [bib_ref] Long non-coding RNAs in stem cell pluripotency, Ng [/bib_ref] [bib_ref] Single-cell transcriptome analysis reveals dynamic changes in lncRNA expression during reprogramming, Kim [/bib_ref] [bib_ref] Deep transcriptome profiling of mammalian stem cells supports a regulatory role for..., Fort [/bib_ref] [bib_ref] Retroviral transcriptional regulation and embryonic stem cells: War and peace, Schlesinger [/bib_ref] [bib_ref] Pivots of pluripotency: The roles of non-coding RNA in regulating embryonic and..., Huo [/bib_ref] [bib_ref] The intertwining of transposable elements and non-coding RNAs, Hadjiargyrou [/bib_ref] [bib_ref] Volatile evolution of long noncoding RNA repertoires: Mechanisms and biological implications, Kapusta [/bib_ref] [bib_ref] The RIDL hypothesis: Transposable elements as functional domains of long noncoding RNAs, Johnson [/bib_ref]. The lncRNAs bind to and tether a variety of epigenetic modification complexes that execute genome reprogramming [bib_ref] the circuitry controlling pluripotency and differentiation, Guttman [/bib_ref] [bib_ref] Modular regulatory principles of large non-coding RNAs, Guttman [/bib_ref] , and the mobile DNA element sequences in each molecule have been proposed to constitute a combinatorial code of RNA domains that link together different genome modification processes (Called the "RIDL hypothesis" for Repeat Insertion Domains of LncRNAs) [bib_ref] The RIDL hypothesis: Transposable elements as functional domains of long noncoding RNAs, Johnson [/bib_ref]. In addition to lncRNAs, mobile DNA elements have also been documented to be sources for cell regulatory miRNAs [bib_ref] Burgeoning evidence indicates that microRNAs were initially formed from transposable element sequences, Roberts [/bib_ref] [bib_ref] Transposable elements as genetic regulatory substrates in early development, Gifford [/bib_ref] [bib_ref] Origin and evolution of human microRNAs from transposable elements, Piriyapongsa [/bib_ref] [bib_ref] Dual coding of siRNAs and miRNAs by plant transposable elements, Piriyapongsa [/bib_ref].
## A 21st century evolutionary principle: cell-mediated variation of read-write (rw) genomes
To recapitulate, a genomics-based view of evolutionary variation introduces novel features to hereditary control of cell biology impossible to predict when Dobzhansky and other evolutionary biologists formulated the neo-Darwinian Modern Synthesis in the middle of the last century:
## ‚
The existence of three distinct realms of cell evolution, Bacteria, Archaea and Eukarya; ‚ Symbiogenetic fusions involving these different realms leading to the formation of eukaryotic cells bearing organelles with multiple genome compartments;
‚ Horizontal organelle, virus and DNA transfers affecting adaptive traits across all cell types;
## ‚
The functional organization of proteins as systems of distinct interacting domains encoded by exons and subject to rapid evolution by exon shuffling and exon origination from non-coding DNA (exonization); ‚ Establishment of adaptive, distributed genome networks integrated by mobile DNA elements dispersing repetitive regulatory signals to multiple loci;
‚ Regulation of cell differentiation in multicellular development by non-coding lncRNA molecules composed largely of mobile repetitive DNA elements that serve as scaffolds for epigenetic modifying activities.
Altogether, the combinatorial coding and regulatory aspects of cell heredity, plus the biochemical abilities cells possess to rearrange DNA molecules, constitute a powerful toolbox for adaptive genome rewriting. Revelations from genomic analysis oblige us to reconsider the simplifying assumptions made in the past two centuries about the nature of evolutionary variation. Rather than single gene traits, we recognize that all phenotypes involve coordinated activity by multiple interacting cell molecules.
As summarized above, we have evidence that genomes contain abundant and functional repetitive components in addition to the unique coding sequences envisaged in the early days of molecular biology [bib_ref] Why repetitive DNA is essential to genome function, Shapiro [/bib_ref]. Instead of the "Constant Genome," subject to accidental modification, we know today that cells possess "Read-Write Genomes" they can alter by numerous biochemical processes capable of rapidly restructuring cellular DNA molecules. Genomics has modernized our understanding of the evolutionary process. Rather than viewing genome evolution as a happenstance series of copying errors, we are now in a position to study it as a complex biological process of active self-modification. [PubMed] |
Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) Equation Pronouncedly Underestimates Glomerular Filtration Rate in Type 2 Diabetes
OBJECTIVEdTo evaluate the performance of the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation to estimate glomerular filtration rate (GFR) in type 2 diabetic patients with GFR .60 mL/min/1.73 m 2 .RESEARCH DESIGN AND METHODSdThis was a cross-sectional study including 105 type 2 diabetic patients. GFR was measured by 51 Cr-EDTA method and estimated by the MDRD and CKD-EPI equations. Serum creatinine was measured by the traceable Jaffe method. Bland-Altman plots were used. Bias, accuracy (P30), and precision were evaluated.RESULTSdThe mean age of patients was 57 6 8 years; 53 (50%) were men and 90 (86%) were white. Forty-six (44%) patients had microalbuminuria, and 14 (13%) had macroalbuminuria. 51 Cr-EDTA GFR was 103 6 23, CKD-EPI GFR was 83 6 15, and MDRD-GFR was 78 6 17 mL/min/1.73 m 2 (P , 0.001). Accuracy (95% CI) was 67% (58-74) for CKD-EPI and 64% (56-75) for MDRD. Precision was 21 and 22, respectively.CONCLUSIONSdThe CKD-EPI and MDRD equations pronouncedly underestimated GFR in type 2 diabetic patients.
T he importance of estimating glomerular filtration rate (GFR) in addition to measuring urinary albumin excretion (UAE) has been recently recognized in individuals with or without diabetes because these two parameters are independent predictors of cardiovascular and renal outcomes and require specific approaches [bib_ref] The definition, classification, and prognosis of chronic kidney disease: a KDIGO Controversies..., Levey [/bib_ref]. However, the accuracy of creatinine-based equations to estimate GFR, including the new Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation, has been questioned for patients with diabetes [bib_ref] Performance of the chronic kidney disease epidemiology collaboration equation to estimate glomerular..., Rognant [/bib_ref]. Recently, demonstrated that CKD-EPI presented a poor performance in diabetic patients with a wide range of renal function, working even worse than MDRD equation. The current study evaluated the performance of the CKD-EPI equation to estimate GFR in individuals with type 2 diabetes and GFR .60 mL/min/1.73 m 2 .
## Research design and
# Statistical analysis
Bias was calculated as the mean difference between measured and estimated GFR. Accuracy was calculated as the percentage of estimates within 30% (P30) of measured GFR. Precision was measured as 1 SD of bias. The agreement between measured GFR and equations was evaluated using Bland-Altman plots, with the calculation of agreement limits (bias 6 2 SD) and CI [bib_ref] Measuring agreement in method comparison studies, Bland [/bib_ref]. According to Bland-Altman, 100 individuals are enough to estimate bias and limits of agreement with a 95% CI of about 34% of SD.
RESULTSdThis cross-sectional study included 105 individuals with type 2 diabetes. Their mean age was 57 6 8 years (42-86); 53 (50%) were men, and 90 (86%) were white. Forty-six (44%) had microalbuminuria, and 14 (13%) had macroalbuminuria. Plasma glucose was 148 6 28 mg/dL, and A1C was 8.3 6 0.3%. Cr-EDTA GFR was 103 6 23, CKD-EPI GFR was 83 6 15, and MDRD GFR was 78 6 17 mL/min/1.73 m 2 (P , 0.001). According to the Kidney Disease: Improving Global Outcomes GFR classification, 75% of the patients in our study were stage 1 (GFR .90 mL/min), and 25% were stage 2 (GFR = 60-89 mL/min) as measured by 51 Cr-EDTA GFR. Misclassification to stage 3a (GFR = 45-59 mL/min) occurred in 7.6 and 9.5% of the cases using the CKD-EPI and MDRD equations, respectively. shows the plots of measured and estimated GFR values. The CKD-EPI and MDRD equations systematically underestimated measured GFR. Bias was 20 and 24 mL/min/1.73 m 2 for CKD-EPI and MDRD, respectively (P = 0.26). Bias was significantly greater in subjects with GFR above versus below the median value (30 vs. 9 for CKD-EPI and 34 vs. 14 mL/min/ 1.73 m 2 for MDRD [P , 0.001 for each pair of comparisons]). Accuracy P30 (95% CI) was 67% (58-74) for CKD-EPI and 64% (56-75) for MDRD. Precision was 21 and 22, respectively.
CONCLUSIONSdThe CKD-EPI and MDRD equations significantly underestimated measured GFR in individuals with type 2 diabetes, especially at higher GFR values.
The Kidney Disease: Improving Global Outcomes, a global nonprofit foundation, has recently proposed a new chronic kidney disease staging system [bib_ref] The definition, classification, and prognosis of chronic kidney disease: a KDIGO Controversies..., Levey [/bib_ref]. This was based on the findings of a remarkable meta-analysis that included 45 cohorts and a total of 1,555,332 participants that confirmed that both GFR and UAE are independent predictors of cardiovascular and renal outcomes. Therefore, a reliable estimation of GFR and UAE is the cornerstone to predict patients' prognosis.
The American Diabetes Association and the National Kidney Foundation recommend annual UAE measurement and GFR estimation using equations that include serum creatinine, such as MDRD and CKD-EPI. However, according to recent studies, these equations have a poor performance for patients with diabetes and markedly underestimate GFR [bib_ref] Performance of the chronic kidney disease epidemiology collaboration equation to estimate glomerular..., Rognant [/bib_ref]. This disappointing performance seems to be associated with specific characteristics of the patients with diabetes, such as hyperglycemia, glomerular hyperfiltration, and obesity, which probably highlight the limitations of creatinine itself as a GFR marker. Hyperglycemia may interfere in two ways. First, it has long been known that glucose levels above 300 mg/dL may affect the performance of the Jaffe reaction to measure creatinine [bib_ref] Estimation of creatinine by the Jaffe reaction. A comparison of three methods, Husdan [/bib_ref]. Indeed, in a previous study we have found that patients with type 2 diabetes presented higher serum creatinine than healthy individuals, despite similar GFR values (4). Another possible explanation could be the hyperglycemia-induced glomerular hyperfiltration and the inability of creatinine to detect this typical phenomenon of diabetes [bib_ref] Glomerular hyperfiltration in NIDDM patients without overt proteinuria, Silveiro [/bib_ref]. This was clearly shown in our hyperfiltering patients, for whom the CKD-EPI and MDRD equations unacceptably underestimated GFR. The poor performance of the formulas in our study was further expressed in the chronic kidney disease misclassification of patients in 8 and 10% of the cases when using the CKD-EPI and MDRD equations, respectively. A recent French study that included a similar number of patients with GFRs .60 mL/min/1.73 m 2 also observed a bad performance of CKD-EPI equation, which also applied for GFRs ,60 mL/min/1.73 m 2 [bib_ref] Performance of the chronic kidney disease epidemiology collaboration equation to estimate glomerular..., Rognant [/bib_ref]. These data stress the need of developing alternative, more precise tools to estimate GFR in individuals with diabetes.
The strengths of our study were the use of a GFR reference method, which ensured a more precise interpretation of the equation results, and a sample size tailored for the specific Bland-Altman statistical analysis. One limitation was that our results were deliberately suited for patients with GFRs .60 mL/min/1.73 m 2 .
In conclusion, both CKD-EPI and MDRD equations present a poor performance to estimate GFR in individuals with diabetes, especially for high-normal GFRs, with a pronounced underestimation. |
Computer assisted diagnosis of Alzheimer’s disease using statistical likelihood-ratio test
The purpose of this work is to present a computer assisted diagnostic tool for radiologists in their diagnosis of Alzheimer's disease. A statistical likelihood-ratio procedure from signal detection theory was implemented in the detection of Alzheimer's disease. The probability density functions of the likelihood ratio were constructed by using medial temporal lobe (MTL) volumes of patients with Alzheimer's disease (AD) and normal controls (NC). The volumes of MTL as well as other anatomical regions of the brains were calculated by the Free-Surfer software using T1 weighted MRI images. The MRI images of AD and NC were downloaded from the database of Alzheimer's disease neuroimaging initiative (ADNI). A separate dataset of minimal interval resonance imaging in Alzheimer's disease (MIRIAD) was used for diagnostic testing. A sensitivity of 89.1% and specificity of 87.0% were achieved for the MIRIAD dataset which are better than the 85% sensitivity and specificity achieved by the best radiologists without input of other patient information.
# Introduction
Alzheimer's disease is the most common cause of dementia affecting ageing population in the world. MRI T1 weighted structural images are recommended [bib_ref] The diagnosis of dementia due to Alzheimer's disease: recommendation from the National..., Mckhann [/bib_ref] and integrated [bib_ref] The clinical use of structural MRI in Alzheimer disease, Frisoni [/bib_ref] in routine diagnosis of Alzheimer's disease (AD). The medial temporal lobe atrophy (MTA) is the hall mark of AD in MRI images [bib_ref] Alzheimer's disease and normal ageing: diagnostic value and neuropsychological correlates, Scheltens [/bib_ref] [bib_ref] Structural volume of Hippocampus and Alzheimer's disease, Balestrieri [/bib_ref]. Radiologists use a coronal section of the T1 weighted MRI images to rate patients' MTA using a 5 points visual grading scale, based on the height of the hippocampal formation, the widths of the choroid fissure and the temporal horn [bib_ref] Alzheimer's disease and normal ageing: diagnostic value and neuropsychological correlates, Scheltens [/bib_ref] [bib_ref] Structural volume of Hippocampus and Alzheimer's disease, Balestrieri [/bib_ref]. A score of 3 or above is considered abnormal [bib_ref] Alzheimer's disease and normal ageing: diagnostic value and neuropsychological correlates, Scheltens [/bib_ref] [bib_ref] Structural volume of Hippocampus and Alzheimer's disease, Balestrieri [/bib_ref]. A diagnostic sensitivity and specificity of 85% can be achieved by the best radiologists using visual scale gradingwithout using the patients' other information.
The hippocampus is the mostly affected region among the sub-regions of the medial temporal lobe [bib_ref] Alzheimer's disease and normal ageing: diagnostic value and neuropsychological correlates, Scheltens [/bib_ref]. Hippocampal atrophy is one of the core biomarkers in the revised National Institute on Aging-Alzheimer's Association (NIA-AA) diagnostic criteria for AD [bib_ref] Steps to standardization and validation of hippocampal volumetry as a biomarker in..., Jack [/bib_ref]. In addition to the atrophy of the hippocampal volume, the asymmetry of the left and right hemispheres as well as shapes and forms of the hippocampus are also of great importance [bib_ref] Automated hippocampal shape analysis predicts the onset of dementia in mild cognitive..., Costafreda [/bib_ref] [bib_ref] Hippocampal volume and asymmetry in mild cognitive impairment and Alzheimer's disease: meta-analysis..., Shi [/bib_ref] [bib_ref] Mild cognitive disorders are associated with different patterns of brain asymmetry than..., Cherbuin [/bib_ref] differences of these characteristics between AD and normal controls (NC) may point to the origin and staging of the Alzheimer's disease [bib_ref] Dynamics of gray matter loss in Alzheimer's disease, Thompson [/bib_ref] [bib_ref] Diagnostic criteria for neuropathologic assessment of Alzheimer's disease. Neurobiology of, Braak [/bib_ref]. Such differences may suggest potential targets for therapeutic drugs or directions for any drug development [bib_ref] Potential predictors of hippocampal atrophy in Alzheimer's disease, Dhikav [/bib_ref] [bib_ref] Detecting regional cerebral blood flow changes in Alzheimer's patients after Milameline treatment:..., Zheng [/bib_ref]. Subfields of the hippocampus have also been investigated in recent years [bib_ref] Structural imaging of hippocampal subfields in healthy aging and Alzheimer's disease, Flores [/bib_ref]. Many machine learning (ML) algorithms have been developed and studied to assist radiologists' diagnosis of Alzheimer's disease using structural MRI images [bib_ref] Multivariate data analysis and machine learning in Alzheimer's disease with a focus..., Falahati [/bib_ref] [bib_ref] Structural brain imaging in Alzheimer's disease and mild cognitive impairment: biomarker analysis..., Ledig [/bib_ref] including advanced deep machine learning algorithms [bib_ref] Deep leaning in medical image analysis, Shen [/bib_ref] [bib_ref] Latent feature representation with stacked auto-encoder for AD/MCI diagnosis, Suk [/bib_ref]. The diagnostic accuracies of these algorithms are similar to that achieved by radiologists using visual scale rating, i.e. between 80-90% [bib_ref] Multivariate data analysis and machine learning in Alzheimer's disease with a focus..., Falahati [/bib_ref] [bib_ref] Structural brain imaging in Alzheimer's disease and mild cognitive impairment: biomarker analysis..., Ledig [/bib_ref] , or, up to 98.8% if additional information is included [bib_ref] Deep leaning in medical image analysis, Shen [/bib_ref] [bib_ref] Latent feature representation with stacked auto-encoder for AD/MCI diagnosis, Suk [/bib_ref] for distinguishing AD from NC. These computers assisted algorithms are yet to be implemented in clinical practice, particularly, in non-research healthcare settings such as those clinics in regional and remote areas. Clinical radiologists/physicians or healthcare professionals would like to have simple yet accurate tools to assist them in their day-to-day patient care and management. The purpose of this work is to develop a simple computer assisted diagnostic tool in detecting Alzheimer's disease using an MRI T1 weighted image without the need for knowing any other information of a patient. This is intended to be a convenient and effective tool for assisting radiologists and physicians as well as healthcare professionals such as radiographers or nurses who are caring aging populations in regional and remote areas.
# Methods and materials
Radiologists make diagnostic decisions based on their probability knowledge of normal vs diseased images which were developed during their specialist training [bib_ref] Statistical analysis in quantitative research, Zheng [/bib_ref]. The decision threshold of a human observer can be biased in their decision making and an ideal observer makes decisions by placing a criterion on the axis of an underlying random variable. The statistical likelihood-ratio test of an ideal observer uses the probability density functions of normal vs diseased images in its decision making:
[formula] LðNÞ ¼ f n ðxÞð1Þ [/formula]
and
[formula] LðDÞ ¼ f d ðxÞð2Þ [/formula]
Where f n (x) and f d (x) are the probability density functions of normal (N) and diseased (D). For a decision variable x, the likelihood ratio observer uses the probability ratio of the normal and diseased or log-likelihood ratio to make decisions:
[formula] y ¼ logR D : N ð Þ ¼ log f ðxÞ d f ðxÞ n ¼ logf d x ð Þ À logf n x ð Þð3Þ [/formula]
Where y>0 is considered to be diseased and y<0 is considered to be normal. The probability density functions of normal and the diseased can be constructed by using clinical data of pathologically confirmed diseased and normal patients. If the patients are randomly drawn from a population, the probability density functions f n (x) and f d (x) can be considered as normal or Gaussian distributions:
[formula] f n x ð Þ ¼ 1 ð2ps n 2 Þ 1 = 2 exp À 1 2 ½x À m n � s n 2 2 � �ð4Þ [/formula]
And
[formula] f d x ð Þ ¼ 1 ð2ps d 2 Þ 1 = 2 exp À 1 2 ½x À m d � s d 2 2 � �ð5Þ [/formula]
Where σ n and σ d are the standard deviations of normal and diseased and μ n and μ d are the mean values of the normal and diseased. The probability density distributions of the normal and diseased can be constructed by using their mean values and standard deviations. On the other hand, if the patients' data are not normally distributed, the probability density functions of the normal and diseased can be constructed by using freely available statistical package Ron the actual data of the patients.
In this work, the probability density distributions of Alzheimer's patients and normal controls are constructed by using patients' MRI T1 weighted images which were downloaded from the database of Alzheimer's disease neuroimaging initiative (ADNI). All MRI images used in this work are in public domain and the accesses of these databases were approved by the database owners and their respective institutions. As such:. the institutional ethics approval was exempted by Charles Sturt University's Ethics Committee; [bib_ref] The diagnosis of dementia due to Alzheimer's disease: recommendation from the National..., Mckhann [/bib_ref]. methods employed in this study are in accordance with the guidelines of Charles Sturt University's Ethics Committee; and (3). informed consent was obtained prior to data collections by the respective database owners. The ADNI data was collected according to good clinical practice guidelines, US21CFR part 50 -protection of human subjects and part 56 -institutional review boards/ research ethics boards (REBs), and pursuant to state and federal HIPA regulations. The ADNI was launched in 2003 as a public-private partnership, led by Principal Investigator Michael W. Weiner, MD. The primary goal of ADNI has been to test whether serial magnetic resonance imaging (MRI), positron emission tomography (PET), other biological markers, and clinical and neuropsychological assessment can be combined to measure the progression of mild cognitive impairment (MCI) and early Alzheimer's disease (AD). For multiple images of the same AD or NC patients, we only downloaded one T1 weighted MRI image of each patient for the reason of randomness in constructing our probability density functions. Repeated images of a patient are correlated and their inclusions may increase the numbers of images but reduce the randomness for statistical analysis. A total 526 individual patients' T1 weighted MRI images (263 AD and 263 NC with randomly mixed gender and ages from 55 to 97 years) were used for the constructions of probability density functions in this study.
Patients' hippocampus volumes as well as volumes of other anatomical regions of the brains including sub-regions of the medial temporal lobe were calculated by using the "FreeSurfer" software. FreeSurfer is a software package for the analysis and visualization of structural and functional neuroimaging data from cross-sectional or longitudinal studies. It is developed by the Laboratory for Computational Neuroimaging at the Athinoula A. Martinos Centrer for Biomedical Imaging. It is freely available for public use. As shown in [fig_ref] Fig 1: Coronal T1 weighted MRI images of [/fig_ref] the Freesurfer software divides the brain into left and right hemispheres for various anatomical regions. The medial temporal lobe is segmented into hippocampus, entorhinal, amygdala and para-hippocampal subregions. All T1 weighted MRI images of AD and NC patients were processed by using SPAN computing facility at Charles Sturt University and the National Computational Infrastructure (NCI) of Australia at Canberra. A dataset of Minimal Interval Resonance Imaging in Alzheimer's disease (MIRIAD) [bib_ref] MIRIAD-Public release of a multiple time point Alzheimer's MR imaging database, Malone [/bib_ref] was used for the diagnostic testing. The MIRIAD study was approved by the research ethics committee and written consent was obtained prior to the data collection [bib_ref] Combining short interval MRI in Alzheimer's disease: Implications for therapeutic trials, Scott [/bib_ref]. The MIRIAD data set consists of a series of T1 weighted MRI images for each of the 46 AD and 23 NC patients. Again, we used one image from each of the 69 individual patients for the reason of randomness. The SPSS statistical package version 25was used to calculate the area under the curve, Az, or the detectability index, for the joint probability density distributions of normal vs Alzheimer's disease. [fig_ref] Table 1: Left and right hippocampus volumes of normal controls [/fig_ref] shows the averaged volumes of left and right hippocampus for both NC and AD patients. The averaged left hippocampal volume is smaller than that of right hippocampus for both AD (97.3 mm 3 ) and NC (116.5 mm 3 ) subjects. The left/right asymmetry of the hippocampal volumes are 3.26% for AD and 3.20% for NC. There is no statistically significant difference in volumetric asymmetry between AD and NC and no evidence to suggest that the neural degeneration of hippocampus is more or started at the left hand side [bib_ref] Dynamics of gray matter loss in Alzheimer's disease, Thompson [/bib_ref]. Both left and right hand side hippocampal volumes of AD patients are reduced by a similar amount from that of the NC patients (left by 629.5 mm 3 or 19.4% and right by 648.7 mm 3 , or 19.3%). [fig_ref] Fig 2: The ROC curves and their areas under the ROC curves of the... [/fig_ref] the areas under the ROC curves, Az, or detectability index for the detection of AD by using left and right as well as total hippocampal volumes. The Az equals to 0.814, 0.819 and 0.826 for using left, right and total hippocampal volumes, respectively. The Az values show that using total volumes of hippocampus performed better than that of using either right or left, and using right hippocampal volumes performed better than that of using the left hippocampus. This might be because the averaged right hippocampal volume is slightly larger (3.23%) than that of the left hippocampus.
# Results and discussion
## Plos one
Computer assisted diagnosis of Alzheimer's disease AD and NC respectively. The hippocampal volume of the AD is 19.3% smaller than that of the NC patients on average. To test the randomness of the patient datasets, similar probability density distributions of [fig_ref] Fig 3: shows the probability density functions of the AD and NC constructed by... [/fig_ref] were also obtained by using the R software on the patients' data. The results suggest that the Gaussian probability density functions are good for the likelihood ratio observer in detecting Alzheimer's disease using the hippocampal volumes of the ADNI dataset. [fig_ref] Table 2: Mean volumes, standard deviations and detectability index Az of various anatomical regions [/fig_ref] shows the mean volumes, standard deviations and detectability indexes of the total medial temporal lobe (MTL) as well as its sub-regions. The volume of the MTL includes hippocampus, amygdala, entorhinal and para-hippocampal volumes. The detectability index or area under the ROC curve for the total MTL volume in detecting Alzheimer's disease is Az = 0.840 which is the highest among the individual sub-regions of the MTL. The detectability indexes for the individual regions of hippocampus, amygdala, entorhinal and para-hippocampal are Az = 0.827, 0.804, 0.786 and 0.734, respectively, as listed in [fig_ref] Table 2: Mean volumes, standard deviations and detectability index Az of various anatomical regions [/fig_ref]. The fact that the detectability index of the total hippocampus is the highest among the indexes of those individual components of MTL is not surprising as the hippocampus atrophy is one of the diagnostic criteria for Alzheimer's disease [bib_ref] Steps to standardization and validation of hippocampal volumetry as a biomarker in..., Jack [/bib_ref]. However, the present finding that the total volume of the MTL performed better than individual regions suggests that the AD involves not only the hippocampus but a number of other related sub-regions of the MTL. Our detectability indexes suggest that the para-hippocampal region (Az = 0.734) is the least affected among the four sub-regions of the medial temporal lobe. [fig_ref] Fig 4: Probability density distributions of total volumes of the medial temporal lobe [/fig_ref] shows the joint probability density distributions of the medial temporal lobe (MTL) volumes of the normal controls (NC) vs Alzheimer's diseases (AD). Again, the probability density functions are constructed by using Eqs (4 and 5) as did for hippocampus volumes. The mean volumes of NC and AD are 17944 ± 2053 mm 3 and 14697 ± 2522 mm 3 , respectively. The likelihood observer uses the joint probability density distributions to calculate the likelihood ratio and determines if the patient should be classified as normal or Alzheimer's disease. Similar joint probability density distributions can also be constructed for each of the individual anatomical regions, i.e. amygdala, entorhinal and para-hippocampal, and each of these anatomical regions can be used to classify the patents as normal or Alzheimer's disease. Our detectability analysis for each of the individual sub-regions suggests that the total volume of the MTL is the best for our likelihood ratio observer in the detection of the AD as it has the highest detectability index (Az = 0.840). The MIRIAD dataset [bib_ref] MIRIAD-Public release of a multiple time point Alzheimer's MR imaging database, Malone [/bib_ref] was used to test the performances of the likelihood observers constructed by the total volumes of MTL and the hippocampus volumes alone for comparison. the ROC curves for the MIRIAD dataset by using total MTL and hippocampal volumes. The detectability indexes are calculated to be Az = 0.935 and 0.926 with an upper bound of 0.993 and 0.985 by using the total MTL and hippocampal volumes, respectively. It suggests that the MIRIAD dataset is a good clinical dataset for testing our likelihood ratio observer. Both sensitivity and specificity of the likelihood ratio observer using hippocampus alone are 82.6% whilst the sensitivity and specificity of the likelihood ratio observer using total volume of the MTL are 89.1% and 87.0% respectively. Our result shows that using total volume of the MTL improves the performance by 5.45% in comparison with that of using the hippocampal volumes alone. An 85% of sensitivity and specificity can be achieved by the radiologists using scores of visual grading on the volumes of MTLwithout other clinical information
## Plos one
such as the MMSE scores. Our likelihood observer outperformed the best radiologists and our likelihood observer makes decisions based on MTL volume alone without any input of other patients' information such as gender or age. Our likelihood ratio observer can, therefore, assist those who do not have specialty knowledge of neuroradiology, which is especially useful for healthcare professionals such as radiographers and nurses caring aging populations in regional and remote areas. Machine learning (ML) algorithms generally divide a large dataset into subsets for training, validation and testing [bib_ref] A comprehensive study of Alzheimer's disease classification using convolutional networks, Guan [/bib_ref] [bib_ref] Correction for scatter and cross-talk contamination in dual radionuclide 99m-Tc and 123-I..., Zheng [/bib_ref]. Our likelihood ratio observer uses the ADNI dataset in constructing the probability density functions which is analogous to the training of ML algorithms or radiologists' specialist training [bib_ref] Statistical analysis in quantitative research, Zheng [/bib_ref]. The ROC analysis and its detectability index Az are analogous to the validation of ML algorithms using the same ADNI dataset. We used an independent dataset, the MIRIAD dataset, for testing which should be better than that of using the same dataset as that for training (overfitting). It is worth noting that the performance of any machine learning algorithms or Alzheimer's diagnosis in general is dependent on a number of factors, such as the numbers of patients included in the study and whether or not other biomarkers or variables are included in the diagnostic decision making. By including other biomarkers or variables, the diagnostic accuracy may be improved, i.e. higher than that of using a single biomarker such as the MTL volume. The numbers of patients used in a study can have a major impact on the performance evaluation. The performance index or diagnostic accuracy using smaller numbers of patients may or may not be higher but the performance results from using large numbers of patients should be more robust and reliable [bib_ref] Multivariate data analysis and machine learning in Alzheimer's disease with a focus..., Falahati [/bib_ref]. The fact that we used independent (no repeat images of the same patient) 263 AD and 263 NC patients in the construction of the probability density functions for our likelihood ratio observer suggests that the 84% detectability of our likelihood ratio observer using the MTL volume calculated by a publicly available software should be a robust and reliable performance index for any clinical setting without knowing patient's any other information such as gender or age. Our likelihood ratio observer is equivalent to the best radiologists without any knowledge of neuroradiology. Our likelihood observer needs only a patient's T1 weighted image to make a diagnostic decision. This is especially useful for healthcare professionals such as radiographers and nurses who don't have specialist training and are caring aging populations in regional and remote areas. Further work is to test the robustness of our likelihood observer in clinical practice which is underway.
A comparison of our results with a recent work of Ledig et al [bib_ref] Structural brain imaging in Alzheimer's disease and mild cognitive impairment: biomarker analysis..., Ledig [/bib_ref] is worthwhile. Ledig et al [bib_ref] Structural brain imaging in Alzheimer's disease and mild cognitive impairment: biomarker analysis..., Ledig [/bib_ref] used a MAPEM software for brain MR image segmentations and volume calculations. They used both Random Forest and Support Vector Machine (SVM) algorithms for disease classifications. Their highest achievable sensitivities and specificities (Random Forest, SVM) are (83%, 86%) and (90%, 92%), respectively, using all features including gender, brain size and age corrections. This contrasts with our likelihood ratio observer of 87% sensitivity and 89. 1% specificity using MTL volume only without gender, brain size and age corrections. The overall accuracy of our likelihood ratio observer (88%) is equivalent to that of the Random Forest (87%) and SVM (90%) [bib_ref] Structural brain imaging in Alzheimer's disease and mild cognitive impairment: biomarker analysis..., Ledig [/bib_ref]. Using the volumes of hippocampus, entorhinal and amygdala and a linear discriminant analysis (LDA) classifier, the overall classification accuracies of Ledig et al [bib_ref] Structural brain imaging in Alzheimer's disease and mild cognitive impairment: biomarker analysis..., Ledig [/bib_ref] are 78%, 80% and 78% with, and 75%, 73% and 75% without gender, brain size and age corrections. This contrasts with our likelihood ratio observer of 82.6%, 80.4% and 78.6% without gender, brain size and age corrections. Our likelihood ratio observer performed better than that of the LDA using individual anatomical volumes. More importantly, our result is consistent with the fact that hippocampus is the most affected area in Alzheimer's disease [bib_ref] Steps to standardization and validation of hippocampal volumetry as a biomarker in..., Jack [/bib_ref] which should have the highest discriminant power for Alzheimer's disease. It also suggests that the correction for gender, brain size and age could improve the overall diagnostic accuracy by about 3%. One of the main limitations of this work is that gender, brain size and age corrections are not included in the calculations of brain volumes. This limitation is also our strength that the likelihood ratio observer is a simple yet highly accurate diagnostic tool to assist not only radiologists but also allied health professionals such as radiographers and nurses caring ageing populations. Further work is to include age, brain size and gender in the brain volume calculations.
# Conclusions
A computer assisted diagnostic tool is developed and tested for radiologists in their diagnosis of Alzheimer's disease. It makes diagnostic decision based on a patient's medial temporal lobe volume which can be calculated by a publicly available software using a T1 weighted MRI images without knowledge of a patient's any other information such as gender or age. It is a simple and robust tool not only in assisting radiologists for their diagnosis of Alzheimer's disease, but also for those healthcare professionals such as radiographers or nurses who have no adequate knowledge of neuroradiology and care for aging populations in regional and remote areas. The AD diagnosis by the likelihood ratio observer using hippocampal volume suggests that the hippocampus is the most affected region of atrophy among the sub-regions of the medial temporal lobe but the total volume atrophy of the MTL is more effective in AD diagnosis. There is an asymmetry of the hippocampal volumes that the left is smaller than right on average for both AD and NC patients yet there is no evidence to suggest left/right asymmetrical reductions of the hippocampal volumes between AD and NC patients. Further work is to test the robustness of the likelihood ratio observer in a wider clinical practice for the diagnosis of the Alzheimer's disease.
## Supporting information
[fig] Fig 1: Coronal T1 weighted MRI images of (a) Alzheimer's patient ID-201 and (b) normal patient ID-200 of the MIRIAD dataset. The FreeSurfer software divides the brain into left and right hemispheres and segments the medial temporal lobe into hippocampus, entorhinal, amygdala and para-hippocampal subregions. https://doi.org/10.1371/journal.pone.0279574.g001 [/fig]
[fig] Fig 3: shows the probability density functions of the AD and NC constructed by using the total mean volumes of the hippocampus assuming a Gaussian distribution of Eqs (4 and 5). The total mean volumes of the hippocampus are 5962.1±1025.3 and 7238.4±870.0 mm 3 for [/fig]
[fig] Fig 2: The ROC curves and their areas under the ROC curves of the left, right and total hippocampal volumes for the detection of Alzheimer's disease. https://doi.org/10.1371/journal.pone.0279574.g002 [/fig]
[fig] Fig 4: Probability density distributions of total volumes of the medial temporal lobe (MTL): Normal controls (NC) vs Alzheimer's disease (AD). The mean volumes (±SD) of NC and AD are 17944 (±2053) mm 3 and 14697 (±2522) mm 3 , respectively. https://doi.org/10.1371/journal.pone.0279574.g004Fig 5. The ROC curves and their areas under the ROC curves for the total MTL and hippocampal volumes for the detection of Alzheimer's disease. https://doi.org/10.1371/journal.pone.0279574.g005 [/fig]
[fig] S1: File. The dataset for the construction of probability density distribution of Alzheimer's disease. (ZIP) S2 File. The dataset for the construction of probability density distribution of normal controls. (ZIP) S3 File. The test dataset of Alzheimer's disease. (ZIP) S4 File. The test dataset of normal controls. (ZIP) [/fig]
[table] , because: PLOS ONEPLOS ONE | https://doi.org/10.1371/journal.pone. [/table]
[table] Table 1: Left and right hippocampus volumes of normal controls (NC) and Alzheimer's diseases (AD). [/table]
[table] Table 2: Mean volumes, standard deviations and detectability index Az of various anatomical regions. [/table]
|
Cytotoxicity of 15-Deoxy-Δ12,14-prostaglandin J2 through PPARγ-independent Pathway and the Involvement of the JNK and Akt Pathway in Renal Cell Carcinoma
Introduction:Agonists of peroxisome proliferator-activated receptor gamma (PPARγ) have been examined as chemopreventive and chemotherapeutic agents. The aim was to investigate the cytotoxicity and action mechanisms of 15-deoxy-Δ 12,14 -prostaglandin J 2 (15d-PGJ 2 ), one of endogenous ligands for PPARγ, in terms of PPARγ-dependency and the mitogen-activated protein kinase (MAPK) and Akt pathway in three human renal cell carcinoma (RCC)-derived cell lines. Methods: 786-O, Caki-2 and ACHN cells were used as human RCC-derived cell lines. Cell viability and caspase-3 activity was detected by fluorescent reagents, and chromatin-condensation was observed with a brightfield fluorescent microscope after staining cells with Hoechst33342. The expression levels of proteins were detected by Western blot analysis. Results: 15d-PGJ 2 showed cytotoxicity in dose-dependent manner. 15d-PGJ 2 induced chromatin-condensation and elevated caspase-3 activity, and the cell viability was restored by co-treatment with a pan-caspase inhibitor, Z-VAD-FMK, indicating the involvement of caspase-dependent apoptosis. The cytotoxicity was not impaired by a PPARγ inhibitor, GW9662, suggesting that 15d-PGJ 2 exerted the cytotoxicity in a PPARγ-independent manner. Some antioxidants rescued cells from cell death induced by 15d-PGJ 2 , but some did not, suggesting that reactive oxygen species (ROS) did not contribute to the apoptosis. 15d-PGJ 2 also increased the expression levels of phospho-c-Jun N terminal kinase (JNK) in Caki-2 cells, and decreased those of phospho-Akt in 786-O cells, indicating that the JNK MAPK and the Akt pathways participated in the anticancer effects of 15d-PGJ 2 in some cell lines. Conclusion: 15d-PGJ 2 exerted cytotoxic effects accompanying caspase-dependent apoptosis, and this effect was elicited in a PPARγ-independent manner in three cell lines. In addition, the JNK MAPK and Akt pathway was involved in the cytotoxicity of 15d-PGJ 2 to some extent in some cell line. Therefore, our study showed the 15d-PGJ 2 to potentially be an interesting approach for RCC treatment.
# Introduction
Renal cell carcinoma (RCC) accounts for about 2% of all tumors, and the most common histological subtype is clear renal cell carcinoma (75%) [bib_ref] Epidemiologic aspects of renal cell carcinoma, Mclaughlin [/bib_ref]. Although nephrectomy is the most effective treatment at an early stage, advanced renal cancer is still associated with a poor prognosis, with a five-year survival rate of less than 10%, as RCC is relatively resistant to chemotherapy and radiotherapy [bib_ref] Renal cell carcinoma. Review of novel single agent therapeutics and combination regimens, Amato [/bib_ref]. Recently, novel agents for RCC targeting cancer-specific pathways have been developed, such as sorafenib and sunitinib. Although they have been shown to be beneficial in patients with advanced RCC [bib_ref] Targeted-therapy in advanced renal cell carcinoma, Pirrotta [/bib_ref] [bib_ref] Building on a foundation of VEGF and mTOR targeted agents in renal..., Flaherty [/bib_ref] , the effect is insufficient and it is therefore necessary to discover new targets for the treatment of RCC. Indeed, there have been several reports of other possible targets by us and others [bib_ref] Building on a foundation of VEGF and mTOR targeted agents in renal..., Flaherty [/bib_ref] [bib_ref] The efficacy of the novel dual PI3-kinase/mTOR inhibitor NVP-BEZ235 compared with rapamycin..., Cho [/bib_ref] [bib_ref] Cytotoxicity of troglitazone through PPARγ-independent pathway and p38 MAPK pathway in renal..., Fujita [/bib_ref].
Peroxisome proliferator-activated receptor gamma (PPARγ) belongs to the superfamily of nuclear hormone receptor transcription factors [bib_ref] Activation of a member of the steroid hormone receptor superfamily by peroxisome..., Issemann [/bib_ref] , playing major roles in adipogenesis [bib_ref] Adipogenesis and fat-cell function in obesity and diabetes, Camp [/bib_ref] , glucose metabolism [bib_ref] Peroxisome proliferator-activated receptor γ in diabetes and metabolism, Rangwala [/bib_ref] , and angiogenesis [bib_ref] Differential regulation of vascular endothelial growth factor expression by peroxisome proliferator-activated receptors..., Fauconnet [/bib_ref]. It has been reported that PPARγ is expressed in normal tissues and its expression is changed at a variety of tumor sites [bib_ref] Peroxisome proliferator-activated receptor γ and cancers, Koeffler [/bib_ref]. It has been also detected in the tissue of renal cell carcinoma patients at the mRNA and protein level [bib_ref] Safe S. 1,1-Bis(3'-Indolyl)-1-(p-Substitutedphenyl)methanes induce apoptosis and inhibit renal cell carcinoma growth, York [/bib_ref] [bib_ref] Proteasome inhibition overcomes the resistance of renal cell carcinoma cells against the..., Von Schwarzenberg [/bib_ref].
15-deoxy-Δ 12,14 -prostaglandin J2 (15d-PGJ2) is an end-product of prostaglandin D2, and known as an endogenous ligand for PPARγ [bib_ref] 15-Deoxy-Δ 12,14 -prostaglandin J2 as a potential endogenous regulator of redox-sensitive transcription..., Kim [/bib_ref] [bib_ref] A prostaglandin J2 metabolite binds peroxisome proliferator-activated receptor γ and promotes adipocyte..., Kliewer [/bib_ref]. showed that tissue levels of 15d-PGJ2 were markedly decreased in tumors and metastatic breast tissue compared with control tissue, suggesting that the modulation in tissue levels of 15d-PGJ2 influence the development of cancer and its progression to metastasis [bib_ref] Expression of cyclooxygenase-2 and peroxisome proliferator-activated receptor-γ and levels of prostaglandin E2..., Badawi [/bib_ref]. Therefore, it is thought to be well worth investigating the effects of 15d-PGJ2 on cancer cells as a potential treatment for cancer. Indeed, 15d-PGJ2 shows not only anti-inflammatory and cytoprotective activities, but also proapoptotic and anti-proliferative properties in many cancer cell lines [bib_ref] Peroxisome proliferator-activated receptor γ and cancers, Koeffler [/bib_ref] , including colorectal [bib_ref] Inhibition of activator protein 1 activation, vascular endothelial growth factor, and cyclooxygenase-2..., Grau [/bib_ref] [bib_ref] 15d-PGJ2 induces apoptosis by reactive oxygen species-mediated inactivation of Akt in leukemia..., Shin [/bib_ref] [bib_ref] PPARγ inhibitors reduce tubulin protein levels by a PPARγ, PPARδ and proteasome-independent..., Schaefer [/bib_ref] [bib_ref] Selective activation of PPARγ in breast, colon, and lung cancer cell lines, Allred [/bib_ref] , lung [bib_ref] Selective activation of PPARγ in breast, colon, and lung cancer cell lines, Allred [/bib_ref] , breast [bib_ref] Selective activation of PPARγ in breast, colon, and lung cancer cell lines, Allred [/bib_ref] [bib_ref] 15-Deoxy-Δ 12,14 -prostaglandin J2-induced apoptosis does not require PPARγ in breast cancer..., Clay [/bib_ref] , hepatocellular [bib_ref] 15-deoxy-Δ-12-14 -PGJ2 regulates apoptosis induction and nuclear factor-κB activation via a peroxisome..., Okano [/bib_ref] , renal cell [bib_ref] Peroxisome proliferator-activated receptor gamma is frequently underexpressed in renal cell carcinoma, Yuan [/bib_ref] [bib_ref] Ligand for peroxisome proliferator-activated receptor gamma have potent antitumor effect against human..., Yuan [/bib_ref] [bib_ref] The effect of peroxisome proliferator-activated receptor-γ ligand on urological cancer cells, Yoshimura [/bib_ref] , prostate, and bladder cancers [bib_ref] The effect of peroxisome proliferator-activated receptor-γ ligand on urological cancer cells, Yoshimura [/bib_ref] [bib_ref] PPARγ-independent induction of growth arrest and apoptosis in prostate and bladder carcinoma, Chaffer [/bib_ref]. These apoptotic effects are thought to be mediated by PPARγ-dependent [bib_ref] Peroxisome proliferator-activated receptor-γ activator 15-deoxy-Δ 12,14 -prostaglandin J2 inhibits neuroblastoma cell growth..., Kim [/bib_ref] [bib_ref] 15-deoxy-Δ 12,14 -prostaglandin J2 induces apoptosis of human hepatic myofibroblasts. A pathway..., Li [/bib_ref] and PPARγ-independent [bib_ref] PPARγ inhibitors reduce tubulin protein levels by a PPARγ, PPARδ and proteasome-independent..., Schaefer [/bib_ref] [bib_ref] 15-Deoxy-Δ 12,14 -prostaglandin J2-induced apoptosis does not require PPARγ in breast cancer..., Clay [/bib_ref] [bib_ref] 15-deoxy-Δ-12-14 -PGJ2 regulates apoptosis induction and nuclear factor-κB activation via a peroxisome..., Okano [/bib_ref] [bib_ref] PPARγ-independent induction of growth arrest and apoptosis in prostate and bladder carcinoma, Chaffer [/bib_ref] [bib_ref] MAP kinase cascades are activated in astrocytes and preadipocytes by 15-deoxy-Δ 12-14..., Lennon [/bib_ref] pathways. Moreover, a number of apoptotic mechanisms have been proposed for 15d-PGJ2 [bib_ref] Peroxisome proliferator-activated receptor γ and cancers, Koeffler [/bib_ref] [bib_ref] 15-Deoxy-Δ 12,14 -prostaglandin J2 as a potential endogenous regulator of redox-sensitive transcription..., Kim [/bib_ref] [bib_ref] 15-Deoxy-Δ 12,14 -prostaglandin J2-induced apoptosis does not require PPARγ in breast cancer..., Clay [/bib_ref] [bib_ref] Peroxisome proliferator-activated receptor-γ activator 15-deoxy-Δ 12,14 -prostaglandin J2 inhibits neuroblastoma cell growth..., Kim [/bib_ref] [bib_ref] Suppression of chondrosarcoma cells by 15-deoxy-Δ 12,14 -prostaglandin J2 is associated with..., Shen [/bib_ref] [bib_ref] 15-deoxy-Δ 12,14 -prostaglandin J2 induces vascular endothelial cell apoptosis through the sequential..., Ho [/bib_ref] [bib_ref] 15-Deoxy-Δ 12,14 -prostaglandin J2: the endogenous electrophile that induces neuronal apoptosis, Kondo [/bib_ref] [bib_ref] 15-deoxy-Δ 12,14 -prostaglandin J2 induces renal epithelial cell death through NF-κB-dependent and..., Kang [/bib_ref] , for example, the involvement of the mitogen-activated protein kinase (MAPK) pathway [bib_ref] Inhibition of activator protein 1 activation, vascular endothelial growth factor, and cyclooxygenase-2..., Grau [/bib_ref] [bib_ref] 15d-PGJ2 induces apoptosis by reactive oxygen species-mediated inactivation of Akt in leukemia..., Shin [/bib_ref] [bib_ref] MAP kinase cascades are activated in astrocytes and preadipocytes by 15-deoxy-Δ 12-14..., Lennon [/bib_ref] [bib_ref] 15-deoxy-Δ 12,14 -prostaglandin J2 induces vascular endothelial cell apoptosis through the sequential..., Ho [/bib_ref] ; however, little information on RCC has been available to date. Furthermore, we already demonstrated that 15d-PGJ2 enhanced the anti-tumor activity of camp-tothecin against Caki-2 cells, one of renal carcinoma cell lines independently of PPARγ pathway [bib_ref] 15-deoxy-Δ 12,14 -prostaglandin J2 enhanced the anti-tumor activity of camptothecin against renal..., Yamamoto [/bib_ref].
In the present study, we investigated whether 15d-PGJ2 induced cell death in three human RCC cell lines, 786-O (Von Hippel-Lindau (VHL) -deficient and primary organ-derived cells), Caki-2 (VHL-expressing and primary organ-derived cells) and ACHN (VHL-expressing and metastatic site-derived cells). We chose them because all three cell lines have been commonly used and we could elucidate whether they would respond to 15d-PGJ2 through a similar mechanism or not. We also examined its mechanisms of action in terms of dependency on PPARγ and the MAPK and Akt pathways, comparing among three cell lines.
# Materials and methods
Chemicals 15d-PGJ2 was purchased from Cayman Chemical (Ann Arbor, MI) and dissolved in DMSO just before use. The final concentration of DMSO in the medium did not exceed 0.1%. GW9662 was purchased from Sigma (St. Louis, MO). N-benzyloxycarbonyl-Val-Ala-Asp fluoromethyl ketone (Z-VAD-FMK) was obtained from Enzo Life Sciences International (Plymouth Meeting, PA). N-acetyl-L-cysteine (NAC), reduced glutathione (GSH), vitamin E, and melatonin were purchased from Nacalai Tesque (Kyoto, Japan). SB202190 (p38 inhibitor), SP600125 (c-Jun N-terminal kinase (JNK) inhibitor), and Akt inhibitor IV were obtained from Merck (Darmstadt, Germany).
## Cells and cell culture
786-O, ACHN, and Caki-2 cells were obtained from Summit Pharmaceuticals International (Tokyo, Japan), the Cell Resource Center for Biomedical Research Institute of Development, Aging and Cancer Tohoku University (Sendai, Japan), and DS Pharma Biomedical (Osaka, Japan), respectively, and were used as human renal carcinoma cell models. Then 786-O and Caki-2 were maintained in a culture medium consisting of RPMI1640 (Nacalai Tesque) and ACHN was cultured in DMEM (Wako Pure Chemical Industries, Osaka, Japan). The media were supplemented with 10% heat-inactivated FBS (Invitrogen, Life Technologies, Carlsbad, CA) and 50 U/mL penicillin-50 µg/mL streptomycin (Nacalai Tesque). Cells were cultured in an atmosphere of 95% air and 5% CO2 at 37˚C, and subcultured every 3 or 4 days.
## Cell viability
## Cell viability was measured with cell
Quanti-Blue TM (BioAssay Systems, Hayward, CA). Briefly, cells were seeded onto 96-well plates (Asahi Glass, Tokyo, Japan) at 5 × 10 3 /well and incubated for 24 h. The cells were treated with 15d-PGJ2 or Akt inhibitor IV in the presence or absence of other chemicals for a further 24 h using FBS-free medium. The assay was performed as described previously [bib_ref] Cytotoxicity of troglitazone through PPARγ-independent pathway and p38 MAPK pathway in renal..., Fujita [/bib_ref]. It utilizes the conversion of alamar blue reagent to resorufin fluorescence by metabolically active cells and resorufin was measured in a CytoFluor ® Series 4000 Fluorescence Multi-Well Plate Reader (PerSeptive Biosystems, Framingham, MA) at an excitation wavelength of 530 nm / emission 580 nm. The 50% growth inhibitory concentrations (IC50) were calculated according to the sigmoid inhibitory effect model E = IC50 γ /(IC50 γ + C γ ). E represents the surviving fraction (% of control), and C and γ represent the drug concentration in the medium and the Hill coefficient, respectively. For co-exposure experiments, 15d-PGJ2 was used at 1.
## Detection of chromatin condensation (fluorescence microscopy)
For nuclear staining, cells were treated with 15d-PGJ2 for 24 h at 1.5, 7 and 3 μM for 786-O, Caki-2 and ACHN cell lines, respectively. Immediately after, the nuclear chromatin of trypsinized cells was stained with 80 μg/mL of Hoechst 33342 (Nacalai Tesque) in the dark at room temperature for 15 min. They were then observed with a brightfield fluorescent microscope (VANOX; Olympus, Tokyo, Japan) under UV excitation. Chromatin-condensed cells were photographed at a 40-fold magnification. In addition, at a 20-fold magnification, more than 120 cells with condensed chromatin were counted in each experiment and their percentage was calculated.
## Fluorimetric assay of caspase-3 activity
Caspase-3 activity was assessed using a Caspase 3 Assay Kit, Fluorimetric (Sigma), according to the manufacturer's instructions. Briefly, cells were seeded in 96-well plates at 5 × 10 3 /well and cultured for 24 h. After exposure to 15d-PGJ2 at 1.5, 7 and 3 μM for 786-O, Caki-2, and ACHN cell lines, respectively, the supernatant was aspirated and cells were harvested with lysis buffer (50 mM HEPES, pH 7.4, 5 mM CHAPS and 5 mM DTT). The reaction buffer, containing 16.6 μM Acetyl-Asp-Glu-Val-Asp-7amido-4-methylcoumarin (Ac-DEVD-AMC), a caspase-3-specific substrate, was added to the wells, and the production of AMC was detected in a CytoFluor ® Plate Reader at 360 nm (excitation) / 460 nm (emission). Enzymatic activity was determined as initial velocity expressed as nmol AMC/min/mL. It was then corrected with the quantity of protein in each well detected by a BCA protein assay kit (Thermo Fisher Scientific, Waltham, MA).
## Assay of lactate dehydrogenase (ldh) leakage
Cells were seeded onto 24-well plates (Asahi glass) at 3 × 10 4 /well and cultured for 24 h. The medium was removed and the cells were washed with fresh serum-free medium 3 times and incubated for 1 h. They were treated with 15d-PGJ2 at 1.5, 7 and 3 μM for 786-O, Caki-2, and ACHN cell lines, respectively, for 24 h, and supernatant was obtained by centrifugation (170 × g, 3 min). The LDH activity in the supernatant was assayed with LDH-Cytotoxic Test Wako (Wako), according to the manufacturer's instructions. Briefly, the samples were reacted with a coloring solution containing nitrotetrazolium blue, and the diformazan produced from nitrotetrazolium blue by oxidation of NADH was detected in a MTP-500 microplate reader (CORONA ELECTRIC, Ibaraki, Japan) at 570 nm. LDH leakage was calculated according to the equation; LDH leakage (% per positive control) = (S-N) / (P-N). S, N and P represent the absorbance in the sample, negative and positive controls, respectively. The positive control was the supernatant of cells treated with 0.2% Tween 20 (Nacalai Tesque), and the negative control was the supernatant incubated with 0.1% DMSO.
## Antibodies
The rabbit monoclonal antibodies against phospho-SAPK/JNK (Thr183/Tyr185), Akt (C67E7) and phospho-Akt (Ser 473) (D9E) XP™ were purchased from Cell Signaling Technology (Danvers, MA) and the mouse monoclonal antibody against β-actin (C4), from Santa Cruz Biotechnology (Santa Cruz, CA). The horseradish peroxidase-linked antibody for donkey anti-rabbit IgG and sheep anti-mouse IgG was obtained from GE Healthcare (Buckinghamshire, UK). All antibodies were diluted with blocking buffer.
# Western blot analysis
ACHN cells were harvested by scraping and washed with ice-cold phosphate-buffered saline (PBS). In the experiment for phosphorylation of JNK MAPK, cells were plated in 100-mm dishes 24 h before treatment and then treated with 15d-PGJ2 at 3 μM for 1, 8 or 24 h, and in the experiment for phosphorylation of Akt, cells were treated with 15d-PGJ2 for 1 and 24 h. Lysis buffer (20 mM Tris (pH 7.5), 150 mM NaCl, 1% Triton X-100, 0.5% sodium deoxycholate, 1 mM EDTA, 0.1% SDS, 1mM NaF, 1 mmol/L Na3VO4 and 0.1% protease inhibitor cocktail (Sigma)) was added to pellets, and the cells were sonicated briefly then incubated on ice for 20 min. Cell extracts were centrifuged at 16,000 × g for 15 min at 4˚C, and the supernatants were transferred to new tubes. Protein concentrations were determined by a BCA protein assay kit (Thermo Fisher Scientific). The samples were mixed with the same volume of Laemmli Sample buffer containing β-mercaptoethanol (Bio-Rad Laboratories, Hercules, CA) and boiled for 5 min, and 15 μg of protein was loaded onto 10% SDS-polyacrylamide gels. After electrophoresis, the proteins were transferred to polyvinylidene difluoride membranes (ATTO, Tokyo, Japan) and blocked with Tris-buffered saline-0.1% Tween 20 (TBS-T) containing 2% ECL Advance™ Blocking Agent (GE Healthcare) for 1 h. Blocked membranes were reacted with primary antibodies (diluted 1:10,000) for 1 h at room temperature and washed five times with TBS-T. After incubation with the secondary antibody (diluted 1:25,000) for 1 h at room temperature, the membranes were again washed five times. The signal was visualized using ECL Advance™ detection reagents (GE Healthcare). Protein expression levels were evaluated on densitometric measurements using ImageJ software. β-actin level was used as a standard for each sample, and protein levels at base line (0 h) were defined as 1.0.
# Statistical analysis
Data are expressed as the mean ± SD. Statistical significance was compared with Student's or Welch's t-test for comparisons of two groups and non-repeated one-way ANOVA followed by Dun-nett's post-hoc test for multiple comparisons. P < 0.05 was considered significant. [fig_ref] Figure 1: Concentration-dependent cytotoxic effects of 15d-PGJ2 in 786-O, Caki-2, and ACHN cells [/fig_ref] shows the effects of 15d-PGJ2 on cell viability in three RCC cell lines. 15d-PGJ2 showed cytotoxicity in a dose-dependent manner with IC50 values of 1.41 ± 0.09, 6.45 ± 0.26, and 3.21 ± 0.87 µM in 786-O, Caki-2, and ACHN cells, respectively, suggesting that 15d-PGJ2 showed cytotoxicity in RCC cell lines.
# Results
## Effects of 15d-pgj2 on cell viability
## Effects of 15d-pgj2 on nuclear morphology and caspase-3 activity
To explore whether the cytotoxicity of 15d-PGJ2 occurred through apoptosis or necrosis, we observed chromatin condensation, a typical morphological change in apoptosis, and the activation of caspase-3, a downstream factor of the caspase cascade. When cells were exposed to 15d-PGJ2 at approximately the IC50 of each cell line for 24 h, chromatin condensation was found after staining with Hoechst 33342 [fig_ref] Figure 2: Apoptosis assays for 15d-PGJ2 in three RCC cell lines [/fig_ref]. The percentage of cells with chromatin condensation was significantly increased by 15d-PGJ2 treatment, from 3 to 24, 4 to 27 and 7 to 30% in 786-O, Caki-2, and ACHN cells, respectively [fig_ref] Figure 2: Apoptosis assays for 15d-PGJ2 in three RCC cell lines [/fig_ref]. Next, we examined the activation of caspase-3. Treatment with 15d-PGJ2 resulted in approximately a 36-, 12-, and 19-fold increase of caspase activity in 786-O, Caki-2, and ACHN cells, respectively [fig_ref] Figure 2: Apoptosis assays for 15d-PGJ2 in three RCC cell lines [/fig_ref]. Moreover, a pan-caspase inhibitor, Z-VAD-FMK, significantly but not completely recovered the viability of cells exposed to 15d-PGJ2, from 25 to 71%, 21 to 47%, and 31 to 63%, respectively [fig_ref] Figure 2: Apoptosis assays for 15d-PGJ2 in three RCC cell lines [/fig_ref]. These results suggested that 15d-PGJ2 induced cell death through caspase-dependent apoptosis to significant extent. Effects of a pan-caspase inhibitor on 15d-PGJ2-induced cell death. Cells were precultured for 24 h at 5 × 10 3 /well in 96-well plates and exposed to 15d-PGJ2 with or without Z-VAD-FMK (100 μM) for 24 h. Cell viability was assessed by fluorescent assay, and data represent the mean ± S.D. from 4 to 5 independent preparations. Statistical significance was assessed by t-test.
## Effects of 15d-pgj2 on ldh leakage
To find the contribution of necrosis to cell death, we next measured the levels of LDH in media as a marker of cell lysis and necrosis [bib_ref] Cellular non-heme iron content is a determinant of nitric oxide-mediated apoptosis, necrosis,..., Kim [/bib_ref]. LDH leakage per positive control were 12.3 ± 2.9%, 13.2 ± 0.7%, and 9.6 ± 0.8% in 786-O, Caki-2, and ACHN cells, respectively, when cells were treated with approximately the IC50 of 15d-PGJ2 [fig_ref] Figure 3: LDH leakage in RCC cell lines treated with 15d-PGJ2 [/fig_ref] , suggesting that necrosis was partly involved in cell death.
## Participation of pparγ in the effect of 15d-pgj2 on cell viability
To investigate the molecular mechanisms of cell death induced by 15d-PGJ2, we first examined PPARγ-dependency. shows the effects of a PPARγ antagonist, GW9662, on 15d-PGJ2-induced cell death. The cytotoxic effect of 15d-PGJ2 was not impaired by GW9662, indicating it to be independent of PPARγ.
## The effects of antioxidants on the cytotoxicity of 15d-pgj2
As 15d-PGJ2 was also reported to generate reactive oxygen species (ROS) [bib_ref] Dissociation of oxidant production by peroxisome proliferator-activated receptor ligands from cell death..., Atarod [/bib_ref] , we next investigated the effects of antioxidants. shows the effect of antioxidants, NAC and GSH, on the cytotoxic action of 15d-PGJ2. NAC or GSH almost completely reversed the 15d-PGJ2-induced cell death in all three cell lines. Conversely, there was no recovery at all from 15d-PGJ2-induced cell death among cells co-treated with other antioxidants, vitamin E (1 µM) and melatonin (1 mM) [fig_ref] Figure 6: Effects of antioxidants without a thiol-group, vitamin E [/fig_ref].
## Participation of the p38 and jnk mapk pathways in the effect of 15d-pgj2 on cell viability
To investigate the involvement of the MAPK pathway in the cytotoxic mechanisms of 15d-PGJ2, two MAPK inhibitors were co-treated with 15d-PGJ2. [fig_ref] Figure 7: Involvement of the p38 and JNK MAPK pathway in 15d-PGJ2-induced cell death [/fig_ref] shows the effects of a p38 inhibitor, SB202190, on 15d-PGJ2-induced cell death. Cell viability was not impaired by co-exposure to SB202190, suggesting that the p38 pathway did not mainly participate in the cytotoxic effect of 15d-PGJ2.
## Figure 4. effects of a pparγ inhibitor, gw9662
, on 15d-PGJ2-induced cell death. Cells were precultured for 24 h at 5 × 10 3 /well in 96-well plates and exposed to 15d-PGJ2 with or without GW9662 (20 μM) for 24 h. Cell viability was assessed by fluorescent assay, and data represent the mean ± S.D. from 5 independent preparations. Statistical significance was assessed by t-test.
## Figure 5. effects of antioxidants with a thiol-group, nac (a) and gsh (b), on 15d-pgj2-induced cell death.
Cells were precultured for 24 h at 5 × 10 3 /well in 96-well plates and exposed to 15d-PGJ2 in the presence or absence of NAC (500 μM) or GSH (1 mM) for 24 h. Cell viability was assessed by fluorescent assay, and data represent the mean ± S.D. from 4 to 5 independent preparations. Statistical significance was assessed by t-test. . Cells were precultured for 24 h at 5 × 10 3 /well in 96-well plates and exposed to 15d-PGJ2 at approximately the IC50 in the presence or absence of SB202190 (3 μM) or SP600125 (0.1, 0.3, 1 μM) for 24 h. Cell viability was assessed by fluorescent assay, and data represent the mean ± S.D. from 4 independent preparations. Statistical significance was assessed by t-test or Dunnett's test. To detect proteins, cells were precultured for 24 h in 100-mm dishes. Cells were then treated with 15d-PGJ2 at 3 μM for 0 and 8 h. Protein (15 μg) was analyzed by Western blotting for the expression of phospho-JNK. Relative protein levels were quantified using ImageJ, and each phospho-JNK signal was normalized to the β-actin signal. The representative bands and the results of densitometric analysis from three independent preparations were described, and data represent the mean ± S.D. from 4 independent preparations.
A JNK inhibitor, SP600125, did not prevent 15d-PGJ2-induced cell death in 786-O and Caki-2 cells, but significantly improved the viability of ACHN cells [fig_ref] Figure 7: Involvement of the p38 and JNK MAPK pathway in 15d-PGJ2-induced cell death [/fig_ref].
As improvement of cell viability by SP600125 (1 µM) was observed only in ACHN cells, we next investigated the dose-dependent effects of SP600125 and the effects of 15d-PGJ2 on the expression of phospho-JNK in ACHN cells. SP600125 improved the viability of ACHN cells in a dose-dependent manner [fig_ref] Figure 7: Involvement of the p38 and JNK MAPK pathway in 15d-PGJ2-induced cell death [/fig_ref]. In addition, the expression levels of phospho-JNK was tend to be increased by exposure to 15d-PGJ2 for 8 h [fig_ref] Figure 7: Involvement of the p38 and JNK MAPK pathway in 15d-PGJ2-induced cell death [/fig_ref]. This suggested the JNK pathway to be involved in 15d-PGJ2-induced cell death to some extent in ACHN cells. To investigate the effects of an Akt inhibitor, cells were precultured for 24 h at 5 × 10 3 /well in 96-well plates and treated with Akt inhibitor IV for 24 h. Cell viability was assessed by fluorescent assay, and data represent the mean ± S.D. from 6 independent preparations. Relative protein levels were quantified using ImageJ, and each Akt and phospho-Akt signal was normalized to the β-actin signal. Relative ratio of normalized Akt and phospho-Akt signal to untreated cells in each time was described.
## Participation of the akt pathways in the effect of 15d-pgj2 on cell viability
The expression levels of phospho-Akt were markedly decreased by exposure to 15d-PGJ2 for 1 and 24 h only in 786-O cells, whereas the levels of total Akt were not changed [fig_ref] Figure 8: Involvement of the Akt pathway in 15d-PGJ2-induced cell death [/fig_ref] , but not in other cells (data not shown). In addition, Akt inhibitor IV decreased the cell viability of 786-O cells in a dose-dependent manner, with an IC50 value of 3.67 ± 0.84 μM [fig_ref] Figure 8: Involvement of the Akt pathway in 15d-PGJ2-induced cell death [/fig_ref]. This suggested that 15d-PGJ2 inhibited Akt activation at the same levels of an Akt inhibitor and exerted cytotoxic effects in 786-O cells.
# Discussion
PPARγ is known to play many roles when activated by a ligand, such as synthetic antidiabetic thiazolidinediones and a natural eicosanoid derivate, 15d-PGJ2. These ligands, apart from their functions in normal tissues, have been reported to induce growth arrest or apoptosis in diverse tumor cells . In the present study, we evaluated the effects of 15d-PGJ2 on the viability of RCC cells and investigated the mechanisms of cytotoxicity.
First, we confirmed the cytotoxicity of 15d-PGJ2 in RCC cell lines as it was demonstrated in Caki-2 cells in the previous study [bib_ref] 15-deoxy-Δ 12,14 -prostaglandin J2 enhanced the anti-tumor activity of camptothecin against renal..., Yamamoto [/bib_ref]. 15d-PGJ2 exhibited dose-dependent cytotoxicity with IC50 values of 1.4-6.5 μM [fig_ref] Figure 1: Concentration-dependent cytotoxic effects of 15d-PGJ2 in 786-O, Caki-2, and ACHN cells [/fig_ref]. This cytotoxicity of 15d-PGJ2 accompanied chromatin condensation [fig_ref] Figure 2: Apoptosis assays for 15d-PGJ2 in three RCC cell lines [/fig_ref] , 2B) and caspase-3 activation [fig_ref] Figure 2: Apoptosis assays for 15d-PGJ2 in three RCC cell lines [/fig_ref]. In addition, a pan-caspase inhibitor, Z-VAD-FMK, significantly restored cell viability [fig_ref] Figure 2: Apoptosis assays for 15d-PGJ2 in three RCC cell lines [/fig_ref] , suggesting that 15d-PGJ2 induced cell death through caspase-dependent apoptosis to some extent in RCC cell lines; however, the recovery achieved by Z-VAD-FMK was incomplete. This suggested some involvement of other forms of death, such as caspase-independent apoptosis, autophagy, or necrosis. Indeed, a caspase-independent mechanism was reported to be partly involved in the 15d-PGJ2-induced apoptosis of hepatoma, chondrosarcoma, and choriocarcinoma cells [bib_ref] 15-deoxy-Δ-12-14 -PGJ2 regulates apoptosis induction and nuclear factor-κB activation via a peroxisome..., Okano [/bib_ref] [bib_ref] Inhibition of human chondrosarcoma cell growth via apoptosis by peroxisome proliferator-activated receptor-γ, Nishida [/bib_ref] [bib_ref] 15-Deoxy-Δ 12,14 -prostaglandin J2, a ligand for peroxisome proliferator-activated receptor-γ, induces apoptosis..., Keelan [/bib_ref]. Moreover, LDH leakage was observed in all three cell lines treated with 15d-PGJ2, albeit to a small extent [fig_ref] Figure 3: LDH leakage in RCC cell lines treated with 15d-PGJ2 [/fig_ref] , suggesting that necrosis was partly but not mainly involved in the cell death.
To investigate the molecular mechanisms of 15d-PGJ2-induced cell death, we first examined the involvement of PPARγ. 15d-PGJ2 has high affinity for human PPARγ [bib_ref] A prostaglandin J2 metabolite binds peroxisome proliferator-activated receptor γ and promotes adipocyte..., Kliewer [/bib_ref] , and is known to induce adipogenesis through PPARγ-activation, but it is unclear whether PPARγ is involved in the cytotoxic effect of 15d-PGJ2. In our results, GW9662, an irreversible PPARγ antagonist, did not improve cell viability at all , although the PPARγ protein is expressed in these three cell lines [bib_ref] Cytotoxicity of troglitazone through PPARγ-independent pathway and p38 MAPK pathway in renal..., Fujita [/bib_ref]. This suggested the cytotoxicity of 15d-PGJ2 in RCC cell lines to be independent of PPARγ, as supported by reports in other cell types [bib_ref] PPARγ inhibitors reduce tubulin protein levels by a PPARγ, PPARδ and proteasome-independent..., Schaefer [/bib_ref] [bib_ref] 15-Deoxy-Δ 12,14 -prostaglandin J2-induced apoptosis does not require PPARγ in breast cancer..., Clay [/bib_ref] [bib_ref] 15-deoxy-Δ-12-14 -PGJ2 regulates apoptosis induction and nuclear factor-κB activation via a peroxisome..., Okano [/bib_ref] [bib_ref] PPARγ-independent induction of growth arrest and apoptosis in prostate and bladder carcinoma, Chaffer [/bib_ref] [bib_ref] MAP kinase cascades are activated in astrocytes and preadipocytes by 15-deoxy-Δ 12-14..., Lennon [/bib_ref]. showed that GW9662 inhibited the cytotoxic effect of 15d-PGJ2 in neuroblastoma cells [bib_ref] Peroxisome proliferator-activated receptor-γ activator 15-deoxy-Δ 12,14 -prostaglandin J2 inhibits neuroblastoma cell growth..., Kim [/bib_ref]. Shen et al. showed that cell death was partly, but not completely, blocked by GW9662 in chondrosarcoma cells, suggesting that 15d-PGJ2 exerted its effect via both PPARγ-dependent and -independent pathways [bib_ref] Suppression of chondrosarcoma cells by 15-deoxy-Δ 12,14 -prostaglandin J2 is associated with..., Shen [/bib_ref]. Accordingly, the involvement of PPARγ in the cytotoxicity of 15d-PGJ2 seems to depend on the cell strain, and it is important to accumulate information for each type of cell as the genomic responses to PPARγ activation are complex [bib_ref] Peroxisome proliferator-activated receptor γ and cancers, Koeffler [/bib_ref].
15d-PGJ2 was also reported to generate ROS and to induce apoptosis in various kinds of cells [bib_ref] Inhibition of activator protein 1 activation, vascular endothelial growth factor, and cyclooxygenase-2..., Grau [/bib_ref] [bib_ref] 15d-PGJ2 induces apoptosis by reactive oxygen species-mediated inactivation of Akt in leukemia..., Shin [/bib_ref] [bib_ref] 15-deoxy-Δ 12,14 -prostaglandin J2 induces apoptosis of human hepatic myofibroblasts. A pathway..., Li [/bib_ref] [bib_ref] Dissociation of oxidant production by peroxisome proliferator-activated receptor ligands from cell death..., Atarod [/bib_ref] , so we next investigated the effects of antioxidants. Interestingly, our data showed that two antioxidants, NAC and GSH, rescued cells from cell death induced by 15d-PGJ2 , although the other antioxidants, vitamin E and melatonin, did not [fig_ref] Figure 6: Effects of antioxidants without a thiol-group, vitamin E [/fig_ref]. These findings suggested that NAC and GSH recovered the cell viability by the mechanisms other than the antioxidative effects, and this discrepancy among the antioxidants might be accounted for by their construction. As 15d-PGJ2 can reportedly form covalent adducts with thiol-containing biomolecules via Michael addition [bib_ref] 15-Deoxy-Δ 12,14 -prostaglandin J2 as a potential endogenous regulator of redox-sensitive transcription..., Kim [/bib_ref] , due to an electrophilic α,β-unsaturated carbonyl group in the cyclopentenone ring, NAC and GSH, possessing a thiol-group, were assumed to bind to 15d-PGJ2 and attenuate the effect of 15d-PGJ2. In addition, Paumi et al. showed that a GSH-conjugate of 15d-PGJ2 was effluxed from cells through multidrug resistance proteins, and the cytotoxic and transactivating effects of 15d-PGJ2 were attenuated in breast cancer cells [bib_ref] Multidrug resistance protein (MRP) 1 and MRP3 attenuate cytotoxic and transactivating effects..., Paumi [/bib_ref]. Accordingly, in this case as well, NAC, a precursor of GSH, was converted to GSH, and could reduce the intracellular concentration of 15d-PGJ2, thereby recovering the cell viability. Therefore, these findings indicated that ROS did not contribute to the apoptosis induced by 15d-PGJ2 in RCC cells.
Next, we focused on the cell signaling pathways involved in cell survival or proliferation. As some reports showed that 15d-PGJ2 also regulated MAPK activation [bib_ref] Inhibition of activator protein 1 activation, vascular endothelial growth factor, and cyclooxygenase-2..., Grau [/bib_ref] [bib_ref] 15d-PGJ2 induces apoptosis by reactive oxygen species-mediated inactivation of Akt in leukemia..., Shin [/bib_ref] [bib_ref] MAP kinase cascades are activated in astrocytes and preadipocytes by 15-deoxy-Δ 12-14..., Lennon [/bib_ref] [bib_ref] 15-deoxy-Δ 12,14 -prostaglandin J2 induces vascular endothelial cell apoptosis through the sequential..., Ho [/bib_ref] , we first investigated the involvement of the p38 and JNK MAPK pathways, known to be activated by various forms of stress, in the cytotoxic effects of 15d-PGJ2 on RCC cells. While a p38 MAPK inhibitor (SB202190) did not affect the cytotoxicity of 15d-PGJ2 in any of the cell lines [fig_ref] Figure 7: Involvement of the p38 and JNK MAPK pathway in 15d-PGJ2-induced cell death [/fig_ref] , a JNK inhibitor (SP600125) improved the viability of 15d-PGJ2-treated ACHN cells in a dose-dependent manner [fig_ref] Figure 7: Involvement of the p38 and JNK MAPK pathway in 15d-PGJ2-induced cell death [/fig_ref] , and the level of phosphorylated JNK tended to be increased after 8 h [fig_ref] Figure 7: Involvement of the p38 and JNK MAPK pathway in 15d-PGJ2-induced cell death [/fig_ref]. These findings suggested that the p38 pathway did not mainly participate in the 15d-PGJ2-elicited cell death, but that the JNK pathway was involved to some extent only in ACHN cells. The reason why the mechanisms of action differed among the three cell lines was unclear; but Okano et al. also reported that the actions of 15d-PGJ2 differed even among different cell lines derived from the organ [bib_ref] 15-deoxy-Δ-12-14 -PGJ2 regulates apoptosis induction and nuclear factor-κB activation via a peroxisome..., Okano [/bib_ref]. In addition, ACHN cells are derived from a metastatic site (pleural effusion), while 786-O and Caki-2 cells originate from a primary organ, and this might partly affect the cell-dependency.
We next investigated the involvement of Akt, which is known to be a downstream of phosphatidylinositol 3-kinase (PI3K) and have potent antiapoptotic and proliferative functions. 15d-PGJ2 markedly decreased the expression level of phosphorylated Akt after 1 and 24 h-exposure in 786-O cells [fig_ref] Figure 8: Involvement of the Akt pathway in 15d-PGJ2-induced cell death [/fig_ref]. This suggested that 15d-PGJ2 inactivated Akt and exerted cytotoxicity in 786-O cells, in which an Akt inhibitor also showed cytotoxicity with a low IC50 value [fig_ref] Figure 8: Involvement of the Akt pathway in 15d-PGJ2-induced cell death [/fig_ref]. The detail mechanisms of inactivating Akt must be further investigated, but, like other reports, the up-regulation of PTEN followed by the reduction of PI3K activity might be involved in this case [bib_ref] Peroxisome proliferator-activated receptor γ (PPARγ) ligands as bifunctional regulators of cell proliferation, Na [/bib_ref]. Therefore, the inactivation of Akt observed in this study might be an important mechanism of 15d-PGJ2 in some cell lines like 786-O cells.
In conclusion, the cytotoxicity of 15d-PGJ2 were mainly due to caspase apoptosis in three RCC cell lines but through different mechanisms in terms of JNK MAPK and Akt pathways. Our study showed the treatment with 15d-PGJ2 to potentially be an interesting approach for RCC though further experimental and clinical investigations are needed. 7-amido-4-methylcoumarin; GSH: Reduced glutathione; LDH: Lactate dehydrogenase; MAPK: Mitogen-activated protein kinase; NAC: N-acetyl-L-cysteine; PPARγ: Peroxisome proliferator-activated receptor gamma; RCC: Renal cell carcinoma; ROS: Generate reactive oxygen species; TBS: Tris-buffered saline; VHL: von Hippel-Lindau; Z-VAD-FMK:
N-benzyloxycarbonyl-Val-Ala-Asp fluoromethyl ketone.
[fig] Figure 1: Concentration-dependent cytotoxic effects of 15d-PGJ2 in 786-O, Caki-2, and ACHN cells. Cells were precultured for 24 h at 5 × 10 3 /well in 96-well plates and treated with 15d-PGJ2 for 24 h. Cell viability was assessed by fluorescent assay, and data represent the mean ± S.D. from 4 independent preparations. [/fig]
[fig] Figure 2: Apoptosis assays for 15d-PGJ2 in three RCC cell lines. (A, B) Induction of chromatin condensation by 15d-PGJ2. Typical fluorescence micrographs of Caki-2 cells stained with Hoechst 33342 (A) and percentages of chromatin-condensed cells (B). Cells were treated with 15d-PGJ2 for 24 h at the IC50 of each cell line and stained with Hoechst 33342 for 15 min at room temperature. They were then observed under a brightfield fluorescence microscope with UV excitation and the percentage of chromatin-condensed cells was determined. Data represent the mean ± S.D. from 4 independent preparations. Statistically significant difference from the control: **p < 0.01 assessed by t-test. (C) Effects of 15d-PGJ2 on caspase-3 activity. Cells were seeded in 96-well plates at 5 × 10 3 /well and cultured for 24 h. After exposure to 15d-PGJ2 at the IC50 of each cell line for 24 h, caspase-3 activity was assessed with a fluorimetric Caspase 3 Assay Kit according to the manufacturer's instructions. Enzymatic activities were determined as initial velocities corrected with the quantity of protein. Data represent the mean ± S.D. from 4 independent preparations. Statistically significant difference from the control: **p < 0.01 assessed by t-test. (D) [/fig]
[fig] Figure 3: LDH leakage in RCC cell lines treated with 15d-PGJ2. Cells were precultured for 24 h at 3 × 10 4 /well in 24-well plates and exposed to 15d-PGJ2 for 24 h. Absorbance at 570 nm in supernatant of the sample (S), negative control (N), and positive control (P) was used for calculation. LDH leakage was determined at follows; LDH leakage (% per positive control) = (S-N) / (P-N), and data represent the mean + S.D. from 3 independent preparations. The positive control was the supernatant of cells treated with 0.2% Tween 20 and the negative control was the supernatant of cells incubated with 0.1% DMSO. [/fig]
[fig] Figure 6: Effects of antioxidants without a thiol-group, vitamin E (A) and melatonin (B), on 15d-PGJ2-induced cell death. Cells were precultured for 24 h at 5× 10 3 /well in 96-well plates and exposed to 15d-PGJ2 in the presence or absence of vitamin E (1 μM) or melatonin (1 mM) for 24 h. Cell viability was assessed by fluorescent assay, and data represent the mean ± S.D. from 4 to 5 independent preparations. Statistical significance was assessed by t-test. [/fig]
[fig] Figure 7: Involvement of the p38 and JNK MAPK pathway in 15d-PGJ2-induced cell death. Effects of a p38 MAPK inhibitor, SB202190 (A), or JNK inhibitor, SP600125 (B, C), on 15d-PGJ2-induced cell death and the effects of 15d-PGJ2 on phosphorylation of JNK (D) [/fig]
[fig] Figure 8: Involvement of the Akt pathway in 15d-PGJ2-induced cell death. (A) To detect proteins, cells were precultured for 24 h in 100-mm dishes. Cells were then treated with 15d-PGJ2 at 3 μM for 1 and 24 h. Protein (15 μg) was analyzed by Western blotting for the expression of Akt and phospho-Akt. (B) [/fig]
|
An Updated Overview on Therapeutic Drug Monitoring of Recent Antiepileptic Drugs
Given the distinctive characteristics of both epilepsy and antiepileptic drugs (AEDs), therapeutic drug monitoring (TDM) can make a significant contribution to the field of epilepsy. The measurement and interpretation of serum drug concentrations can be of benefit in the treatment of uncontrollable seizures and in cases of clinical toxicity; it can aid in the individualization of therapy and in adjusting for variable or nonlinear pharmacokinetics; and can be useful in special populations such as pregnancy. This review examines the potential for TDM of newer AEDs such as eslicarbazepine acetate, felbamate, gabapentin, lacosamide, lamotrigine, levetiracetam, perampanel, pregabalin, rufinamide, retigabine, stiripentol, tiagabine, topiramate, vigabatrin, and zonisamide. We describe the relationships between serum drug concentration, clinical effect, and adverse drug reactions for each AED as well as the different analytical methods used for serum drug quantification. We discuss retrospective studies and prospective data on the serum drug concentration-efficacy of these drugs and present the pharmacokinetic parameters, oral bioavailability, reference concentration range, and active metabolites of newer AEDs. Limited data are available for recent AEDs, and we discuss the connection between drug concentrations in terms of clinical efficacy and nonresponse. Although we do not propose routine TDM, serum drug measurement can play a beneficial role in patient management and treatment individualization. Standardized studies designed to assess, in particular, concentration-efficacy-toxicity relationships for recent AEDs are urgently required.Key PointsSeizures occur sporadically, so antiepileptic drug therapy is generally experiential and prophylactic.Therapeutic drug monitoring can help establish an individual's optimal serum/plasma concentration range and benchmark the serum concentrations at which seizures are restrained or at which antiepileptic drug-specific adverse effects occur.
# Introduction
The fundamental objective of therapeutic drug monitoring (TDM) for antiepileptic drugs (AEDs) is the prevention of seizures and the minimization of negative effects on general well-being, including cognition, mood, and endocrine function. The International League Against Epilepsy (ILAE) determines seizure type on the basis of clinical outcomes and electroencephalograms. Around 34 AEDs have been prescribed to manage seizures over the last century. Many enzyme-inducing AEDs are cytochrome P450 (CYP) mixed function oxidase, glucuronyl transferase, or epoxy hydrolysis enzyme inducers. First-generation 'old' AEDs such as carbamazepine, phenobarbital, and phenytoin are enzyme inducers. [fig_ref] Table 1: Pharmacokinetic parameters of current antiepileptic medications [/fig_ref] lists the pharmacokinetic parameters of the newer or 'second-generation' AEDs (e.g., lamotrigine, gabapentin), the newest or 'third-generation' AEDs (e.g., retigabine, perampanel), and the newest orphan drugs (e.g., rufinamide, stiripentol).
## Factors influencing the selection of antiepileptic drugs
The selection of AEDs for a particular seizure type depends on drug-specific (e.g., adverse effects, toxicity, drug interactions), patient-specific (e.g., sex, age, use of contraception, genetics), and country-specific (e.g., availability, cost) variables [bib_ref] Harnessing the clinical potential of antiepileptic drug therapy: dosage optimisation, Perucca [/bib_ref].
## Therapeutic drug monitoring or target concentration interpretation
The aim of TDM is to contribute a 'reference concentration range' that laboratories can cite and clinicians can use as a benchmark. The 'therapeutic concentration range' of an individual patient is the range that achieves the best possible response and should be selected on the basis of symptoms and associated risks. The flaw of this strategy is that, in a few patients, optimal benefit will only be attained above minimum toxic concentrations, with associated risks of adverse reactions. Serum drug concentrations (SDC) achieved over months or years can provide invaluable information regarding the clinical scenario for each patient to enable interpretation of a change in response [bib_ref] Is there a role for therapeutic drug monitoring of new anticonvulsants?, Perucca [/bib_ref] [bib_ref] Pharmacokinetic variability of newer antiepileptic drugs: when is monitoring needed?, Johannessen [/bib_ref]. Serum samples collected within a few hours of a sudden seizure can provide information that may lead to clinical reasons or a definite cause of the seizure [bib_ref] Antiepileptic drugs as a cause of worsening seizures, Perucca [/bib_ref]. SDC measurement can distinguish between highly variable concentrations as a result of poor compliance and low concentrations because of erratic absorption, rapid metabolism, or drug interactions. The ability to discriminate between insufficient seizure control as a result of inadequate dosing or drug overload is invaluable [bib_ref] Antiepileptic drugs as a cause of worsening seizures, Perucca [/bib_ref]. Polypharmacy can also lead to complex intoxication despite concentrations being within the therapeutic window [bib_ref] Overtreatment in epilepsy: adverse consequences and mechanisms, Perucca [/bib_ref]. SDC is particularly relevant in pediatric and psychiatric patients in whom clinical assessment can be difficult. It is also useful for dosage adjustment in complex epilepsy that may need multiple drug therapy. The reference concentration range may not always be 'therapeutic', 'effective', or 'target', so the reporting method should be factual, with a statement such as 'the SDC lies between/ above/below the reference range'. Clinical judgment should not be based solely on SDC, because adverse drug reactions can occur even at low concentrations.
2 Newest Orphan Antiepileptic Drugs
## Rufinamide
Rufinamide is a well-tolerated and effective treatment for drug-resistant epilepsies and is also used as a second-line adjuvant for routine neurologic practice [bib_ref] Efficacy of rufinamide in drug-resistant epilepsy: a meta-analysis, Verrotti [/bib_ref]. A recently published clinical report [bib_ref] Successful treatment of super-refractory tonic status epilepticus with rufinamide: first clinical report, Thompson [/bib_ref] also indicated that rufinamide was successful in the treatment of super-refractory tonicstatus epilepticus. Rufinamide was granted orphan drug status by the European Medicines Agency (EMA) and the US FDA in 2004 for the treatment of Lennox-Gastaut syndrome (LGS). Mean serum drug concentration was found to be 10 ± 6.5 mg/l in children aged \12 years, 10.6 ± 7.0 mg/l in adolescents aged 12-17.9 years, and 15.9 ± 8.5 mg/l in adults aged C18 years [bib_ref] Rufinamide: clinical pharmacokinetics and concentration-response relationships in patients with epilepsy, Perucca [/bib_ref]. Approximately 85 % of the metabolites are eliminated via the kidneys, with an average half-life of 8 h. Metabolism is induced by enzyme inducers such as carbamazepine and rifampin [bib_ref] Rufinamide: clinical pharmacokinetics and concentration-response relationships in patients with epilepsy, Perucca [/bib_ref]. Impaired renal function does not affect clearance, so patients undergoing hemodialysis may require larger doses. Rufinamide exhibits dose-dependent gastrointestinal (GI) absorption with good bioavailability (&85 %) and high peak plasma drug concentration (C max ) especially when administered with food [bib_ref] Dose-dependent pharmacokinetics and brain penetration of rufinamide following intravenous and oral administration..., Gall [/bib_ref]. Therefore, patients should be adequately counselled to ensure they understand the need to follow the strict dosage regimen with respect to meals. Serum rufinamide concentration was high (&22 %) when co-administered with valproic acid (VPA). Reductions in seizure frequency have been found to be dose dependent. Mean concentration was 10 ± 6.5 mg/l in children aged \12 years, 10.6 ± 7.0 mg/l in those aged 12-17.9 years, and 15.9 ± 8.5 mg/l in patients aged C18 years [bib_ref] Pharmacokinetics of eslicarbazepine acetate in patients with moderate hepatic impairment, Almeida [/bib_ref]. Patients receiving a dose of 40 mg/ kg were found to have a serum concentration range of 0-45 lg/ml, indicating considerable pharmacokinetic variability. Therefore, TDM may be beneficial in clinical practice, albeit no reference range is available. Adverse reactions such as dizziness, fatigue, nausea, vomiting, diplopia, and somnolence have also been found to increase with concentration [bib_ref] Rufinamide: clinical pharmacokinetics and concentration-response relationships in patients with epilepsy, Perucca [/bib_ref]. Enzyme-inducing AEDs such as oxcarbazepine, and particularly methsuximide, have been demonstrated to decrease the serum concentration of rufinamide [bib_ref] Serum concentrations of rufinamide in children and adults with epilepsy:the influence of..., May [/bib_ref]. TDM for rufinamide correlates well with seizure control and therefore can be helpful in treatment individualization [bib_ref] Third-generation antiepileptic drugs: mechanisms of action, pharmacokinetics and interactions, Luszczki [/bib_ref]. Monitoring is useful for patients receiving enzyme inducers or those who are concurrently undergoing hemodialysis. Information regarding the pharmacokinetics of rufinamide in pregnancy and patients with hepatic impairment is inadequate. Drug interactions data are comparable to those for (PER) perampanel. A current report suggests a reference range of 126-168 lmol/l (30-40 mg/l) in LGS, which presumably would be lower for other seizure types [bib_ref] Successful treatment of super-refractory tonic status epilepticus with rufinamide: first clinical report, Thompson [/bib_ref]. High-performance liquid chromatography (HPLC) [bib_ref] Liquid chromatography-mass spectrometric determination of rufinamide in low volume plasma samples, Gall [/bib_ref] and liquid chromatography-mass spectrometry (LC-MS) [bib_ref] Rapid assay of rufinamide in dried blood spots by a new liquid..., La Marca [/bib_ref] can be used to determine plasma/serum concentrations.
## Stiripentol
Stiripentol (STP) has been granted orphan drug status by EMEA for PGTCS in 2001 and approved additionally for the treatment of severe myoclonic epilepsy in infancy (SMEI) or Dravet's syndrome [bib_ref] Stiripentol and vigabatrin current roles in the treatment of epilepsy, Chiron [/bib_ref] [bib_ref] Pharmacological considerations in the use of stiripentol for the treatment of epilepsy, Verrotti [/bib_ref]. Stiripentol has been reported to elevate brain c-aminobutyric acid (GABA) levels and to interfere with uptake and metabolism [bib_ref] A novel antiepileptic drug, Trojnar [/bib_ref]. It is bound strongly to plasma proteins (&99 %) with extensive hepatic metabolism and low bioavailability. Major metabolites are excreted renally.
The half-life of stiripentol increases with dose because of its dose-dependent pharmacokinetics. Stiripentol exhibits a significant reduction in clearance with dose escalation [bib_ref] Pharmacokinetics of stiripentol in normal man: evidence of nonlinearity, Levy [/bib_ref]. The reference serum concentration for stiripentol is not distinct, but a range of 4-22 mg/l may be associated with control of absence seizures in children, and a range of 8-12 mg/l may be associated with control of Dravet syndrome [bib_ref] Stiripentol in atypical absence seizures in children: an open trial, Farwell [/bib_ref]. The pharmacokinetics of stiripentol are complex as a result of non-linearity, strong protein binding, and considerable metabolism [bib_ref] Third-generation antiepileptic drugs: mechanisms of action, pharmacokinetics and interactions, Luszczki [/bib_ref]. Stiripentol inhibits many CYPs (CYP3A4, 1A2, 2C19) and interacts with many drugs, including AEDs. Stiripentol significantly decreases phenobarbital and phenytoin concentrations, whereas clobazam moderately increased stiripentol serum concentrations [bib_ref] Concentrations of stiripentol in children and adults with epilepsy: the influence of..., May [/bib_ref]. No data are available regarding the pharmacokinetics of stiripentol in pregnancy, or in patients with hepatic and renal impairment. TDM seems to be beneficial when using stiripentol because of the broad fluctuation in its concentration-dose ratio, age-dependent pharmacokinetics, non-linear relationship, extensive clearance, and drug-drug interactions [bib_ref] Therapeutic drug monitoring of stiripentol, Verdier [/bib_ref]. Quantification of plasma/serum stiripentol via HPLC has been reported [bib_ref] A stability-indicating HPLC-DAD method for determination of stiripentol: development, validation, kinetics, structure..., Darwish [/bib_ref].
## Newest antiepileptic drugs
## Perampanel
Perampanel is prescribed for refractory partial onset seizures, tonic-clonic seizures in patients aged [12 years, and PGTCS [bib_ref] Perampanel for tonic-clonic seizures in idiopathic generalized epilepsy. A randomized trial, French [/bib_ref]. Perampanel received regulatory approval from the FDA and the EMA in 2012. The minimum effective dose is 4 mg once daily, and larger doses provide a greater therapeutic effect with a comparable increase in adverse events [bib_ref] Perampanel serum concentrations in adults with epilepsy: effect of dose, age, gender..., Patsalos [/bib_ref]. Perampanel is not an enzyme inducer and acts as a selective non-competitive antagonist of ionotropic glutamate receptors. Oral absorption of the drug is delayed up to 2 h with food. The high protein binding (96 %) and volume of distribution (77 l/kg) may lead to displacement interactions with other AEDs. The drug has a long half-life of 70-110 h, and steady state may be achieved after 14 days. Its primary route of hepatic metabolism is via enzyme CYP3A4. The bulk of the dose is excreted in the feces and the rest in the urine, with \2 % excreted unchanged in the urine. Although the drug is well tolerated, the prevalence of adverse events rises with increasing dosages. Serious psychiatric and behavioral reactions such as dizziness, fatigue, irritability, obesity, vertigo, ataxia, gait disturbance, anxiety, blurred vision, dysarthria, asthenia, and hypersomnia have been reported [bib_ref] Adverse effects and safety profile of perampanel: a review of pooled data, Rugg-Gunn [/bib_ref]. A recent investigation demonstrated a linear dose-concentration relationship with serum perampanel concentrations, independent of age and sex [bib_ref] Perampanel serum concentrations in adults with epilepsy: effect of dose, age, sex,..., Patsalos [/bib_ref]. Carbamazepine and oxcarbazepine can significantly and dose-dependently reduce perampanel concentrations, presumably through CYP3A4-induced metabolism. Perampanel dosage should be monitored carefully during pregnancy and after childbirth, with adjustments made on a clinical basis. Anticonvulsant effects are possibly antagonized by antipsychotics and antimalarials such as mefloquine. Antifungals such as ketoconazole can elevate plasma concentrations, and anxiolytics and hypnotics such as midazolam can reduce them. Concomitant administration with orlistat, an obesity drug, may lead to an increased risk of convulsions [bib_ref] Orlistat and possible drug interactions that can affect over-the-counter sales, Baxter [/bib_ref]. No adequate well-controlled studies have investigated the pharmacokinetics of perampanel in pregnancy, possible drug interactions, or TDM, and a reference range has not yet been established. HPLC with fluorescence detection and HPLC-MS has been used to assay serum perampanel [bib_ref] High-performance liquid chromatography-tandem mass spectrometry method for the determination of perampanel, a..., Mano [/bib_ref].
## Retigabine
Retigabine (also known as ezogabine) is used as an adjunctive treatment for partial epilepsies in adults and was approved by both the EMA and the FDA in 2011 [bib_ref] Third-generation antiepileptic drugs: mechanisms of action, pharmacokinetics and interactions, Luszczki [/bib_ref] [bib_ref] Ezogabine (retigabine) and its role in the treatment of partial-onset seizures: a..., Splinter [/bib_ref]. The mechanism of action of retigabine is due to the activation of voltage-gated potassium channels in the brain [bib_ref] 2/Kv7.3-channel opener, attenuates drug-induced seizures in knock-in mice harboring Kcnq2 Mutations, Ihara [/bib_ref]. A placebo-controlled clinical trial has suggested that seizure frequency may be significantly reduced at higher doses [bib_ref] Retigabine: in partial seizures, Plosker [/bib_ref]. The most frequently used dose in an open-label add-on study was 600 mg per day [bib_ref] Efficacy and safety of ezogabine/retigabine as adjunctive therapy to specified single antiepileptic..., Lerche [/bib_ref]. Ezogabine/retigabine was demonstrated to be effective as adjunctive therapy to specified monotherapies such as carbamazepine/oxcarbazepine, lamotrigine, levetiracetam, and VPA using a flexible dosing regimen in adults with partial-onset seizures. Retigabine increases lamotrigine (LTG) metabolism, whereas carbamazepine and phenytoin enhance the clearance of retigabine [bib_ref] Third-generation antiepileptic drugs: mechanisms of action, pharmacokinetics and interactions, Luszczki [/bib_ref]. The drug and metabolites are excreted almost completely by the kidneys [bib_ref] Third-generation antiepileptic drugs: mechanisms of action, pharmacokinetics and interactions, Luszczki [/bib_ref]. Determination of retigabine and its acetyl metabolite via solid-phase extraction LC-MS has been reported [bib_ref] Determination of retigabine and its acetyl metabolite in biological matrices by on-line..., Knebel [/bib_ref]. A recent post-authorization safety study recognized the importance of TDM for adverse episodes of retinal pigmentation and alteration of vision in addition to established risks of urinary retention, central nervous system effects, and QTc prolongation [bib_ref] Survey of physicians' understanding of specific risks associated with retigabine, Daniluk [/bib_ref]. An observational study suggested that retigabine is useful in patients with treatmentrefractory seizures as a drug of reserve [bib_ref] Adjunctive retigabine in refractory focal epilepsy: postmarketing experience at four tertiary epilepsy..., Nass [/bib_ref]. Data are limited regarding the pharmacokinetics of retigabine in pregnancy, possible drug interactions, and TDM. Identification and quantification of retigabine via HPLC has been described [bib_ref] Identification, characterization, synthesis and HPLC quantification of new process-related impurities and degradation..., Douša [/bib_ref].
## Eslicarbazepine acetate
Eslicarbazepine acetate (ESL) is a prodrug that is structurally related to oxcarbazepine; it was approved by the FDA as a monotherapy and adjunct treatment for partial onset seizures in 2009 [bib_ref] Eslicarbazepine acetate monotherapy: a review in partial-onset seizures, Shirley [/bib_ref]. It was recently proposed that oxcarbazepine be switched to ESL with a dose ratio of 1-1.5:1 and that carbamazepine be switched to ESL with a dose ratio of 1-1.3:1 [bib_ref] A proposal for a model to replace carbamazepine or oxcarbazepine by eslicarbazepine..., Poza-Aldea [/bib_ref] [bib_ref] Overnight switching from oxcarbazepine to eslicarbazepine acetate: an observational study, Schmid [/bib_ref]. ESL is extensively metabolized to S-licarbazepine (95 %), which, similar to oxcarbazepine, inhibits voltage-gated sodium channels [bib_ref] Inhibition of glutamate release by BIA 2-093 and BIA 2-024, two novel..., Ambrosio [/bib_ref]. A linear relationship exists between dose and serum concentration. It also induces CYPs and increases clearance (12-16 %) of carbamazepine, lamotrigine, and topiramate. A stable dose-response relationship was recognized between ESL serum concentrations and reductions in seizure frequency that were not altered by other AEDs. ESL has been reported as having minimum drug-drug interactions. The pharmacokinetics of ESL are not influenced by enzyme-inducing AEDs or VPA, and ESL does not change the metabolism of lamotrigine [bib_ref] The impact of pharmacokinetic interactions with eslicarbazepine acetate versus oxcarbazepine and carbamazepine..., Johannessen Landmark [/bib_ref]. ESL metabolites are primarily eliminated renally, and dose adjustment is particularly necessary in patients with a clearance rate of \ 60 ml/min. Hemodialysis efficiently removes ESL and its metabolites from serum [bib_ref] Effect of renal impairment on the pharmacokinetics of eslicarbazepine acetate, Maia [/bib_ref]. Moderate hepatic impairment has limited clinical effects [bib_ref] Pharmacokinetics of eslicarbazepine acetate in patients with moderate hepatic impairment, Almeida [/bib_ref]. This was endorsed in a single-dose study of ESL 800 mg once daily over a period of 1 week, which demonstrated no change in pharmacokinetic parameters [bib_ref] Single-dose and steady-state pharmacokinetics of eslicarbazepine acetate (BIA 2-093) in healthy elderly..., Almeida [/bib_ref]. Concomitant administration of drugs that prolong the PR interval should be avoided. ESL adversely interacts with oral contraceptives. There is a possible risk of developing Stevens-Johnson syndrome in the presence of human leukocyte antigen (HLA)-B 1502 allele. No clinically relevant data are available with respect to pregnancy, and no data support the usefulness of TDM in ESL although it is expected to be similar to oxcarbazepine. An enantioselective HPLC-UV detector [bib_ref] Enantioselective assay for therapeutic drug monitoring of eslicarbazepine acetate: no interference with..., Alves [/bib_ref] and LC [bib_ref] Liquid chromatography separation of the chiral prodrug eslicarbazepine acetate and its main..., Servais [/bib_ref] have been used to analyze ESL and its metabolites.
## Vigabatrin
Vigabatrin has been indicated as adjunct therapy for adults and children aged [10 years with refractory complex partial seizures and to control infantile spasms. Vigabatrin was granted initial approval from the FDA in 2009 and received orphan drug status from the EMA in 2000. Vigabatrin is currently available in more than 50 countries. Its mechanism of action is based on the irreversible inhibition of GABA-transaminase (GABA-T), thus elevating the concentration of GABA in the brain [bib_ref] Clinical pharmacokinetics, Rey [/bib_ref]. Vigabatrin S (?) isomer is pharmacologically active and associated with irreversible visual field defects in 44 % of patients with epilepsy, which has resulted in prescription limitations [bib_ref] Clinical pharmacokinetics, Rey [/bib_ref] [bib_ref] Prevalence of visual field loss following exposure to vigabatrin therapy: a systematic..., Maguire [/bib_ref]. Clearance in children is greater than in adults, and therefore higher doses are required to attain comparable serum concentrations [bib_ref] Determination of vigabatrin by capillary electrophoresis with laser-induced fluorescence detection, Chang [/bib_ref]. As the drug is eliminated renally, toxicity might occur in patients with renal impairment; therefore dosage adjustment is required [bib_ref] Clinical pharmacokinetics, Rey [/bib_ref]. Pharmacokinetic interactions are minimal since vigabatrin is neither metabolized nor protein bound. Due to irreversible inhibition of GABA-T, there is no rationale for TDM of vigabatrin, although it may be helpful in assessing compliance [bib_ref] New antiepileptic drugs, Patsalos [/bib_ref]. Serum GABA and GABA-T concentrations are not an indicator/marker for clinical response because the blood-brain barrier is relatively impermeable to GABA, and serum concentrations do not necessarily reflect cerebral spinal fluid concentrations [bib_ref] Changes in plasma GABA concentration during vigabatrin treatment of epilepsy: a prospective..., Erdal [/bib_ref]. Information about the pharmacokinetics of vigabatrin in pregnancy is inadequate. The use of nuclear magnetic resonance (NMR) spectroscopy to analyze brain GABA concentrations with vigabatrin treatment has been described [bib_ref] Effects of vigabatrin on brain GABA?/CR signals in patients with epilepsy monitored..., Mueller [/bib_ref]. Compliance can be verified by checking the anticipated trough serum vigabatrin target range (6-278 mol/l) at a dose level of 1000-3000 mg/day [bib_ref] Serum concentrations and effects of gabapentin and vigabatrin: observations from a dose..., Lindberger [/bib_ref]. Serum vigabatrin has been quantified with HPLC, LC-MS [bib_ref] Vigabatrin in dried plasma spots: validation of a novel LC-MS/MS method and..., Kostic [/bib_ref] , and gas chromatography (GC)-MS [bib_ref] Determination of vigabatrin by capillary electrophoresis with laser-induced fluorescence detection, Chang [/bib_ref] [bib_ref] Quantitative determination of vigabatrin and gabapentin in human serum by gas chromatography-mass..., Borrey [/bib_ref]. Simultaneous HPLC analysis of vigabatrin, gabapentin, pregabalin, and topiramate has been reported as useful for monitoring polytherapy regimens including these drugs [bib_ref] Fast isocratic high-performance liquid chromatographic assay method for the simultaneous determination of..., Chollet [/bib_ref] [bib_ref] Simultaneous determination of gabapentin, pregabalin, vigabatrin, and topiramate in plasma by HPLC..., Martinc [/bib_ref].
## Lacosamide
Lacosamide has been approved as an adjunctive treatment for partial-onset seizures and focal epilepsies in adult patients [bib_ref] Lacosamide: new adjunctive treatment option for partial-onset seizures, Chung [/bib_ref]. It was approved by both the FDA and the EMA in 2008. It enhances the inactivation of voltage-gated sodium channels, leading to the stabilization of hyper-excitable neuronal membranes [bib_ref] Pharmacotherapy of the third-generation AEDs: lacosamide, retigabine and eslicarbazepine acetate, Patsalos [/bib_ref]. Lacosamide has been proposed for the treatment of neuropathic pain of various etiologies in patients who do not respond to standard treatments [bib_ref] Lacosamide and neuropathic pain, a review, Alcantara-Montero [/bib_ref]. Lacosamide has a high level of oral bioavailability (&100 %) and a short time to C max (t max ), is unaffected by food, and has linear pharmacokinetics, which are all advantages in the clinical use of lacosamide [bib_ref] Pharmacotherapy of the third-generation AEDs: lacosamide, retigabine and eslicarbazepine acetate, Patsalos [/bib_ref]. The presence of high free drug concentrations of lacosamide led to the hypothesis that the drug should be effectively eliminated by hemodialysis, although published data to support this are limited [bib_ref] Optimizing therapy of seizures in patients with renal or hepatic dysfunction, Lacerda [/bib_ref]. It is primarily metabolized via the hepatic route by demethylation to an inactive O-desmethyl metabolite by CYP2C19 (30 %) [bib_ref] No pharmacokinetic interaction between lacosamide and carbamazepine in healthy volunteers, Cawello [/bib_ref]. Approximately 40 % of lacosamide is eliminated unchanged via renal excretion mechanisms. Population pharmacokinetic data from phase II clinical studies of lacosamide have shown that lamotrigine, levetiracetam, topiramate, oxcarbazepine, and VPA have no significant effects on the pharmacokinetics of lacosamide [bib_ref] 640 Low potential for drug-drug-interaction of lacosamide, Thomas [/bib_ref]. Enzyme inducers such as carbamazepine and phenytoin can significantly reduce plasma lacosamide concentrations and thereby enhance clearance [bib_ref] Lacosamide therapeutic monitoring in patients with epilepsy: effect of concomitant antiepileptic drugs, Contin [/bib_ref] [bib_ref] Pharmacodynamic and pharmacokinetic evaluation of coadministration of lacosamide and an oral contraceptive..., Cawello [/bib_ref]. Although data are limited, overall, lacosamide does not affect the pharmacokinetics of many AEDs [bib_ref] Lacosamide: new adjunctive treatment option for partial-onset seizures, Chung [/bib_ref]. Frequent adverse effects include diplopia, nausea, vomiting, dizziness, abnormal coordination, and blurred vision. Daytime fluctuations can be lowered by increasing the frequency of administration from twice daily to three times daily [bib_ref] Fluctuationof lacosamide serum concentrations during the day and occurrence of adverse drug..., Sattlera [/bib_ref]. Lacosamide serum levels are elevated in patients with hepatic or renal impairment. Dose-dependent escalation of serum lacosamide concentrations have been reported to be independent of age and higher in women than in men [bib_ref] Lacosamide serum concentrations in adult patients with epilepsy: the influence of gender,..., Markoula [/bib_ref]. No clinically relevant information is available regarding serum lacosamide concentrations in pregnancy. A rapid sensitive HPLC and GC-MS method for the measurement of lacosamide and desmethyl lacosamide has been reported [bib_ref] A sensitive and rapid method for quantification of lacosamide and desmethyl lacosamide..., Payto [/bib_ref] [bib_ref] A fully validated method for the determination of lacosamide in human plasma..., Nikolaou [/bib_ref].
## Newer antiepileptic drugs
## Pregabalin
Pregabalin is a short-acting (elimination half-life [t ] 5-7 h) analogue of gabapentin used as an adjunct treatment for focal epilepsies, neuropathic pain, fibromyalgia, and anxiety disorder in adult patients [bib_ref] Pregabalin (Pfizer), Selak [/bib_ref] [bib_ref] The safety and efficacy of pregabalin for treating subjects with fibromyalgia and..., Clair [/bib_ref]. Initial approval was granted by the FDA and EMA in 2004. Although pregabalin is patent protected in the USA until 2018, a generic version is currently available in the UK and Canada. Pregabalin has been investigated for use in cancerinduced bone pain [bib_ref] Does pregabalin still have a role in treating cancer-induced bone pain?, Raman [/bib_ref] and as a premedication to provide postoperative analgesia [bib_ref] Efficacy of pregabalin as premedication for post-operative analgesia in vaginal hysterectomy, Rajappa [/bib_ref] [bib_ref] Effect of oral pregabalin as preemptive analgesic in patients undergoing lower limb..., Sebastian [/bib_ref]. Pregabalin can be an effective treatment for peripheral neuropathic pain provided adverse reactions are carefully monitored and dosages are correctly determined [bib_ref] Efficacy and safety of pregabalin for the treatment of neuropathic pain in..., Otsuki [/bib_ref]. It demonstrates good bioavailability (&98 %), minimal drug interactions, and low protein binding [bib_ref] Pregabalin pharmacology and its relevance to clinical practice, Ben-Menachem [/bib_ref]. The poor binding of pregabalin means it can be effectively cleared by hemodialysis techniques [bib_ref] Pharmacokinetics of pregabalin in subjects with various degrees of renal function, Randinitis [/bib_ref]. A major portion of the drug is excreted unchanged in the urine, with a clearance equal to the glomerular filtration rate. Thus, the dosage regimen should be adjusted in geriatric patients as well as those with renal impairment [bib_ref] Pharmacological treatment of neuropathic pain in older persons, Haslam [/bib_ref]. TDM of pregabalin is restricted to dosage adjustment and to assess compliance. Although no specific reference concentration range has been established, a comparative value of 3-8 mg/l has been suggested [bib_ref] Antiepileptic drugs-best practice guidelines for therapeutic drug monitoring: a position paper by..., Patsalos [/bib_ref]. Pregabalin trough concentrations are subject to high intersubject variability [bib_ref] Serum concentrations of pregabalin in patients with epilepsy: the influence of dose,..., May [/bib_ref]. Significant adverse drug reactions are sedation, obesity, cerebellar symptoms, and peripheral edema. Enzyme-inducing AEDs can moderately lower pregabalin concentrations [bib_ref] Pharmacological treatment of neuropathic pain in older persons, Haslam [/bib_ref]. An increased risk of major birth defects after first-trimester exposure to pregabalin has been recently documented [bib_ref] Pregnancy outcome following maternal exposure to pregabalin may call for concern, Winterfeld [/bib_ref]. An initial clinical study has demonstrated the use of pregabalin as an adjuvant to labor pain relief associated with late termination of pregnancy [bib_ref] Evaluation of pregabalin as an adjuvant to patient-controlled epidural analgesia during late..., Lavand'homme [/bib_ref]. Plasma drug concentrations can be determined via HPLC [bib_ref] Analysis of pregabalin at therapeutic concentrations in human plasma/serum by reversed-phase HPLC, Berry [/bib_ref] , LC-MS/MS [bib_ref] Analysis of pregabalin at therapeutic concentrations in human plasma/serum by reversed-phase HPLC, Berry [/bib_ref] [bib_ref] Quantitation of pregabalin in dried blood spots and dried plasma spots by..., Kostic [/bib_ref] , GC-MS [bib_ref] Gabapentin, pregabalin and vigabatrin quantification in human serum by GC-MS after hexyl..., Hlozek [/bib_ref] , or a fluorometric method [bib_ref] A validated fluorometric method for the rapid determination of pregabalin in human..., Yoshikawa [/bib_ref].
## Zonisamide
Zonisamide is used as an adjunct treatment for partial seizure epilepsy and off label for bipolar disorder [bib_ref] Zonisamide for bipolar disorder, mania or mixed states: a randomized, double blind,..., Dauphinais [/bib_ref] , chronic pain, migraine [bib_ref] Zonisamide for migraine prophylaxis in refractory patients, Ashkenazi [/bib_ref] , and myoclonic dystonia [bib_ref] A randomized, controlled, double-blind, crossover trial of zonisamide in myoclonus-dystonia, Hainque [/bib_ref]. It was initially approved by the FDA in 2000, and the EMA granted a marketing authorization valid throughout the EU in 2005. It exhibits low capacity strong binding to red cell components and carbonic anhydrase enzymes. The dual mechanism of action is due to weak inhibition of enzymes and modulation of GABAergic and glutamatergic neurotransmission via alteration of voltage sensitive sodium and calcium channels . Zonisamide exhibits linear pharmacokinetics with metabolism via phase I and phase II biotransformation pathways to produce inactive metabolites [bib_ref] The clinical pharmacokinetics of the newer antiepileptic drugs. Focus on topiramate, zonisamide..., Perucca [/bib_ref]. The t decreased from 60 h to 30 h in patients concomitantly receiving enzyme inducers. However, enzyme inhibitors such as cimetidine, erythromycin, ketoconazole, and VPA can extend the half-life of zonisamide [bib_ref] The clinical pharmacokinetics of the newer antiepileptic drugs. Focus on topiramate, zonisamide..., Perucca [/bib_ref]. Inter-individual variability in patients who are also receiving CYP enzyme inducers or inhibitors should be monitored. Zonisamide is cleared efficiently from serum via hemodialysis [bib_ref] Dialyzability of the antiepileptic drug zonisamide in patients undergoing hemodialysis, Ijiri [/bib_ref]. In general, children need larger doses than adults, calculated on the basis of weight [bib_ref] Clinical pharmacokinetics of new-generation antiepileptic drugs at the extremes of age, Perucca [/bib_ref]. A serum/plasma target range of 10-40 mg/l has been suggested for seizure management [bib_ref] Clinical pharmacology and therapeutic drug monitoring of zonisamide, Mimaki [/bib_ref]. Many analytical methods to assay zonisamide in biological fluids have been reported, including HPLC [bib_ref] Development and validation of an analytical method based on high performance thin..., Antonilli [/bib_ref] , LC-MS [bib_ref] High-speed simultaneous determination of nine antiepileptic drugs using liquid chromatography-mass spectrometry, Subramanian [/bib_ref] , and solidphase extraction techniques [bib_ref] Monitoring of zonisamide in human breast milk and maternal plasma by solidphase..., Shimoyama [/bib_ref].
## Levetiracetam
Levetiracetam is an enantioselective (S) isomer used for focal epilepsies in adult patients as monotherapy and as adjunctive treatment in children. It is also used for the management of juvenile myoclonic and generalized tonicclonic seizures in patients aged C12 years [bib_ref] Population pharmacokinetic modeling of levetiracetam in pediatric and adult patients with epilepsy..., Ito [/bib_ref]. Levetiracetam obtained initial approval from the FDA in 1999. It may be effective for prophylaxis of early traumatic brain injury [bib_ref] Low-dose levetiracetam for seizure prophylaxis after traumatic brain injury, Patanwala [/bib_ref]. Its t is increased in patients with renal impairment; therefore dosage adjustments will be necessary. The clearance value of the drug is greater in children than in adults [bib_ref] Pharmacokinetic variability of four newer antiepileptic drugs, lamotrigine, levetiracetam, oxcarbazepine, and topiramate:..., Johannessen Landmark [/bib_ref]. Correlation between efficacy and serum levels of levetiracetam has been demonstrated in children with refractory epilepsy [bib_ref] Correlation between efficacy of levetiracetam and serum levels among children with refractory..., Sheinberg [/bib_ref]. Levetiracetam has all the optimal characteristics of an AED and follows linear pharmacokinetics [bib_ref] Pharmacokinetic profile of levetiracetam: toward ideal characteristics, Patsalos [/bib_ref]. It undergoes minimal metabolism, meaning drug interactions with enzyme inducers or inhibitors are unlikely. SDC monitoring is not yet established for levetiracetam by virtue of its wide therapeutic index and minimal side effects. For a daily dosage of 1-3 g, a target reference range of 12-46 mg/l is suggested for monitoring for compliance, drug overdose, and dose adjustment. A reduction in SDC of %60 % has been reported in pregnancy [bib_ref] Pharmacokinetics of levetiracetam during pregnancy, delivery, in the neonatal period, and lactation, Tomson [/bib_ref]. A 12-year comparative study of AED use in pregnancy indicated that levetiracetam was more effective and better tolerated than older AEDs and lamotrigine [bib_ref] Comparative study of antiepileptic drug use during pregnancy over a period of..., Martinez Ferri [/bib_ref]. Levetiracetam can be measured in biological fluids using either HPLC [bib_ref] High-performance liquid chromatographic determination of Levetiracetam in human plasma: comparison of different..., Pucci [/bib_ref] , GC [bib_ref] High-performance liquid chromatographic and megabore gas-liquid chromatographic determination of levetiracetam (ucb L059)..., Vermeij [/bib_ref] , or GC/ MS [bib_ref] A fully validated method for the determination of lacosamide in human plasma..., Nikolaou [/bib_ref].
## Tiagabine
Tiagabine is used as add-on treatment for focal epilepsies not responding to other AEDs [bib_ref] A single-dose study to define tiagabine pharmacokinetics in pediatric patients with complex..., Gustavson [/bib_ref] and was approved by the FDA in 1997. Tiagabine follows linear pharmacokinetics with extensive oxidative hepatic metabolism [bib_ref] Tiagabine for complex partial seizures: a randomized, add-on, dose-response trial, Uthman [/bib_ref]. The half-life of tiagabine decreases from 7 h to 3 h in patients receiving concomitant enzyme inducers [bib_ref] The pharmacokinetic profile of tiagabine, Wang [/bib_ref]. Higher clearance was observed in children than in adults [bib_ref] A single-dose study to define tiagabine pharmacokinetics in pediatric patients with complex..., Gustavson [/bib_ref]. Metabolism is gradual in patients with hepatic impairment; consequently, the half-life in these patients is &14 h [bib_ref] Pharmacokinetics and safety of tiagabine in subjects with various degrees of hepatic..., Lau [/bib_ref]. Therefore, dose individualization may be necessary in hepatic dysfunction but not in renal dysfunction [bib_ref] Pharmacokinetics and safety of tiagabine in subjects with various degrees of hepatic..., Lau [/bib_ref]. A placebo-controlled clinical trial demonstrated that the frequency of complex partial seizures was concentration dependent [bib_ref] Distribution of topiramate in a medical examiner's case, Mozayani [/bib_ref]. Accordingly, a broad reference range of 20-200 ng/ml (53-532 nmol/l) has been proposed [bib_ref] Sensitive analytical method for Topiramate in human serum by HPLC with pre-column..., Bahrami [/bib_ref]. It has been demonstrated that tiagabine may induce generalized non-convulsive status epilepticus (NCSE) in patients with focal lesional epilepsy in addition to those with generalized syndromes [bib_ref] Tiagabine-induced generalised non convulsive status epilepticus in patients with lesional focal epilepsy, Vinton [/bib_ref]. Data regarding the pharmacokinetics of tiagabine during pregnancy are limited. TDM may be helpful in cases of noncompliance and toxicity, provided assay techniques are sensitive and robust [bib_ref] Interlaboratory variability in the quantification of new generation antiepileptic drugs based on..., Williams [/bib_ref]. Strong protein binding and high inter-subject variability due to hepatic metabolism makes TDM necessary. Ideally, the serum sample for TDM should be taken before the morning dose because the half-life of the drug is short. Tiagabine in serum can be quantified with HPLC [bib_ref] High-performance liquid chromatographic procedure for the determination of tiagabine concentrations in human..., Gustavson [/bib_ref] , LC/MS [bib_ref] The pharmacokinetic interrelationship of tiagabine in blood, cerebrospinal fluid and brain extracellular..., Wang [/bib_ref] , and GC/MS [bib_ref] The pharmacokinetic interrelationship of tiagabine in blood, cerebrospinal fluid and brain extracellular..., Wang [/bib_ref] [bib_ref] Gas chromatography-mass spectrometry assay method for the therapeutic drug monitoring of the..., Chollet [/bib_ref].
## Topiramate
Topiramate is licensed for the treatment of epilepsies in adults and children, for adjunct treatment in polytherapy,
LGS, and migraine prophylaxis [bib_ref] Neuroprotective effects of various doses of topiramate against methylphenidate induced oxidative stress..., Motaghinejad [/bib_ref] [bib_ref] The new antiepileptic drugs: clinical applications, Laroche [/bib_ref]. The FDA initially approved topiramate in 1996 and endorsed it for migration prophylaxis in 2003. The mechanism of action includes sodium channel blockade, GABAergic, anti-glutamergic, and carbonic anhydrase inhibition. Increased clearance of the drug was found in children aged \10years and in the presence of enzyme inducers [bib_ref] Clinical pharmacokinetics of new-generation antiepileptic drugs at the extremes of age, Perucca [/bib_ref] [bib_ref] Influence of dosage, age, and co-medication on plasma topiramate concentrations in children..., Ferrari [/bib_ref]. The half-life of topiramate is 25 h, which reduces to 12 h in patients receiving concomitant enzyme-inducing AEDs [bib_ref] Comparison of lacosamide concentrations in cerebrospinal fluid and serum in patients with..., May [/bib_ref]. Topiramate clearance is decreased by lithium, propranolol, amitriptyline, and sumatriptan, thereby elevating serum topiramate concentrations [bib_ref] Pharmacotherapy of the third-generation AEDs: lacosamide, retigabine and eslicarbazepine acetate, Patsalos [/bib_ref]. Food can delay oral absorption of the drug (&2 h), but serum topiramate concentrations are unaffected [bib_ref] Single-dose pharmacokinetics and effect of food on the bioavailability of topiramate, a..., Doose [/bib_ref]. Average drug concentrations did not differ significantly between responders, nonresponders, and patients with adverse reactions [bib_ref] Topiramate: a prospective study on the relationship between concentration, dosage and adverse..., Froscher [/bib_ref] , but did decrease during pregnancy [bib_ref] Pharmacokinetics of topiramate during pregnancy, Ohman [/bib_ref]. A higher topiramate concentration of [10 mg/l [bib_ref] Clinical experience with topiramate: correlation of serum levels with efficacy and adverse..., Penovich [/bib_ref] and according to another study an average serum concentration of 7 mg/l has been reported to result in marked improvements in seizure frequency [bib_ref] Prognostic factors affecting long-term retention of topiramate in patients with chronic epilepsy, Lhatoo [/bib_ref]. A topiramate concentration of 2.4-8 mg/l was reported with dose ranges between 125 and 400 mg in combination with other AEDs [bib_ref] Concentrations of stiripentol in children and adults with epilepsy: the influence of..., May [/bib_ref]. Serum concentrations between 5 and 20 mg/l have also been reported with therapeutic doses [bib_ref] Therapeutic drug monitoring of the newer antiepileptic drugs, Johannessen [/bib_ref]. The available literature indicates routine TDM is only necessary in patients with hepatic and renal impairment; it may also be used to differentiate between non-convulsive epilepsy and drug intoxication [bib_ref] Topiramate overdose: a case report of a patient with extremely high topiramate..., Brandt [/bib_ref]. Salivary drug concentrations can be considered an alternative TDM method in patients receiving chronic therapy [bib_ref] Topiramate concentration in saliva: an alternative to serum monitoring, Miles [/bib_ref]. Fluorescence polarization immunoassay (FPIA) is commercially available to estimate serum concentrations [bib_ref] Comparison of topiramate concentrations in plasma and serum by fluorescence polarization immunoassay, Berry [/bib_ref]. Many analytical methods are available for estimating serum topiramate concentrations, including GC [bib_ref] An improved gas chromatography assay for topiramate monitoring in pediatric patients, Tang [/bib_ref] , HPLC [bib_ref] Sensitive analytical method for Topiramate in human serum by HPLC with pre-column..., Bahrami [/bib_ref] , LC/MS [bib_ref] Analysis of topiramate and its metabolites in plasma and urine of healthy..., Britzi [/bib_ref] , and immunoassay [bib_ref] Comparison of a new serum topiramate immunoassay to fluorescence polarization immunoassay, Snozek [/bib_ref]. Simultaneous HPLC-MS/MS analysis of topiramate, zonisamide, lamotrigine, carbamazepine, and levetiracetam in serum has been discussed [bib_ref] Simultaneous quantitation of lamotrigine, levetiracetam, 10-hydroxycarbazepine, topiramate, and zonisamide in serum using..., Carlow [/bib_ref].
## Lamotrigine
Lamotrigine is effective as monotherapy or adjunctive therapy in partial, generalized tonic-clonic, and absence seizures [bib_ref] Lamotrigine monotherapy for newly diagnosed typical absence seizures in children: a multi-center,..., Yasumoto [/bib_ref] , and in LGS [bib_ref] Lamotrigine as monotherapy in clinical practice: efficacy of various dosages in epilepsy, Warshavsky [/bib_ref]. It has also been investigated for off-label use in treatment-resistant depressive disorders [bib_ref] Prediction of an optimal dose of lamotrigine for augmentation therapy in treatment-resistant..., Nakamura [/bib_ref]. Lamotrigine received initial FDA approval in 1994 and approval for bipolar disorder in 2003. Estradiol-containing oral contraceptives and pregnancy can drastically decrease serum lamotrigine concentrations [bib_ref] Oral contraceptives induce lamotrigine metabolism: evidence from a double-blind, placebo-controlled trial, Christensen [/bib_ref] ; however, this effect was not observed in women also receiving VPA [bib_ref] Valproate effects on kinetics of lamotrigine in pregnancy and treatment with oral..., Tomson [/bib_ref]. Transient declines in lamotrigine serum concentrations has also been observed after epilepsy surgery; thus, TDM will be helpful in the prevention of early postoperative seizures [bib_ref] Perioperative fluctuations of lamotrigine serum levels in patients undergoing epilepsy surgery, Paul [/bib_ref]. TDM appears beneficial both for the finalization of individual reference ranges and for the optimization of individual dosage regimens [bib_ref] Lamotrigine therapeutic thresholds, Sondergaard Khinchi [/bib_ref]. TDM and dose adjustments may be necessary in patients undergoing hemodialysis or in those with severe liver impairment [bib_ref] Lamotrigine and hemodialysis in bipolar disorder: case analysis of dosing strategy with..., Kaufman [/bib_ref] [bib_ref] Influence of cirrhosis on lamotrigine pharmacokinetics, Marcellin [/bib_ref]. Lamotrigine should be introduced with slow dose titrations because of the possibility of severe dermatological reactions, including Stevens-Johnson syndrome and toxic epidermal necrolysis. Enzymeinducing AEDs can halve the t of lamotrigine , resulting in decreased lamotrigine concentrations. The halflife of lamotrigine increased to 60 h when combined with VPA [bib_ref] Pharmacokinetics, toxicology and safety of lamotrigine in epilepsy, Biton [/bib_ref]. In polytherapy regimens with VPA, its inhibitory effect counteracts the inducing effect of enzymeinducing AEDs [bib_ref] An update of its pharmacology and therapeutic use in epilepsy, Fitton [/bib_ref]. Co-medication with methsuximide lowers lamotrigine concentration by 70 % [bib_ref] Influence of oxcarbazepine and methsuximide on lamotrigine concentrations in epileptic patients with..., May [/bib_ref]. Lamotrigine serum concentrations have been shown to increase by 90 % from pre-pregnancy baseline to third-trimester [bib_ref] Antiepileptic drug therapy in pregnancy I: gestation-induced effects on AED pharmacokinetics, Pennell [/bib_ref] ; however, this was not observed in patients receiving concomitant VPA [bib_ref] Influence of oxcarbazepine and methsuximide on lamotrigine concentrations in epileptic patients with..., May [/bib_ref]. Therefore, TDM is important in the prevention of seizure episodes during pregnancy, and dose adjustments are based on trough serum concentrations at least once a month [bib_ref] Influence of oxcarbazepine and methsuximide on lamotrigine concentrations in epileptic patients with..., May [/bib_ref]. Trans-placental transfer of lamotrigine in maternal blood, amniotic fluid, and cord blood has been measured and correlated [bib_ref] Lamotrigine in pregnancy-therapeutic drug monitoring in maternal blood, amniotic fluid, and cord..., Paulzen [/bib_ref]. Lamotrigine clearance has been reported to be higher in children and the elderly [bib_ref] Clinical pharmacokinetics of new-generation antiepileptic drugs at the extremes of age, Perucca [/bib_ref]. TDM is valuable for dosage adjustments in pregnancy, and to check inter-individual variability and drug interactions. In general, reference concentrations of 2.5-15 mg/l have been suggested in patients receiving therapeutic doses [bib_ref] Is there a role for therapeutic drug monitoring of new anticonvulsants?, Perucca [/bib_ref] [bib_ref] Pharmacokinetic variability of newer antiepileptic drugs: when is monitoring needed?, Johannessen [/bib_ref]. The prevalence of toxicity increases significantly with concentrations [15 mg/l [bib_ref] Lamotrigine therapeutic thresholds, Sondergaard Khinchi [/bib_ref]. TDM of lamotrigine may be further complicated by pharmacodynamic interactions as demonstrated for co-medication with carbamazepine [bib_ref] Clinical pharmacokinetics of new-generation antiepileptic drugs at the extremes of age, Perucca [/bib_ref]. Assay of lamotrigine based on HPLC [bib_ref] New highperformance liquid chromatographic method for plasma/serum analysis of lamotrigine, Croci [/bib_ref] [bib_ref] A rapid and simple assay for lamotrigine in serum/plasma by HPLC, and..., Morgan [/bib_ref] , turbidimetric immunoassay [bib_ref] Seradyn quantitative microsphere system lamotrigine immunoassay on a Hitachi 911 analyzer compared..., Westley [/bib_ref] , GC-MS [bib_ref] Lamotrigine analysis in plasma by gas chromatography-mass spectrometry after conversion to a..., Dasgupta [/bib_ref] , and GC and densitometric methods [bib_ref] Identification and assay of lamotrigine in human milk with gas chromatography and..., Wyszomirska [/bib_ref] have been reported.
## Felbamate
Felbamate is used as add-on therapy in LGS in patients aged [4 years [bib_ref] Treatment of Lennox-Gastaut syndrome: overview and recent findings, Van Rijckevorsel [/bib_ref]. It was approved by the FDA in 1993. The half-life decreases from &19 to &14 h in patients receiving concomitant enzyme inducers [bib_ref] Population pharmacokinetics of felbamate in children, Kelley [/bib_ref]. Felbamate serum concentrations vary widely between patients and are affected by age; they are higher in patients with renal impairment, and the longer half-life depends on the degree of renal function [bib_ref] Single-dose pharmacokinetics of felbamate in patients with renal dysfunction, Glue [/bib_ref]. The clearance value is quite high in children (20-65 %) compared with adults [bib_ref] Clinical pharmacokinetics of new-generation antiepileptic drugs at the extremes of age, Perucca [/bib_ref]. Atropaldehyde, an intermediate metabolite, is responsible for serious progressive organ toxicity [bib_ref] Treatment of Lennox-Gastaut syndrome: overview and recent findings, Van Rijckevorsel [/bib_ref]. The drug is contraindicated in patients with liver impairment [bib_ref] Pharmacokinetic variability of newer antiepileptic drugs: when is monitoring needed?, Johannessen [/bib_ref]. Data on the pharmacokinetics of felbamate during pregnancy are insufficient. Life-threatening adverse reactions such as liver toxicity and aplastic anemia strongly restrict the use of this drug. Serum concentrations of felbamate can be decreased by enzyme-inducing AEDs [bib_ref] Felbamate: therapeutic range and other kinetic information, Troupin [/bib_ref] and increased by enzyme inhibitors such as VPA [bib_ref] Felbamate: therapeutic range and other kinetic information, Troupin [/bib_ref]. Clinical investigators have divided felbamate concentrations into lower (9-36 mg/l), intermediate (37-54 mg/l), and higher (54-134 mg/l) ranges [bib_ref] Felbamate levels in patients with epilepsy, Harden [/bib_ref] , and a therapeutic reference range of 210-462 lmol/l has been proposed [bib_ref] Felbamate: therapeutic range and other kinetic information, Troupin [/bib_ref] in patients with epilepsy. The drug is less well-tolerated in elderly populations. The serum concentration of felbamate can be determined via HPLC-MS [bib_ref] Simultaneous liquid chromatographic determination of lamotrigine, oxcarbazepine monohydroxy derivative and felbamate in..., Contin [/bib_ref] [bib_ref] Measurement of felbamate by wide-bore capillary gas chromatography and flame ionization detection, Rifai [/bib_ref] , HPLC-UV [bib_ref] Drug monitoring and toxicology: a simple procedure for the monitoring of felbamate..., Tang [/bib_ref] , LC-MS [bib_ref] Quantification in patient urine samples of felbamate and three metabolites: acid carbamate..., Thompson [/bib_ref] , and GC with flame ionization detection [bib_ref] Measurement of felbamate by wide-bore capillary gas chromatography and flame ionization detection, Rifai [/bib_ref] methods.
## Gabapentin
Gabapentin is approved as an adjunctive in the management of focal epilepsies in patients aged [6 years and as monotherapy in patients aged [12 years. It is also used for the management of peripheral neuropathy in adults [bib_ref] Crossover, open-label trial of the effects of gabapentin versus pregabalin on painful..., Atalay [/bib_ref]. Gabapentin was granted initial FDA approval in 1993. Gabapentin is rapidly absorbed (t max 2-3 h) from the GI tract by capacity-limited L-amino acid transport systems. It has an indirect effect on voltage-gated calcium channels and increases brain GABA concentrations [bib_ref] Effects of gabapentin on brain GABA, homocarnosine, and pyrrolidinone in epilepsy patients, Petroff [/bib_ref]. This was demonstrated by linearity of concentration with doses up to 1800 mg/day and non-linearity at higher doses [bib_ref] A saturable transport mechanism in the intestinal absorption of gabapentin is the..., Stewart [/bib_ref]. The half-life of gabapentin (5-9 h) increases with renal impairment [bib_ref] Gabapentin in the management of convulsive disorders, Mclean [/bib_ref]. Dose adjustment is mandatory in patient with creatinine clearance \60 ml/min and in the elderly because of reduced renal function. The concentration/dose ratio progresses with age, and inter-subject variability is common at any dose [bib_ref] Association between patient age and gabapentin serum concentration-to-dose ratio: a preliminary multivariate..., Armijo [/bib_ref]. Capacity-limited GI transport means gabapentin can be administered safely and effectively either twice or three times daily as steady state fluctuations are minor. Data are limited regarding the pharmacokinetics of gabapentin during pregnancy and in terms of drug-drug interactions. Antacids reduce gabapentin absorption from the GI tract, and H-2 receptor blockers decrease the renal clearance of gabapentin. The main adverse reactions are sedation, dizziness, headache, nausea, and obesity. Therapeutic concentrations of gabapentin in treatment-refractory patients with partial seizures range from 2 to 20 mg/l [bib_ref] Serum concentrations and effects of gabapentin and vigabatrin: observations from a dose..., Lindberger [/bib_ref]. HPLC [bib_ref] Fast isocratic high-performance liquid chromatographic assay method for the simultaneous determination of..., Chollet [/bib_ref] , LC/MScoupled GC [bib_ref] Gabapentin quantification in human plasma by highperformance liquid chromatography coupled to electrospray..., Ifa [/bib_ref] , and GC-electron ionization-MS [bib_ref] Gas chromatography-electron ionization-mass spectrometry quantitation of valproic acid and gabapentin, using dried..., Ikeda [/bib_ref] methods have been used to quantify serum gabapentin.
# Conclusion
TDM is a valuable tool in the optimization of individual dosage regimens to maximize clinical efficacy while minimizing adverse drug reactions. Although reference concentration ranges indicate the range in which most patients display therapeutic response, many patients may require target concentrations outside the reference range. When interpreting SDC, situations or factors that may modify the relationship between serum AED and clinical effect should be considered, such as type and severity of epilepsy, pathological and physiological states, drug interactions, and the variability of pharmacogenetics. It is worthwhile remembering that clinical decisions should not be made on the basis of drug concentrations alone. In short, dosage regimens must be individualized based on patient history, clinical signs and symptoms, pharmacogenetics, and interpretation of clinical laboratory data.
Compliance with Ethical Standards Shery Jacob and Anroop Nair have no conflicts of interest that are directly related to the content of this work.
Funding No sources of funding were used to conduct this study or prepare this manuscript.
Open Access This article is distributed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.
[table] Table 1: Pharmacokinetic parameters of current antiepileptic medications [/table]
|
A selective cyclic integrin antagonist blocks the integrin receptors αvβ3 and αvβ5 and inhibits retinal pigment epithelium cell attachment, migration and invasion
Background: Proliferative vitreoretinopathy (PVR) is a leading cause of blindness after failed retinal reattachment surgery. PVR is characterized by the proliferation, migration and contraction of retinal pigmented epithelial cells (RPE), and these cellular responses are influenced by the expression and function of integrin receptors. The effect of a cyclic integrin antagonist containing the amino acid sequence Arg-Gly-Asp-D-Phe-Val (RGDfV), specific for the integrin receptors α v β 3 and α v β 5 , was investigated on basic fibroblast growth factor (bFGF), platelet derived growth factor-BB (PDGF-BB), and serum induced human RPE proliferation, migration, invasion and attachment to the extracellular matrix. Furthermore, the effects of bFGF and PDGF-BB regulated expression of integrins α v β 3 and α v β 5 on RPE cells was examined.Methods:The effect of a cyclic integrin antagonist and a control peptide (0.01 µg/ml to 300 µg/ ml) was investigated on serum or cytokine (bFGF or PDGF-BB pretreatment) induced human fetal RPE cell proliferation by H 3 -thymidine uptake. The effect of the cyclic integrin antagonist on RPE cell attachment onto different extracellular matrices (laminin, collagen IV, fibronectin), RPE cell invasion stimulated by PDGF-BB or serum, and migration stimulated by PDGF-BB, vascular endothelial growth factor (VEGF) or serum was explored. PDGF-BB and bFGF modulation of the integrin receptors α v β 3 and α v β 5 was evaluated by flow cytometry.Results:The integrin antagonist did not inhibit DNA synthesis stimulated by serum, bFGF, or PDGF-BB treatment. RPE attachment onto fibronectin was inhibited in a concentration range of 1-10 µg/ml (p < 0.05). Attachment of the RPE cells onto collagen IV and laminin was inhibited in a range of 3-10 µg/ml (p < 0.05). Serum and PDGF-BB stimulated migration was inhibited by the cyclic integrin antagonist in a concentration range of 1-10 µg/ml (p < 0.05). Furthermore, the cyclic integrin antagonist inhibited PDGF-BB stimulated RPE cell invasion through fibronectin (3µg/ml: 66% inhibition, p < 0.001). In each of these experiments, the control peptides had no significant
effects. PDGF-BB and bFGF pretreatment of RPE cells increased the expression of integrin receptors α v β 3 (bFGF: 1.9 fold, PDGF-BB: 2.3 fold) and α v β 5 (bFGF: 2.9 fold, PDGF-BB: 1.5 fold).
# Conclusion:
A selective inhibition of the integrin receptors α v β 3 and α v β 5 through a cyclic integrin antagonist is able to inhibit RPE cell attachment, migration and invasion. Since these steps are of importance for the progression of PVR, a cyclic integrin antagonist should be further evaluated for the treatment of this disease.
# Background
Integrins are a family of heterodimeric, non-covalently bound cell surface receptors, which mediate cell-cell and cell to extracellular matrix (ECM) adhesion. They are transmembrane glycoproteins consisting of a larger α and a smaller β subunit. In mammals, 18 αand 8 β integrin genes encode polypeptides that combine to form 24 αβ heterodimeric receptors [bib_ref] Integrins: bidirectional, allosteric signaling machines, Hynes [/bib_ref]. By mediating cell-cell and cell-matrix contact, the integrins are involved in a spectrum of physiologic and pathologic processes. Integrin expression is modulated by several cytokines, growth factors and the extracellular matrix in different cell types [bib_ref] The regulation of expression of integrin receptors, Kim [/bib_ref].
The integrins α v β 3 and α v β 5 have been shown to be of particular significance for the progression of neovascularization. Both integrin receptors contain RGD sequences containing the amino acid sequence Arg-Gly-Asp. Vitronectin receptor integrins of the α v -chain subfamily interact with their target proteins via the tripeptide sequence RGD (NH2-arginine-glycine-aspartic acid-COOH) [bib_ref] Integrin (alpha v beta 3)-ligand interaction. Identification of a heterodimeric RGD binding..., Smith [/bib_ref] [bib_ref] New perspectives in cell adhesion: RGD and integrins, Ruoslahti [/bib_ref]. The blockage of these RGD receptor sequences induces endothelial cell apoptosis by inhibiting their binding to the ECM [bib_ref] Integrin alpha v beta 3 antagonists promote tumor regression by inducing apoptosis..., Brooks [/bib_ref] [bib_ref] Subcutaneous injection of a cyclic peptide antagonist of vitronectin receptor-type integrins inhibits..., Hammes [/bib_ref]. In addition fibroblast attachment onto fibrinogen, a substep involved in wound healing, is inhibited by blockage of the α v β 3 integrin receptor [bib_ref] Human fibroblasts bind directly to fibrinogen at RGD sites through integrin alpha(v)beta3, Gailit [/bib_ref]. Therefore, the integrin α v β 3 plays a critical role during wound repair as an adhesion receptor and as a signaling receptor [bib_ref] Protein kinase mediators of integrin signal transduction, Hannigan [/bib_ref]. As an adhesion receptor, it has the exceptional ability to bind vitronectin, fibronectin and fibrinogen, the three major proteins present in the wound provisional matrix [bib_ref] Transient functional expression of alphaVbeta 3 on vascular cells during wound repair, Clark [/bib_ref].
In proliferative vitreoretinopathy (PVR), a disease of the posterior eye characteristically induced by trauma or failed retinal reattachment surgery, integrins are also of importance for the progression of the disease [bib_ref] Immunolocalization of integrins in proliferative retinal membranes, Robbins [/bib_ref] [bib_ref] Inhibition of retinal pigment epithelial cell-induced tractional retinal detachment by disintegrins, a..., Yang [/bib_ref]. PVR, an exaggerated wound healing process, is characterized by the proliferation, migration and contraction of retinal pigment epithelial cells (RPE), fibroblasts, glial cells and macrophages. PVR membranes are formed on, or under the sensory retina; their contraction may lead to retinal detachment and blindness. RPE cells play a dominant role in the pathobiology of the disease [bib_ref] Proliferative vitreoretinopathy: pathobiology, surgical management, and adjunctive treatment, Charteris [/bib_ref]. Growth factors including basic fibroblast growth factor (bFGF), platelet derived growth factor-BB (PDGF-BB), and vascular endothelial growth factor (VEGF), and the ECM mole-cule fibronectin, regulate proliferation, migration and invasion of the RPE cell and are expressed in PVR membranes [bib_ref] Proliferative vitreoretinopathy: pathobiology, surgical management, and adjunctive treatment, Charteris [/bib_ref] [bib_ref] Proliferative vitreoretinopathy membranes. An immunohistochemical study, Jerdan [/bib_ref]. Fibronectin is an adhesive substrate for both the integrin receptors α v β 3 and α v β 5 [bib_ref] The integrin superfamily and the eye, Elner [/bib_ref]. Several studies have identified integrins α v β 3 , α v β 5 , integrin alpha subunits 2,3,4,5,6, V, and integrin beta subunits 1, 2, 3 on RPE cells and cells within PVR membranes [bib_ref] Vitronectin receptor expression and distribution at the photoreceptor-retinal pigment epithelial interface, Anderson [/bib_ref] [bib_ref] Identification and localization of a beta-1 receptor from the integrin family in..., Anderson [/bib_ref] [bib_ref] Vitronectin and proliferative intraocular disorders. II. Expression of cell surface receptors for..., Weller [/bib_ref]. The ECM-cell interactions are thought to be mediated largely through this family of cell surface receptors. Because cell-ECM interaction is important in the process of wound healing, one could predict that integrins would be actively involved in the pathogenesis of PVR. For PVR, it was shown in vivo, that tractional retinal detachment could be inhibited [bib_ref] Inhibition of retinal pigment epithelial cell-induced tractional retinal detachment by disintegrins, a..., Yang [/bib_ref] by blockage of integrin-ECM binding by an RGD containing disintegrin.
Peptide antagonists specific for individual integrins have been developed, including cyclic RGD peptides selective for the α v integrin [bib_ref] Selective recognition of cyclic RGD peptides of NMR defined conformation by alpha..., Pfaff [/bib_ref] , the α v β 3 integrin [bib_ref] Peptide ligands for integrin alpha v beta 3 selected from random phage..., Healy [/bib_ref] , and the α v β 5 integrin [bib_ref] Selective recognition of cyclic RGD peptides of NMR defined conformation by alpha..., Pfaff [/bib_ref] [bib_ref] Peptide ligands for integrin alpha v beta 3 selected from random phage..., Healy [/bib_ref] [bib_ref] Selection of peptides binding to the alpha 5 beta 1 integrin from..., Koivunen [/bib_ref] [bib_ref] Isolation of a highly specific ligand for the alpha 5 beta 1..., Koivunen [/bib_ref]. Sequence alterations in the backbone of the RGD-containing cyclic peptides that result in conformational differences in the interatomic distance between the C β atoms of Arg and Asp, have been shown to correlate with selective peptide recognition [bib_ref] Selective recognition of cyclic RGD peptides of NMR defined conformation by alpha..., Pfaff [/bib_ref] [bib_ref] An angiogenic laminin site and its antagonist bind through the alpha(v)beta3 and..., Ponce [/bib_ref].
In this study we investigated the modulatory effect of the cytokines PDGF-BB and bFGF on the expression of the RGD containing integrin receptors α v β 3 and α v β 5 on human fetal RPE cells. Furthermore, the effect of a specific cyclic integrin antagonist that blocks the integrin receptors α v β 3 and α v β 5 , was investigated for its effects on RPE attachment, proliferation, migration and invasion. A specific cyclic integrin antagonist was chosen for its higher affinity and stronger binding capacity to the α v β 3 and α v β 5 integrins compared to a non-cyclic integrin antagonist [bib_ref] The pharmacology of the integrins, Cox [/bib_ref] [bib_ref] Selective recognition of cyclic RGD peptides of NMR defined conformation by alpha..., Pfaff [/bib_ref].
# Methods
All cell culture solutions, media and ECM molecules (fibronectin, laminin and vitronectin) were purchased from Sigma (Sigma, St. Louis, MO).
## Isolation and culture of human rpe cells
Human RPE cells were isolated from fetal donor eyes (gestation time over 22 weeks) which were obtained from the Anatomic Gift foundation (Woodbine, GA) as has been described [bib_ref] Verapamil inhibits proliferation, migration and protein kinase C activity in human retinal..., Hoffman [/bib_ref]. RPE cells were seeded onto laminin-coated 6 well plates (Fisher Scientific, Tustin, CA) using DMEM with 10 % fetal bovine serum, 1 % penicillin/streptomycin and glutamine. Cells from passage 3 to 5 were used in all experiments.
## Synthesis of the cyclic integrin antagonist and its control peptide
The cyclic integrin antagonist RGDfV, specific for the inhibition of integrins α v β 3 and α v β 5 [bib_ref] Selective recognition of cyclic RGD peptides of NMR defined conformation by alpha..., Pfaff [/bib_ref] [bib_ref] Peptide ligands for integrin alpha v beta 3 selected from random phage..., Healy [/bib_ref] [bib_ref] Selection of peptides binding to the alpha 5 beta 1 integrin from..., Koivunen [/bib_ref] [bib_ref] Isolation of a highly specific ligand for the alpha 5 beta 1..., Koivunen [/bib_ref] and the control peptide RADfV were synthesized as described [bib_ref] Subcutaneous injection of a cyclic peptide antagonist of vitronectin receptor-type integrins inhibits..., Hammes [/bib_ref] at the USC/ Norris Cancer Peptide synthesis facility. The peptides were dissolved in Hank's Balanced Salt Solution (HBSS), and stored at -80°C. HPLC analysis revealed a purity > 99% of the synthesized peptide.
## Flow cytometry
RPE cells (30,000 cells/well) were treated for 5 days with the cytokines bFGF and PDGF-BB at a concentration of 10 ng/ml in 6 well plates. Cells were detached, fixed with 4 % paraformaldehyde, blocked with 5 % goat serum for 30 minutes, centrifuged at 4°C, and incubated with a monoclonal mouse antibody against α v β 3 or α v β 5 (5 µg/ml) (Chemicon, CA) for one hour. After washing, a fluorescein conjugated secondary antibody (Chemicon, CA) directed against the mouse antibodies was added in a dilution of 1:100 for one hour. The expression of the integrin receptors α v β 3 or α v β 5 was measured using a fluorescent-activated cell sorter (FACStar plus, Becton Dickinson, Mountain View, CA). Forward and side light scatter was used to gate the desired scattered events (RPE cells) from dead cells and debris. The fluorescent index was determined by multiplying the percentage of cells by their mean fluorescence. At least 10,000 cells were evaluated/ experiment and all experiments were performed in triplicate.
## Cell counting and viability studies
Subconfluent, bFGF (10 ng/ml) or PDGF-BB (10 ng/ml) pre-stimulated RPE cells were seeded at a density of 8 × 10 4 cells per well in 6 well plates containing DMEM with 10 % FBS and the cyclic integrin antagonist or the control peptide (0.01 µg/ml, 0.1 µg/ml, 1 µg/ml, 30 µg/ml and 300 µg/ml). The cells were exposed during the proliferation assay to 10% FBS or the cytokines bFGF and PDGF-BB (10 ng/ml). After 3 days of culture, the cell number was determined using a hemocytometer. The experiments were performed six times, each in triplicate. Cell viability was estimated by observation of cell morphology and by exclusion of 0.4 % trypan blue solution.
## H-thymidine uptake
RPE cells (10,000 cells/well) were seeded into DMEM with 10 % FBS in 24 well plates with the cyclic integrin antagonist or the control peptide (0.01 µg/ml, 0.1 µg/ml, 1 µg/ml, 30 µg/ml and 300 µg/ml). To determine a possible stronger effect of the cyclic integrin antagonist on RPE cell proliferation, cells were pre-stimulated for 5 days with bFGF (10 ng/ml) or PDGF-BB (10 ng/ml) (R&D Systems, Minneapolis, MN) to induce an upregulation of the integrin receptors α v β 3 or α v β 5. After the pre-incubation, cells were seeded in DMEM with 10 % FBS, with the cyclic integrin antagonist or control peptide, and bFGF (10 ng/ ml) or PDGF-BB (10 ng/ml). Cells were treated for three days with the peptides, then 3 H-thymidine uptake was measured as described previously [bib_ref] Verapamil inhibits proliferation, migration and protein kinase C activity in human retinal..., Hoffman [/bib_ref]. Results are presented as the percentage of control values.
## Migration assay
Migration was performed in a modified Boyden chamber (Falcon, Becton Dickinson Labware, Franklin Lakes, NJ). Subconfluent RPE cells, which have been pre-exposed to the cyclic integrin antagonist peptide (0.001 ng/ml, 0.01 ng/ml, 0.1 ng/ml, 1 ng/ml, 3 ng/ml and 10 ng/ml) or the control peptide (10 µg/ml) over night were seeded into DMEM with 1 % FBS into the upper part of a Boyden chamber (50,000 cells/chamber). The cyclic integrin antagonist was added directly to the upper part of the chamber (0.001 ng/ml, 0.01 ng/ml, 0.1 ng/ml, 1 ng/ml, 3 ng/ml, or 10 ng/ml). To evaluate nonspecific activity of the integrin antagonist, a control peptide (10 µg/ml) was used. The lower part of the chamber was filled with DMEM with 1 % FBS containing the chemoattractant PDGF-BB (10 ng/ml) or the ECM molecule fibronectin (30 ng/ml). RPE cells were incubated for 8 hours in a humidified CO 2 -incubator. Then the cells in the upper part were scraped away and the membrane was fixed with 1/2 strength Karnovsky's solution and stained with 0.25 % Richardson's solution. RPE cells on the lower side of the membrane were counted with the help of an inverted microscope and a grid (Carl Zeiss, Jena, Germany) in five random high power fields (200 X) in triplicate.
## Cell attachment
Ninety-six well plates were coated with the ECM molecules laminin, fibronectin or collagen IV (25 µg/ml). RPE cells were pre-incubated with the cyclic integrin antagonist (concentrations as above) or control peptide (10 µg/ ml) over night. Then 15 × 10 3 cells, suspended in 100 µl DMEM with the cyclic or control peptide were seeded in each well and allowed to attach for 60 minutes. A tetrazolium MTT [3-(4,5-dimethylthiazolyl-2)-2,5-diphenyltetrazolium bromide] solution (20 µl/well) (Boehringer Mannheim, Indianapolis, IN, USA) was added for 5 hours of incubation. Then the supernatants were decanted, 150 µl of 100% DMSO added, and the cells were placed on a shaker for 10 minutes. Absorbance of the cells was measured at 550 nm with a Dynatech MR 600 spectrophotometric microplate reader.
## Invasion assay
RPE cell invasion was examined in a modified Boyden chambers. The insert membrane was coated overnight with fibronectin (25 µg/ml). Subconfluent RPE cells were pretreated overnight with the cyclic integrin antagonist (3 µg/ml) or the control peptide (3 µg/ml). An RPE cell suspension (50,000 cells/0.1 ml) in DMEM with 0.4 % FBS and 3 µg/ml cyclic integrin antagonist or control peptide was added into the upper part of the chamber. 600 µl DMEM with 0.4% FBS and PDGF-BB (10 µg/ml) was added into the lower part of the chamber. Invasion was measured with and without PDGF-BB stimulation after 5 hours at 37°C (95 % air/ 5 % Co 2 ). After PBS washing and methanol fixation (10 minutes at 4°C), counterstaining with hematoxylin was performed. Cells in the upper chamber were wiped away. RPE cells that invaded the lower surface of the membrane were quantified by cell counting (320 × magnification).
# Statistical analysis
All experiments were repeated in triplicate. Standard deviations and averages were calculated. The data obtained were analyzed for significance using one way analysis of variance (ANOVA). Data with a p-value of less then 0.05 were considered to be statistically significant.
# Results
## The cyclic integrin antagonist does not inhibit rpe cell proliferation
The RPE cells were stimulated to incorporate thymidine by either 10% serum or by 10 ng/ml bFGF or 10 ng/ml PDGF-BB (p < 0.01). Neither the integrin antagonist, nor control peptide showed any significant effect on serum, PDGF-BB or bFGF stimulated DNA synthesis. (data not shown). In addition, cell counting showed no effects of the integrin antagonist or the control peptide on serum or cytokine stimulated proliferation. Furthermore, a lack of toxicity for the integrin antagonist and the control peptide in the concentration range tested (0.01 µg/ml, 0.1 µg/ml, 1 µg/ml, 30 µg/ml and 300 µg/ml) was determined by use of the trypan blue exclusion assay (data not shown).
## Bfgf and pdgf-bb upregulate surface expression of α v β 3 and α v β 5
Pre-stimulation of RPE cells for 5 days with the cytokines PDGF-BB and bFGF (10 ng/ml) increased the expression of the integrin receptors α v β 3 and α v β 5 representative flow cytometric study). Incubation with 10 ng/ml bFGF upregulated the expression of α v β 3 (1.9 fold) and strongly enhanced the expression of α v β 5 (2.9 fold). PDGF-BB (10 ng/ml) strongly enhanced α v β 3 (2.3 fold), and less strongly upregulated surface expression of α v β 5 (1.5 fold).
## The cyclic integrin antagonist inhibits rpe attachment onto ecm
RPE cell attachment onto fibronectin was inhibited significantly in the range of 1 to 10 µg/ml of cyclic integrin antagonist (p < 0.05). A concentration of 10 µg/ ml control peptide showed no effect. RPE cell attachment was inhibited 18 % with 1 µg/ml (p = 0.03), 23 % with 3 µg/ml (p < 0.005) and 27 % with 10 µg/ml (p < 0.001) onto fibronectin. An ID 50 of 18.5 µg/ml cyclic integrin antagonist was calculated for the RPE cell attachment onto fibronectin. The attachment onto collagen IV was inhibited 25 % with 3 µg/ml (p < 0.001), and 29 % with 10 µg/ml (p < 0.001). Lower concentrations showed no significant effect. An ID 50 of 17.2 µg/ml cyclic peptide was calculated for RPE cell attachment onto collagen IV. Attachment onto laminin was inhibited by the cyclic integrin antagonist in the range of 3-10 µg/ml (3 µg/ml, 27 % inhibition (p < 0.001); 10µg/ml: 23 % inhibition (p < 0.001). An ID 50 of 21.7 µg/ml was calculated for the inhibition of RPE cell attachment onto laminin.
## The cyclic integrin antagonist inhibits rpe cell migration
The ECM component fibronectin (30 ng/ml) induced chemotaxis when placed as a soluble agent in the lower part of the Boyden chamber. This migratory response was inhibited by the cyclic integrin antagonist but not by the control peptide. RPE cell migration was inhibited in a concentration range of 1-10 µg/ml (1µg/ ml, 18 % inhibition (p < 0.001); 3 µg/ml, 23 % inhibition (p < 0.001); 10 mg/ml, 27 % inhibition (p < 0.001)). An ID 50 of 18.5 µg/ml integrin antagonist was calculated for the serum induced migration. The migratory response induced by PDGF-BB (10 ng/ml) (93 % stimulation of migration) was diminished by the cyclic integrin antagonist. (1µg/ml, 26 % inhibition (p < 0.001); 3 µg/ml, 66 % inhibition (p < 0.001); and 10µg/ml, 74 % inhibition (p < 0.001)). An ID 50 of 5.8 µg/ml was calculated for the inhibition of PDGF-BB induced migration of RPE cells. VEGF induced a migratory response of RPE cells by 39 %. This migratory response was inhibited by the cyclic integrin antagonist in a concentration range of 3-10 µg/ ml (p < 0.01).
## The cyclic integrin antagonist inhibits pdgf-bb stimulated invasion through fibronectin
PDGF-BB strongly stimulated RPE cell invasion (77 %, (p < 0.001)) through fibronectin coated onto the lower surface of a membrane insert in a modified Boyden chamber. This invasion was inhibited by the cyclic integrin antagonist (3 µg/ml, 66 % inhibition (p < 0.001)). The control peptide had no significant effect on invasion. An ID 50 of 2.3 µg/ml was calculated for the effects of the cyclic integrin antagonist on PDGF-BB induced RPE invasion through fibronectin.
>A. Expression of the integrin receptors α v β 3 and α v β 5 after pretreatment with the cytokine bFGF in a concentration of 10 ng/ ml for 5 days A. Expression of the integrin receptors α v β 3 and α v β 5 after pretreatment with the cytokine bFGF in a concentration of 10 ng/ ml for 5 days. B. Expression of the integrin receptors α v β 3 and α v β 5 after pretreatment with the cytokine PDGF-BB in a concentration of 10 ng/ml for 5 days.
A. RPE attachment on fibronectin, laminin and collagen IV
# Discussion
These data provide support for the suggestion that inhibition of the integrin receptors α v β 3 and α v β 5 by a specific cyclic integrin antagonist is feasible for inhibiting RPE cell attachment, migration and invasion and may serve as an adjunctive therapy for disorders such as PVR. Integrins have been shown to influence cell signaling pathways that promote cell proliferation in concert with cytokines [bib_ref] Integrins and signal transduction pathways: the road taken, Clark [/bib_ref]. Interestingly, our data show that cyclic integrin inhibition of α v β 3 and α v β 5 had no effect on serum, PDGF-BB or bFGF induced proliferation of RPE. Since we also show that bFGF and PDGF-BB upregulate both α v β 3 and α v β 5 , and that proliferation is not inhibited in these stimulated cells, we can infer that the lack of effect of the cyclic peptide on proliferation is not due to lack of integrin expression. Consistent with the lack of effect on proliferation, we found that the cyclic integrin inhibitor did not induce significant cell death. This suggests that in the conditions studied here, the effect of integrin inhibition by the cyclic peptide on survival and proliferation of RPE cells may differ from previously described effects on endothelial cells [bib_ref] The extracellular matrix as a cell survival factor, Meredith [/bib_ref] , specific tumor cells [bib_ref] Laug WE: alpha v-Integrin antagonist EMD 121974 induces apoptosis in brain tumor..., Taga [/bib_ref] and mammary epithelial cells [bib_ref] Suppression of ICE and apoptosis in mammary epithelial cells by extracellular matrix, Boudreau [/bib_ref].
Cell attachment of RPE cells onto fibronectin, laminin and collagen IV was inhibited by the cyclic integrin antagonist. Because substrate attachment and interaction is important for activated RPE function [bib_ref] Reattachment to a substrate prevents apoptosis of human retinal pigment epithelium, Tezel [/bib_ref] , it is reasonable to predict that anti-adhesive therapy holds promise for the therapy of PVR. The cyclic integrin antagonist has been shown previously to inhibit specific ligand binding or cell adhesion events mediated by α v β 3 or α v β 5 [bib_ref] Selective recognition of cyclic RGD peptides of NMR defined conformation by alpha..., Pfaff [/bib_ref]. Blockage of RPE attachment onto laminin, collagen IV and fibronectin is of specific interest, because of the strong expression of these matrix proteins in PVR [bib_ref] Proliferative vitreoretinopathy: pathobiology, surgical management, and adjunctive treatment, Charteris [/bib_ref]. The ECM fibronectin reacts with the α v β 3 receptor with high specificity and affinity, supporting cell adhesion to this matrix protein. The α v β 5 receptor binds in a similar way to fibronectin [bib_ref] Cell surface receptors for extracellular matrix components, Akiyama [/bib_ref]. The blockage of α v β 3 and α v β 5 thus provides a reasonable explanation for the decreased attachment of RPE onto collagen IV, fibronectin and laminin in the presence of the inhibitory cyclic peptide. RPE cell migration is another important process in the development of PVR. Without migration, RPE cells would not gain access to the vitreous and form vitreoretinal membranes. For RPE cells, it was shown previously that monoclonal antibody inhibition of integrin subunit β 1 inhibits RPE cell migration from wound edges in organ culture [bib_ref] The integrin superfamily and the eye, Elner [/bib_ref]. Integrin signaling results in actin cytoskeleton polymerization, which in turn regulates cell shape and motility [bib_ref] Integrin cytoplasmic domains: mediators of cytoskeletal linkages and extra-and intracellular initiated transmembrane..., Sastry [/bib_ref]. The α v β 3 receptor appears to be involved in pathways that regulate intracellular pH and intracellular Ca 2+ [bib_ref] Integrin beta 1-and beta 3-mediated endothelial cell migration is triggered through distinct..., Leavesley [/bib_ref] and may therefore be involved in migration in response to motility factors and extracellular matrix components [bib_ref] Structure-function analysis of signal and growth inhibition by carboxyamido-triazole, CAI, Kohn [/bib_ref]. The same range of concentration of the cyclic peptide inhibited migration induced by both soluble fibronectin and PDGF-BB. This result may in part be due to the finding that PDGF-BB induced migration is mediated by signaling pathways that are also used by integrin receptors [bib_ref] Definition of two angiogenic pathways by distinct alpha v integrins, Friedlander [/bib_ref] [bib_ref] Insulin-like growth factor receptor cooperates with integrin alpha v beta 5 to..., Brooks [/bib_ref] [bib_ref] Mitogen-activated protein kinase activation mediates PDGF-directed migration of RPE cells, Hinton [/bib_ref].
The cyclic integrin antagonist had a particularly strong effect on RPE invasion into fibronectin. This may be due to the concurrent effects of the peptide on the inhibition of adhesion, migration and proteolytic activity since each of these steps may play a critical role in the invasion of the ECM.
The cyclic integrin antagonist described here blocks both α v β 3 and α v β 5 . In contrast, highly selective integrin antagonists such as LM609 block only α v β 3 . Therefore, a stronger inhibitory effect of the cyclic integrin antagonist can be expected as has been shown in tumors [bib_ref] Definition of two angiogenic pathways by distinct alpha v integrins, Friedlander [/bib_ref]. While increased specificity could result in lower toxicity, we did not find that the cyclic integrin antagonist inhibited thymidine incorporation, or that it increased trypan blue uptake in the concentrations tested.
[fig] Figure 2 A: RPE attachment on fibronectin, laminin and collagen IV. The cells were pretreated before the assay with the cyclic integrin antagonist, or the control peptide (10 µg/ml). Untreated RPE cells were used in addition as a standard (ST). B. RPE cell migration induced by PDGF-BB, VEGF and fibronectin. The effect of the cyclic integrin antagonist or the control peptide (cp) on migration is shown. (ST = untreated RPE cells) C. Effect of the cyclic integrin antagonist, (3 ug/ml) or the control peptide on RPE invasion through fibronectin. Cell invasion through fibronectin (F) was stimulated with PDGF-BB (10 ng/ml). Nonspecific effects were excluded by the use of a cyclic control peptide. DMEM without PDGF-BB was used as a control in the assay. [/fig]
|
Attitudes towards Violence in Adolescents and Youth Intimate Partner Relationships: Validation of the Spanish Version of the EAV
# Introduction
An increasing interest is being devoted to intimate partner violence (IPV), due to the severe physical, psychological, and social consequences associated, as well as the growing prevalence through the world. This specific kind of violence that includes physical, verbal, and sexual violence has become a serious and prevalent problem, not only in adulthood but also in adolescence and into emerging adulthood. For instance, The WHO Multi-country Study on Women's Health and Domestic Violence against women revealed that women between 15 and 49 years old had suffer some kind of partner violence in percentages that ranged from 13 to 61 percent. These rates are worrying, especially taking into account that some of the mental and behavioural problems, as well as violence manifestations that develops during adolescence, tend to perpetuate to adulthood. This type of violence, involving an actual or former partner, is at this time the more prevalent way of violence affecting women, being a major public health problem. This situation is affecting women, and has become a major cause of social disengagement, economy burden, and mental health problems, including depression, substance abuse, or trauma. Thus, nowadays, it is possible to talk of a global epidemic of violence against women, with a total life prevalence in the world of one third of women that have experience some type of violence related to a romantic relationship.
Worth noting, although most of the victims addressing IPV are females, it is known that males are also victims of females perpetrators, with a total of up to in between 20 to 30 percent of men recognizing being victims of this kind of behaviours in heterosexual relationships. Moreover, this problem is also affecting to same-sex relationships. In the case of adolescents, recent data suggest that IPV is affecting more than half of those that have engaged or that have experienced a relationship either as a victim or as perpetrator. This is even more alarming considering that teenagers with maltreatment history have a greater risk of being involve in a relationship with IPV both as a victim and as a perpetrator.
Taking into account the before mentioned, it seems relevant to devote resources aimed to understand the inner mechanisms that is behind IPV. Thus, understanding the attitudes towards IPV that adolescents show could be relevant. Different previous studies have focused on the violence behaviours, but not on the attitudes, nor in the relationship between these attitudes and IPV. Among the different instruments devoted to assess these problematics, we can find the Attitude Scale towards Intimate Violence (Escala de Actitudes hacia la Violencia Intima, EAV). The EAV encompass ten items addressing the degree to which the individual considers appropriate the use of violence towards the partner in different situations. There is, thus, a general question asking: In what circumstances do you consider the use of intimate partner violence justifie? And then, ten different options such as: when a member of the couple is unfaithful (ítem 1) or when a member of the couple disqualifies the other in front of his/her family (ítem 3).
Nonetheless, the psychometric properties of this recent version have not been, yet, reported. Therefore, a question needs still to be solve. Is it possible to use the EAV in its Spanish version as an instrument with adequate evidences of reliability of the scores and validity? Moreover, recent research reveals that different variables, including education, attitudes, and ideas about gender roles and expectations of a relationship are connected to IPV. With this regard, and considering that the ideas and attitudes of adolescents towards IPV could be related to emotional symptoms, new attention is being devoted to the quality of romantic relationships and to this, in order to promote healthy relationships.
Considering the previous background, and the fact that the EAV whose psychometric properties has not been yet validated in Spanish population, outside South America, the main goal of the present study was, therefore, to analyze the psychometric properties of the EAV in a large sample of Spanish adolescents. We, thus, gather evidences about the structure of the questionnaire, study the measurement invariance (MI) by gender and age, and analyze evidences of validity with external variables. We hypothesized that a onedimensional model would reveal adequate goodness-of-fit indices and that this factorial structure would be invariant across gender and age. In addition, we hypothesized that measures of EAV would be related to violence in the context of intimate relationships and depressive symptoms.
# Method
## Participants
In order to obtain a representative community sample, we recruited participants from different cities and different types of secondary schools (e.g., public, funded, and private) and vocational/technical schools belonging to Navarra and La Rioja (two regions located at the north of Spain). Both rural and urban areas were represented, as well as a range of socioeconomic levels. We recruited students from ten schools, including educational and training centres. The initial sample included 1305 students, and we discarded data from participants who presented: (a) omissions of any demographics or items without responding (n = 37); and (b) scores in the range of outliers (n = 20) (e.g., scores higher than 2.5 standard deviations in the subscales of the measures used). The final sample consisted of 1248 students, of which 483 were male (38.7%). The age of the participants ranged from 13 to 21 years-old (M = 16.12 years; SD = 2.12). The age distribution of the sample was the following: 13 years (n = 65; 5.2%), 14 years (n = 216; 17.3%), 15 years (n = 336; 26.9%), 16 years (n = 231; 18.5%), 17 years (n = 147; 11.8%), 18 years (n = 129; 10.3%), and 19-21 years (n = 92; 7.3%). The EAVis an instrument devoted to assess attitudes towards violence in the context of intimate and romantic relationships. The EAV is composed of 10 items in a Likert response format with five options (1 = "totally disagree", 2 = "disagree", 3 = "neutral", 4 = "agree", and 5 = "totally agree". The items ask about under which circumstances the use of violence is justified in a relationship. For example, the use of violence is justified "When a member of the couple insult to the other" or "When one member of the people does not agree to have sexual intercourse". The scale has shown evidences of internal consistency of the scores in previous studies with alpha values over 0.90.
## Instruments
## The modified conflict tactics scale (m-cts)
The M-CTSis one of the most widely used instrument to measure the way in which individuals deal issues with their partners. It addresses behaviours when arguing with the actual partner or the most recent relationship. The M-CTS is composed of 18 bidirectional items addressing behaviours as victim and aggressor. The items are in a 5-point Likert scale ranging from 1 (never) to 5 (very often). The validated Spanish version was used in the present study. The M-CTS revealed adequate evidences of internal consistency of the scores in the present study with a McDonald's Omega value of 0.82. In addition, CR value was 0.823 and AVE was 0.713.
## The reynolds adolescent depression scale (rads)
The RADSassesses the severity of depressive symptoms in adolescents. It is composed of 30 items in a Likert response format with four options (1 = "almost never", 2 = "hardly ever", 3 = "sometimes", 4 = "most of the time"). The RADS encompasses four empirically derived scales: Anhedonia, Somatic complaints, Negative self-evaluation, and Dysphoric mood. The validated Spanish version of the RADS was used in the present study. The RADS revealed adequate evidences of internal consistency of the scores in the present study with a McDonald's Omega value of 0.84. The CR (0.815) and AVE (0.727) values were also adequate.
## Procedures
The questionnaires were administered collectively, in groups of 10 to 35 students, during normal school hours and in a classroom specially prepared for this purpose. For participants under 18, parents were asked to provide a written informed consent in order for their child to participate in the study. Participants were informed of the confidentiality of their responses and of the voluntary nature of the study. No incentive was provided for their participation. Administration took place under the supervision of the researchers. The study was approved by the research an ethic committee at the University of La Rioja.
## Data analyses
First, we calculated the internal consistency of the EAV scores. To obtain a measure of the reliability of the scores, we calculated McDonald's Omega. In addition, we gather evidences of convergence validity by means of the composite reliability (CR) and the average variance extracted (AVE). Values greater than 0.6 and 0.5 are considered adequate for CR and AVE respectively. Moreover, we analyzed the differential validity studying the square root of the AVE value. The data for the diagonal position are the square root of the mean variance extraction rate (AVE value) for each study variable If the square root of the mean variance extraction rate (AVE value) of each question is greater than the correlation coefficient between the variables, it indicates that there is a strong discriminant coefficient between the variables, that is, the difference between each measurement variable is better. Second, in order to analyze the internal structure of the EAV, we performed several confirmatory factor analysis (CFA) in the second subsample. We tested a onedimensional factor model, a two-factor model, derived from the results of the EFA, and a bifactor solution with a general factor and two group factors. The WLSMV estimator for dichotomous items was used. The following goodness-of-fit indices were used: Chi-square (χ 2 ), Comparative Fit Index (CFI), Tucker-Lewis Index (TLI), Root Mean Square Error of Approximation (RMSEA), and Weighted Root Mean Square Residual (WRMR). Hu and Bentlersuggested that RMSEA should be. 06 or less for a good model fit and CFI and TLI should be 0.95 or more, though any value over 0.90 tends to be considered acceptable. For WRMR, values less than 0.95 indicate good model fit (for dichotomous outcomes). Fourth, in order to test measurement invariance across gender, successive multigroup CFAs were conducted. Using Delta parameterization in Mplus, two steps on measuring invariance need to be considered: configural and strong invariance models. The ∆CFI were used to determine in cases where nested models were practically equivalent.
Third, and with the aim to test measurement invariance (MI) by gender and age, successive multigroup CFAs were conductedIn order to compare age, we established two different groups: younger adolescents (12-15 years old) and older adolescents (16-19 years old),.Basically, a hierarchical set of steps are followed when MI is tested, typically starting with the determination of a well-fitting multigroup baseline model and continuing with the establishment of successive equivalence constraints in the model parameters across groups. The analyzed dimensional models can be seen as nested models to which constraints are progressively added. Due to the limitations of the ∆χ 2 regarding its sensitivity to sample size, Cheung and Rensvoldproposed a more practical criterion, the change in CFI (∆CFI), to determine if nested models are practically equivalent. In this study, when ∆CFI is greater than 0.01 between two nested models, the more constrained model is rejected since the additional constraints have produced a practically worse fit. However, if the change in CFI is less than or equal to 0.01, it is considered that all specified equal constraints are tenable and, therefore, it is possible to continue with the next step in the analysis of MI. Latent mean differences across gender and age were estimated, fixing the latent mean values to zero in the male and in the younger group respectively. For comparisons among groups in the latent means, statistical significance was based on the z statistic. The group in which the latent mean was fixed to zero was considered as the reference group. Fourth, the associations between self-reported EAV scores and other measures including the EAV and the RADS subscales were examined using Pearson's correlations. In addition, we conducted a mediation analysis. To this purpose, we followed a two-step procedure, adapted to analyses the mediation effect in order to confirm the structural relations of the latent variables. SPSS 24.0, Mplus 7.4, and FACTOR 10.0were used for data analysis.
# Results
## Descriptive statistics for all the measuring instruments and evidences of reliability of the eav scores
First, descriptive statistics for the subscales and total scores of the measuring instruments used were calculated (see. Descriptive statistics of the EAV items are depicted in . As can be seen in , the prevalence of attitudes towards violence ranged from 1.52 (item 9 = "when a member of the couple presents excessive consume of substance like alcohol or drugs") to 1.27 (item 10 = when a member of the couple refuse to have sexual intercourse). No statistically significant differences were found by gender in the EAV total score (t = 0.910; p = 0.363).shows the results of the analysis of the McDonald's Omega, the CR, the AVE and the root square of the AVE. As it can be seen, CR was above the recommended 0.7 value in all the variables. Also, AVE was higher than 0.50, revealing good evidences of convergent validity. In addition, the correlation between the constructs was lower than the root square of AVE, indicating adequate evi-dences of discriminant validity. Finally, the McDonald's Omega was 0.862 and 0.872 for the two hypothesized factors, revealing adequate internal consistency of the scores.
## Validity evidences of factorial structure
The analysis of the EFA in the first subsample revealed statistically significant values of Bartlett's Sphericity Index (2539.8), being statistically significant (p < 0.001). Also, Kaiser-Meyer-Olkin (KMO) indices were above 0.85 in all cases. The GFI values found were in all the dimension above. 95. In addition, the RMSR was under 0.08. A two-factor solution explained more than 35% of the variance in all the dimensions. Factor 1 was composed of items 1, 2, 3, 4, and 5 that are related to justification of the violence due to misbehave of the partner. On the other hand, Factor 2 was integrated by items 6, 7, 8, 9, and 10 which relate to justification of the violence because a history of problems (e.g., emotional problems) of the partner.
After the EFA, we conducted different CFA at the item level. shows the goodness-of-fit indices for the different factor models tested. The one-dimensional model yielded adequate CFI and TLI values over. 95, however RMSEA values were over the recommended. 06 cut off value, as well as the WRMR values. Moreover, the bifactor solution revealed poor goodness-of-fit indices. Therefore, we decided to retain the twofactor model as the most adequate solution (see. The standardized factor loadings for the whole sample as well as for males and females are shown in . The range of the factor loadings, for the total sample was from 0.36 (item 1) to 0.72 (item 3). All standardized factor loadings estimated were statistically significant (p < 0.01).
## Int. j. environ. res. public health 2020, 17, x 6 of 11
Note. AVE = average variance extract. * = square root of the variance shared between the constructs and their measures. ** = correlations among the constructs. For evidences of discriminant validity, the correlations among constructs should be lower than the square root of the variance shared.
## Validity evidences of factorial structure
The analysis of the EFA in the first subsample revealed statistically significant values of Bartlett's Sphericity Index (2539.8), being statistically significant (p < 0.001). Also, Kaiser-Meyer-Olkin (KMO) indices were above 0.85 in all cases. The GFI values found were in all the dimension above. 95. In addition, the RMSR was under 0.08. A two-factor solution explained more than 35% of the variance in all the dimensions. Factor 1 was composed of items 1, 2, 3, 4, and 5 that are related to justification of the violence due to misbehave of the partner. On the other hand, Factor 2 was integrated by items 6, 7, 8, 9, and 10 which relate to justification of the violence because a history of problems (e.g., emotional problems) of the partner.
After the EFA, we conducted different CFA at the item level. shows the goodness-of-fit indices for the different factor models tested. The one-dimensional model yielded adequate CFI and TLI values over. 95, however RMSEA values were over the recommended. 06 cut off value, as well as the WRMR values. Moreover, the bifactor solution revealed poor goodness-of-fit indices. Therefore, we decided to retain the two-factor model as the most adequate solution (see. The standardized factor loadings for the whole sample as well as for males and females are shown in . The range of the factor loadings, for the total sample was from 0.36 (item 1) to 0.72 (item 3). All standardized factor loadings estimated were statistically significant (p < 0.01).
## Measurement invariance of the eav scores across gender
Given that the one-factor model evidenced good model fit, we therefore tested the measurement invariance of the EAV scores across gender and age. Prior to the analysis of
## Measurement invariance of the eav scores across gender
Given that the one-factor model evidenced good model fit, we therefore tested the measurement invariance of the EAV scores across gender and age. Prior to the analysis of measurement invariance across gender and age, we tested whether the two-factor model showed a reasonably good fit to the data in each group separately (see . Goodnessof-fit indices for males and females, as well as for the two age's groups were adequate. The configural invariance model in which no equality constraints across groups were imposed showed an adequate fit to the data. Next, a strong invariance model was tested with the item thresholds and factor loadings constrained to be equal across groups. The ∆CFI between the constrained and unconstrained models was under 0.01, indicating that strong measurement invariance across gender and age was supported for the EAV scores . Goodness-of-fit indices for the hypothetical models tested and measurement invariance across gender and age.
## Model
## Evidences of relation with other variables
We calculated the Pearson's correlation between EAV total score and the RADS, the M-CTS subscales, and the RADS subscales. As shown in, statistically significant associations were found between the EAV scores and the M-CTS and the RADS. Specifically, the EAV showed a positive and significant association between the EAV and the M-CTS subscales related to psychological aggression as a victim and all the physical aggression both as a victim and as a perpetrator. In addition, positive significant correlations were found between the EAV and the Anhedonia and Negative subscales of the RADS.
# Mediation analysis
With the aim to analyze the mediation effect, we used structural equation modeling (SEM). First, the direct effect of the M-CTS scores on attitudes towards violence without mediators was tested. The directly standardized path (β = −0.46, p < 0.001) was significant. Then, a partially-mediated model containing a mediator (depression) and a direct path from scores on the M-CTS to attitudes towards violence was tested. All the path coefficients were statistically significant. The results showed an acceptable fit of the model to the data [χ 2 (df = 15) = 20.35, χ 2 /df = 1.19; RMSEA = 0.036; SRMR = 0.051 and CFI = 0.981. These results revealed that scores on the M-CTS and depression have significant effects on attitudes towards violence among adolescents and youths.
Then, the mediating effects of depression on attitudes towards violence and scores of the M-CTS were tested for significance by adopting the Bootstrap estimation procedure in AMOS (a bootstrap sample of 1000 was specified). shows the indirect effects and their associated 95% confidence intervals. The indirect effect of the M-CTS on attitudes towards violence through depression was significant.
# Discussion
To date, violence against women and violence manifestations within the context of romantic relationships is becoming a world global issue. Specially worrying is the fact the IPV is starting earlier, with a larger number of adolescents involved in this kind of violence. Adolescence intimate partner violence rates are increasing affecting now to more than half of all dating youth. Nonetheless, to date, little is known about attitudes towards intimate partner violence across the world and, specifically, in Spain. Moreover, there is a lack of adequate and sound instruments measuring this.
The present study aimed, thus, to examine the prevalence, factorial structure, measurement invariance across gender, and reliability of the EAV scores, as well as its associations with intimate partner violence and depression symptoms, in a large sample of non-clinical adolescents. The study of evidences of an instrument such as the EAV allows generating and assess profiles of possible adolescents and youth that are more likely to engage in IPV. With this regard, present study reveals that the EAV is a short instrument with adequate evidences of validity and internal consistency of the scores for its use in educational settings like school or university, as well as clinical settings. This is particularly relevant, as it seems reasonable to think that early detection and promotion of positive attitudes towards intimate relationships may prevent IPV.
The results indicated that the EAV is an easy, simple, and brief tool in order to screen for violence attitudes. The study revealed adequate psychometric properties in Spanish adolescents. The internal consistency of the scores estimated by means of ordinal alpha was good. In addition, the results the EAV should be considered as a unidimensional factor structure. Furthermore, this structure was equivalent by gender, after the study of the Measurement Invariance. To date, no previous studies have analyzed, to the best of our knowledge, the factorial structure of the EAV scores in a non-clinical sample of adolescents, being the first study analyzing and the psychometric goodness of the EAV in a Spanish sample. Therefore, future studies should further analyze the extent to which this result are similar in other samples, in order to validate the results found in the present study. A previous study analyzing abuse in non-married couples used the EAV, assuming a unidimensional structure, but this assumption had still to be confirmed. Nowadays, there is still, not surprisingly, a lack of studies about the psychometric data on self-report measures, being in their nascent stage. Thus, new empirical studies need to replicate the findings established here.
The results revealed that attitudes toward violence were moderate associated with intimate partner violence and with depression symptoms. Specifically, higher rates of attitudes toward violence were associated to psychological aggression as a victim and to both medium and severe physical aggression both as a victim and as a perpetrator. Also, a positive relationship was found between the EAV and Anhedonia and Negative subscales of the RADS. This is somehow consistent with the idea that IPV is related to different issues like trauma and mental health problems, including depression. Worth noting, the correlations found between the RADS and the EAV were low. One possible explanation is that the EAV measures attitudes towards violence instead of IPV per se. Future studies should analyze the exact relation between these two constructs. Nonetheless, this study is one of the first posing the association between attitudes towards violence and these mental health issues. In addition, previous studies have established the relation between attitudes toward violence and explicit IPV. With regards to the mediation analysis, the results of the SEM revealed that depression mediated the relationship between attitudes towards violence and scores on the M-CTS. Moreover, the scores on the M-CTS had a statistically significant effect on attitudes towards violence. This is consistent with the idea that those adolescents who justify violence are more likely to engage in IPV, being depression a variable that may affect the outcome. More studies could further analyze this association.
The results of the present study should be interpreted in the light of the following limitations. First, measurement of violence attitudes, as well as depressive symptoms and violence manifestations were based solely on self-report and there are well-known inherent problematics like the effect of stigmatization, the possibility of misunderstanding of some items or the lack of introspection of some participants. Therefore, future studies should consider the use of external informants, interviews or even bio-behavioral and/or biological markers. Second, adolescence is a developmental period in which personality is still consolidating. Thus, the present results must be further evaluated in order to understand their natural developmental course. Third, no information was gathered regarding the participants' psychiatric morbidity or the use or abuse of substances, aspects that may partially influence the results. Finally, our data was cross-sectional in nature.
Despite the noted limitations, the present study allows confirming the adequate psychometric properties of the EAV, an instrument devoted to assess attitudes towards violence in intimate relationships in a large sample of Spanish adolescents. In addition, these attitudes seem to be related with depressive symptoms and violence manifestations during adolescence. The results have clear implications for the construct validity of the EAV and for its use in school populations in order to study intimate violence attitudes in adolescent populations. In addition, this study contributes relevant information to further understand the structure of and relations of attitudes towards violence, allowing the implementation of future preventive treatments. More research is needed in order to advance in the study of attitudes towards violence in intimate relationships settings and the role that they play in adolescents. Also, the study of measurement invariance of the EAV across other relevant variables like race or culture could also be relevant. In addition, the role of the attitudes toward violence in the prediction and transition to actual violence in intimate relationships during adolescence should continue to be explored in greater depth through independent longitudinal studies. Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.
# Data availability statement:
The data is not available as it contains sensible information about human people and consent was not obtained for its dissemination.
## Conflicts of interest:
The authors declare no conflict of interest. |
Somatic Mutation Profiles of MSI and MSS Colorectal Cancer Identified by Whole Exome Next Generation Sequencing and Bioinformatics Analysis
Background: Colorectal cancer (CRC) is with approximately 1 million cases the third most common cancer worldwide. Extensive research is ongoing to decipher the underlying genetic patterns with the hope to improve early cancer diagnosis and treatment. In this direction, the recent progress in next generation sequencing technologies has revolutionized the field of cancer genomics. However, one caveat of these studies remains the large amount of genetic variations identified and their interpretation.Methodology/Principal Findings:Here we present the first work on whole exome NGS of primary colon cancers. We performed 454 whole exome pyrosequencing of tumor as well as adjacent not affected normal colonic tissue from microsatellite stable (MSS) and microsatellite instable (MSI) colon cancer patients and identified more than 50,000 small nucleotide variations for each tissue. According to predictions based on MSS and MSI pathomechanisms we identified eight times more somatic non-synonymous variations in MSI cancers than in MSS and we were able to reproduce the result in four additional CRCs. Our bioinformatics filtering approach narrowed down the rate of most significant mutations to 359 for MSI and 45 for MSS CRCs with predicted altered protein functions. In both CRCs, MSI and MSS, we found somatic mutations in the intracellular kinase domain of bone morphogenetic protein receptor 1A, BMPR1A, a gene where so far germline mutations are associated with juvenile polyposis syndrome, and show that the mutations functionally impair the protein function.Conclusions/Significance: We conclude that with deep sequencing of tumor exomes one may be able to predict the microsatellite status of CRC and in addition identify potentially clinically relevant mutations.
# Introduction
Colorectal cancer is the third most common cancer with about 1 million cases worldwide. Over the last decades it has become clear that CRC evolves through multiple pathways and that these pathways can be roughly defined on the basis of molecular patterns such as the integrity of the mismatch repair system (MMR) or mutational and epigenetic patterns. Deficiency in the MMR is reflected in DNA microsatellite instability (MSI) which has also been associated with treatment outcome, but which needs to be further validated in additional clinical studies [bib_ref] Isolation and characterization of allelic losses and gains in colorectal tumors by..., Peinado [/bib_ref] [bib_ref] Microsatellite instability in cancer of the proximal colon, Thibodeau [/bib_ref] [bib_ref] Ubiquitous somatic mutations in simple repeated sequences reveal a new mechanism for..., Ionov [/bib_ref] [bib_ref] Clues to the pathogenesis of familial colorectal cancer, Aaltonen [/bib_ref] [bib_ref] Review of the Lynch syndrome: history, molecular genetics, screening, differential diagnosis, and..., Lynch [/bib_ref] [bib_ref] Microsatellite instability is a predictive factor of the tumor response to irinotecan..., Fallik [/bib_ref].
High-throughput Sanger sequencing studies on the other hand have shown that the mutation frequency of candidate cancer genes might be much higher than expected, and that the particular combination of mutations might influence the tumor's properties [bib_ref] The consensus coding sequences of human breast and colorectal cancers, Sjoblom [/bib_ref] [bib_ref] Core signaling pathways in human pancreatic cancers revealed by global genomic analyses, Jones [/bib_ref] [bib_ref] Patterns of somatic mutation in human cancer genomes, Greenman [/bib_ref] [bib_ref] The genomic landscapes of human breast and colorectal cancers, Wood [/bib_ref]. With the development of next-generation sequencing (NGS) technologies the sequencing throughput has dramatically increased and the costs have decreased. In addition, and especially important for clinical settings, NGS can be applied to formalin-fixed and paraffin embedded FFPE tissue material as well as highly degraded DNA which is routinely prepared in pathology departments or found in ancient DNA [bib_ref] Direct multiplex sequencing (DMPS)-a novel method for targeted high-throughput sequencing of ancient..., Stiller [/bib_ref] [bib_ref] Genome-wide massively parallel sequencing of formaldehyde fixedparaffin embedded (FFPE) tumor tissues for..., Schweiger [/bib_ref]. Several studies have used NGS technologies for the identification of the underlying mutation in monogenetic diseases [bib_ref] Targeted capture and massively parallel sequencing of 12 human exomes, Ng [/bib_ref] [bib_ref] Identity-by-descent filtering of exome sequence data identifies PIGV mutations in hyperphosphatasia mental..., Krawitz [/bib_ref]. However, only a limited number of studies report on next-generation sequencing to identify new candidate cancer genes; one of the earliest studies examined cytogenetically normal acute myeloid leukemia, and breast cancer genomes [bib_ref] DNA sequencing of a cytogenetically normal acute myeloid leukaemia genome, Ley [/bib_ref] [bib_ref] Mutational evolution in a lobular breast tumour profiled at single nucleotide resolution, Shah [/bib_ref]. In addition, studies on malignant melanoma and small-cell lung cancer cell lines have provided first insights into genomic alterations induced by ultraviolet light exposure or tobacco smoke [bib_ref] A comprehensive catalogue of somatic mutations from a human cancer genome, Pleasance [/bib_ref] [bib_ref] A small-cell lung cancer genome with complex signatures of tobacco exposure, Pleasance [/bib_ref].
To gain insight into the genomes of microsatellite stable and instable colorectal cancers and to identify functional relevant mutational patterns we used a hybridization based whole exome DNA capturing approach followed by 454 next generation sequencing [bib_ref] Exome sequencing of a multigenerational human pedigree, Hedges [/bib_ref]. Applying stringent bioinformatics analyses, we narrowed down the amount of functionally significant somatic mutations in MSI to 359 and 45 in MSS cancers, thus highlighting specific mutation patterns depending on the microsatellite status. We were able to confirm our results by sequencing the exomes of four additional CRC cases (one MSI, three MSS) using a different enrichment and sequencing technology. Among these mutations are BRAF in the MSI cancer and KRAS and TP53 in the MSS cancer, further underscoring the validity of our selection approach [bib_ref] Integrated genetic and epigenetic analysis identifies three different subclasses of colon cancer, Shen [/bib_ref]. Further functional characterizations identified recurrent somatic mutations in BMPR1A, a protein which has been associated so far with juvenile polyposis syndrome, a cancer predisposition syndrome.
# Results
## Sequence-specific enrichment and sequencing strategy
We sequenced tumor and matching normal colon tissues from two patients with high grade adenocarcinoma of the colon (G3), patient 1 with a microsatellite instable and patient 2 with a microsatellite stable tumor (Table 1, [fig_ref] Figure 1: Qualities of the targeted whole exome sequencing approach [/fig_ref]. For the determination of germline mutations we sequenced in addition to the tumor tissues from each patient adjacent not affected normal colonic tissue. Using Illumina sequencing and SNP arrays we determined that the tumor of patient 1 is copy number stable whereas patient 2 showed variations which we used for the reevaluation of identified high stringency mutations .
We analyzed the complete exomes of more than 135,000 exons with single-read shotgun 454 sequencing [fig_ref] Figure 1: Qualities of the targeted whole exome sequencing approach [/fig_ref] , [fig_ref] Figure 1: Qualities of the targeted whole exome sequencing approach [/fig_ref] , [fig_ref] Figure 2: Identification process of somatic relevant SNVs [/fig_ref] , [fig_ref] Table 2: Tumor and normal genome coverages from MSI and MSS cancer patients [/fig_ref]. To assess the effect of coverage depth on the sensitivity and specificity of sequence variant detection, genotype calls of the Affymetrix SNP array 6.0 were compared step-wise to the called nucleic acid positions and resulted in an accuracy of more than 99% [fig_ref] Figure 3: Characterization of primary identified SNVs [/fig_ref]. In addition to the SNP array, we used Sanger sequencing to confirm 23 selected mutations , .
## Identification of somatic mutations in coding sequences for a msi crc
Searching for variants in coding regions we found 12,767 and 12,518 small nuclear variations in 6,428 and 6,205 genes, for tumor and normal respectively. Of these variants 1,428 for control and 2,404 for tumor have an average heterozygosity or minor allele frequency lower than 1% or have not been previously reported in dbSNP or the 1000 Genomes Project [fig_ref] Figure 2: Identification process of somatic relevant SNVs [/fig_ref]. Since indels (small insertions and deletions) at homopolymeric sites are a major source of sequencing errors of the 454 platform we ignored this type of alteration in our analyses.
Our somatic variant detection strategy was designed to minimize false positive somatic variant calls rather than to determine zygosity. We used a two-step approach with two different stringency levels to detect variants in tumor and benign tissues similar to Pleasance et al. [bib_ref] A comprehensive catalogue of somatic mutations from a human cancer genome, Pleasance [/bib_ref]. In the first step, tumor variants were called under stringent criteria, which were determined by comparison to the SNP genotyping array data. The second step ascertained whether the tumor variant was germline or somatic. To keep the false negative rate in the benign tissue at less than 10%, we set the coverage cut-off at 5-fold, below which no conclusions were drawn regarding whether a variant was somatic or germline. If the coverage cut-off was met, a single read showing the same variant in benign and tumor tissue resulted in the categorization of the variant as germline.
Using this strategy we identified 915 somatic non-synonymous mutations affecting 864 genes. The majority of somatic mutations were missense mutations (65%). However, many (7%) are located within untranslated regions of genes and might therefore result in altered expression or increased decay of mRNA species. Furthermore, approximately 0.5% of these mutations are found at splice sites and could influence splicing events, leading to an altered transcriptome structure. In addition, three somatic variants were identified in miRNA regions . These mutations are of particular interest because miRNAs have been implicated as master regulators of tumor homeostasis. Analysis of the specific types of nucleic acid variations, including known and unknown variants, showed essentially the expected rates of nucleotide exchanges, as determined by calculations using dbSNP130 .
## Functional analysis of mutations for a msi colon cancer
Since not all of the 1,304 somatic mutations are likely to be pathologically relevant, we sought to identify those that probably destroy protein function or affect highly conserved amino acids and might therefore be functionally important. We used Polyphen and MutationTaster classification tools to predict the functional consequences of amino acid changes or frameshift mutations and found that 359 genes had at least one potentially destructive mutation [fig_ref] Table 1: Colorectal cancer patients selected for NGS [/fig_ref] [bib_ref] A method and server for predicting damaging missense mutations, Adzhubei [/bib_ref] [bib_ref] MutationTaster evaluates disease-causing potential of sequence alterations, Schwarz [/bib_ref].
Of the potentially destructive somatic mutations, 309 were located in positions highly conserved in 44 different species, including opossum (Monodelphis domestica), chicken (Gallus gallus) and lamprey (Petromyzon marinus). Of these, 259 were located in genes expressed in the colon, of which 47 were repair, receptor, or kinase genes. Visualization of selected mutations on protein structures indicates that these nucleotides are on the protein surface, potentially resulting in disrupted protein-protein interactions.
The Catalog of Somatic Mutations in Cancer (COSMIC) database is a comprehensive collection of cancer-related muta- tions. Approximately 39% of our mutated genes were already described in this database, and 13% were found by Wood et al [bib_ref] The genomic landscapes of human breast and colorectal cancers, Wood [/bib_ref]. As did these previous databases, we found the BRAF p.V600E mutation in the MSI case and we identified KRAS and TP53 mutations in the MSS tumor. BRAF mutations are found in approximately 10% of CRCs, predominantly MSI and 30 to 35% of all patients with sporadic colorectal cancers carry somatic KRAS and TP53 mutations. These findings further demonstrate the sensitivity of our classification strategy.
## Mutational landscape of a mss colon cancer
For the MSS colon cancer we identified 10,622 small nuclear variations. After the same filtering processes as for the MSI cancer using the dbSNP database and the data from the 1000 Genomes Project 1,288 variants remained which either had low prevalence or were unknown. Of these, 198 were somatic and 45 were predicted to alter gene function based on MutationTaster and Polyphen calculations [fig_ref] Figure 2: Identification process of somatic relevant SNVs [/fig_ref] , [fig_ref] Table 2: Tumor and normal genome coverages from MSI and MSS cancer patients [/fig_ref] [bib_ref] A method and server for predicting damaging missense mutations, Adzhubei [/bib_ref] [bib_ref] MutationTaster evaluates disease-causing potential of sequence alterations, Schwarz [/bib_ref]. In regard to copy number variations five of the 45 identified mutations are located within amplified regions, and, as expected, none in regions with deletions. The ratios of reads with reference sequence to mutated sequence are not exceeding ratios in copy number stable areas which supports the SNV-calling algorithm.
In contrast to 1,304 somatic mutations in the MSI tumor we found 198 somatic mutations in the MSS tumor which demonstrates that the defective MMR system in MSI tumors results in a significant increase in mutation rates in colorectal cancer.
Furthermore, looking at intersections between both cancer types we found BMPR1A, WDTC1 (WD and tetratricopeptode repeats 1) and EHD3 (EH-domain containing 3) mutated in both tumors. The selection was based on functional impairment with high probability in Polyphen and MutationTaster [bib_ref] A method and server for predicting damaging missense mutations, Adzhubei [/bib_ref] [bib_ref] MutationTaster evaluates disease-causing potential of sequence alterations, Schwarz [/bib_ref]. All mutations are located on the surface of the protein and are highly conserved. Since we found significant cancer-related pathways associated only with BMPR1A but not with WDTC1 or EHD3, and in addition germline mutations in BMPR1A are a known risk factor for juvenile polyposis syndrome, we chose BMPR1A for additional functional assays [fig_ref] Figure 3: Characterization of primary identified SNVs [/fig_ref] , . Using reporter assays with wild type and mutated BMPR1A proteins we were able to show that the mutated proteins are strongly impaired in their signalling function and that stimulation with BMP2 results in a reduced maximum activity [fig_ref] Figure 3: Characterization of primary identified SNVs [/fig_ref]. The analyses presented so far have been based on 454 whole exome sequencings of one colorectal cancer patient for each microsatellite status. To further confirm that the increased amount of coding mutations in MSI cancers can be generalized and is not due to the technology used ('array' enrichment and 454 sequencing) we sequenced the exomes of four additional colorectal cancers, each with matching normal tissues. This time we used 'in solution hybridization' for capturing of DNA followed by SOLiD sequencing. After the same filtering procedures as described for the first two patients we again determined up to 8-fold higher mutation rates for the MSI colorectal cancer than for MSS cancers . In this regard we found 532 non-synonymous somatic SNVs in the additional MSI CRC and only 65, 74 and 76 in the three MSS CRC cases. Thus, the differences in mutation rates are reproducible and independent of the sequencing technology used.
# Discussion
Using next-generation sequencing, we sequenced the exomes of MSS and MSI colon cancer patients, with mean coverages of approximately 20-fold. We applied a two-sided classification algorithm to uncover functionally relevant mutations. Using this approach, we demonstrate for the first time that an array-capture The functional impact of the somatic variations was predicted using two functional prediction algorithms, Polyphen and MutationTaster, and we found 359 somatic mutations for the MSI and 45 for the MSS cancer that are highly likely to cause functional impairment [bib_ref] A method and server for predicting damaging missense mutations, Adzhubei [/bib_ref] [bib_ref] MutationTaster evaluates disease-causing potential of sequence alterations, Schwarz [/bib_ref]. The heterogeneity of mutated genes suggests that not a specific gene per se but the affected pathway plays a major role for tumor development. In this regard we find a significant enrichment of mutations in cancer-related pathways such as cancer, cellular development and DNA replication, recombination and repair . Interestingly, we find 50% of the most significant enriched pathways in the MSS cancer also as significantly enriched pathways for the MSI cancer, indicating that even though MSI cancers harbour an increased number of mutations both cancers might develop through overlapping pathomechanisms.
Historically, microsatellite testing in colorectal cancers was the first predictive test for the identification of an underlying mismatch repair (MMR) mutation. Since more than 90% of hereditary nonpolyposis colorectal cancers (HNPCC) show MSI, the microsat- Activity of wt mBMPR1A, mBMPR1A E502G and mBMPR1A W487R was determined in C2C12 cells using a SMAD-responsive Luciferase reporter gene assay. Induced Luciferase activity was normalized to Renilla acitivty. The activity of untransfected cells was set to 0% and the activity of wt mBmpr1a was set to 100%. Significant differences were calculated with a two-tailed t-test and marked as: * p#0.05, ** p#0.01, *** p#0.001. doi:10.1371/journal.pone.0015661.g003 . Distribution of SNVs in MSI and MSS tumors. ellite status has become a common diagnostic marker of MMR. In addition, survival advantages and therapeutic consequences have been reported for patients with MSI tumors [bib_ref] Review of the Lynch syndrome: history, molecular genetics, screening, differential diagnosis, and..., Lynch [/bib_ref] [bib_ref] Ultradeep sequencing of a human ultraconserved region reveals somatic and constitutional genomic..., De Grassi [/bib_ref]. Using ultradeep sequencing of one conserved region, UCR41, de Grassi and colleagues show that this region has higher mutation rates in HNPCC samples than in healthy controls and suggest that this might be used as sensitive molecular assay of genomic instability [bib_ref] Ultradeep sequencing of a human ultraconserved region reveals somatic and constitutional genomic..., De Grassi [/bib_ref]. We have extended their analyses to whole exomes and found 8-fold differences in the numbers of somatic mutation of MSI and MSS colorectal cancers. In comparison, a study by which reported on the sequencing of 518 protein kinase genes in 210 diverse human cancers found an approximately 25-fold higher mutation rate for MMR-deficient cancers [bib_ref] Patterns of somatic mutation in human cancer genomes, Greenman [/bib_ref]. However, these are extrapolations and in contrast to our study they included tumors from different origins and examined all somatic mutations irrespective of their functional relevance. With our sequencing approach we also detected a somatic MLH3 mutation in the MSI tumor, which, even though MLH3 mutations do not belong to the classical MMR mutations in CRC, might contribute to the microsatellite instability phenotype [bib_ref] Germline and somatic mutation analyses in the DNA mismatch repair gene MLH3:..., Lipkin [/bib_ref]. Furthermore, the combination of MSI and BRAF mutation, as detected for the MSI tumor described, is most frequently found for CpG island methylation phenotype 1 (CIMP1) tumors which are associated with MLH1 promoter methylations [bib_ref] Integrated genetic and epigenetic analysis identifies three different subclasses of colon cancer, Shen [/bib_ref] [bib_ref] CpG island methylator phenotype in colorectal cancer, Toyota [/bib_ref]. The promoter methylation in turn is associated with gene silencing mechanisms which is suggestive as an explanation for the MSI status of the tumors. On the other side, it has been proposed that chromosomal instability (CIN) and CIMP represent two independent and inversely related mechanisms of instability [bib_ref] The CpG island methylator phenotype and chromosomal instability are inversely correlated in..., Goel [/bib_ref]. CIMP-negative cases are associated with p53 (71%) and KRAS (33%) mutations, but are rarely found with BRAF (2%) mutations. Since we found large copy number variations as well as KRAS and TP53 mutations in the MSS tumor analyzed this tumor is most likely CIMPnegative [bib_ref] Review of the Lynch syndrome: history, molecular genetics, screening, differential diagnosis, and..., Lynch [/bib_ref] [bib_ref] Ultradeep sequencing of a human ultraconserved region reveals somatic and constitutional genomic..., De Grassi [/bib_ref]. The sequencing and analysis strategy we have presented might be the basis for future classification tools for colorectal cancers because it may allow a parallel detection of an increased mutation frequency in MSI tumors as well as the detection of the underlying MMR defect. In addition, we were able to detect mutations of genes frequently associated with certain subtypes of colorectal cancers such as BRAF, KRAS and TP53. Within our high priority genes, encompassing all genes which pass all selection filters, BMPR1A stands out as mutated in both cases. The overall structure reveals that the mutated amino acids are all located at the C-terminal intracellular helix bundle at the protein kinase domain and suggests that protein-protein interactions are destroyed [bib_ref] The prevalence of MADH4 and BMPR1A mutations in juvenile polyposis and absence..., Howe [/bib_ref]. Germline BMPR1A mutations predispose to juvenile polyposis syndrome; however, our findings indicate that also somatic mutations might play an important role in sporadic colorectal cancer development [bib_ref] Inactivation of the type II TGF-beta receptor in colon cancer cells with..., Markowitz [/bib_ref] [bib_ref] Germline mutations of the gene encoding bone morphogenetic protein receptor 1A in..., Howe [/bib_ref] [bib_ref] Mutations in the SMAD4/DPC4 gene in juvenile polyposis, Howe [/bib_ref] [bib_ref] DPC4, a candidate tumor suppressor gene at human chromosome 18q21, Hahn [/bib_ref] [bib_ref] Cancer genetics: colorectal cancer as a model, Bodmer [/bib_ref].
Besides these mutations we have also identified several mutated cancer drug targets or genes that are associated with treatment outcome, including BRAF, KRAS, FGFR2 and MTOR, which might help to choose optimal drug combinations. As such similar targeted re-sequencing approaches and bioinformatics filtering strategies might become a gold standard for individually tailored colorectal cancer treatment in the future.
# Materials and methods
# Ethics statement
The study has been approved by the Ethical Committee of the Medical University of Graz. For new samples patients have given their written informed consent. For old samples (15 years old) no informed consent was available, therefore all samples and medical data used in this study have been irreversibly anonymized.
## Case presentation and tissue sample collection
Patient 1 had a high-grade (G3) adenocarcinoma of the proximal colon, staged pT-3C, pN-0, pM-X, pR-0, microsatellite instable [fig_ref] Figure 1: Qualities of the targeted whole exome sequencing approach [/fig_ref]. In addition, this case was selected because of its chromosomal stability, as determined using genome-wide next generation sequencing (NGS) [fig_ref] Figure 1: Qualities of the targeted whole exome sequencing approach [/fig_ref]. Patient 2 had a highgrade (G3) adenocarcinoma of the proximal colon, staged pT-4B, pN-2, pM-X, microsatellite stable.
Human tissue obtained during surgery was snap-frozen in liquid nitrogen. Cryosections (3 mm thick) were prepared and stained with haematoxylin and eosin to evaluate tumor cell content. Dissections were performed under the microscope to achieve a tumor cell content of .80%. DNA isolation was performed using the QIAamp DNA Mini Kit (Qiagen), according to the manufacturer's instructions.
## Whole exome dna enrichment and genome sequencer flx sequencing
Genomic DNA of both tissues was subjected to whole exome sequence capture using Roche/NimbleGen's 2.1M Human Exome Array. This array is based on build 36.3 of the human genome sequence, and captures the coding regions of 16,755 NCBI RefSeq genes (approximately 180,000 coding exons) as well as 493 miRNA regions. Tumor and normal tissue DNA were subjected to whole exome sequence capturing according to the manufacturer's protocol. DNA was sheared by nebulization to fragment sizes below 800bp, cleaned (Zymo Research) and endpolished using T4 DNA Polymerase and T4 Polynucleotide Kinase. Linker adapters pSel3 (59 -CTCGAG AAT TCT GGA TCC TC -39) and pSel4-P (59 -Phos/GAG GAT CCA GAA TTC TCG AGT T -39) were ligated and size selection was performed using AMPure DNA Purification Beads (Agencourt). Quality was controlled with the Bioanalyzer system. LM-PCR was performed with LMPCR3 primers (59 -CUC GAG AAU UCU GGA UCC UC -39) before the library was used for hybridization at 42uC for 72h. The arrays were washed two times at 47.5uC, two times at room temperature and two times at 42uC with washing buffers as recommended. Bound genomic DNA was eluted with 125 mM NaOH for 10 min at room temperature and amplified by LM-PCR using primers LMPCR3. Captured amplified samples were subjected to quantitative PCR to measure the relative enrichment.
The enriched Nimblegen DNA was used to construct singlestranded Genome Sequencer FLX (454/Roche) libraries. After emulsion PCRs sequencing primers were annealed to the template and beads were incubated with Bst DNA polymerase, apyrase, and single-stranded binding protein. Pyrosequencing was performed on a 70675 mm picotiter plate in 13 separate sequencing runs. After default raw data processing, a resequencing trimming filter was used to increase the data output. (Parameters used: doValleyFilterTrim-Back = false, vfBadFlowThreshold = 6, vfLastFlowToTest = 168, errorQscoreWindowTrim = 0.01).
For the sequencing we performed 13 Genome Sequencer FLX runs, which produced over 558 million bases and 1.43 million reads per run. Reads were aligned to the human reference genome, NCBI build 36 (http://hgdownload.cse.ucsc.edu/ goldenPath/hg18/), using GS Reference Mapper Version 2.0.0.12 (Roche). The best matches in the genome were used as the location for the reads with multiple matches. Only unique reads with a minimum length of 50 bp were used for further analysis (see run statistics Tab. 2).
Detection of variants was performed with the GS Reference Mapper Version 2.0.0.12 (Roche). Redundant reads were subtracted before variant callings. Only the HCDiff (high confidence differences) of the GS Mapper software were used as basis of variant detection [bib_ref] Exome sequencing of a multigenerational human pedigree, Hedges [/bib_ref]. HCDiff callings presume at least three reads with the variant with both forward and reverse reads included; alternatively the quality scores at the variable positions must be over 20 (or over 30 if a homopolymer of five or more bases is involved). As additional quality criteria we used only variants with a coverage .106 of high quality reads.
## Whole exome 'in solution' dna enrichment and solid sequencing
Enrichments and SOLiD library preparation were performed according to Agilent's SureSelect Target Enrichment protocol for the Applied Biosystems SOLiD system. In brief, whole genomic DNA was sheared and end repaired. For adapter ligations 30x excess of the adapters were used. Size selections for 150-200 bp DNA fragments were performed followed by a nick-translation and amplification step with Platinum polymerase (Invitrogen) and Pfu-Polymerase (Fermentas). For hybrid selection the libraries were adjusted to 500 ng in 3.4 ml volume and added to the SureSelect Block solutions. Hybridizations were performed for 24 h at 65uC, hybrids were extracted with 500 ng M-280 streptavidin Dynabeads (Invitrogen) and finally eluted with 50 ml Elution buffer. After amplification with Platinum polymerase the libraries were quantified by qPCR and DNA concentration was titrated to achieve a fraction of 10-20% monoclonal template beads in the emulsion PCR using in total 0.7 to 1 billion beads. Successive bead enrichment and deposition of 130 million beads per quarter slide (quad) was followed by standard 50 bp fragment runs. Each of the four patient samples was analyzed on a single quad.
Mapping was performed with the Bioscope alignment pipeline using the seed & extend algorithm with a mismatch penalty score 22.0. Single Nucleotide Variants (SNV) were called with the DiBayes algorithm integrated in the Bioscope package.
## Single nucleotide variant (snv) detection
Tissue materials were genotyped on the Affymetrix 6.0 array, according to the manufacturer's protocol. Array positions with a quality score (p-value) ,0.1 were used for comparison with the sequencing data. Sequencing data positions were used if their coverage exceeded 3-fold. This generated 46,000 and 49,000 positions for tumor and benign tissue, respectively, that were eligible for comparison. To determine false positive and false negative rates, we set the array data as standard and distinguished between reference call and SNP call dependence on the array data.
For the detection of somatic variants, a bimodal strategy was applied with tumor variants called under stringent criteria, whereas variants in control tissue were called using less stringent criteria: A minimum threshold for reads was set with variants of 15% of all reads at a given position in tumor. Less stringent criteria were used for calling control tissue variants with a minimum of one variant read and a minimum coverage cutoff of 5. For coverages above 30-fold one variant read was accepted.
## Capillary sequencing
Follow-up confirmation of identified SNVs was performed on an ABI 3730 (Applied Biosystems) capillary sequencing instrument following standard procedures.
## Determination of copy number variations
Preparation of single read libraries and sequencing were performed using the Solexa sequencing platform (GenomeAnalyzer IIx, Illumina) following the manufacturer's instructions. Image Analysis and base calling were performed using Firecrest 1.9.5_14 and Bustard 1.9.5_14 and reads were aligned to the human genome (NCBI36) using Bowtie 0.9.7.1 [bib_ref] Ultrafast and memoryefficient alignment of short DNA sequences to the human genome, Langmead [/bib_ref]. Copy number analysis was done in R using the DNAcopy package [bib_ref] A faster circular binary segmentation algorithm for the analysis of array CGH..., Venkatraman [/bib_ref]. In short, DNA read frequencies were determined for bins of 50 Kb. The log2 frequency ratio of corresponding bins was calculated for tumor versus normal tissue. Median of ratios was centered to zero experiment wise. Log ratios were smoothed by DNAcopy using default values and copy number variation was detected by DNAcopy using a threshold of two standard deviations.
Genotyping on the Affymetrix 6.0 array was performed according to the manufacturer's protocol. Regions of copy number gain and loss were determined by paired and analysis using the Hidden Markov Model (HMM) of the Partek Genomics Suite software (Partek Inc, St.Louis, MO) with default parameter settings. For paired analysis, copy number values were generated by comparing tumor and benign tissue profiles from the same patient.
## Bioinformatics workflow
For each tissue, variations were annotated using the gene models generated by Ensembl (ensembl 54.36, www.ensembl.org). All variations were mapped to all transcript models, which led to multiple annotations for several loci. For instance, a variant can lead to an amino acid change in one transcript and appear in the UTR of another. Comparison of tumor and benign tissue variants to dbSNP130 and 1000 Genomes Project data was carried out with the subset of dbSNP130 and 1000 Genomes Project positions with minor allele frequencies or average heterozygosity .0.01. Variants were subjected to many comparisons with external data sources. Most data sources are integrated in the UCSC genome browser (http://hgdownload.cse.ucsc.edu/goldenPath/hg18/ database/) or were derived from websites like the 1000 Genomes Project (ftp://ftp.1000genomes.ebi.ac.uk/vol1/ftp/release/2009_ 04/) the gene ontology data (http://archive.geneontology. org/full/2009-10-01/go_200910-termdb.obo-xml.gz), the cosmic database version 46 (http://www.sanger.ac.uk/genetics/CGP/ cosmic/) or the cancer gene census database (http://www. sanger.ac.uk/genetics/CGP/census/). Functional classifications were performed using Polyphen and MutationTaster classification tools (http://genetics.bwh.harvard.edu/pph/, http://neurocore. charite.de/MutationTaster/) [bib_ref] A method and server for predicting damaging missense mutations, Adzhubei [/bib_ref] [bib_ref] MutationTaster evaluates disease-causing potential of sequence alterations, Schwarz [/bib_ref]. Base conservation among 44 species was tested using the phyloP track of the UCSC browser (http://genome.cshlp.org/content/early/2009/10/26/gr. 097857.109.abstract). Bases were considered highly conserved if their conservation score was greater or equal 2.0 (0.975 quantil of all conservation scores). For gene expression healthy control samples from http://www.ncbi.nlm.nih.gov/sites/GDSbrowser?acc = GDS2609 were used. We calculated genewise mean expression values across all samples and used the first quartile as threshold to determine gene expression. Pathway analyses were performed with the ingenuity pathway analysis tool (http://www.ingenuity.com).
All new data from this study has been deposited at NCBI dbSNP database of genetic variation (user-name MPIMGCancerogenomics). Accession numbers are included as [fig_ref] Table S4: Ingenuity pathway analysis of the MSI and MSS CRC [/fig_ref].
## Protein structure modeling
Models for BMPR1A were obtained from SwissModel and ModBase. Very similar models except for some loops (total Calpha rmsd 0.66) were rendered in PyMol [bib_ref] The SWISS-MODEL Repository and associated resources, Kiefer [/bib_ref] [bib_ref] MODBASE, a database of annotated comparative protein structure models, and associated resources, Pieper [/bib_ref].
## Reporter assays
The activity of the wildtype mouse protein (wtmBMPR1A) and its mutants was determined by measuring induced Luciferase activity in the transiently transfected pre-myoblastic mouse cell line C2C12 (ATCC). WtmBMPR1A was amplified from mouse cDNA and cloned in the expression vector pCS2+. Both mutations W487R and E502G were inserted by Quikchange mutagenesis (Stratagene) using the following primer pairs: mBmpr1a_W487R_fwd caatcgtgtctaaccgcCggaacagcgatgaatg; mBmpr1a_W487R_rev cattcatcgctgttccGgcggttagacacgattg and mBmpr1a_E502G_fwd gttttgaagctaatgtcagGatgttgggcccataatc; mBmpr1a_E502G_rev gattatgggcccaacatCctgacattagcttcaaaac. C2C12 cells were cultured in DMEM glucose 4,5 g/L with 10% FCS were co-transfected with each Bmpr1a expression construct, a Smad Binding Element (SBE) luciferase construct [bib_ref] Identification and functional characterization of a Smad binding element (SBE) in the..., Jonk [/bib_ref] and the normalization vector pRL-Tk (Promega Corporation, Madison, WI, USA) using Turbofect (Fermentas GmbH, St. Leon-Rot, Germany). Luciferase activity was determined as described previously [bib_ref] A protocol for combined Photinus and Renilla luciferase quantification compatible with protein..., Hampf [/bib_ref].
## Supporting information
[fig] Figure 1: Qualities of the targeted whole exome sequencing approach. (A) Venn diagram of captured exons of normal and tumor samples. Captured exons with at least one read were counted. (B) Representative normalized coverage-distribution plot. The fraction of bait-covered exons in the genome achieving coverages equal or lower than the normalized coverage is indicated on the x-axis. The mean coverage per exon was divided by the mean coverage of all exons. doi:10.1371/journal.pone.0015661.g001 [/fig]
[fig] Figure 2: Identification process of somatic relevant SNVs. (A) Schematic of the bioinformatics SNV detection workflow. (B) Extraction of functionally relevant somatic mutations for MSI and MSS colorectal cancers. Variants were detected with the GS Reference Mapper before they were filtered for their localization, annotation in dbSNP130 or the 1000genomes, somatic and functionally impairment. From dbSNP130 or the 1000genomes variants with frequencies above 1% were used. For MSI CRC 359 variants and for MSS CRC 45 with predicted altered protein functions were identified. doi:10.1371/journal.pone.0015661.g002 NGS one-step work flow is a powerful tool for deep characterization of solid tumors and show that MSI tumors carry eight times more functional relevant mutations than MSS tumors (Figure 2). [/fig]
[fig] Figure 3: Characterization of primary identified SNVs. (A) Proportional Venn diagram. Fractions of called SNVs identical to the Genomes Project data and dbSNP130. Only data for which the minor allele frequency or the average heterozygosity was known and below 1% were used for comparison. (B) Distribution of synonymous, missense, nonsense and mutations affecting the start or stop codon are shown in relation to all somatic mutations. (C) BMPR1A mutations p.W487R and p.E502G are located at the protein kinase domain of BMPR1A. Reference amino acids are in green, the mutated forms are shown in red. The net structure at the left lower side indicates the ATP binding domain. (D) BMPR1A mutations show decreased signaling acitivity. [/fig]
[fig] Figure S1: Quality controls of the colon cancer case 1 and experimental performances. (A) Visualization of Copy number variations (CNV) using Illumina sequencing for MSI and MSS cancers. Chromosomal coverage ratio of tumor versus benign tissue sample. Each chromosome was divided into 50-kb bins. The log2 ratios of unique reads per bin are plotted across all chromosomes. The red lines depict the local averages as calculated by DNAcopy [35]. (B) Influence of sequencing depth on exon capture coverage (left) and SNV detection (right). Exon coverage and SNVs in the enrichment regions were determined after each sequencing run. The numbers of exons covered and the number of SNVs detected at different coverage levels were compared for tumor and benign tissue separately. Sigmoid functions Y = c+ (d-c)/(1+exp(b)*(log(X)-log(e)) were used to fit the data and extrapolate the saturation level. (TIF) Figure S2 Sequence coverage along a contiguous target. (A) The base-by-base sequence coverage along a typical 80-kb segment (BRCA1 gene) in the UCSC browser is shown. The 10-fold coverage level is highlighted by a black line. (B) Coverage profiles of exon targets depending on exon size. Exons have been divided into four groups depending on exon size. Coverages were calculated in relation to the relative position on the exon and averaged by the mean over all exons of the group. (TIF) Figure S3 Comparison between the SNP array and NGS. About one million known SNP positions have been investigated using the Affymetrix human whole genome SNP array 6.0. Array positions with a quality score (p-value),0.1 and sequencing positions with coverage exceeding 3-fold coverage were used for comparison. Forty thousand and thirty-six thousand positions for tumor and benign tissue, respectively, were eligible for comparison. To determine false positive and false negative rates, the array data was set as standard and between reference call and SNP call dependence on the array data was distinguished. (A) homo-and heterozygous SNVs were discerned (B) for the calculation of the haploid concordances heterozygous positions were counted as homozygous non-reference positions. (TIF) Figure S4 Nucleotide exchange rates in DNA from tumor and benign tissue, as compared to dbSNP130 data for patient 1. (A) Using the GS Reference Mapper Version 2.0.0.12 (Roche), software nucleotide exchanges were calculated for all possible transitions (e.g. A,.G, A,.C). Dark grey: tumor, grey: benign, light grey: dbSNP (B) Dinucleotide context for single nucleotide variants from tumor and benign tissue. Dark grey: tumor, grey: benign, light grey: dbSNP (TIF) Figure S5 Validation, visualization and pathway analyses. (A) Visualization of the Sanger and 454 next generation sequencing result of BMPR1A. Red arrows indicate the location of the mutation. (B) Ingenuity pathway analysis of BMPR1A, WDTC1, EHD3 and CTR9 (top) and visualization of the conservation of BMPR1A p.W487 and p.E502 across human, mouse, chicken, zebrafish and other organisms (bottom). (TIF) Table S1 List of 359 somatic candidate genes with functionally relevant mutations for the MSI colorectal cancer case. Column headings are as follows: (A) Location, (B) Mutation, (C) coverage in tumor, (D) number of reads with SNV in tumor, (E) coverage in normal, (F) number of reads with SNV in normal, (G) amino acid position, (H) amino acid, (I) mutated amino acid, (J) gene name, (K-M) Ensembl transcript ID, Ensembl gene ID, Ensembl protein ID, (N) nucleotide conservation (PhyloP), (O) protein domain, (P) Polyphen, (Q) MutationTaster, (R) mean expression in colon, (S) described in Wood et al.2007, (T) listed in COSMIC database, (U) listed as CancerGeneCensus (dom = dominant (oncogene), rec = recessive (tumor suppressor)), (V-Y) listed in GO database as repair gene, kinase, receptor, transmembrane receptor, (Z-AJ): annotated within dbSNP130 (rs numbers indicate mutations with frequencies above 1%), 1000 genomes, Venter genome, Watson genome, Yoruban genome, Corean genome, Han genome, genome 12891, genome 12878, genome 12892, genome 19240. (XLS) Table S2 List of 45 somatic candidate genes with functionally relevant mutations for the MSS colorectal cancer case. Column headings are as follows: (A) Location, (B) Mutation, (C) coverage in tumor, (D) number of reads with SNV in tumor, (E) coverage in normal, (F) number of reads with SNV in normal, (G) amino acid position, (H) amino acid, (I) mutated amino acid, (J) gene name, (K-M) Ensembl transcript ID, Ensembl gene ID, Ensembl protein ID, (N) nucleotide conservation (PhyloP), (O) protein domain, (P) Polyphen, (Q) MutationTaster, (R) mean expression in colon, (S) described in Wood et al.2007, (T) listed in COSMIC database, (U) listed as CancerGeneCensus (dom = dominant (oncogene), rec = recessive (tumor suppressor)), (V-Y) listed in GO database as repair gene, kinase, receptor, transmembrane receptor, (Z-AJ): annotated within dbSNP130 (rs numbers indicate mutations with frequencies above 1%), 1000 genomes, Venter genome, Watson genome, Yoruban genome, Corean genome, Han genome, genome 12891, genome 12878, genome 12892, genome 19240. (XLS) Table S3 Copy number variations (CNVs) of the MSS colon cancer case. Somatic amplifications and deletions were determined with the Affymetrix 6.0 array followed by a paired Hidden Markov Model analysis with the Partek genomics Suite software. Column headings are as follows: Chromosome, start and end position of the CNV, cytoband, copy number, length in bp and CNV state. (XLS) [/fig]
[table] Table 1: Colorectal cancer patients selected for NGS. [/table]
[table] Table 2: Tumor and normal genome coverages from MSI and MSS cancer patients. [/table]
[table] Table S4: Ingenuity pathway analysis of the MSI and MSS CRC. Selected were pathways listed in the top 25 significantly enriched pathways for the MSS cancer which were also found to be highly significant in the MSS cancer. (XLS) [/table]
[table] Table S7: NCBI numbers of the identified SNVs. (XLS) [/table]
|
Eosinophilic Esophagitis in Children: Clinical Findings and Diagnostic Approach
Eosinophilic esophagitis (EoE) is an emerging chronic immune and antigen-mediated clinicopathologic disease. During the last 2 decades, the incidence of this condition in children has increased significantly, thanks to practitioners for creating the awareness and higher use of diagnostic endoscopy. We have analysed paediatric literature on EoE focusing on the epidemiology, pathophysiology, clinical findings and diagnostic approach.EoE is pathogenically related to a Th2 inflammation characterized by a mixed IgE and non-IgEmediated reaction to food and/or environmental agents. This leads to esophageal dysfunction and remodeling accompanied by subepithelial fibrosis. EoE can be presented with several range of gastrointestinal symptoms, including regurgitation, vomiting, feeding difficulties or feeding refusal in infants and toddlers, as well as heartburn, dysphagia and food bolus impaction in older children and adults. The diagnostic suspicion is based on the presence of chronic symptoms of esophgeal dysfunction and esophageal eosinophilia characterised histologically by a significant eosinophilic infiltration of the oesophageal mucosa (>15 eosinophils per high powered field). In this review, we will provide an update on clinical presentation and diagnostic approach to EoE in children. We emphasized on the relevant aspects of the new clinical condition termed "PPI responsive esophageal eosinophilia", as entities distinct from EoE and the role of PPI trial in the diagnostic workup, therefore we proposed a new diagnostic algorithm.
# Introduction
Eosinophilic esophagitis (EoE) is an emerging chronic immune and antigen-mediated clinicopathologic disease, affecting both children and adults. This condition is characterized by severe eosinophil-predominant inflammation into the esophageal epithelium resulting in esophageal disfunction. EoE was first described as a disease entity in 1995 by Kelly et al.. Over the last decade, EoE has become increasingly recognized with an incidence considered to be similar to that of Chron's disease. In 2007, a multidisciplinary group of experts published the first consensus guidelines on the diagnosis and treatment of EoE, which was updated in 2011and recently revised in 2017.
Here, we will revise new insights on clinical presentation and diagnostic approaches to EoE in children. Specifically the increasing incidence, the new clinical conditions (such as PPI responsive esophageal eosinophilia) and the revaluation of the role of PPI in the diagnostic workup, will be addressed.
## Epidemiology
In recent years, the incidence and prevalence of eosinophilic esophagitis (EoE) in children and adults, have increased significantly. A recent systematic review showed that population-based prevalence of EoE in children is 19.1 cases per 100.000 children/year, with a wide geographic variation from 2.3 in Denmarkto 50.5 per 100.000 children in the United States. Likewise, the population-based incidence varies among the Westernised countries, with over-all incidence rate estimates of 5.1 per 100.000 children/year. Possible biases of these variations may depend on the studied population, study design and diagnostic criteria for EoE. Population-based data on the epidemiology of EoE in developing countries are scarce. There is a predominance of EoE among male patientswith a male to female ratio of 3:1 in both children and adults. The increasing of incidence and prevalence, during the last few decades, may be explained with improved recognition of EoE, more awareness of practitioners and higher use of diagnostic esophagogastroduodenoscopy (EGDS) with biopsies in children. Data showing that increasing incidence closely matches the increase in endoscopy volume and biopsies rates supported this hypothesis. Further factors may affect the increasing frequency of EoE such as a change in food allergens, increasing aeroallergens and other environmental factors, the decrease of H. pylori and microbiome changes.
## Pathophysiology
In the last years, there has been a rapid advancement in understanding the mechanisms involved in EoE generally attributed to an interplay between genetic, immunologic and environmental factors. EoE is now considered as a chronic atopic disorder, histologically characterized by a dense epithelial eosinophilic infiltrate. Pathogenically related to a Th2 inflammation characterized by a mixed IgE and non-IgE-mediated reaction to food and/or environmental agents, which is drivenby Thymic Stromal Lymphopoietin (TLSP) secreted by esophageal epithelial cells under the influence of genetic predisposition.
The role of TSLP as a strong inducer of a Th2 response is well-known and it has been linked to atopic diseases. TLSP is similar to IL-7 and regulates host adaptive immune responses through dendritic cells and T-cell interactions. Rothenberg et al. found an association between a single nuclear polymorphism (SNP) in the gene encoding TSLP and the risk for EoE. In addition, Sherill et al. identified a significant association between an SNP located in the TSLP receptor and male EoE patients. Another identified risk factor linked to EoE is a polymorphism in eotaxin-3 that is a chemokine and a potent eosinophil and mast cell chemoattractant. The most recently described is an SNP located in the proximity of CAPN 14, a gene specifically expressed in esophagus and upregulated by Th2 cytokines. However, a recent study, focusing on the risk of developing EoE in monozygotic and dizygotic twins, suggests a minor genetic contribution in favor of more relevant environment factors.
Environmental exposures such as birth by cesarean section, premature delivery, antibiotic exposure during infancy, food allergy, lack of breastfeeding, lack of early exposure to microbes and altered microbiome can create an epigenetic signature that may increase the risk for EoE onset.
In most patients, food has been identified as a trigger of EoE inflammation with non-IgE-mediated mechanismand his role is supported by the response to dietary elimination of food antigens and relapse with the reintroduction of similar food antigens. However, the role of aeroallergens still remains unclear.
Most studies have shown that EoE inflammation has the typical characteristics of Th2 atopic inflammation. Esophageal biopsies and blood samples of patients with active EoE, have high levels of Th2 prototypical cytokines and chemokines such as interleukin IL-5, IL-4, IL-13, IL-5, TSLP, eotaxin-3 secreted by the typical cells involved in allergic inflammation: mast cells, T cells, basophils, invariant natural killer T cells (iNKTs).
Eosinophils are pathognomonic of EoE inflammation in the absence of GERD and although they represent the histologic mark of the disease, do not guarantee the EoE pathogenesis being most likely a consequence of the local Th2 inflammation. Recent clinical trials with antibodies against IL-5, which is the most important eosinophils growth factor, have shown a partial reduction of esophageal inflammation, and minimal symptomatic relief in patients with EoE.
Finally, the esophageal epithelium plays a major role in EoE by promoting a local Th2 inflammation. As for atopic dermatitis, EoE patients have an altered epithelial barrier function, and a down-regulation of proteins associated with barrier function (filaggrin and zonulin-1)and adhesion molecules (desmoglein-1), that favors antigen penetration and sensitization. Altered epithelial permeability can lead to a permissive environment that enhances antigen presentation, which in turn leads to recruitments of eosinophils and pro-inflammatory cytokines.
The major long-term consequence of chronic inflammation in EoE is esophageal remodeling followed by the development of irreversible structures. Aceves et alobserved basal thickening and increased vascular activation in untreated patients with EoE, associated with high levels of TGF-β 1 that stimulates myofibroblast differentiation and extracellular matrix remodeling.
## Clinical findings
Eosinophilic esophagitis (EoE) can affect individuals at any age and clinical presentation depends on the patient's abilities to report symptoms associated with esophageal dysfunction. Symptoms can be present for a long time (mean of 3-5 years) before reaching a diagnosis of EoE, especially if the disease appears progressively.
In pediatric population, vomiting, abdominal pain, dysphagia and bolus impaction are the most prevalent symptoms.
Clinical features of EoE are also different based on the age of children. In toddlers and infants, the most frequent symptoms are feeding difficulties and failure to thrive (median age 2,8 years); vomiting and sleeping disturbance are also described. In EoE, vomit rarely appears before 6 months of life; it's sporadic and not associated with meals, as in the case of food protein-induced enterocolitis or IgEmediated food allergy. School-aged children are present with nausea, vomiting, regurgitation and abdominal pain. In a study that included 43 patients with EoE, 100% of children under 7 years presented these symptoms. Abdominal pain in this age is mainly epigastric. Mainly in infants, toddlers or young children EoE can be present with GERD like symptoms, such as heartburn, regurgitation and vomit. As demonstrated in some studies, there are cases of PPI refractory GERD that underwent fundoplication without clinical improvement, some of these patients receive a post-operative diagnosis of EoE. The misdiagnosis of EoE in GERD like symptoms patients can be related to the lack of routine esophageal biopsies as part of the standard preoperative evaluation before anti-reflux surgery.
In adolescents, the most common symptoms are dysphagia, heartburn, food impaction and chest pain. The prevalence of EoE among EGDS performed in children with esophageal food impaction and/or dysphagia is high (63-88%), so children with these presenting symptoms should be rapidly tested with EGDS and multiple esophageal biopsies. A studyevidenced an increasing prevalence of EoE in patients with dysphagia; moreover, they were more likely to have higher eosinophils peak mucosal counts. Furthermore, the incidence of EoE in patients having these symptoms is probably underestimated, because biopsies are not always performed.Food impaction is a gastrointestinal emergency, requiring endoscopic intervention to remove the impacted food: considering the close link with EoE biopsies during endoscopic disimpaction are recommended.Food impaction is often associated with acute severe retrosternal or chest pain. When symptoms do not respond to medical treatments for GERD, EoE should be strongly considered. Chest pain, which is spasmodic, can be severe enough to lead patients to seek emergency evaluation and lead to cardiac evaluation. Dysphagia in children is not always related to esophageal anatomical damage, but also it can be related to secondary esophageal dysmotility.
It is important to realize that children may develop longterm coping strategies to avoid symptoms, including taking small bites, eating slowly with excessive chewing, and drinking fluids after each bite. The patients may also avoid certain kinds of food due to their not easy swallowing.
Children with EoE have a higher rate of atopy compared with normal children. The rates of asthma, allergic rhinitis and atopic dermatitis are approximately three times higher than the general population.
Adolescents with EoE may be misdiagnosed as having eating disorders, because of symptoms of food-related anxiety, vomiting and food aversion.
EoE should be considered also in children with chest pain and cough that does not respond to aggressive asthma treatment.
Given the variability of the symptom patterns in EoE patients, the measurement of disease activity based on clinical symptoms remains a challenging task. As several clinical trials showed, there is a poor correlation between histological abnormalities and symptoms. A possible explanation for this discrepancy might be that symptoms can be caused not only by esophageal inflammation but also by fibrotic changes in esophageal motility even in the absence of inflammation, or even by psychological factors.
## Diagnosis
EoE diagnosis is difficult because clinical symptoms are variable and sometimes unspecific and the presence of esophageal eosinophilia can be found in several diseases.
The new diagnostic criteria for EoE are the result of the collaboration between paediatrics, gastroenterologists, allergo-immunlogists and pathologists in order to identify clinical and histological characteristics useful for having a clinical suspicion associated with supports elements to confirm the diagnosis.
Guidelines of 2017 and recent updates on EoE showed that diagnostic suspicion is based on the presence of chronic symptoms of esophageal dysfunction and esophageal eosinophilia. The absence of symptoms even in the presence of esophageal eosinophilia does not allow the diagnosis of EoE.
Accurate anamnesis can be useful for making the diagnosis of EoE: in addition to clinical symptoms and esophageal eosinophilia recently, independent predictors of EoE have been identified in order to distinguish it from other diseases, first of all, GERD.
The early age of onset of symptoms is more suggestive of EoE. Up to 75% of patients with EoE have a personal or family history of atopic disease (asthma, eczema, allergic rhinitis or food allergies).
Testing for allergic sensitization may be considered: skin prick test, blood testing for allergen-specific IgE or atopy patch testing are useful for making the diagnosis of allergy but they cannot identify EoE trigger and their positive predicted value remains poor.
If the clinical evaluation is suggestive for EoE, EGDS with biopsies examination has to be made. At least six biopsies should be taken from different esophageal segments, focusing on areas with endoscopic mucosal abnormalities: a study demonstrated that six biopsies increase diagnostic sensibility to 99%. Hematoxilin-eosin staining is sufficient for histological assessment. The principal diagnostic element for the diagnosis of EoE is the presence of at least 15 eosinophils per high power field. Although GERD can increase eosinophilic infiltration in the distal esophagus, eosinophils associated with GERD generally occur at a lower density. Additional histological features consistently associated with EoE may include eosinophils microabscesses (32-64,8% -defined as aggregates of four or more eosinophils in a cluster), basal zone hyperplasia (35-86,4% -more severe in patients with EoE than in those with GERD), dilated intercellular spaces, eosinophil surface layering and papillary elongation (75,6%).
Endoscopy in EoE can be macroscopically normal, with a range from 4 to 23% [57, . There are also some endoscopic characteristics associated with the diagnosis of EoE in order to support the diagnosis of EoE. In adults, their sensibility range from 50% to 90% while in the pediatric population, it is not known. These findings include: fixed rings(6%), exudates(28,5%), furrows(21,6-80%), oedema(26%)and mucosal alterations.The presence of strictures in children is rare and less common than in adults. The alterations can be quantified using the EoE Endoscopic Reference Score (EREFS).
In the case of esophageal eosinophilia, clinical evaluation and endoscopy allow to distinguish EoE from other diseases. The diagnosis of EoE is confirmed if there are no other conditions that justify clinical symptoms and esophageal eosinophilia. GERD is the main disease to consider differential diagnosis with EoE. They often have similar symptoms, and both clinic and endoscopic findings have the low sensibility and specificity: they were felt to be mutually exclusive disorders where a response to PPI/pathologic pH exposure was consistent with GERD and non-reponse/normal pH confirmed EoE. Nowadays, without a definitive method for defining GERD, no single test (including PPI trial) can exclude the presence of GERD.
# Conclusion
Multiple prospective and retrospective studies found that esophageal eosinophilia (≥ 15 eos/HPF) in patients with symptoms and endoscopic sign suggestive of EoE responded to PPI therapy both in adults and children at a rate of 28-82%. From 2011, diagnostic guidelines defined a new condition termed PPI-responsive esophageal eosinophilia (PPI-REE). It is a debate if PPI-REE represents a subtype of EoE or GERD, but the latest updates tend to consider it a pattern of EoE. At baseline (prior to a PPI trial) patients with EoE and PPI-REE have similar clinical, endoscopic and histologic features, in addition, RNA expression profiles were largely similar between EoE and PPI-REE but distinct from GERD. Finally, patients with PPI-REE can also have a response to dietary elimination or topical steroid therapy.
The complex relationship between GERD and EoE and the identification of PPI-REE questioning the role of PPI trial for diagnosis of EoE, even if PPI trial potentially reduces the number of endoscopies required, helps concomitant GERD and provides a stepwise approach for EoE diagnosis. Most recently, the response of esophageal eosinophilia to PPI has been defined as a treatment option for EoE rather than an exclusion criterion: therefore, we proposed a new diagnostic algorithm.
## List of abbreviations
# Funding
None.
## Conflict of interest
The authors declare no conflict of interest, financial or otherwise. |
The Role of Isoflavones in the Prevention of Breast Cancer and Prostate Cancer
Citation: Pejčić, T.; Zeković, M.; Bumbaširević, U.; Kalaba, M.; Vovk, I.; Bensa, M.; Popović, L.; Tešić, Ž. (I.V.) † These authors contributed equally to this work.Abstract: This narrative review summarizes epidemiological studies on breast cancer and prostate cancer with an overview of their global incidence distribution to investigate the relationship between these diseases and diet. The biological properties, mechanisms of action, and available data supporting the potential role of isoflavones in the prevention of breast cancer and prostate cancer are discussed. Studies evaluating the effects of isoflavones in tissue cultures of normal and malignant breast and prostate cells, as well as the current body of research regarding the effects of isoflavones attained through multiple modifications of cellular molecular signaling pathways and control of oxidative stress, are summarized. Furthermore, this review compiles literature sources reporting on the following: (1) levels of estrogen in breast and prostate tissue;(2) levels of isoflavones in the normal and malignant tissue of these organs in European and Asian populations; (3) average concentrations of isoflavones in the secretion of these organs (milk and semen). Finally, particular emphasis is placed on studies investigating the effect of isoflavones on tissues via estrogen receptors (ER).
# Introduction
At least three facts support the thesis that the Western dietary pattern featuring high consumption of processed red meat and fat and limited intake of plant-based foods may be associated with the increased risk of breast cancer and prostate cancer. First, these oncopathologies are most common in geographical areas with predominant Western-style nutrition and quite rare in parts of Asia where soy and unrefined plant foods are the main dietary components. Second, extensive literature suggests markedly low incidence rates of breast cancer and prostate cancer among great apes, who are herbivores and the closest evolutionary cousins of modern humans. Third, Homo sapiens was obligatory herbivore until it began to introduce meat into the diet. This significant change in evolutionary dietary trajectory is now hypothesized to be an important factor in the rising incidence of breast cancer and prostate cancer in humans [bib_ref] Similarities of prostate and breast cancer: Evolution, diet, and estrogens, Coffey [/bib_ref] [bib_ref] Diet and the evolution of the earliest human ancestors, Teaford [/bib_ref]. Exploration of nutritional factors and diet-derived mediators is a rapidly growing area of anticancer research. Remarkable advancements in high-throughput genomic technologies, molecular epidemiology and biochemical techniques have paved the way for a more holistic and precise approach to Since polyphenols have different forms [bib_ref] Polyphenolic profile of the fruits grown in Serbia, Tešić [/bib_ref] [bib_ref] The effects of polyphenols and other bioactives on human health, Fraga [/bib_ref] , their biological activity can be manifested in different ways according to the group to which they belong. Polyphenols can be considered to be valuable natural antioxidants that reduce free radicals through several mechanisms [bib_ref] The effects of polyphenols and other bioactives on human health, Fraga [/bib_ref]. Free radicals-molecules with an unpaired electron(s)-can damage other biomolecules, create new radicals and adversely affect cell functionality, which is related to different metabolic processes and reactions in the human body. Even small concentrations of antioxidants can prevent or slow down oxidation, which results in cell protection [bib_ref] The characterization of antioxidants, Halliwell [/bib_ref]. An excess of free radicals, which often arises during various cell mutations and pathological processes, leads to oxidative stress, an important cause of diseases [bib_ref] Oxidative stress, inflammation, and cancer: How are they linked? Free Radic, Reuter [/bib_ref]. Antioxidant activity is important as antioxidative compounds activate the body's defense Flavonoids, as good electron donors, are well-known antioxidants. They have the ability to scavenge free radicals by creating less reactive species. Antioxidant activity depends on the number and distribution of hydroxyl groups in their structures. The basic structural backbone formula is C 6 -C 3 -C [bib_ref] Effects of isoflavones on breast tissue and the thyroid hormone system in..., Hüser [/bib_ref] , where two rings, A and C, present chroman moiety and ring B phenyl moiety. Therefore, more OHgroups bonded to the B ring achieve lower redox potential and, thus, more effective antioxidants. The saturation of the heterocyclic C ring stabilizes the radical and reduces the antioxidant activity of polyphenols (i.e., flavones and flavonols have higher antioxidant activity compared to flavans and flavanols). The subclass of isoflavones is characterized by ring B attached at C-3 [fig_ref] Figure 1: Schematic representation of the phenolic compounds and their chemical structures [/fig_ref].
The antioxidant activity of isoflavones was documented in the literature [bib_ref] Phenolic compounds of soybean seeds from two European countries and their antioxidant..., Król-Grzymała [/bib_ref] [bib_ref] Antioxidant properties of soybean isoflavone extract and tofu in vitro and in..., Liu [/bib_ref] [bib_ref] The antioxidant activity of daidzein metabolites, O-desmethylangolensin and equol, in HepG2 cells, Choi [/bib_ref] [bib_ref] Current knowledge and future direction of research on soy isoflavones as a..., Kalaiselvan [/bib_ref] [bib_ref] The antioxidant role of soy and soy foods in human health, Rizzo [/bib_ref] and largely depends on their structure. Combinations of phenolic compounds with each other or with other molecules can have a synergistic effect [bib_ref] Antioxidant properties of soybean isoflavone extract and tofu in vitro and in..., Liu [/bib_ref] [bib_ref] Sulforaphane synergizes with quercetin to inhibit self-renewal capacity of pancreatic cancer stem..., Srivastava [/bib_ref]. Polyphenols are an important group of phytochemicals with a high potential for a positive impact on human health [bib_ref] Polyphenols: Chemoprevention and therapeutic potentials in hematological malignancies, Izuegbuna [/bib_ref] [bib_ref] Dietary polyphenols and their role in oxidative stress-induced human diseases: Insights into..., Rudrapal [/bib_ref].
## Isoflavones
Based on their structure, isoflavones are divided into three groups: (1) O-substituted derivatives (e.g., hydroxy, methoxy and methylenedioxy); (2) prenylated derivatives; (3) glycosylated derivatives [bib_ref] Isoflavonoids of the leguminosae, Veitch [/bib_ref]. For example, O-substituted isoflavones (and their glycosides, [fig_ref] Figure 2: Structures of isoflavones glycosides [/fig_ref] are daidzein (daidzin, [fig_ref] Figure 2: Structures of isoflavones glycosides [/fig_ref] , genistein (genistin, [fig_ref] Figure 2: Structures of isoflavones glycosides [/fig_ref] and glycitein (glycitin, [fig_ref] Figure 2: Structures of isoflavones glycosides [/fig_ref] [fig_ref] Figure 1: Schematic representation of the phenolic compounds and their chemical structures [/fig_ref]. Most often, isoflavones appear in the form of glycosides. In the human body, they are transformed into effective aglycones such as daidzein (7hydroxy-3-(4-hydroxyphenyl)-4H-1-benzopyran-4-one) and genistein (5,7-dihydroxy-3-(4hydroxyphenyl)-4H-1-benzopyran-4-one) and glycitein (7-hydroxy-3-(4-hydroxyphenyl)-6methoxy-4H-1-benzopyran-4-one).
Flavonoids, as good electron donors, are well-known antioxidants. They have the ability to scavenge free radicals by creating less reactive species. Antioxidant activity depends on the number and distribution of hydroxyl groups in their structures. The basic structural backbone formula is C6-C3-C6, where two rings, A and C, present chroman moiety and ring B phenyl moiety. Therefore, more OHgroups bonded to the B ring achieve lower redox potential and, thus, more effective antioxidants. The saturation of the heterocyclic C ring stabilizes the radical and reduces the antioxidant activity of polyphenols (i.e., flavones and flavonols have higher antioxidant activity compared to flavans and flavanols). The subclass of isoflavones is characterized by ring B attached at C-3 [fig_ref] Figure 1: Schematic representation of the phenolic compounds and their chemical structures [/fig_ref].
The antioxidant activity of isoflavones was documented in the literature [bib_ref] Phenolic compounds of soybean seeds from two European countries and their antioxidant..., Król-Grzymała [/bib_ref] [bib_ref] Antioxidant properties of soybean isoflavone extract and tofu in vitro and in..., Liu [/bib_ref] [bib_ref] The antioxidant activity of daidzein metabolites, O-desmethylangolensin and equol, in HepG2 cells, Choi [/bib_ref] [bib_ref] Current knowledge and future direction of research on soy isoflavones as a..., Kalaiselvan [/bib_ref] [bib_ref] The antioxidant role of soy and soy foods in human health, Rizzo [/bib_ref] and largely depends on their structure. Combinations of phenolic compounds with each other or with other molecules can have a synergistic effect [bib_ref] Antioxidant properties of soybean isoflavone extract and tofu in vitro and in..., Liu [/bib_ref] [bib_ref] Sulforaphane synergizes with quercetin to inhibit self-renewal capacity of pancreatic cancer stem..., Srivastava [/bib_ref]. Polyphenols are an important group of phytochemicals with a high potential for a positive impact on human health [bib_ref] Polyphenols: Chemoprevention and therapeutic potentials in hematological malignancies, Izuegbuna [/bib_ref] [bib_ref] Dietary polyphenols and their role in oxidative stress-induced human diseases: Insights into..., Rudrapal [/bib_ref].
## Isoflavones
Based on their structure, isoflavones are divided into three groups: (1) O-substituted derivatives (e.g., hydroxy, methoxy and methylenedioxy); (2) prenylated derivatives; (3) glycosylated derivatives [bib_ref] Isoflavonoids of the leguminosae, Veitch [/bib_ref]. For example, O-substituted isoflavones (and their glycosides, [fig_ref] Figure 2: Structures of isoflavones glycosides [/fig_ref] are daidzein (daidzin, [fig_ref] Figure 2: Structures of isoflavones glycosides [/fig_ref] , genistein (genistin, [fig_ref] Figure 2: Structures of isoflavones glycosides [/fig_ref] and glycitein (glycitin, [fig_ref] Figure 2: Structures of isoflavones glycosides [/fig_ref] [fig_ref] Figure 1: Schematic representation of the phenolic compounds and their chemical structures [/fig_ref]. Most often, isoflavones appear in the form of glycosides. In the human body, they are transformed into effective aglycones such as daidzein (7-hydroxy-3-(4-hydroxyphenyl)-4H-1-benzopyran-4-one) and genistein (5,7-dihydroxy-3-(4hydroxyphenyl)-4H-1-benzopyran-4-one) and glycitein (7-hydroxy-3-(4-hydroxyphenyl)-6-methoxy-4H-1-benzopyran-4-one). The USDA Database for the Isoflavone Content of Selected Foods provides an informative overview of isoflavone food sources. Isoflavones represent polyphenols found in high content in soybean and other legumes as well as their products, with soybeans having higher contents of isoflavones compared to other legumes [bib_ref] Isoflavonoids of the leguminosae, Veitch [/bib_ref]. In soybean seeds, malonyl glycosides forms (malonyl daidzein and malonyl genistein) were present in a higher amount than aglycons (daidzein and genistein), while glucosides (daidzin and genistin) [fig_ref] Figure 2: Structures of isoflavones glycosides [/fig_ref] were present in the lowest amount [bib_ref] Phenolic compounds of soybean seeds from two European countries and their antioxidant..., Król-Grzymała [/bib_ref]. Glycitin [fig_ref] Figure 2: Structures of isoflavones glycosides [/fig_ref] , an O-methylated isoflavone glycoside (glycitein 7-O-glucoside), was also found in soy [bib_ref] Bioavailability, disposition, and dose-response effects of soy isoflavones when consumed by healthy..., Setchell [/bib_ref].
The content of isoflavones differs through the products and regions. Daily consumption [bib_ref] Genistein: Dual role in women's health, Yu [/bib_ref] in Asian countries is much higher than in other countries [bib_ref] Flavonoids and other polyphenols act as epigenetic modifiers in breast cancer, Selvakumar [/bib_ref]. The consumption of food containing isoflavones is related to the traditional foods of countries such as China, Japan and South Korea. Higher dietary intake of isoflavones can be related to health benefits due to the biological activity of isoflavones [bib_ref] Antioxidant properties of soybean isoflavone extract and tofu in vitro and in..., Liu [/bib_ref]. The USDA Database for the Isoflavone Content of Selected Foods provides an informative overview of isoflavone food sources. Isoflavones represent polyphenols found in high content in soybean and other legumes as well as their products, with soybeans having higher contents of isoflavones compared to other legumes [bib_ref] Isoflavonoids of the leguminosae, Veitch [/bib_ref]. In soybean seeds, malonyl glycosides forms (malonyl daidzein and malonyl genistein) were present in a higher amount than aglycons (daidzein and genistein), while glucosides (daidzin and genistin) [fig_ref] Figure 2: Structures of isoflavones glycosides [/fig_ref] were present in the lowest amount [bib_ref] Phenolic compounds of soybean seeds from two European countries and their antioxidant..., Król-Grzymała [/bib_ref]. Glycitin [fig_ref] Figure 2: Structures of isoflavones glycosides [/fig_ref] , an O-methylated isoflavone glycoside (glycitein 7-O-glucoside), was also found in soy [bib_ref] Bioavailability, disposition, and dose-response effects of soy isoflavones when consumed by healthy..., Setchell [/bib_ref].
The content of isoflavones differs through the products and regions. Daily consumption [bib_ref] Genistein: Dual role in women's health, Yu [/bib_ref] in Asian countries is much higher than in other countries [bib_ref] Flavonoids and other polyphenols act as epigenetic modifiers in breast cancer, Selvakumar [/bib_ref]. The consumption of food containing isoflavones is related to the traditional foods of countries such as China, Japan and South Korea. Higher dietary intake of isoflavones can be related to health benefits due to the biological activity of isoflavones [bib_ref] Antioxidant properties of soybean isoflavone extract and tofu in vitro and in..., Liu [/bib_ref].
## Health-related properties of isoflavones
Isoflavones are known to be beneficial for human health due to their many functional characteristics. Isoflavones are a type of phytoestrogens, which are estrogen-like nonsteroid substances found in plants. Isoflavones represent phytochemicals with identical physiological activity as estrogens, becoming physiologically active through the intestinal microflora [bib_ref] Isoflavones, their glycosides and glycoconjugates. Synthesis and biological activity, Szeja [/bib_ref]. For example, S-equol, a metabolite of daidzein and genistein, has a strong antioxidative effect, effects on cardiometabolic risk factors [bib_ref] Cardiometabolic risk and gut microbial phytoestrogen metabolite phenotypes, Frankenfeld [/bib_ref] , high estrogenic activity, as well as the highest antioxidant activity among isoflavones, with an effect on carcinoma cells. Levels of S-equol > 5-10 ng/mL have been associated as a therapeutic agent with a positive effect against prostate cancer and other health benefits [bib_ref] The antioxidant activity of daidzein metabolites, O-desmethylangolensin and equol, in HepG2 cells, Choi [/bib_ref] [bib_ref] Emerging evidence of the health benefits of S-equol, an estrogen receptor β..., Jackson [/bib_ref]. Although isoflavone aglycones have better bioavailability and bioactivity than their glycosides [bib_ref] Bioavailability and health benefits of major isoflavone aglycones and their metabolites, Hsiao [/bib_ref] , hydrolysis of glycosides to aglycones before consumption of a nonfermented soy drink did not enhance the bioavailability of isoflavones in postmenopausal women [bib_ref] Hydrolysis of isoflavone glycosides to aglycones by beta-glycosidase does not alter plasma..., Richelle [/bib_ref]. The structural similarity with estrogen indicates the possible biological effects of isoflavones and their potential for cardiovascular and hormone replacement therapy in postmenopausal women [bib_ref] Genistein: Dual role in women's health, Yu [/bib_ref]. Other research has shown great interest in the health benefits of the association of S-equol and the gut microbial community shown by improvement in cardiometabolic risk factors [bib_ref] Cardiometabolic risk and gut microbial phytoestrogen metabolite phenotypes, Frankenfeld [/bib_ref]. Furthermore, the importance of isoflavones is reflected in investigations of their influence on cancer [bib_ref] The antioxidant activity of daidzein metabolites, O-desmethylangolensin and equol, in HepG2 cells, Choi [/bib_ref] [bib_ref] Current knowledge and future direction of research on soy isoflavones as a..., Kalaiselvan [/bib_ref] [bib_ref] Isoflavonoids of the leguminosae, Veitch [/bib_ref] [bib_ref] Flavonoids and other polyphenols act as epigenetic modifiers in breast cancer, Selvakumar [/bib_ref] , described in detail in Sections 3 and 4. The usefulness of polyphenols synergism, their relationship with other phytochemicals, as well as food intake, are of critical importance to understanding the overall impact of isoflavones.
## Breast cancer and isoflavones
## Breast cancer-epidemiology and risk factors
With extraordinary heterogeneity in terms of occurrence rates, clinical presentation, available preventive and curative interventions, and therapeutic response, the overall magnitude of the public health burden related to cancer incidence and mortality is substantially increasing worldwide. The increasing cancer occurrence and mortality are the results of the cumulative effects of aging and the growth of the global population, as well as the transition of distribution and prevalence of major risk factors. Breast cancer is the most frequently occurring invasive cancer in the female population. With an annual estimation of 2.3 million new cases based on the Global Cancer Statistics-GLOBOCAN report for 2020, breast cancer surpassed lung cancer as the most commonly diagnosed malignancy accounting for 11.7% of total cases. Furthermore, breast cancer is considered to be the leading cause of cancer death among women and ranks fifth in combined-sex mortality estimates [bib_ref] Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for..., Sung [/bib_ref]. There is prominent geographic variability in breast cancer registry data. Although the age-standardized incidence rates are approximately 88% higher in transitioned countries (i.e., Australia, New Zealand, and other countries in North America, Western and Northern Europe) compared to transitioning countries (55.9 versus 29.7 per 100,000, respectively), mortality rates are 19% higher in regions with a lower level of socioeconomic development [bib_ref] Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for..., Sung [/bib_ref]. These patterns may be attributed to both variances in exposure to risk factors and to worldwide disparities across the continuum of healthcare. Furthermore, global temporal-trend figures show that incidence rates have been increasing in virtually all regions over the last three decades. Although overall lethality indices continuously decline, case-fatality proportions remain markedly higher in lower-income countries, particularly in Africa and among small island nations in Melanesia, Micronesia, Polynesia and the Caribbean [bib_ref] Global burden and trends in premenopausal and postmenopausal breast cancer: A population-based..., Heer [/bib_ref] [bib_ref] Westernizing" women's risk? Breast cancer in lower-income countries, Porter [/bib_ref]. Current patterns in breast cancer incidence and survival trends are affected by: alterations in environmental and individual-level variables and risk factors, public health determinants (such as widespread mammography screening programs), and general advancements in diagnostic and treatment infrastructure [bib_ref] Trends in female breast cancer incidence, mortality, and survival in Austria, with..., Ilic [/bib_ref] [bib_ref] A historic and scientific review of breast cancer: The next global healthcare..., Becker [/bib_ref]. Pronounced differences in mortality-to-incidence ratios warrant concern and an urgent need for methodical and harmonized dissemination and implementation of evidence-based, contextually suitable policies and interventions targeting awareness, education, risk identification, and early detection, as well as reinforcement of equitable resources and capacities for high-quality multimodal treatment and survivorship care [bib_ref] The role of dissemination and implementation science in global breast cancer control..., Rositch [/bib_ref]. Nevertheless, the rising incidence in developing countries in Africa, South America, and Asia, as well as in certain affluent Asian countries, where rates are traditionally low, is associated with the so-called "Westernization" phenomenon. This term refers to dramatic changes in sociocultural circumstances, economic development, urbanization, and demographic and workforce shifts that are accompanied by behavioral and lifestyle modifications. Similar incidence trend variations were observed in migration studies. Sedentary culture with less physical activity, increased prevalence of obesity, nutritional transition towards high-fat and (ultra) processed foods, postponement of childbirth and lower parity, and accessibility to hormonal birth control and replacement therapy are all contributing to the convergence towards the industrialized Western countries breast cancer risk-profile and the narrowing of the international gap between historically low and high morbidity areas [bib_ref] Westernizing" women's risk? Breast cancer in lower-income countries, Porter [/bib_ref].
Having a remarkably heterogeneous oncopathology, breast cancer comprises a wide spectrum of tumors with a high level of diversity in molecular alterations, cellular composition, histological features, clinical presentation, metastatic potential and dissemination patterns, treatment sensitivity and patient outcomes [bib_ref] Deconstructing the molecular portraits of breast cancer, Prat [/bib_ref]. The classification of breast cancer is challenging due to the versatility of cancer cell phenotypes and their perplexing interaction with a plethora of molecular and cellular mediators of both tumoral and microenvironmental plasticity. The complexity of inter and intra-tumor variability is only partially epitomized by routinely applied principal clinical parameters (patient age, lymph node involvement, neoplasm size, tumor stage and grading) and pathobiological markers such as estrogen receptors (ER), progesterone receptors (PR), and human epidermal growth factor receptor 2 (HER2/ERBB2) [bib_ref] Heterogeneity in breast cancer, Polyak [/bib_ref]. The traditional framework for clinical breast cancer classification encompasses three major types: (1) luminal ER-positive and PR-positive (additionally subdivided into luminal types A and B based on histochemical staining for the proliferation marker protein Ki-67 (MKI67) or genetic profiling); (2) HER2-positive;
(3) triple-negative breast cancer (TNBC). Nevertheless, the appreciation of the tumorimmune dynamic relationship, breakthroughs in omics technologies and next-generation sequencing, and inclusion of predictive and prognostic genomic and immunomarkers have enabled better comprehension of the malignant ecosystem and progression pathways, refined diagnostic and patient-stratification algorithms, and most importantly, revolutionized the therapeutic landscape, thus providing novel opportunities for breast cancer management [bib_ref] Treatment landscape of triple-negative breast cancer-Expanded options, evolving needs, Bianchini [/bib_ref].
Etiopathology of breast cancer is rather complex. A variety of intricately interrelated modifiable and non-modifiable factors, including genetics, environmental exposures, and physiological, sociobiological and lifestyle determinants, may increase susceptibility or encourage disease development [bib_ref] Breast cancer-Epidemiology, risk factors, classification, prognostic markers, and current treatment strategies-An updated..., Łukasiewicz [/bib_ref]. Female sex is one of the major risk factors, predominantly due to elevated hormonal stimulation and increased vulnerability of breast tissue to estrogen-progesterone balance disruption. Male breast cancer accounts for approximately 1% of all cases, and the most common being the ER-positive type [bib_ref] Breast cancer in men, Giordano [/bib_ref]. Currently, nearly 80% of breast cancer patients are older than 50, while more than 40% of affected individuals are older than 65 years. Epidemiological data reveal associations between age at diagnosis and certain cancer molecular subtypes. Aggressive, recurrence-prone, resistant TNBC is most frequently detected in populations less than 40 years of age, whereas among patients above 70 years the luminal A type is most frequently detected [bib_ref] Effects of age on the detection and management of breast cancer, Mcguire [/bib_ref]. A significantly higher incidence rate of breast cancer is observed among individuals with a positive family history, with 13-19% of patients having a first-degree relative affected with the same or related malignancy [bib_ref] Emerging trends in family history of breast cancer and associated risk, Shiyanbola [/bib_ref]. Familial clustering of breast cancer was estimated to be approximately 73% attributed to heritable components and 27% to environmental triggers. Most inherited cases due to predisposing genetic mutations are associated with two autosomal dominant high-risk genes: BRCA1 (BReast CAncer gene 1) and BRCA2 (BReast CAncer gene 2), located on chromosomes 17(17q21.31) and 13(13q13.1), respectively. Several pathogenic variants in highly penetrant (e.g., TP53, CDH1, PTEN and STK11) and moderately penetrant susceptibility genes (e.g., CHECK2, PALB2, ATM and BRIP1) were also reported to be involved in breast carcinogenesis [bib_ref] Identifying breast cancer susceptibility genes-A review of the genetic background in familial..., Wendt [/bib_ref]. Although large-scale genotyping analysis within genome-wide association studies (GWAS) provided significant advancement in this area, a considerable proportion remains unclear. Further research is needed to elucidate the intricate breast cancer genetic background. Extensive scientific evidence corroborates the relationship between an individual's reproductive history and the risk of breast cancer. Timing, duration and accompanying hormonal fluctuations of particular entities, such as menstruation, pregnancy and breastfeeding may have a considerable impact on the potential induction of oncogenesis in the breast microenvironment. Early menarches, advanced age of full-term pregnancy, absence or short-period breastfeeding, nulliparas or low parity, and late menopause (with a high total lifetime count of menstrual cycles) elevate the breast cancer risk. Additionally, certain breast properties such as high tissue density, personal history of cancerous and non-cancerous lesions, as well as previous chest exposure to radiation were correlated with increased risk of breast cancer occurrence. Other environmental and lifestyle risk factors include smoking, a high-fat diet, excessive alcohol consumption, hormonal replacement therapy, oral contraceptives, obesity and low physical activity [bib_ref] Risk factors and preventions of breast cancer, Sun [/bib_ref] [bib_ref] A Review of modifiable risk factors in young women for the prevention..., Daly [/bib_ref].
## Breast cancer chemoprevention and dietary isoflavone intake
Accumulating evidence and enhanced comprehension of intricate mechanisms involved in cancerogenesis have directed scientific attention toward the immense potential of chemopreventive strategies and, in particular, compounds that may suppress, arrest or reverse various molecular pathways associated with cancer development and progression. Chemoprevention encompasses a myriad of molecular targets related to distinct stages of the carcinogenic process (i.e., initiation, promotion, and progression) and may be broadly categorized into three major types: primary, secondary and tertiary [bib_ref] Cancer chemoprevention: A rapidly evolving field, Steward [/bib_ref]. Primary prevention refers to the exposure of healthy individuals or populations exerting particular risk features but without the overt disease to certain natural or synthetic agents in order to prevent cancer occurrence. Secondary prevention pertains to premalignant lesions and the administration of selected agents to perturb the transition to invasive cancer. Tertiary prevention applies to patients who successfully underwent cancer treatment in order to prevent recurrence or the development of a new primary tumor [bib_ref] Cancer chemoprevention: A strategic approach using phytochemicals, Shankar [/bib_ref]. Compounds that impede the initial neoplastic transformation are termed as "blocking agents", whereas those that counteract the progression of benign lesions to advanced-stage cancers are denominated as the "suppressing agents". Mechanistically, these agents effectuate their chemopreventive properties by a wide spectrum of functions, including the inhibition of carcinogen uptake, induction of cellular detoxifying enzymes and genomic repair pathways, antioxidant activity and free radical scavenging, modulation of epigenetic alterations, downregulation of chronic inflammation, as well as interaction with cellular signaling networks responsible for maintaining the delicate balance between cell survival, differentiation, proliferation, senescence, and death [bib_ref] Dietary cancer-chemopreventive compounds: From signaling and gene expression to pharmacological effects, Chen [/bib_ref].
A multitude of epidemiological studies has associated isoflavones and their dietary sources with various health benefits, including protection against hormone-dependent oncopathologies (such as breast and prostate cancer), cardiovascular disease, disturbances in glycoregulation, osteoporosis, support to weight management/obesity reduction and attenuation of physical and emotional menopausal symptoms [bib_ref] Biological effect of soy isoflavones in the prevention of civilization diseases, Pabich [/bib_ref]. Available dietary surveys indicate a great level of variability regarding the intake of total phytoestrogens, particular subclasses of these plant-derived secondary metabolites, and their food sources with regard to geographical and sociocultural determinants. Consumption of foods containing high levels of isoflavones, such as soybeans and soy-based products and other legume seeds (beans, peas, lentils, clover), is commonly associated with traditional Asian dietary patterns rather than with nutritional preferences of the Western world. Despite the considerable inter-study variations, a substantially higher average daily intake of isoflavones per capita is reported for South and East Asian countries (i.e., 20-50 mg/day) than for European countries (0.37-4.5 mg/day) and the USA (0.73-3.3 mg/day) [bib_ref] Effect of the intake of isoflavones on risk factors of breast cancer-A..., Finkeldey [/bib_ref] [bib_ref] Soy, soy foods and their role in vegetarian diets, Rizzo [/bib_ref] [bib_ref] Dietary intakes and food sources of phytoestrogens in the European Prospective Investigation..., Zamora-Ros [/bib_ref] [bib_ref] Intakes and sources of soya foods and isoflavones in a UK population..., Mulligan [/bib_ref]. Interestingly, however, in the last couple of decades, noteworthy changes in isoflavone-related dietary habits were observed: simultaneously with the decreased consumption of soy in favor of animal-sourced protein in Asia due to the adoption of Westernized nutrition, the intake of soy is growing among health-conscious individuals in the USA and Europe. Based on dietary analyses, vegans and vegetarians are a Western-world subgroup with the highest isoflavone intake of approximately 7-12 mg/day [bib_ref] Analytical and compositional aspects of isoflavones in food and their biological effects, Mortensen [/bib_ref]. Dominant isoflavone sources in traditional Asian diet are fermented and non-fermented soybeans and first-generation soy products, including tofu, tempeh, miso, soy milk, natto and cheonggukjang. In contrast, soy-based meat and dairy substitutes, as well as various food derivatives with soy flour or soy proteins added as fillers or extenders, prevail in modern Western habitual nutrition.
Given the biological properties of isoflavones and a striking concomitant geographical discrepancy in isoflavone intake and breast cancer prevalence, researchers have hypothesized the potential role of these compounds in the prevention of mammary tumorigenesis among susceptible populations and among women with or without previous history. In addition to epidemiological-level research, numerous studies explored the benefits of isoflavones in cell cultures, animal and human breast cancer scientific models trying to: (1) elucidate mechanisms of action and tissue-specific dose-dependent response; (2) validate relevant biomarkers; (3) determine efficacy, safety, opportunities, and challenges of controlled preventive and curative interventions. Meta-analyses indicated a positive association between the consumption of high levels of isoflavones and a reduced breast cancer risk in Asian populations, but failed to detect a significant correlation in Western populations, presumably due to low isoflavone intake levels [bib_ref] Isoflavone consumption and risk of breast cancer: A dose-response meta-analysis of observational..., Xie [/bib_ref] [bib_ref] Soyfood intake in the prevention of breast cancer risk in women: A..., Qin [/bib_ref] [bib_ref] Association between soy isoflavone intake and breast cancer risk for pre-and post-menopausal..., Chen [/bib_ref] [bib_ref] Epidemiology of soy exposures and breast cancer risk, Wu [/bib_ref]. Nevertheless, in these meta-analyses, there was considerable between-study heterogeneity that may have compromised the reported observations. Key limitations comprise the case-control design in the majority of included studies, inherent drawbacks of the employed exposure evaluation methods, varied range definitions and intake cut-off values, as well as residual confounding related to unreported yet relevant factors. Contrary to other reports, the China Kadoorie Biobank (CKB) study, which enrolled more than 300,000 women residing in 10 diverse regions in China, found no association between soy intake and breast cancer incidence overall. However, based on the further dose-response meta-analysis integrating the CKB study with other prospective studies, each 10 mg/day of soy isoflavone led to a 3% risk reduction [bib_ref] Soy intake and breast cancer risk: A prospective study of 300,000 Chinese..., Wei [/bib_ref]. Literature reports accounting for menopausal status and hence endogenous estrogen environment are also conflicting. Although a significant inverse association between the isoflavone intake and breast cancer risk in both pre-and post-menopausal Asian women was reported [bib_ref] Association between soy isoflavone intake and breast cancer risk for pre-and post-menopausal..., Chen [/bib_ref] [bib_ref] Epidemiology of soy exposures and breast cancer risk, Wu [/bib_ref] , the menopausal status may be an important effect modifier with the protective effect of soy isoflavone intake existing among postmenopausal subjects only [bib_ref] Isoflavone consumption and risk of breast cancer: A dose-response meta-analysis of observational..., Xie [/bib_ref] [bib_ref] Soy isoflavones consumption and risk of breast cancer incidence or recurrence: A..., Dong [/bib_ref]. Recent meta-analyses summarizing findings from eight prospective studies from 2003 to 2021 concluded that elevated soy consumption and, consequently, isoflavone intake are associated with better breast cancer prognosis and lower occurrence risk regardless of menopause status. According to the systematic review and meta-analysis by combining data derived from 12 scientific articles and 37,275 cases of women with breast cancer, pre-diagnosis soy and isoflavone intake were favorably correlated with the overall survival and reduced recurrence risk among post-menopausal patients exclusively [bib_ref] Soy and isoflavones consumption and breast cancer survival and recurrence: A systematic..., Qiu [/bib_ref]. A number of epidemiological studies evaluated the potential association between isoflavone intake and breast cancer risk stratified by hormone receptor (ER and PR) and HER2 status in the general population. The Shanghai Breast Cancer Study reported a superior risk reduction for ER+/PR+ tumors than for other subtypes [bib_ref] Population-based case-control study of soyfood intake and breast cancer risk in Shanghai, Dai [/bib_ref]. However, in several case-control and cohort studies the protective effect of soy products against breast cancer remained comparable across all ER/PR status subtypes. Concerning the HER2 status, a Japanese case-control study [bib_ref] Effect of soybean on breast cancer according to receptor status: A case-control..., Suzuki [/bib_ref] found that elevated intake of soy-based products significantly reduced the risk of HER2-negative breast cancer by 21%, whereas no significant differences with respect to HER2 status were found in another study [bib_ref] The association of soy food consumption with the risk of subtype of..., Baglia [/bib_ref]. A recent study examined the association between the isoflavone intake and breast cancer risk by molecular subtype in 1,709 Korean females featuring a high risk of hereditary breast cancer (i.e., BRCA1/2 mutation carriers and non-carriers with family history and early-onset breast cancer). The study indicated that the high isoflavone intake in this population might act as a preventive factor, particularly in BRCA2-mutated luminal A type and BRCA1-mutated triple-negative breast cancer [bib_ref] Isoflavone intake on the risk of overall breast cancer and molecular subtypes..., Sim [/bib_ref]. A summary of discussed studies is provided in [fig_ref] Table 1: Summary of discussed epidemiological studies/meta-analysis on isoflavones intake and breast cancer in... [/fig_ref]. Prospective study and a dose-response meta-analysis No association was found between soy intake and BCa incidence in the prospective study. Meta-analysis of prospective studies estimated that each 10 mg increment in daily soy isoflavones intake was associated with a 3% RR of BCa. Pre-diagnosis isoflavones intake was significantly associated with reduced risk of total mortality and recurrence. RR was predominantly observed among post-menopausal patients. Despite the extraordinary scientific interest, evidence on the association between isoflavone dietary intake and breast cancer incidence, recurrence and survival remains rather inconclusive. Additional investigation is needed to substantiate and further elucidate the postulated isoflavone health benefits. The inconsistency in the reported results on both single-study and meta-analysis levels may be attributed to a number of factors: (1) methodological limitations; (2) variability in experimental design (retrospective vs. prospective); (3) sample size and population characteristics; (4) differences in overall exposure level and timing (childhood/adolescence vs. adulthood); (5) environmental and sociocultural circumstances; (6) the lack of information on the nature, type, source and bioavailability of isoflavones. In general, although promising, findings yielded in observational studies warrant confirmation and further clarification in future research with enhanced data standardization and documentation, reliable sensitivity analysis, larger sample size, and prolonged follow-up duration, precise dietary intake assessment, and carefully controlled confounding factors.
## Isoflavone mechanisms of action and impact on breast cancer
Isoflavones can mediate the majority of their biological effects via estrogen receptor (ER) signaling pathways due to a significant structural resemblance with the mammalian 17-β-estradiol (E2). Although the importance of estrogens and their receptors in breast cancerogenesis and subsequent tumor development is well-established, the implication of various environmental and synthetic receptor modulators (i.e., phytoestrogens and xenoestrogens) is still a controversial subject and a matter of debate. Isoflavones may interact with both alpha (ERα) and beta receptor (ERβ) isoforms but exert markedly stronger affinity towards the ERβ. Even though the ligand binding cavities of ERα and ERβ differ by only two amino acid positions (replacement of Leu-384 and Met-421 in ERα with Met-336 and Ile-373, respectively, in ERβ), the sequence diversity beyond the pocket residues confers the ERβ to the topology more suitable for the isoflavones [bib_ref] Interactions of isoflavones and other plant derived estrogens with estrogen receptors for..., Leclercq [/bib_ref].
In general, the ER binding potency of isoflavones ranges from 10 −4 to 10 −3 compared to the primary natural ligand, i.e., estradiol (E2). Appropriate orientation, hydrophobic core and planar chemical structure featuring p-hydroxy-substituted aromatic ring approximately 12 Å distant from the 2 nd planar hydroxyl group enable isoflavones to mimic endogenous estrogen interaction with the hormone binding pockets of ERs. Specific structural elements shaping spatial, electronic, and topological characteristics of particular isoflavone compounds determine their potency as phytoestrogenic ligands and convey their qualitative and functional properties [bib_ref] Molecular aspects of phytoestrogen selective binding at estrogen receptors, Turner [/bib_ref]. Quantitative structure-activity relationship (QSAR) and comparative molecular field research highlighted pharmacophore features significant for receptor binding and activation, including partial charges on atoms C2 , C4 (ring B) and C7 (ring A), conformational rigidity determinants and molecular orientation [bib_ref] QSAR modeling on binding affinity of diverse estrogenic flavonoids: Electronic, topological and..., Mukherjee [/bib_ref]. Given that isoflavones may display both estrogenic and anti-estrogenic effects, they are sometimes regarded as natural selective ER modulators (SERMs). The ultimate actions are predisposed by a multitude of factors, including ambient estrogen concentrations and environmental hormonal milieu, relative levels of ERα and ERβ, the local concentration of the studied isoflavone compound(s), the presence of other phyto/xenoestrogens as well as the interference of present co-activators and co-repressors [bib_ref] Isoflavones-Mechanism of action and impact on breast cancer risk, Stubert [/bib_ref]. Nevertheless, the findings of various studies are rather inconsistent, and comprehensive characterization of the intricate bioactivity of isoflavones warrants further research. These incoherencies may be at least partly attributed to substantial qualitative and quantitative variability of isoflavones content and composition in food sources, the complexity of their metabolic transformations as well as inter and intra-individual genetic and physiological differences. Several preclinical rodent tumor models were applied in animal studies exploring the isoflavone influence on breast cancer. Experimental induction of breast cancer may be obtained through genetic engineering leading to spontaneous cancer development by administration of a chemical carcinogen (most commonly 1-methyl-1-nitrosourea (MNU) for protocols aiming to investigate more aggressive mammary oncopathologies or 7,12-dimethylbenz[a]anthracene (DMBA) or by exposure to excess estrogen levels [bib_ref] Effects of isoflavones on breast tissue and the thyroid hormone system in..., Hüser [/bib_ref]. The inoculation model refers to the transplantation of tumor cells or explants into susceptible murine hosts. The xenograft model is generated in immunodeficient hosts with immortalized human cancer cell lines (commonly MCF-7 or MDA-MB-231) or tumor explants of patient origin, whereas allograft models are created when allogenic cells/tissues are transplanted into immunocompetent syngeneic mice [bib_ref] Understanding genistein in cancer: The "good" and the "bad" effects: A review, Russo [/bib_ref]. Each of these models is of great scientific significance, albeit their inherent imperfections and species-specific peculiarities limit the translational relevance of findings to humans. The deficiency of target-site exposure data undoubtedly remains the critical knowledge gap to be addressed in future studies. These data would contribute to better interpretation of in vitro results, translation of such findings into in vivo relevant context, and eventually their integration with human observational and interventional research in comprehensive risk-assessment analyses.
Although the interaction with ER is the most commonly studied isoflavone mechanism of action, there is a growing body of evidence indicating that additional pathways and alternative cascades may also be affected. In addition to ER-mediated signaling, isoflavones may modulate the local hormone levels in breasts and ovaries in a tissue-specific and concentration-dependent manner by affecting the activity or expression of steroidogenic enzymes implicated in estrogen synthesis and metabolism (e.g., cytochrome P450-aromatase (CYP19A1), 3β-hydroxysteroid dehydrogenase and 17β-hydroxysteroid dehydrogenase, cytochrome P450 1A1 and 1B1 and sulfotransferase, thereby altering the conversion of androgenic precursors to estrogens and the dehydrogenation of estrone to estradiol [bib_ref] Modulation of estrogen synthesis and metabolism by phytoestrogens in vitro and the..., Van Duursen [/bib_ref]. Furthermore, isoflavones deliver antitumorigenic effects through several ER-independent activities, including apoptosis induction, inhibition of cell proliferation and angiogenesis, support for antioxidant defense and DNA repair mechanisms, inflammation downregulation, and interference in other signal-transduction systems [fig_ref] Figure 3: Graphical overview of the isoflavones' possible molecular targets and cellular signaling pathways... [/fig_ref].
Antioxidants 2023, 12, x FOR PEER REVIEW 12 of 28 aromatase (CYP19A1), 3β-hydroxysteroid dehydrogenase and 17β-hydroxysteroid dehydrogenase, cytochrome P450 1A1 and 1B1 and sulfotransferase, thereby altering the conversion of androgenic precursors to estrogens and the dehydrogenation of estrone to estradiol [bib_ref] Modulation of estrogen synthesis and metabolism by phytoestrogens in vitro and the..., Van Duursen [/bib_ref]. Furthermore, isoflavones deliver antitumorigenic effects through several ERindependent activities, including apoptosis induction, inhibition of cell proliferation and angiogenesis, support for antioxidant defense and DNA repair mechanisms, inflammation downregulation, and interference in other signal-transduction systems [fig_ref] Figure 3: Graphical overview of the isoflavones' possible molecular targets and cellular signaling pathways... [/fig_ref]. Studies indicated dose-sensitive dual (i.e., estrogenic and anti-estrogenic) effects of isoflavones via multimodal mechanistic pathways. Contrasting effects (cell proliferation and apoptosis) on breast cancer cells were detected both in vitro and in vivo. Due to isoflavone pleiotropic properties, exceptional heterogeneity of breast cancer molecular features and a multitude of modulating factors it is not always clear under what circumstances these plant-derived compounds act favorably against tumors and when they exert adverse effects by promoting cancer cell proliferation [bib_ref] Understanding genistein in cancer: The "good" and the "bad" effects: A review, Russo [/bib_ref]. At low doses (≤10 μmol/L), major soy isoflavone genistein displayed estrogen-like effects and was found to stimulate the proliferation of hormone-sensitive breast cancer cell lines and tumors in mouse models [bib_ref] Individual factors define the overall effects of dietary genistein exposure on breast..., Liu [/bib_ref] [bib_ref] Molecular mechanisms of anticancer effects of phytoestrogens in breast cancer, Hsieh [/bib_ref] [bib_ref] Soy diets containing varying amounts of genistein stimulate growth of estrogen-dependent (MCF-7)..., Allred [/bib_ref]. The observed pro-tumorigenic effect is likely mediated by the activation of the ERα pathway and attributed to genistein acting as a weak estrogen. However, at higher doses genistein was reported to antagonize the progression of breast neoplasms by triggering Bax, Bcl-2-associated X-protein; E2, Estradiol; EGF, epidermal growth factor; EGFR, Epidermal growth factor receptor; ERα, Estrogen receptor alpha; ERβ, Estrogen receptor beta; ERK, Extracellular-signal-regulated kinase; IGF, insulin-like growth factor; IGFR, insulin-like growth factor receptor; IKKα, IKKβ and IKKγ, IκB ki-nases; MAPK, Mitogen-activated protein kinases; MEK, Mitogen-activated protein kinase kinase; mTOR, mammalian target of rapamycin; NF-κB, nuclear factor-κB; PI3K, Phosphoinositide 3-kinase; PTEN, Phosphatase and tensin homolog; Raf, Rapidly accelerated fibrosarcoma; Ras, Rat sarcoma, TKR; Tyrosine kinase receptor.
Studies indicated dose-sensitive dual (i.e., estrogenic and anti-estrogenic) effects of isoflavones via multimodal mechanistic pathways. Contrasting effects (cell proliferation and apoptosis) on breast cancer cells were detected both in vitro and in vivo. Due to isoflavone pleiotropic properties, exceptional heterogeneity of breast cancer molecular features and a multitude of modulating factors it is not always clear under what circumstances these plant-derived compounds act favorably against tumors and when they exert adverse effects by promoting cancer cell proliferation [bib_ref] Understanding genistein in cancer: The "good" and the "bad" effects: A review, Russo [/bib_ref]. At low doses (≤10 µmol/L), major soy isoflavone genistein displayed estrogen-like effects and was found to stimulate the proliferation of hormone-sensitive breast cancer cell lines and tumors in mouse models [bib_ref] Individual factors define the overall effects of dietary genistein exposure on breast..., Liu [/bib_ref] [bib_ref] Molecular mechanisms of anticancer effects of phytoestrogens in breast cancer, Hsieh [/bib_ref] [bib_ref] Soy diets containing varying amounts of genistein stimulate growth of estrogen-dependent (MCF-7)..., Allred [/bib_ref]. The observed pro-tumorigenic effect is likely mediated by the activation of the ERα pathway and attributed to genistein acting as a weak estrogen. However, at higher doses genistein was reported to antagonize the progression of breast neoplasms by triggering apoptosis, inducing cell cycle arrest, and impeding angiogenesis employing estrogen-independent pathways, including the activation of caspase and several endoplasmic reticulum stress-regulators, disabling of VEGF signaling, downregulation of matrix metalloproteinases, inhibition of the MEK5/ERK5/NF-kB, protein tyrosine kinase (PTK), mitogen-activated protein kinases (MAPK) and epigenetic modification [bib_ref] Dual effects of isoflavonoids from Pueraria lobata roots on estrogenic activity and..., Ahn [/bib_ref] [bib_ref] Genistein induces cell apoptosis in MDA-MB-231 breast cancer cells via the mitogen-activated..., Li [/bib_ref] [bib_ref] Anti-angiogenic genistein inhibits VEGF-induced endothelial cell activation by decreasing PTK activity and..., Yu [/bib_ref] [bib_ref] Lifetime genistein intake increases the response of mammary tumors to tamoxifen in..., Zhang [/bib_ref] [bib_ref] Epigenetic reactivation of estrogen receptor-α (ERα) by genistein enhances hormonal therapy sensitivity..., Li [/bib_ref].
## Prostate cancer and isoflavones
## Prostate cancer-epidemiology and risk factors
Androgens are essential for the regulation of cell survival, development, differentiation, and proliferation within the prostate gland. However, the development of both normal and abnormal prostatic glandular structures is determined by the complex reciprocal stroma-epithelial interactions at endocrine and paracrine levels.
In benign conditions called benign prostatic hyperplasia (BPH), the intensified proliferation of fibroblasts is stimulated by the increased local conversion of testosterone (T) to estradiol (E2) in the stroma. The enlarged cell population of fibroblasts produces larger amounts of dihydrotestosterone (DHT) and tissue growth factors (GF), which activate the proliferation of the epithelium. In this way, a vicious cycle is created, followed by progressive growth of total prostate volume (TPV) and increased prostate-specific antigen (PSA) synthesis [bib_ref] Molecular and cellular pathogenesis of benign prostatic hyperplasia, Lee [/bib_ref]. A better comprehension of the steroid pathways in prostatic disorders, with an emphasis on intraprostatic androgen levels in particular, may contribute to the personalization and optimization of treatment algorithms and presumably new therapeutic targets. In prostate cancer, stroma-epithelium interaction is maintained. However, multiple mutations and defects alter the cellular signaling pathways, androgen and estrogen receptors, as well as hormonal levels in the tissue.
The vast majority of prostate malignant neoplasms have epithelial origin and differentiation and are considered carcinomas based on histopathological and immunophenotypic attributes. The oncogenesis and pathobiology of prostate cancer are associated with multifarious interactions between inherent germline lineage susceptibility, acquired somatic mutations, and both micro and macro environmental factors. Localized prostate cancer often comprises multiple foci featuring specific genetic alterations with diverse capacities for metastatic colonization and treatment resistance. The identity of cancer-initiating cells in prostatic adenocarcinomas remains controversial, but it is generally acknowledged that during tumorigenesis the epithelium undergoes a series of phenotypic modifications, including cell signaling perturbations, which facilitates the transformation from benign to malignant disease [bib_ref] Pathological and molecular mechanisms of prostate carcinogenesis: Implications for diagnosis, detection, prevention,..., De Marzo [/bib_ref]. The most common type of prostate cancer is adenocarcinoma of the glandular acini of the prostate. Given that the largest percentage of glandular acini are located in the peripheral zone of the prostate, prostate cancer most often occurs in the peripheral zone and, in the initial stages, rarely causes urethral obstruction and dysuric disorders [bib_ref] Prostate zones and cancer: Lost in transition?, Ali [/bib_ref]. During growth, cancer occupies a larger part of the prostate and, if not detected and treated in time, spreads beyond the prostate borders and metastasizes to the pelvic lymph glands and bones.
Prostate cancer risk substantially increases with aging, and it is predominantly diagnosed in men over the age of 65 and very rarely in those younger than 50 [bib_ref] Epidemiology and prevention of prostate cancer, Gandaglia [/bib_ref]. Consequently, prostate cancer incidence is exceptionally high in regions with long life expectancies. Prostate cancer is the most common cancer in men in Europe, America, Australia and Sub-Saharan Africa, accounting globally for approximately 7% of newly diagnosed cancer cases in the male population (15% in developed regions). Additionally, worldwide estimation of prostate cancer-related deaths exceeds 375,000 annually, qualifying it as the fifth leading cause of cancer-associated mortality in men. The general burden of prostate cancer is supposed to increase due to the overall population aging trends and economic growth. Based on the GLOBOCAN 2020 Report [bib_ref] Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for..., Sung [/bib_ref] estimated region-specific age-standardized incidence rates range from 6.3 to 83.4/100,000 men. The highest incidence levels (exceeding 59.0/100,000 males) are reported for Northern and Western Europe, the Caribbean, Australia/New Zealand, Northern America and Southern Africa, while the lowest rates are found in Asia and Northern Africa. Nevertheless, geographic variability of mortality estimates deviates from the incidence patterns with the highest rates being noted for the Caribbean, Africa and Oceania. Remarkable international variation of prostate cancer epidemiological statistics and temporal trends may be attributed to both genetic and environmental influences, including socioeconomic status, lifestyle and nutrition factors, screening programs and healthcare disparities.
Extensive research endeavors are being conducted with the aim to determine and describe variables that affect the risk of prostate cancer. Identified or postulated risk factors for the development of this malignancy are numerous and heterogeneous. Currently acknowledged determinants that modulate the susceptibility to prostate carcinogenesis include genetic, androgen-associated, infectious, inflammatory, ethnic, lifestyle and nutrition pathways [bib_ref] Risk factors for prostate cancer, Patel [/bib_ref] [bib_ref] Human prostate cancer risk factors, Bostwick [/bib_ref] [bib_ref] Risk factors for prostate cancer, Gann [/bib_ref]. A better comprehension of modifiable risk factors may focus and guide the creation and implementation of preventative measures and targeted interventions on the public health and individual level.
Epidemiologic studies indicated a familial clustering of prostate cancer, which proposed that risk is significantly increased among men with an affected first-degree relative. The prostate cancer relative risk increases with the increasing number of family members with prostate cancer in medical history, the grade of genetic relatedness of affected individuals, and the decreasing age at diagnosis [bib_ref] Mendelian inheritance of familial prostate cancer, Carter [/bib_ref] [bib_ref] Familial prostate cancer, Giri [/bib_ref]. Both common environmental exposures and genetic influences may contribute to the familial aggregation of prostate cancer. Higher risk is determined in individuals with an affected brother in comparison with those having an affected father [bib_ref] Empiric risk of prostate carcinoma for relatives of patients with prostate carcinoma:..., Zeegers [/bib_ref]. These generational discrepancies may be attributed to alterations in environmental, nutritional and occupational exposures over the course of time, especially during childhood and adolescence, and the contribution of maternal genetic inheritance.
Based on paleopathologistic research, Homo sapiens was an obligatory herbivore for over 300,000 years of existence until the last 12,000-15,000 years, when it began to introduce the meat of domestic animals into the habitual diet, concomitantly reducing the intake of plant foods. This significant change in evolutionary dietary trajectory is now hypothesized to be an important factor in the rising incidence of breast and prostate cancer in humans. Although discrepancies in national screening practices may impact the detection rate of prostate cancer, the cross-country comparisons of epidemiologic data support the importance of nutrition and dietary pattern as risk or protective factors. The incidence of prostate cancer in Chinese immigrants who live in Western countries is significantly higher compared to Chinese living in China [bib_ref] Comparative epidemiology of cancers of the colon, rectum, prostate and breast in..., Yu [/bib_ref] [bib_ref] East meets West: Ethnic differences in prostate cancer epidemiology between East Asians..., Kimura [/bib_ref] [bib_ref] Cancers of the prostate and breast among Japanese and white immigrants in..., Shimizu [/bib_ref]. Albeit the differences in medical systems, screening programs, and cancer registries cannot be discarded, it is reasonable to postulate that the abandonment of traditional diet and lifestyle may also be significant. Furthermore, a similar trend of increasing incidence of prostate cancer is noted in large Chinese cities, where the intake of meat and animal fats increased by almost 600% in the period from 1960 to 2000 [bib_ref] Trends in mortality from cancers of the breast, colon, prostate, esophagus, and..., Zhang [/bib_ref].
The association between infections, infection-mediated chronic inflammatory process, and oncogenic transformation is well-acknowledged. Accumulating evidence from basic science, clinical research and epidemiological studies suggests that inflammation may affect prostate carcinogenesis. Although causative agents and drivers of intraprostatic tumorpromoting inflammation remain elusive, it is hypothesized that proliferative inflammationinduced atrophy featuring cell infiltrates elicited by unknown stimuli, which potentially include microbial pathogens, might serve as a precancerous lesion leading to prostatic intraepithelial neoplasia and/or prostate cancer [bib_ref] Global burden of cancers attributable to infections in 2008: A review and..., Martel [/bib_ref].
Copious scientific evidence supports the notion of the vitamin D signaling-cascade role in prostate cancer and other oncopathologies. Epidemiologic data suggest that vitamin D status may be significant in the causation of prostate cancer and a plausible target for the prevention and treatment of this malignancy. A higher risk for prostate cancer and/or poor prognosis was reported for population groups with low vitamin D intake and 25(OH)D 3 serum concentration, indicating (subclinical) vitamin D deficiency, including elderly, men with limited sun exposure or living in higher latitudes, and African-Americans whose skin pigmentation is associated with decreased intracutaneous vitamin D synthesis. Vitamin D receptor is found in both normal and cancerous prostate tissue, and it is postulated that its expression and certain polymorphisms may represent predictive markers of prostate cancer progression and outcome beyond standard clinical indicators [bib_ref] Expression patterns of vitamin D receptor in human prostate, Krill [/bib_ref] [bib_ref] Vitamin D and the epidemiology of prostate cancer, Schwartz [/bib_ref]. The potential translation of these observations into preventive or prostate cancer management strategies warrants further research and carefully designed clinical trials.
The increase in human lifespan is another factor that enables the cumulative effect of carcinogens, thus imposing significant global health challenges. In the Neolithic period, the average life expectancy of humans was approximately 20 years, and nowadays, it is estimated to be 70-80 years. To accurately assess the social burden caused by a particular disease and thereby effectively allocate healthcare resources, it is essential to comprehend the impact of shifting population age structure. The prevalence of prostate cancer undeniably and substantially increases with age, and based on autopsy studies, the estimated rate varies from 5% among men under 30 years of age to almost 59% among men older than 79 years [bib_ref] Prevalence of incidental prostate cancer: A systematic review of autopsy studies, Bell [/bib_ref]. Cancer statistics and population projections indicate that prostate cancer will probably become more prevalent in a globally aging population. Nevertheless, aging-adjusted morbidity rates in the Asian region and the developing Western countries are anticipated to rise considerably more than in the developed Western world, where they have already plateaued and may even be declining with time.
Several studies demonstrated an association between obesity and an increased risk of developing prostate cancer. Furthermore, obesity may affect tumor growth, the prognosis of the disease, recurrence rates and overall mortality estimates [bib_ref] The impact of body mass index on prostate cancer: An updated systematic..., Tzenios [/bib_ref] [bib_ref] Prospective study of adiposity and weight change in relation to prostate cancer..., Wright [/bib_ref] [bib_ref] Body mass index and mortality in prostate cancer patients: A dose-response meta-analysis, Zhong [/bib_ref] [bib_ref] Body mass index and weight change in men with prostate cancer: Progression..., Bonn [/bib_ref] [bib_ref] Weight change and prostate cancer incidence and mortality, Bassett [/bib_ref]. Although the association between prostate cancer and the alteration of hormone status mediated by obesity and endocrine activity of adipose tissue is still controversial, it has been shown that obese men exert increased estradiol, insulin, leptin and free IGF-1 levels concomitantly with the decreased free T concentration, and declining luteinizing hormone (LH) pulse amplitude [bib_ref] Obesity and prostate cancer: Epidemiology and clinical implications, Buschemeyer [/bib_ref]. Moreover, the multifaceted interplay between obesity-induced low-grade systemic inflammation and immune response may be a significant factor in the progression of prostate cancer. It is hypothesized that several chemokines and cytokines secreted from adipose depots may cause alterations of the local immune cellular profile in the tumor microenvironment, including myeloid-derived suppressor cells, cancer-associated neutrophils, macrophages, B-cells, immunoglobulins and the complement system, thereby promoting proliferation, invasion and metastatic seeding. Another possible biological vehicle of obesity-related prostate cancerogenesis refers to host immune-system modulation triggered by intestinal microbiome alterations. Given the scarcity of available data, additional research is needed to fully elucidate precise mechanisms in the background of this phenomenon.
## The effects of isoflavones on prostate cancer
Several investigators reported that genistein might suppress the PSA expression in prostate cancer cell lines and arrest the cell cycle progression, presumably by interfering with tyrosine kinases vital for mitogenic signal transduction [bib_ref] Genistein and biochanin A inhibit the growth of human prostate cancer cells..., Peterson [/bib_ref] [bib_ref] Genistein-induced upregulation of p21WAF1, downregulation of cyclin B, and induction of apoptosis..., Davis [/bib_ref]. A histoculture study of surgical specimens of human BPH and prostate cancer was performed using the preestablished hormone-sensitive three-dimensional collagen gel matrix. It was demonstrated that at a dosage of 1.25-10 µg/mL genistein inhibits the growth of both BPH and prostate cancer tissues in a dose-dependent manner as measured by 3 H-thymidine DNA incorporation per mg protein, with a negligible additional effect in higher concentrations [bib_ref] Genistein inhibits the growth of human-patient bph and prostate cancer in histoculture, Geller [/bib_ref].
The primary action of genistein on prostate cancer cells at pharmacological doses appears to be the induction of apoptosis. Substantial evidence derived from in vitro studies suggests that genistein inhibits the growth and metastatic dissemination of diverse types of both hormone-dependent and hormone-independent prostate and breast tumors, displaying the half-maximal inhibitory concentration (IC 50 ) ranging from 5 to 40 nM (i.e., 2-10 mg/mL) [bib_ref] Genistein and biochanin A inhibit the growth of human prostate cancer cells..., Peterson [/bib_ref] [bib_ref] A mini-review of flavone isomers apigenin and genistein in prostate cancer treatment, Ji [/bib_ref] [bib_ref] Genistein-induced apoptosis of prostate cancer cells is preceded by a specific decrease..., Kyle [/bib_ref] [bib_ref] Genistein inhibits proliferation and in vitro invasive potential of human prostatic cancer..., Santibáñez [/bib_ref]. The mechanism by which genistein acts as an anticancer agent is probably based on targeting multiple oncogenic pathways, including PLK1, EGFR, Wnt, Akt, JAK/STAT, epithelial-mesenchymal transition, and ER modulation [fig_ref] Figure 4: Graphical overview of the isoflavones' possible molecular targets and cellular signaling pathways... [/fig_ref]. Furthermore, genistein inhibits the activity of DNA topoisomerase II and increases in vitro concentration of TGFb (transforming growth factor b), which can inhibit the growth of malignant cells [bib_ref] Molecular targets of genistein and its related flavonoids to exert anticancer effects, Chae [/bib_ref] [bib_ref] A novel anti-cancer effect of genistein: Reversal of epithelial mesenchymal transition in..., Zhang [/bib_ref].
playing the half-maximal inhibitory concentration (IC50) ranging from 5 to 40 nM (i.e., 2-10 mg/mL) [bib_ref] Genistein and biochanin A inhibit the growth of human prostate cancer cells..., Peterson [/bib_ref] [bib_ref] A mini-review of flavone isomers apigenin and genistein in prostate cancer treatment, Ji [/bib_ref] [bib_ref] Genistein-induced apoptosis of prostate cancer cells is preceded by a specific decrease..., Kyle [/bib_ref] [bib_ref] Genistein inhibits proliferation and in vitro invasive potential of human prostatic cancer..., Santibáñez [/bib_ref]. The mechanism by which genistein acts as an anticancer agent is probably based on targeting multiple oncogenic pathways, including PLK1, EGFR, Wnt, Akt, JAK/STAT, epithelial-mesenchymal transition, and ER modulation [fig_ref] Figure 4: Graphical overview of the isoflavones' possible molecular targets and cellular signaling pathways... [/fig_ref]. Furthermore, genistein inhibits the activity of DNA topoisomerase II and increases in vitro concentration of TGFb (transforming growth factor b), which can inhibit the growth of malignant cells [bib_ref] Molecular targets of genistein and its related flavonoids to exert anticancer effects, Chae [/bib_ref] [bib_ref] A novel anti-cancer effect of genistein: Reversal of epithelial mesenchymal transition in..., Zhang [/bib_ref]. The growth-inhibitory potential of genistein and vitamin D compounds on prostate epithelial cells was examined. Genistein and the hormonally active form of cholecalciferol synergistically suppressed the proliferation of primary human prostatic epithelial cells (HPEC) and prostate cancer cells, according to the isobolographic study. Based on the flow-cytometry analysis, genistein induced an arrest in the G(2)M phase of the cell cycle, while 1α,25-dihydroxycholecalciferol and low-calcemic 25-hydroxycholecalciferol caused an arrest in the G(1/0) phase [bib_ref] Genistein and vitamin D synergistically inhibit human prostatic epithelial cell growth, Rao [/bib_ref].
Another study demonstrated that in an androgen-depleted LNCaP cell system, genistein dose-dependently inhibits the DHT-induced expression of the prostate androgen-regulated transcript 1 gene (PART-1) with complete inhibition obtained at the concentration of 50 mM. These results suggested that PART-1 may serve as a promising marker for assessing the effect of isoflavones (and soy products) in prostate cancer prevention. Nevertheless, additional research efforts, in particular animal studies, are required to explore whether these compounds inhibit prostate tumor growth via PART-1 expression inhibition in vivo [bib_ref] Genistein and daidzein downregulate prostate androgen-regulated transcript-1 (PART-1) gene expression induced by..., Yu [/bib_ref]. Studies indicated that genistein inhibits tumor growth and induces apoptosis at concentrations ≤20 mM in both androgen-dependent and independent human prostate adenocarcinoma cell lines, such as LNCaP, DU-145 and PC3 The growth-inhibitory potential of genistein and vitamin D compounds on prostate epithelial cells was examined. Genistein and the hormonally active form of cholecalciferol synergistically suppressed the proliferation of primary human prostatic epithelial cells (HPEC) and prostate cancer cells, according to the isobolographic study. Based on the flow-cytometry analysis, genistein induced an arrest in the G(2)M phase of the cell cycle, while 1α,25-dihydroxycholecalciferol and low-calcemic 25-hydroxycholecalciferol caused an arrest in the G(1/0) phase [bib_ref] Genistein and vitamin D synergistically inhibit human prostatic epithelial cell growth, Rao [/bib_ref].
Another study demonstrated that in an androgen-depleted LNCaP cell system, genistein dose-dependently inhibits the DHT-induced expression of the prostate androgenregulated transcript 1 gene (PART-1) with complete inhibition obtained at the concentration of 50 mM. These results suggested that PART-1 may serve as a promising marker for assessing the effect of isoflavones (and soy products) in prostate cancer prevention. Nevertheless, additional research efforts, in particular animal studies, are required to explore whether these compounds inhibit prostate tumor growth via PART-1 expression inhibition in vivo [bib_ref] Genistein and daidzein downregulate prostate androgen-regulated transcript-1 (PART-1) gene expression induced by..., Yu [/bib_ref]. Studies indicated that genistein inhibits tumor growth and induces apoptosis at concentrations ≤20 mM in both androgen-dependent and independent human prostate adenocarcinoma cell lines, such as LNCaP, DU-145 and PC3 [bib_ref] Phytoestrogens in common herbs regulate prostate cancer cell growth in vitro, Shenouda [/bib_ref] [bib_ref] Low-dose genistein induces cyclin-dependent kinase inhibitors and G1 cell-cycle arrest in human..., Shen [/bib_ref] [bib_ref] Genistein induces cell growth inhibition in prostate cancer through the suppression of..., Ouchi [/bib_ref]. Moreover, in vitro research suggested that genistein acts by inhibiting NF-κB (nuclear factor kappa-light-chainenhancer of activated B cells) in various cells, suppresses metalloproteinases associated with cancer and inhibits telomerase activity via transcriptional downregulation of human telomerase reverse transcriptase (hTERT) [bib_ref] Inactivation of nuclear factor kappaB by soy isoflavone genistein contributes to increased..., Li [/bib_ref] [bib_ref] Genistein suppresses the invasive potential of human breast cancer cells through transcriptional..., Kousidou [/bib_ref].
Given that chemopreventive compounds should be administered at biologically effective doses over extended periods of time without causing toxicity and health-compromising adverse effects, research on genistein considering mechanisms of action at physiologically relevant concentrations is of particular importance. It was found that physiological concentrations of genistein (comparable to those observed in the sera of Asian male habitual soy consumers) decrease the androgen receptor (AR) expression on mRNA and protein levels in the androgen-sensitive LNCaP line, coupled with a dose-dependent reduction of PSA secretion. Pure anti-estrogen caused the abrogation of the inhibitory effect of genistein, confirming the hypothesis that the downregulation of AR was mediated by the ERβ [bib_ref] Androgen receptor regulation by physiological concentrations of the isoflavonoid genistein in androgen-dependent..., Bektic [/bib_ref]. Tumor-suppressive features of genistein on PC3 and DU145 cells were investigated by conducting a functional assay cluster analysis. Genistein was reported to inhibit growth of prostate cancer cells by downregulating the HOX transcript antisense RNA (HOTAIR) gene located on chromosome 12. Based on luciferase reporter assay, real-time PCR and long non-coding RNA profiling, HOTAIR was directly targeted by miR-34a, significantly regulated by genistein, and compared to normal prostate cells its expression was elevated in castration-resistant cancer cell lines. Knockdown of HOTAIR induced apoptotic cascade, diminished prostate cancer cell proliferation, migration potential and invasiveness. Additionally, bioinformatic analysis and mRNA array data implied that MMP9 and VEGF genes, key components of the KEGG pathway, are among genistein's targets [bib_ref] Genistein inhibits prostate cancer cell growth by targeting miR-34a and oncogenic HOTAIR, Chiyomaru [/bib_ref]. Genistein also modulates the expression of several microRNAs (miRNAs), thus contributing to a better comprehension of its pleiotropic biological effects and antitumorigenic potential. Genistein may up-regulate the expression of the tumor suppressor miR-574-3p in clinical prostate cancer samples and prostate cancer cell lines, thereby causing apoptosis through Bcl-xL, caspase-9 and caspase-3 pathways regulation [bib_ref] Genistein up-regulates tumor suppressor microRNA-574-3p in prostate cancer, Chiyomaru [/bib_ref].
The effects of isoflavonoids and other phytoestrogens on castration-resistant prostate cancer were discussed in a review article [bib_ref] Phytoestrogens selective for the estrogen receptor beta exert anti-androgenic effects in castration..., Thelen [/bib_ref]. Androgen ablation represents the goldstandard therapy for the advanced non-organ-confined illness with survival-prolonging potency but lacking a curative perspective. The AR-positive human prostate cancer cell line LNCaP (derived from a metastatic prostate cancer lymph node lesion featuring androgensensitive growth and a mutation in the ligand binding domain of the AR) is the principal culture model used for investigating the complex interaction of ER and AR axis in prostate cancer. The gain-of-function T877A mutation expands the range of activating ligands for the AR and assigns a specific key castration-resistant prostate cancer pathway in these cells. Functional studies of isoflavones on this representative model indicated that these compounds exert ER ligand capacity with a preference for the ERβ subtype leading to tumor-suppressing activity with an anti-androgenic signature. Apparently, competitive binding to the AR is not how phytoestrogens mediate their effects. The ERβ is virtually silenced during prostate carcinogenesis and phytoestrogen treatment causes its up-regulation. ER-knock-down essentially mimics the androgen stimulation, and in an intricate hormone receptors' cross-talk, the re-expression of the ERβ results in lower expression of the following: recognized androgenic markers PSA and PCA3, the androgen co-activator PDEF and the insulin-like growth factor-1 receptor (IGF-1). Given that ERβ release from epigenetic silencing elicits beneficial antiandrogenic effects in prostate cancer models, the subtypeselective ER ligands of both natural and synthetic origin may have a significant role in novel preventive and curative strategies for this oncopathology [bib_ref] The evolving role of oestrogens and their receptors in the development and..., Bonkhoff [/bib_ref] [bib_ref] The insulin and insulin-like growth factor receptor family in neoplasia: An update, Pollak [/bib_ref].
Critically analyzing available literature sources and acknowledging their chemical features, dietary sources, bioavailability and safety profile, Spagnuolo et al. provided a summarized overview of the genistein chemopreventive molecular mechanisms of action [bib_ref] Genistein and cancer: Current status, challenges, and future directions, Spagnuolo [/bib_ref]. Copious evidence supports the notion of its multifaceted effects on apoptosis induction, cell cycle arrest, and inhibition of angiogenesis and metastasis. Target signaling pathways include caspases, Bax (B cell lymphoma 2 (Bcl-2)-associated X protein), KIF20A (Bcl-2, kinesin-like protein 20A), ERK1/2 (extracellular signal-regulated kinase 1/2), NF-κB (nuclear transcription factor κB), MAPK (mitogen-activated protein kinase), IκB (inhibitor of NF-κB), Wnt/β-catenin (Wingless and integration 1 β-catenin) and PI3K/Akt (phosphoinositide 3 kinase/Akt) [fig_ref] Figure 4: Graphical overview of the isoflavones' possible molecular targets and cellular signaling pathways... [/fig_ref]. Furthermore, genistein may act synergistically with some well-known chemotherapeutics, such as adriamycin, docetaxel, and tamoxifen and potentiate their efficacy without causing adverse effects. Despite evident benefits, genistein has two limiting characteristics-high pleiotropy coupled with relatively low bioavailability. Therefore, there is a need for further research in order to clarify genistein's pharmacody-namics and pharmacokinetics, elucidate supplementary molecular targets, define effective cancer-specific therapeutic dosage, explore approaches to increase its bioavailability and evaluate possible interactions with other pharmacological entities [bib_ref] Genistein and cancer: Current status, challenges, and future directions, Spagnuolo [/bib_ref].
## Animal studies
A dose of 1 mg genistein/g food was reported to significantly reduce prostate weight in rats and downregulate the EGF pathway and ErbB2/Neu receptor without toxic/adverse effects for the host in a 3-week treatment [bib_ref] Genistein, a component of soy, inhibits the expression of the EGF and..., Dalu [/bib_ref]. Genistein was suggested to inhibit cancer cell growth by modulating transforming growth factor (TGF)β-1 [bib_ref] Mechanisms of action of the soy isoflavone genistein: Emerging role for its..., Kim [/bib_ref]. Another mechanism by which genistein counteracts prostate growth in rats is the inhibition of increased 5αR activity induced by exposure to a high-fat diet [bib_ref] S. 17α-Estradiol and genistein inhibit high fat diet induced prostate gene expression..., Cai [/bib_ref].
The transgenic adenocarcinoma mouse prostate (TRAMP) model was created as a crucial research tool for better comprehension of the development of adenocarcinomas. New opportunities for chemoprevention studies and the development of methods to counteract particular genetic determinants of cancer susceptibility were facilitated by these genetically modified animals. In situ and invasive prostate cancer are both developed in TRAMP mice, mimicking the comprehensive gamut of human prostate cancer progression from prostatic intraepithelial neoplasia to massive multinodular malignancy, as well as androgen-independent illness. Studies conducted on these autochthonous prostate cancer rodent systems proved that isoflavone-abundant diet leads to a significant decrease in tumor growth, incidence of spontaneous tumor formation and the occurrence of prostate cancer metastases [bib_ref] Prevention of spontaneous prostate-related cancer in Lobund-Wistar rats by a soy protein..., Pollard [/bib_ref] [bib_ref] Genistein in the diet reduces the incidence of poorly differentiated prostatic adenocarcinoma..., Mentor-Marcel [/bib_ref] [bib_ref] Genistein chemoprevention of prostate cancer in TRAMP mice, Wang [/bib_ref]. Studies showed that high concentrations of genistein administered subcutaneously in prostate cancer-implanted animals might exert inhibitory action towards a number of proteins involved in primary tumor growth and apoptosis, which was not seen at doses relevant to the human diet [bib_ref] The inhibitory effect of genistein on the growth and metastasis of a..., Schleicher [/bib_ref]. Given that mortality from prostate cancer is attributable to advanced metastatic disease, regulation of the metastatic cascade (i.e., the sequence of steps malignant cells must undergo to escape the primary organ, disseminate, implant, survive and divide in a distant location, thus forming a secondary mass) is of paramount importance. According to several in vitro discoveries, corroborated by both in vivo animal research and early-phase human clinical trials, genistein can modulate metastatic potential markers and reduce human cancer metastasis. At lower concentrations comparable to those obtained through dietary consumption, genistein may suppress the prometastatic processes of cancer cellular detachment, adhesion, migration, invasion, and finally, implantation and growth at the distant site through a number of intricately interrelated pathways [bib_ref] Inhibition of cancer cell invasion and metastasis by genistein, Pavese [/bib_ref] [bib_ref] Genistein inhibits human prostate cancer cell detachment, invasion, and metastasis, Pavese [/bib_ref].
A study investigating the impact of genistein on sex steroid receptor expression in rats demonstrated that both short-term and life-long exposure to amounts that approximate physiological concentrations in humans consuming a soy-rich diet cause a significant reduction of AR, ER-α and ER-β mRNAs in the dorsolateral prostate, without evident histomorphological and functional toxicity to the male reproductive system. Evidence of combined down-regulation of sex steroid receptor expression and growth factors signaling cascades imply that genistein may concomitantly modulate endocrine and paracrine tissue growth [bib_ref] Dietary genistein down-regulates androgen and estrogen receptor expression in the rat prostate, Fritz [/bib_ref].
Cumulative results derived from seven studies featuring transgenic adenocarcinoma of mouse prostate (TRAMP) mice explored the extent of genistein effects that could be attributed to estrogenic activity. Although the results corroborated the premise of protective ERβ and prooncogenic ERα, dietary genistein reduced the incidence of cancer only in the ER wild-type (WT)/TRAMP mice but not in transgenic mice lacking functional ERα or ERβ, i.e., ER knockout (KO)/TRAMP mice. These findings suggest that genistein may require both receptors to exert protective effects on prostate carcinogenesis [bib_ref] Aggressive prostate cancer is prevented in ERαKO mice and stimulated in ERβKO..., Sĺusarz [/bib_ref].
## Human studies
Several studies have addressed the potential hormonal impact of soy consumption among men exploring both beneficial and adverse effects. Monitoring of plasma levels of steroid compounds in men who consumed 1.06 L of soy milk per day for a one-month period showed a decrease in androgen metabolism, manifested as the decline of DHT metabolite 3α,17β-ADG (androstanediol glucuronide) concentration [bib_ref] Effects of soya consumption for one month on steroid hormones in premenopausal..., Lu [/bib_ref]. The absence of clinically significant behavioral or physical evidence of genistein toxicity for doses up to 16 mg/kg body weight was found in a safety study featuring single administration of purified unconjugated isoflavones to healthy men. Despite exceeding manifold dietaryrelevant intake levels these concentrations did not cause alteration of the hypothalamicpituitary-gonadal axis, estrogenic or antiestrogenic symptoms. Pharmacokinetic evaluation of the applied isoflavone mixture confirmed rapid plasma clearance of both genistein and daidzein and predominant urine excretion. The estimated mean elimination half-lives were 3.2 h and 4.2 h for free forms of genistein and daidzein, respectively [bib_ref] Clinical characteristics and pharmacokinetics of purified soy isoflavones: Single-dose administration to healthy..., Busby [/bib_ref].
The level of isoflavones in prostate tissue was obtained after TURP (transurethral resection of the prostate) and radical cystoprostatectomy using gas chromatography-mass spectrometry (GS-MS). Prostate volumes were ≥40 mL and <25 mL in the TURP group and in the cystoprostatectomy control group, respectively. No difference was found in the concentration of enterodiol, enterolactone, equol and daidzein in BPH tissue and normal prostate tissue. However, a significantly lower concentration of genistein was found in BPH than in the control group [bib_ref] Comparative study of concentration of isoflavones and lignans in plasma and prostatic..., Hong [/bib_ref].
Over a 12-month intervention a soy protein drink was administered to a group of 81 older healthy men with low PSA. In this double-blinded, parallel-arm trial, subjects were randomly assigned to either consume a soy protein drink providing 83 mg of isoflavones in a daily dose or a beverage with negligible isoflavone content. Based on the radioimmunometric assay, no statistically significant differences in PSA levels were recorded after 12 months [bib_ref] Soy isoflavones do not modulate prostate-specific antigen concentrations in older men in..., Adams [/bib_ref].
In a prospective study conducted in Hong Kong, 176 participants with BPH and lower urinary tract symptoms were randomly allocated to either intervention group (receiving 40 mg of isoflavones per day; Soylife 40) or placebo-control group in a double-blind manner. Patients tolerated isoflavones well over 12 months, but they were only slightly superior to placebo in terms of clinical improvement and laboratory analyses.
The striking international and interethnic disparity in the prostate cancer incidence rates, coupled with the increased risk in migrants moving from low to high-risk areas, suggest that modifiable environmental determinants, such as certain dietary agents, might modulate the risk of disease development and progression. A wide variety of molecular signaling pathways are implicated in prostate carcinogenesis and disease progression, and multiple lines of evidence have emerged concerning the anticancer and therapeutic benefits of nutritional factors. Substantial epidemiological data indicate a significantly lower occurrence of prostate cancer in Asia, where soy-based foods are traditionally major diet staples [bib_ref] Soy consumption and prostate cancer risk in men: A revisit of a..., Yan [/bib_ref] [bib_ref] Risks and benefits of dietary isoflavones for cancer, Andres [/bib_ref]. These studies aroused research interest and inspired the hypothesis that soy isoflavones may exert a chemopreventive potential relevant to prostate cancer. The fact that latent prostate cancer is twice as common in the USA as it is in Japan, but the incidence of clinically manifested prostate cancer is up to 15 times higher, supports this hypothesis by indicating that there may be some factors that prevent or delay the onset of clinically significant prostate cancer in Japanese men [bib_ref] Cancers of the prostate and breast among Japanese and white immigrants in..., Shimizu [/bib_ref]. Approximately a 100-fold variation was determined in age-adjusted prostate cancer mortality rates between certain geographic and racial groups underpinning the prominence of environmental exposure in the etiology of this oncopathology.
In 1998 a large multinational cross-sectional study exploring predictive variables for prostate cancer mortality found that regular consumption of soy products is significantly protective [bib_ref] Nutritional and socioeconomic factors in relation to prostate cancer mortality: A cross-national..., Hebert [/bib_ref]. Despite encouraging epidemiological and preclinical data, it is challenging to draw definite conclusions regarding the clinical efficacy of isoflavones due to great inter-study variability in terms of research designs, limited samples, insufficient treatment duration and the absence of standardized pharmaceutical formulations [bib_ref] Soy isoflavone genistein in prevention and treatment of prostate cancer, Perabo [/bib_ref]. A metaanalysis conducted in 2014 to assess the efficacy and safety of soy isoflavones among men with histologically confirmed prostate cancer or at identified risk for the development of this malignancy indicated that there might be scientific support for epidemiological findings implying the role of soy isoflavones in prostate cancer risk reduction, although most studies included were underpowered and had considerable limitations. Nonetheless, understanding of isoflavone impact on sex hormone endpoints and PSA levels remained unclear [bib_ref] Soy and soy isoflavones in prostate cancer: A systematic review and meta-analysis..., Van Die [/bib_ref].
Another meta-analysis showed that high consumption of genistein and daidzein, as well as elevated serum enterolactone concentration, were correlated with a significant prostate cancer risk reduction [bib_ref] Phytoestrogens and risk of prostate cancer: A meta-analysis of observational studies, He [/bib_ref]. A population-based case-control study found phytoestrogen subclass inconsistencies regarding the effect on prostate cancer. Specifically, an inverse association was determined for isoflavones and genistein, whereas a high intake of lignans (particularly lariciresinol, pinoresinol, matairesinol and secoisolariciresinol) was related to an increased prostate cancer risk [bib_ref] Association between dietary phytoestrogens intakes and prostate cancer risk in Sicily, Russo [/bib_ref].
A trial designed to investigate the effects of acute exposure to a dietary supplement containing 160 mg of standardized red clover-derived dietary isoflavones in men with clinically significant prostate cancer prior to radical prostatectomy found significantly higher cancer cell apoptosis in the intervention group [bib_ref] Induction of apoptosis in low to moderate-grade human prostate carcinoma by red..., Jarred [/bib_ref]. In another trial, patients with confirmed prostate neoplasia were treated with multiple-dose soy formulations containing 300/600 mg of genistein and 150/300 mg of daidzein for three months. Oral administration of these soy formulations led to a 32% decrease in serum DHT concentration and relatively mild estrogenic effects [bib_ref] Clinical characteristics and pharmacokinetics of purified soy isoflavones: Multiple-dose administration to men..., Fischer [/bib_ref]. In a prospective multicenter trial, male subjects with localized prostate cancer were randomized to receive either cholecalciferol (200,000 IU in a single dose at the starting point of the study) and genistein (600 mg daily) or placebo during the pre-prostatectomy period. Quantitative immunohistochemistry analyses revealed increased apoptosis and AR expression in the treatment-receiving group compared to the placebo arm. Given that AR signaling plays a critical role in prostate cell growth, differentiation, and function, further research is warranted to elucidate the therapeutic potential and the bioactivity of the applied combination [bib_ref] Effects of a genistein-rich extract on PSA levels in men with a..., White [/bib_ref].
# Conclusions
A broad spectrum of molecular signaling pathways and oncogenic cascades are implicated in carcinogenesis and disease progression in prostate and breast tissue. Diverse research avenues have explored the isoflavone capacity to prevent the neoplastic transformation or disrupt the biology of malignant cells leading to multiple lines of evidence supporting the notion of their chemopreventive potential. In addition to most extensively studied tissue-specific, concentration-sensitive hormone-mediated activities, substantial evidence emerging from both in vitro and in vivo studies imply that isoflavones may deliver anticancer effects by several alternative hormone-independent mechanisms, including apoptosis induction, cell-cycle hindrance, inhibition of angiogenesis, inflammation downregulation, antioxidant defense promotion, support for DNA repair mechanisms, and interference with other signal-transduction cascades. Due to isoflavone pleiotropic properties, heterogeneous molecular signature of breast and prostate oncopathologies, and a plethora of additional relevant risk-modulating factors, comprehensive characterization of the intricate mechanistic bioactivity of isoflavones warrants further research.
## Institutional review board statement: not applicable.
## Informed consent statement: not applicable.
Data Availability Statement: Data sharing is not applicable to this article.
## Conflicts of interest:
The authors declare no conflict of interest.
[fig] Figure 1: Schematic representation of the phenolic compounds and their chemical structures. [/fig]
[fig] Figure 2: Structures of isoflavones glycosides: daidzin, genistin and glycitin. [/fig]
[fig] Figure 3: Graphical overview of the isoflavones' possible molecular targets and cellular signaling pathways in the breast cancer model. Abbreviations: AKT, Protein kinase-B; Bcl-2, B-cell lymphoma 2; Bcl-xL, B-cell lymphoma-extra large; Bax, Bcl-2-associated X-protein; E2, Estradiol; EGF, epidermal growth factor; EGFR, Epidermal growth factor receptor; ERα, Estrogen receptor alpha; ERβ, Estrogen receptor beta; ERK, Extracellular-signal-regulated kinase; IGF, insulin-like growth factor; IGFR, insulin-like growth factor receptor; IKKα, IKKβ and IKKγ, IκB ki-nases; MAPK, Mitogenactivated protein kinases; MEK, Mitogen-activated protein kinase kinase; mTOR, mammalian target of rapamycin; NF-κB, nuclear factor-κB; PI3K, Phosphoinositide 3-kinase; PTEN, Phosphatase and tensin homolog; Raf, Rapidly accelerated fibrosarcoma; Ras, Rat sarcoma, TKR; Tyrosine kinase receptor. [/fig]
[fig] Figure 4: Graphical overview of the isoflavones' possible molecular targets and cellular signaling pathways in the prostate cancer model. Abbreviations: AKT, Protein kinase-B; AR, Androgen receptor; DHT, Dihydrotestosterone; E2, Estradiol; ERα, Estrogen receptor alpha; ERβ, Estrogen receptor beta; ERK, Extracellular-signal-regulated kinase; GPCR, G protein-coupled receptor; IKKα, IKKβ and IKKγ, IκB kinases; MEK, Mitogen-activated protein kinase kinase; mTOR, mammalian target of rapamycin; NF-κB, nuclear factor-κB; PI3K, Phosphoinositide 3-kinase; PTEN, Phosphatase and tensin homolog; Raf, Rapidly accelerated fibrosarcoma; Ras, Rat sarcoma; TKR, Tyrosine kinase receptor. [/fig]
[fig] Author: Contributions: Conceptualization, Ž.T., T.P., L.P. and I.V.; writing-original draft preparation, T.P., M.Z., U.B., M.K., I.V., M.B., L.P. and Ž.T.; writing-review and editing, T.P., M.Z., U.B., M.K., I.V., M.B., L.P. and Ž.T.; visualization, M.Z. and M.K.; supervision, Ž.T.; funding acquisition, Ž.T., I.V. and M.K. All authors have read and agreed to the published version of the manuscript. Funding: The authors acknowledge the financial support from the Ministry of Education, Science and Technological Development of the Republic of Serbia (Contracts Nos: 451-03-68/2022-14/200168 (Živoslav Tešić); 451-03-68/2022-14/200051 (Milica Kalaba); Agreement on Realization and Financing of Scientific Research work of NIO for 2022, No.451-03-68/2022-14/200015 (Milica Zeković)) and from the Slovenian Research Agency (research program P1-0005 "Functional food and food supplements" (Irena Vovk)). [/fig]
[table] Table 1: Summary of discussed epidemiological studies/meta-analysis on isoflavones intake and breast cancer in women.For the high-dose category intake of isoflavones reduced the BCa risk in Asian populations (combined RR/OR = 0.68, 95% CI: 0.52-0.89) but not in Western populations (combined RR/OR = 0.98, 95% CI: 0.8-1.11) Among Asian women results implied a dose-dependent, statistically significant association between BCa RR and habitual soy food intake (approximately 16% risk decrease per 10 mg of isoflavones intake/day). No association was found in low-soy consuming Western populations. [/table]
|
Seeing Mixed Emotions: The Specificity of Emotion Perception From Static and Dynamic Facial Expressions Across Cultures
Although perceivers often agree about the primary emotion that is conveyed by a particular expression, observers may concurrently perceive several additional emotions from a given facial expression. In the present research, we compared the perception of two types of nonintended emotions in Chinese and Dutch observers viewing facial expressions: emotions which were morphologically similar to the intended emotion and emotions which were morphologically dissimilar to the intended emotion. Findings were consistent across two studies and showed that (a) morphologically similar emotions were endorsed to a greater extent than dissimilar emotions and (b) Chinese observers endorsed nonintended emotions more than did Dutch observers. Furthermore, the difference between Chinese and Dutch observers was more pronounced for the endorsement of morphologically similar emotions than of dissimilar emotions. We also obtained consistent evidence that Dutch observers endorsed nonintended emotions that were congruent with the preceding expressions to a greater degree. These findings suggest that culture and morphological similarity both influence the extent to which perceivers see several emotions in a facial expression.
more nonintended emotion words than Westerners to describe facial expressions . This may reflect Easterners' interdependent, holistic view of the self in relation to the social environment. Specifically, Easterners emphasize the relation of the self with others, tend to attend to the entire field, and make relatively little use of categories. In contrast, Westerners emphasize the independent self and pay attention primarily to the focal object and the categories to which it belongs [bib_ref] Culture, control, and perception of relationships in the environment, Ji [/bib_ref] [bib_ref] Culture and the self: Implications for cognition, emotion, and motivation, Markus [/bib_ref] [bib_ref] Culture and systems of thought: Holistic versus analytic cognition, Nisbett [/bib_ref]. As a result, Easterners may be more attuned to the presence of multiple different emotions (and their interrelations) within an expression compared with Westerners, who would be attuned primarily to the most salient emotion category.
In addition to cultural differences in terms of interdependence/independence and holistic/analytic cognitive processes, Easterners are considered to be more dialectical thinkers than Westerners [bib_ref] Culture, dialectics, and reasoning about contradiction, Peng [/bib_ref] [bib_ref] Naïve dialecticism and the Tao of Chinese thought, Peng [/bib_ref]. This means that Easterners are less troubled by apparent contradictions in their own thoughts, emotions, and behaviors as well as those of others [bib_ref] When expressions of fake emotions elicit negative reactions: The role of observers'..., Hideg [/bib_ref]. In contrast, Westerners typically look to reconcile inconsistencies and are more disturbed by discrepancies (Spencer-Rodgers, [bib_ref] Dialectical self-esteem and East-West differences in psychological well-being, Spencer-Rodgers [/bib_ref]. As specific emotions are associated with distinct combinations of appraisals and action tendencies [bib_ref] From appraisal to emotion: Differences among unpleasant feelings, Ellsworth [/bib_ref] [bib_ref] Relations among emotion, appraisal, and emotional action readiness, Frijda [/bib_ref] [bib_ref] On the nature and function of emotion: A component process approach, Scherer [/bib_ref] , endorsing concurrent distinct emotions is likely to represent a contradictory belief. Easterners should therefore be more likely to endorse multiple emotions than Westerners. Based on these interrelated accounts, we predicted that nonintended emotions would be perceived as stronger by Eastern compared with Western perceivers (Hypothesis 2). It is worth noting that this inference is different from the argument that Eastern perceivers cannot reliably distinguish certain facial expressions of emotion in forced-choice tasks due to, for example, perceptual biases (see [bib_ref] Cultural confusions show that facial expressions are not universal, Jack [/bib_ref] [bib_ref] Internal representations reveal cultural diversity in expectations of facial expressions of emotion, Jack [/bib_ref]. We return to this issue in the "General Discussion" section.
Furthermore, we examined whether the tendency of Easterners to endorse nonintended emotions to a greater degree than Westerners would vary as a function of the morphological similarity of emotions. Since an interdependent/holistic view is reflected in the perception of overlapping of elements [bib_ref] Culture, control, and perception of relationships in the environment, Ji [/bib_ref] [bib_ref] Culture and the self: Implications for cognition, emotion, and motivation, Markus [/bib_ref] [bib_ref] Culture and systems of thought: Holistic versus analytic cognition, Nisbett [/bib_ref] , groups that differ in terms of interdependent/holistic thought may vary more strongly on the endorsement of elements that are similar to the target (see . In the present context, Easterners, who hold an interdependent/holistic view, may thus be expected to perceive more nonintended emotions that are close to the intended emotion (i.e., morphologically similar emotions) than Westerners, who hold an independent/analytic view. Given the limited theoretical and empirical support for this account, we conducted an explorative analysis to test whether cultural differences would be more pronounced for morphologically similar emotions than for morphologically dissimilar emotions.
## Dynamic facial expressions
Though the majority of research on facial expressions of emotion employs static images, in real life, facial expressions are typically dynamic, changing from one state (henceforth the start emotion) to another state (the end emotion; [bib_ref] Effects of dynamic aspects of facial expressions: A review, Krumhuber [/bib_ref]. We therefore tested the above-mentioned hypotheses using dynamic emotion displays as well as static emotional expressions, to improve both ecological validity and generalizability. In addition, the inclusion of dynamic displays allowed for an internal replication of our findings from static displays.
Unlike static emotion displays, dynamic emotional expressions include a preceding expression as well as an end expression. Despite the ubiquity of successive contexts in real-life emotion perception, surprisingly, little research has studied the influence of a preceding expression on the perception of an end facial expression. One study by [bib_ref] Relativity in the perception of emotion in facial expressions, Russell [/bib_ref] provided initial evidence for an effect of successive contextual information. In their experiment, participants saw two photographs of different facial expressions in succession. The authors found that viewing a preceding expression shifted the judgment of the end expression in the opposite direction to that of the preceding expression. For example, a relatively neutral face was perceived as more sad when presented after a happy face. More recent work has used morphs of emotional changes to present facial expressions sequentially [bib_ref] The perception of changing emotion expressions, Sacharin [/bib_ref] , again highlighting that how an end expression is decoded can be influenced by its context, in this case, the preceding expression.
Related to interdependence/independence and holistic/analytic cognitive processes, the contextual effects may vary across cultures because Easterners and Westerners tend to assign attention differently to contextual information. In general, Easterners attend more to contextual information, whereas Westerners tend to pay more attention to the target and to perceive the target independently from the background [bib_ref] Culture and systems of thought: Holistic versus analytic cognition, Nisbett [/bib_ref]. This difference in attention to context has been linked to the perception of emotional expressions in Eastern and Western observers. A study by found that when seeing a cartoon figure displaying an emotional facial expression surrounded by other expressions, Easterners', but not Westerners', perceptions were influenced by the emotion of the surrounding figures. [bib_ref] Cultural differences in gaze and emotion recognition: Americans contrast more than Chinese, Stanley [/bib_ref] suggested that Westerners saw emotions as individual feelings, whereas Eastern observers saw them as inseparable from the feelings of the group. The temporal context used in the present research (the expresser's preceding expression) is different from the concurrent context manipulations (e.g., other individuals' expressions) used in previous cross-cultural research on contextual influences on emotion perception. Attending to the relationships between the self and others might make Easterners focus more on others' expressions but may not necessarily lead to Easterners paying more attention to their own preceding expressions. However, research in cognitive psychology suggests that contextual effects can occur across domains, including spatial and temporal contexts (e.g., [bib_ref] Effects of spatial and temporal context on color categories and color constancy, Hansen [/bib_ref] [bib_ref] The spatial and temporal signatures of word production components, Indefrey [/bib_ref]. Cultural differences in contextual effects exerted by the concurrent context may thus generalize to the temporal context. Based on these considerations, we exploratively examined whether the temporal context effects on the perception of nonintended emotions would vary across cultures (Easterners vs. Westerners).
## The present research
We conducted two experimental studies to test the above hypotheses. Since perceivers from both East and West were to be tested, both Eastern and Western facial expressions of emotion were included. In each study, participants were asked to rate the intensities of three specified emotions from both static expressions (static task) and dynamic expressions (dynamic task). In Study 1, the emotions included were anger, disgust, and fear, which are all negative in valence and high in arousal [bib_ref] Core affect, prototypical emotional episodes, and other things called emotion: Dissecting the..., Russell [/bib_ref]. Among them, anger and disgust are associated with morphologically similar facial expressions, whereas anger versus fear and disgust versus fear have morphologically dissimilar expressions [bib_ref] Angry, disgusted, or afraid? Studies on the malleability of emotion perception, Aviezer [/bib_ref] [bib_ref] Human and computer recognition of facial expressions of emotion, Susskind [/bib_ref]. In the static task, participants were presented with photographs of facial expressions, while in the dynamic task, participants were presented with morphs of dynamic facial expressions that changed between two emotions. In Study 2, we used a different set of emotions to examine whether the results of Study 1 were robust. Specifically, Study 2 examined the perception of facial expressions of anger, fear, and surprise, a set that also contains morphologically similar facial expressions (fearsurprise) and morphologically dissimilar facial expressions (anger-surprise and anger-fear; [bib_ref] Angry, disgusted, or afraid? Studies on the malleability of emotion perception, Aviezer [/bib_ref] [bib_ref] Human and computer recognition of facial expressions of emotion, Susskind [/bib_ref].
It is noteworthy that the focus of the present research is on the perception of multiple emotions, particularly nonintended emotions, from facial expressions. Although there is a great deal of cross-cultural research on emotion recognition from facial expressions (e.g., [bib_ref] Universals and cultural differences in the judgments of facial expressions of emotion, Ekman [/bib_ref] [bib_ref] Cultural confusions show that facial expressions are not universal, Jack [/bib_ref] [bib_ref] American-Japanese differences in multiscalar intensity ratings of universal facial expressions of emotion, Yrizarry [/bib_ref] , that work has emphasized the accuracy of emotion recognition; that is, the correct identification of intended emotions from facial expressions. Here, emotion perception is described as the extent to which emotions are endorsed from facial expressions, rather than considered in terms of accuracy. One of our goals with this research is to underscore the importance of investigating the perception of nonintended emotions from facial expressions, particularly in cross-cultural studies.
## Development of stimuli and pilot study
Six Dutch actors (three male) posing facial expressions of anger, disgust, fear, and surprise were selected from the Radboud Faces Database [bib_ref] Presentation and validation of the Radboud faces database, Langner [/bib_ref]. Six Chinese actors expressing the same emotions were selected from the Taiwan Corpora of Chinese Emotions [bib_ref] Taiwan corpora of Chinese emotions and relevant psychophysiological data-A college-student database of..., Shyi [/bib_ref] , based on one-to-one correspondence to the Dutch stimuli in terms of gender, appearance, and approximate age (see supplementary [fig_ref] Figure 1: The procedure and materials adopted in the static and dynamic tasks of... [/fig_ref] for examples of facial expression stimuli). Both facial expression databases were produced using instructions based on the Facial Action Coding System (FACS;. Thus, the Eastern and Western facial expression stimuli used in the present research were matched in terms of the muscles ("action units") activated for each emotion (see for an overview of action units and supplementary for a test of the morphological similarities between the facial expression stimuli).
A pilot study was conducted to ensure a match of intensity between the Dutch and Chinese stimuli. A photograph of one of the Dutch facial expressions was presented on the left side of the screen, while the corresponding Chinese facial expression (morphed to change from neutral to a specific emotion in 26 frames) was presented on the right side of the screen. Participants were asked to drag the slider bar under the Chinese morph to choose the frame that was most similar in terms of intensity to the Dutch stimulus on the left side. Each comparison between a Chinese stimulus and a Dutch stimulus included two trials with different initial positions of the slider bar under the Chinese morph, with one starting from the first frame (minimum intensity, that is, fully neutral with no emotion) and the other one starting from the last frame (maximum intensity, that is, fully emotional with no neutral). In total, each participant completed 48 trials (6 actors × 4 emotions × 2 initial positions of the slider bar) in a random order. Twenty Dutch participants from the University of Amsterdam and 20 Chinese participants from Zhejiang University were recruited for the pilot test. No significant differences were found between the two groups of perceivers, so the average frame across the two groups of perceivers was used as the final stimulus for each individual and emotion in the Chinese set.
We used a digital morphing program (FantaMorph; http://www.fantamorph.com/) to produce 26-frame morphs, in which a start emotional expression changed into an end expression for the same individual's face (see [fig_ref] Figure 1: The procedure and materials adopted in the static and dynamic tasks of... [/fig_ref] for an illustration). The morphs were presented at the speed of 30 frames per second, which has been found to adequately reflect natural changes in the dynamic facial expressions [bib_ref] Deciphering the enigmatic face: The importance of facial dynamics in interpreting subtle..., Ambadar [/bib_ref] [bib_ref] The dynamic aspects of emotional facial expressions, Sato [/bib_ref]. Consequently, each frame lasted for 33.33 ms. Based on the unfolding time of natural dynamic emotional expressions [bib_ref] Perceived realism of dynamic facial expressions of emotion: Optimal durations for the..., Hoffmann [/bib_ref] , the exposure time of the first frame (i.e., the start emotion) and the last frame (i.e., the end emotion) were extended to 600 ms each. Thus, each clip lasted for 2,000 ms. In total, 72 morphs (3 actors × 2 actor ethnicities × 2 genders × 3 emotion pairs × 2 directions) were included for anger-disgust-fear (Study 1) and the same number of morphs was included for anger-fear-surprise (Study 2).
The emotion terms in the current research were taken from the respective facial expression databases (i.e., the terms are used to denote the various facial expressions in each stimulus set; Verrassing 惊讶 see [fig_ref] Table 1: Chinese and Dutch Translations of Emotion Terms [/fig_ref]. They were back-translated into English by a native speaker of each language to verify accurate translation. The instructions were translated from English into Chinese and Dutch by means of the standard translation/back-translation procedure.
## Study 1
# Method
Participants. The study included 222 individuals, 11 participants were excluded because they merely clicked through the trials without performing the task. Note. (a) In the static task, each trial started with a fixation cross displayed for 500 ms, followed by a photograph shown for 1,000 ms, and ending with a screen with three scales asking participants to indicate the emotions they perceived on the face. In the dynamic task, each trial started with a fixation cross lasting for 500 ms, followed by a dynamic (morphed) clip lasting a total of 2,000 ms, and ending with a screen with three scales asking participants to indicate the last emotional expression they perceived on the face. (b) A facial expression of anger, as used in the static task (above), and a subset of the morphs shown in a dynamic clip of disgust-to-anger, as used in the dynamic task (below).
Apparatus. The tests were controlled by a custom-written PsychoPy program (Psychophysics software in Python; [bib_ref] PsychoPy-Psychophysics software in Python, Peirce [/bib_ref] and implemented on a Windows 7 computer. The monitor was 24 inch, with a screen resolution of 1920 × 1080 pixels, and a screen refresh rate of 60 Hz.
Procedure. In the static task, each trial started with a fixation cross displayed in the center of the screen for 500 ms, followed by a photograph shown for 1,000 ms. Participants were seated approximately 60 cm from the screen, and the photographs subtended 8° × 12° of their visual angle. Immediately following the photograph, a gray screen appeared with a prompt to judge the emotion expressed in the photograph. Ratings were made by moving sliders on three scales (anger, disgust, and fear), ranging from 0 (not at all) to 100 (very much). All scales were displayed on a single screen, with the order consistent for each subject, but counterbalanced between subjects. Participants completed three practice trials, followed by four blocks of 18 trials each. Each stimulus appeared once in each block, and the order was random.
In the dynamic task, the procedure was identical to that of the static task except that the photographs were replaced with movie clips, which were displayed for 2,000 ms. Participants were asked to judge the last emotion expressed in the video. Participants completed six practice trials, followed by four blocks of 36 trials each. Each stimulus appeared once in each block, and the order was random (see [fig_ref] Figure 1: The procedure and materials adopted in the static and dynamic tasks of... [/fig_ref].
The total study consisted of a static task (15 min), a dynamic task (30 min), and an unrelated filler task in between to separate the two conditions (15 min). 4 The order of the static and dynamic tasks was counterbalanced between participants. On average, participants took 60 min to finish the experimental session.
# Results and discussion
Static task. For facial expressions of anger and disgust, in addition to the intended emotion scale, ratings were made of both morphologically similar (disgust/anger) and dissimilar (fear) emotions. In contrast, for facial expressions of fear, in addition to the intended emotion scale, ratings were only made of two nonintended emotion scales of morphologically dissimilar emotions (anger and disgust). 5 As we were interested in possible intensity differences between perceptions of morphologically similar and dissimilar emotions, our analyses focused only on the nonintended ratings of anger and disgust expressions (48 out of 72 trials per participant).
We examined the nonintended emotion ratings in a 2 (Scale Similarity: Similar, Dissimilar) × 2 (Culture of Perceiver: Chinese, Dutch) × 2 (Culture of Expresser: Chinese, Dutch) mixeddesign analysis of variance (ANOVA), with participants' intensity ratings being the dependent variable. Culture of Perceiver and Culture of Expresser were between-subjects variables, and Scale Similarity was a within-subjects variable. A complete overview of effects can be found in supplementary [fig_ref] Table 1: Chinese and Dutch Translations of Emotion Terms [/fig_ref] ; the results pertaining to our hypotheses are presented below. To rule out possible cultural differences in response styles, we followed [bib_ref] Multimethod probes of individualism and collectivism, Triandis [/bib_ref] recommendation to standardize all the ratings across variables and participants within each culture. We then conducted the same analyses on the standardized scores, and the results were identical (see for the results).
In support of Hypothesis , left panel). This pattern of results indicates that the effect of culture was more pronounced for morphologically similar emotions than for dissimilar emotions.
Dynamic task. In the dynamic task, the last expression shown on the face was the expression to be evaluated. Therefore, only morphs to anger/disgust (i.e., disgust-to-anger, fear-to-anger, anger-to-disgust, and fear-to-disgust) yield ratings for both morphologically similar (disgust/ anger) and dissimilar emotions (fear). Similar to the static task, only morphs to expressions of anger/disgust (96 out of 144 trials per participant) were therefore included in the analysis.
In addition to the factors included in the analyses of the data from the static task, another factor was included in the analysis of dynamic task data, namely, whether the temporal context (i.e., the start expression) was morphologically similar or different to the end expression. A 2 (Scale Similarity: Similar, Dissimilar) × 2 (Context Similarity: Similar, Dissimilar) × 2 (Culture of Perceiver: Chinese, Dutch) × 2 (Culture of Expresser: Chinese, Dutch) mixed-design ANOVA was conducted. Both hypotheses supported in the static task were further supported in the dynamic task (see for all effects based on raw scores and 281. Furthermore, we tested temporal context effects across cultures. A marginal three-way interaction was found between Scale Similarity, Context Similarity, and Culture of Perceiver, F(1, 207) = 2.98, p = .086, h 2 p = .014. The three-way interaction was broken down for Chinese and Dutch observers separately. For Chinese observers, the two-way interaction between Scale Similarity and Context Similarity was not significant, F(1, 76) = 0.01, p = .935, h 2 p < .001. For Dutch observers, however, the interaction between Scale Similarity and Context Similarity was significant, F(1, 133) = 11.10, p = .001, h 2 p = .077. This two-way interaction was analyzed further, revealing that ratings on the morphologically similar emotion were higher when pre- . This pattern suggests a congruency effect in Dutch observers, with an increase in judgments on the rating scales of the preceding emotions. However, the temporal context did not influence Chinese observers' ratings.
In summary, both hypotheses were supported in Study 1, with morphologically similar emotions being endorsed to a greater extent than dissimilar emotions (Hypothesis 1), and Chinese participants endorsing nonintended emotions to a greater extent than Dutch participants (Hypothesis 2). Furthermore, the difference between Chinese and Dutch was greater for morphologically similar emotions than for dissimilar emotions. In addition, the present research showed that the temporal context affected Dutch, but not Chinese, observers' judgments.
In Study 2, we set out to replicate these findings with a different set of emotions to test whether the pattern of results would be robust across distinct emotion constellations.
## Study 2
# Method
Participants. The study included 210 individuals, nine participants were excluded because they merely clicked through the trials without performing the task. The final sample consisted of 122 Dutch participants (M age = 21.5, SD = 3.01; 100 females) and 79 Chinese participants (M age = 19.2, SD = 1.78; 51 females), who were recruited in the same way as in Study 1. Among the Dutch participants, 84 (M age = 21.7, SD = 2.85; 62 females) were presented with Dutch facial expressions and 38 (M age = 21.0, SD = 3.48; 32 females) were presented with Chinese facial expressions. Among the Chinese people, 39 (M age = 19.7, SD = 1.76; 30 females) were presented with Chinese facial expressions and 40 (M age = 18.6, SD = 1.48; 20 females) were presented with Dutch facial expressions. In return for participation, Dutch participants received one course credit or €10, and Chinese participants received 30 yuan (about €4).
Apparatus and procedure. Besides the use of different emotion categories (anger-fear-surprise) and the accompanying rating scales, the apparatus and procedure were identical to Study 1.
# Results and discussion
Static task. Like in Study 1, the analyses focused on static expressions for which both morphologically similar and dissimilar rating scales were included (i.e., fear and surprise for Study 2). The data were analyzed identically to those of Study 1, with a 2 (Scale Similarity: Similar, Dissimilar) × 2 (Culture of Perceiver: Chinese, Dutch) × 2 (Culture of Expresser: Chinese, Dutch) mixed-design ANOVA. A complete overview of effects based on raw scores can be found in supplementary [fig_ref] Table 1: Chinese and Dutch Translations of Emotion Terms [/fig_ref] (results based on standardized scores can be found in supplementary ; the pattern of results was identical.). The results pertaining to the hypotheses are presented below.
In support of Hypothesis We further explored whether cultural differences would be greater for morphologically similar emotions or morphologically dissimilar emotions. The anticipated interaction between Scale Similarity and Culture of Perceiver was not significant, F(1, 197) = 1.77, p = .185, h 2 p = .009. However, a three-way interaction between Scale Similarity, Culture of Perceiver, and Culture of Expresser was found, F(1, 197) = 4.23, p = .041, h 2 p = .021. This three-way interaction was broken down for Chinese and Dutch expressers separately since we were interested in the interaction Note. A greater influence of context was found among Dutch perceivers than among Chinese perceivers. (a) In Study 1, context made Dutch perceivers endorse nonintended emotions (congruent with the preceding expressions) to a greater extent for ratings on both morphologically similar and dissimilar emotions, but context did not affect Chinese perceivers' judgments. (b) In Study 2, context made Dutch perceivers endorse nonintended emotions (congruent with the preceding expressions) to a greater extent for ratings on the morphologically dissimilar emotions but not for ratings on the morphologically similar emotions; context did not affect Chinese perceivers' judgments. The error bars represent confidence interval.
between Scale Similarity and Culture of Perceiver. When judging Chinese stimuli, ratings on the morphologically similar emotions were higher than ratings on the dissimilar emotions, F(1, 75) = 341.78, p < .001, h 2 p = .820, and Chinese observers' ratings were higher than Dutch observers' ratings, F(1, 75) = 20.55, p < .001, h 2 p = .215. Furthermore, these two effects interacted, F(1, 75) = 4.61, p = .035, h 2 p = .058, such that the effect of Culture of Perceiver was greater for ratings on the morphologically similar emotions (M Chinese = 58.56, M Dutch = 41.13; difference = 17.43, 95% CI = [8.76, 26.10]) than for ratings on the dissimilar emotions (M Chinese = 16.10, M Dutch = 7.51; difference = 8.59, 95% CI = [3.66, 13.52]; see , right panel). When judging Dutch stimuli, however, no significant interaction was found, F(1, 122) = 0.33, p = .567, h 2 p = .003. Thus the difference between Chinese and Dutch observers was greater for ratings on the morphologically similar emotions than ratings on the dissimilar emotions for Chinese, but not Dutch, stimuli.
Dynamic task. As in Study 1, only end emotions for which both morphologically similar and dissimilar rating scales were included (i.e., anger-to-fear, surprise-to-fear, anger-to-surprise, and fear-to-surprise).
The data were analyzed in the same way as those of Study 1, using a 2 (Scale Similarity: Similar, Dissimilar) × 2 (Context Similarity: Similar, Dissimilar) × 2 (Culture of Perceiver: Chinese, Dutch) × 2 (Culture of Expresser: Chinese, Dutch) mixed-design ANOVA (see supplementary for all effects based on raw scores and for all effects based on standardized scores). Both hypotheses supported in the static task were further supported in the dynamic task. Specifically 048. To test whether contextual influence varies depending on the culture of perceiver, we examined the three-way interaction between Scale Similarity, Context Similarity, and Culture of Perceiver, which was significant, F(1, 197) = 6.53, p = .011, h 2 p = .032. The three-way interaction was broken down for Chinese and Dutch observers separately. As in Study 1, for Chinese observers, the two-way interaction between Scale Similarity and Context Similarity was not significant, F(1, 78) = 2.56, p = .114, h 2 p = .032, whereas for Dutch observers, the interaction was significant, F(1, 121) = 9.13, p = .003, h 2 p = .070. Follow-up simple effects analyses showed no difference between morphologically similar and dissimilar contexts for ratings on the morphologically similar emotions, t(121) = 1.53, p = .128, Cohen's d = .14 (difference = 2.61, 95% CI = [-0.76, 5.99]), but a significant difference was found for ratings on the morphologically dissimilar emotions, t(121) = 4.18, p < .001, Cohen's d = .38, with ratings on the morphologically dissimilar emotions being higher when preceded by a morphologically dissimilar (M = 16.86) rather than similar (M = 8.38) emotion for Dutch observers . Thus Study 2, like Study 1, showed that start expressions did not influence Chinese observers' perception of end expressions, whereas start expressions elicited a congruence effect in Dutch observers. The pattern of the Dutch participants was largely consistent with the results of Study 1, with a congruency effect emerging in ratings on the morphologically dissimilar emotions, that is, ratings on the dissimilar emotions being higher in dissimilar contexts. Unlike in Study 1, however, no congruence effect was found for ratings on the morphologically similar emotions in Dutch participants in Study 2.
In sum, three main findings from Study 1-morphologically similar emotions being endorsed to a greater extent than dissimilar emotions, Chinese endorsing nonintended emotions to a greater extent than Dutch, and the start expressions influencing Dutch, but not Chinese, observers-were replicated in Study 2. In addition, the finding from Study 1 that the cultural differences were greater for similar as compared with dissimilar emotions was partly replicated in Study 2, with this pattern found for the perception of Chinese, but not Dutch, faces.
# General discussion
Although a considerable body of research has studied the recognition of intended emotions, we know little about the perception of nonintended emotions from facial expressions. The present research investigated the perception of multiple concurrent emotions from both static and dynamic expressions in Chinese and Dutch observers. Four main findings emerged consistently across the two studies (we also conducted a total analysis on the two studies together and obtained the same results; see supplementary Tables S1, S2, S3, and S4). First, morphologically similar emotions were endorsed to a greater extent than dissimilar emotions. For example, facial expressions of anger were rated higher on perceived disgust (which is morphologically similar to anger) than on perceived fear (which is morphologically dissimilar to anger). Second, Chinese observers endorsed nonintended emotions to a greater extent than Dutch observers. When observing a facial expression of anger, for instance, Chinese participants reported perceiving disgust and fear to a greater degree than did Dutch participants. Third, the difference between Chinese and Dutch observers was more pronounced for morphologically similar emotions than for dissimilar emotions (e.g., for expressions of anger, cultural differences were greater for judgments of disgust compared with judgments of fear). Fourth, the start emotions had a greater influence on Dutch than on Chinese observers. Specifically, Dutch observers endorsed nonintended emotions that were congruent with the preceding expressions to a greater degree. The consistency of these findings in Studies 1 and 2 across two different emotion constellations suggests that these results are robust and persist across distinct emotional expressions.
Unlike previous emotion perception studies using multiscalar intensity ratings of emotional expressions [bib_ref] Universals and cultural differences in the judgments of facial expressions of emotion, Ekman [/bib_ref] [bib_ref] Gender differences in judgments of multiple emotions from facial expressions, Hall [/bib_ref] [bib_ref] American-Japanese differences in multiscalar intensity ratings of universal facial expressions of emotion, Yrizarry [/bib_ref] , the current research focused on the perception of nonintended emotions. In using this paradigm, we found that Chinese endorsed nonintended emotions to a greater extent than Dutch and that this cultural difference was moderated by the morphological similarity between emotional expressions. Below we consider some theoretical implications of the present findings, discuss the strengths and limitations of our approach, and outline directions for future research.
## Theoretical implications
Previous research has found that anger and disgust, and fear and surprise, share some morphological features, which make them relatively more difficult for perceivers to distinguish from both static and dynamic emotional expressions [bib_ref] Cultural confusions show that facial expressions are not universal, Jack [/bib_ref] [bib_ref] The perception of changing emotion expressions, Sacharin [/bib_ref]. However, participants in previous research could only specify one emotion category (or a "none of the above" option) when asked to indicate which emotion they perceived from a given expression. Instead of using this categorical classification, the current paradigm allowed participants to indicate that they had seen several concurrent emotions (or no emotion) in an expression. In Study 1, for example, disgust was endorsed to a greater extent than fear for facial expressions of anger, and in Study 2, surprise was endorsed to a greater extent than anger for facial expressions of fear. This implies that observers can perceive several concurrent emotions in a given expression and that the perceived intensity of each emotion depends on its morphological similarity to the intended emotion.
The differential emotion perception we observed here appears to be driven mainly by the morphological (dis)similarities between emotional expressions, rather than by (dis)similarities between emotions in terms of valence, arousal, or approach/avoidance tendencies. Facial expressions of anger, disgust, and fear are typically perceived as high in arousal and negative in valence (see [bib_ref] A circumplex model of affect, Russell [/bib_ref]. Yet anger was endorsed to a larger extent than fear from a disgusted face, even though arousal and valence ratings of disgust and fear expressions were more similar than those of disgust and anger expressions (see for valence and intensity ratings). In terms of approach/avoidance tendencies, anger is typically thought to be approachrelated, while disgust and fear are thought to be avoidance-related [bib_ref] Relations among emotion, appraisal, and emotional action readiness, Frijda [/bib_ref]. Thus, similarities in terms of approach/avoidance cannot account for the finding that anger was endorsed to a larger extent than fear from a disgusted facial expression.
The present finding, that Chinese participants endorsed nonintended emotions to a greater extent than the Dutch participants did, is consistent with previous findings showing that Easterners make a less clear distinction than Westerners between anger and disgust and between fear and surprise [bib_ref] Cultural confusions show that facial expressions are not universal, Jack [/bib_ref] [bib_ref] Internal representations reveal cultural diversity in expectations of facial expressions of emotion, Jack [/bib_ref]. Previous accounts have suggested that this may be due to Easterners and Westerners using different perceptual strategies when viewing facial expressions: Easterners' perception emphasizes the eye region, whereas Westerners attend more to the mouth area. Facial expressions of anger and disgust (and those of fear and surprise) show few differences around the eye region. This explanation, however, cannot account for the cultural differences in the ratings of morphologically dissimilar emotions found in the current study. For example, despite expressions of anger and fear differing greatly in the eye region, Chinese still perceived a greater degree of fear from facial expressions of anger relative to Dutch observers. Thus, cultural differences in the endorsement of multiple emotions cannot be fully accounted for by attentional biases. It is worth noting that participants from both cultures reliably perceived the intended emotions conveyed by the facial expressions studied here, as evidenced by that fact that, in every condition, ratings for the intended emotion were significantly higher than ratings for all nonintended emotions. 6 This pattern of results suggests that Easterners' stronger inclination of perceiving multiple emotions from a face is not due to them failing to perceive the intended emotion.
Previous research has also shown that Easterners experience more complex blends of emotions than Westerners, which is in line with our finding that Chinese observers perceived multiple emotions to a greater degree than did Dutch observers. Studies in which participants were asked to report their feelings in different situations have revealed positive correlations between positive and negative emotions in Asian cultures but negative correlations in Western cultures; that is, mixed emotions co-occurring in Easterners but not in Westerners [bib_ref] Culture and mixed emotions: Co-occurrence of positive and negative emotions in Japan..., Miyamoto [/bib_ref]. The present finding extends this research from an intrapersonal level (perceiving self-emotional experience) to an interpersonal level (perceiving other's emotions). Notably, the present research is the first to establish that Easterners are able to perceive mixed emotions consisting of multiple distinct negative emotions, besides blends of positive and negative emotions.
The finding that cultural differences were more pronounced for morphologically similar emotions than for dissimilar emotions suggests that the cultural differences in emotional complexity discussed above were primarily driven by the perception of morphologically similar emotions. For example, both Chinese and Dutch rated expressions of anger low on the fear scale, with fear being morphologically dissimilar to anger. In contrast, judgments of disgust, which share a stronger morphological similarity with anger, entail larger overlapping elements to the target. Accordingly, Chinese observers, who tend to perceive close elements as more overlapping (interdependent/holistic), rated anger expressions higher on the disgust scale than did Dutch observers, who are more independent and analytical. Here, the effects of culture and morphological similarity interacted to determine the extent to which participants perceived several emotions in a facial expression, with greater cultural differences in the endorsement of morphologically similar rather than dissimilar emotions.
In contrast to documented East-West differences in concurrent context effects (e.g., [bib_ref] Placing the face in context: Cultural differences in the perception of facial..., Masuda [/bib_ref] [bib_ref] Cultural differences in gaze and emotion recognition: Americans contrast more than Chinese, Stanley [/bib_ref] , we consistently found that the temporal context influenced Dutch observers' but not Chinese observers' judgments. Specifically, we found a congruency effect with preceding expressions in Dutch observers. This temporal context effect on Dutch observers' perception of nonintended emotion is, to some extent, consistent with previous findings that concurrent contexts (gender and job status) influence the perception of nonintended emotions from facial expressions in Western perceivers [bib_ref] Gender and job status as contextual cues for the interpretation of facial..., Algoe [/bib_ref]. In contrast, Chinese observers' judgments of nonintended emotions did not differ across preceding expressions. Two explanations may account for this phenomenon. First, relative to Dutch observers, Chinese observers already endorsed a greater degree of nonintended emotions from static expressions, and context may therefore not have caused a substantial increase in ratings of nonintended emotions. However, this explanation is undermined by the fact that Chinese observers' ratings on the nonintended scales of static expressions were far from ceiling in the current studies (see [fig_ref] Figure 3: Nonintended emotion ratings as a function of similarity of rating scale, similarity... [/fig_ref].
Instead, we propose a second possibility, namely, that cultural differences vary across different types of contexts. The cultural differences found in previous research on concurrent context effects (i.e., that Easterners are more strongly influenced by context than Westerners) do not generalize to cultural differences in temporal context effects. A recent study has suggested that the agency of the contextual information may have differing degrees of influence on how faces are evaluated across cultures [bib_ref] Agency and facial emotion judgment in context, Ito [/bib_ref]. Ito and colleagues found that when other figures' emotional expressions served as context, Westerners' judgments of the target person were not influenced by surrounding people's expressions. In contrast, when there was no agency information in the context (e.g., affectively salient landscape scenery), Westerners were influenced by the context just like Easterners. They argued that this difference might be due to Western perceivers considering individuals as agents of their own emotions, while Easterners construe emotions more as products of an individual's social context. In the present study, the context was the expresser's own preceding expression, reflecting the history of the expresser's own feelings rather than those of others. This may have resulted in the start expression having a more marked influence on Dutch observers' judgments.
Notably, Chinese participants endorsed nonintended emotions to a greater extent than did Dutch participants, both in the static and dynamic tasks. This suggests that Easterners' higher endorsement of nonintended emotions was not driven by temporal context, given that there was no preceding expression in the static task. This pattern of results implies that we should be cautious in concluding that Easterners are overall more strongly influenced by context than Westerners. Cultural differences may vary across different types of contextual information (e.g., concurrent vs. temporal context) and perhaps also as a function of the relationship between context and expression.
## Limitations and future directions
One limitation of the current paradigm is that the change from one expression to another does not necessarily reflect the natural emotional change in emotional facial expressions that we come across in real life. For example, with natural emotional change from anger to disgust, the mouth area may change ahead of the eyes, which would not be captured in computer-generated morphs in which changes in the upper and lower face happen simultaneously. Despite this criticism, morphs have clear merits, such as the easily adjustable rate of change and the lack of noise, and they are commonly used in dynamic emotion research (e.g., [bib_ref] The perception of changing emotion expressions, Sacharin [/bib_ref] [bib_ref] Detection of emotional facial expressions and anti-expressions, Sato [/bib_ref]. Nevertheless, this methodology could be improved in future research using videos of changing emotional expressions as they happen in real time. Although databases with naturalistic dynamic emotional facial expressions exist for changes to and from neutral expressions (e.g., [bib_ref] Moving faces, looking places: Validation of the Amsterdam Dynamic Facial Expression Set..., Van Der Schalk [/bib_ref] , comparable stimulus materials for expression changes between emotions have not yet been developed. Even greater ecological validity would be gained by examining emotional expressions occurring in ongoing interactions. But it is worth noting that establishing the intended emotion in a genuine expression and ensuring the absence of additional emotions in the expresser are a challenge. These research approaches complement each other, and future studies should establish whether the current results hold for genuine dynamic expressions.
Another limitation of the current research is that only emotions of negative or neutral valence were included in this study. Positive emotions, such as happiness, may be particularly interesting to examine cross-culturally. In Eastern cultures, smiles do not necessarily function as a positive signal, as smiling is deemed an appropriate reaction in social interaction even when feeling angry with the interaction partner. Westerners are comparatively more encouraged to express what they are feeling, and smiling is consequently more often regarded as a positive social signal [bib_ref] Cultural influences on facial expressions of emotion, Matsumoto [/bib_ref]. Future work could evaluate whether the current pattern of results (e.g., Easterners perceiving emotions as more mixed than Westerners) extends to expressions of positive emotions.
Other potential limitations include the number and type of emotion scales used as response categories; we do not know whether the conclusion that Easterners perceive nonintended emotions to a larger extent than Westerners also holds for other emotions besides the emotion scales used in the present studies. It should also be noted that the facial expressions were Western-based prototypes and we studied only two cultures. As a result, the current findings do not necessarily generalize to other facial expressions and other populations. Regardless of these limitations, however, the present results challenge previous work in this area and raise important questions about the influence of culture and morphological similarity on the judgments of emotional expressions.
Finally, although there is a wealth of research showing that Easterners are more interdependent, holistic, and dialectical than Westerners (e.g., [bib_ref] Culture and systems of thought: Holistic versus analytic cognition, Nisbett [/bib_ref] [bib_ref] Naïve dialecticism and the Tao of Chinese thought, Peng [/bib_ref] , we did not include these measures in the current research. Future research may benefit from measuring interdependence/independence, holistic/analytic cognition, and dialectical thinking to establish the mechanism(s) underlying the cultural differences in emotion perception. It would also be interesting to investigate whether Easterners also produce more mixed expressions of emotion than Westerners. This would complement previous findings showing that Easterners are more likely to experience mixed feelings [bib_ref] Culture and mixed emotions: Co-occurrence of positive and negative emotions in Japan..., Miyamoto [/bib_ref] as well as the current finding that this mixedness extends to perceptual judgments.
# Conclusion
Two studies provided consistent evidence that morphologically similar emotions were endorsed to a greater extent than dissimilar emotions, that Chinese observers endorsed nonintended emotions to a greater extent than Dutch observers, and that this cultural difference was more pronounced for morphologically similar emotions than for dissimilar emotions in both static and dynamic expressions. We also found that temporal contextual cues led Dutch observers to endorse nonintended emotions that were congruent with the preceding expressions to a greater extent but that context did not affect Chinese observers' judgments.
Research on emotion perception tends to focus on the perception of intended emotions, often ignoring the degree to which other emotions are perceived. The present research underscores the importance of investigating the perception of nonintended emotions from facial expressions, particularly in cross-cultural studies. The increasingly popular idea that emotional expressions serve to facilitate social interaction by providing information to observers about an expresser's inner feelings and future intentions hinges on the assumption that emotional expression and emotion perception align (for a comprehensive discussion, see. The current data indicate that such alignment varies as a function of characteristics of the expression (morphological similarity to other emotions) as well characteristics of the perceiver (cultural background).
[fig] Figure 1: The procedure and materials adopted in the static and dynamic tasks of Studies 1 and 2. [/fig]
[fig] 213, Figure 2: all effects based on standardized scores, the results of which were identical). Specifically, morphologically similar emotions (M = 39.87) were endorsed to a greater extent than morphologically dissimilar emotions (M = 16.90; Hypothesis 1), F(1, 207) = 600.33, p < .001, h 2 p = .744 (difference = 22.97, 95% CI = [20.53, 25.41]), Chinese (M = 39.10) endorsed nonintended emotions to a greater extent than Dutch (M = 22.23; Hypothesis 2), F(1, 207) = 56.11, p < .001, h 2 p = .Nonintended emotion ratings as a function of similarity of rating scale and culture of perceiver for static facial expressions of emotion (Studies 1 and 2). [/fig]
[fig] Figure 3: Nonintended emotion ratings as a function of similarity of rating scale, similarity of context, and culture of perceiver for dynamic facial expressions of emotion (Studies 1 and 2). [/fig]
[table] Table 1: Chinese and Dutch Translations of Emotion Terms. [/table]
|
Morgellons disease: a filamentous borrelial dermatitis
Morgellons disease (MD) is a dermopathy characterized by multicolored filaments that lie under, are embedded in, or project from skin. Although MD was initially considered to be a delusional disorder, recent studies have demonstrated that the dermopathy is associated with tickborne infection, that the filaments are composed of keratin and collagen, and that they result from proliferation of keratinocytes and fibroblasts in epithelial tissue. Culture, histopathological and molecular evidence of spirochetal infection associated with MD has been presented in several published studies using a variety of techniques. Spirochetes genetically identified as Borrelia burgdorferi sensu stricto predominate as the infective agent in most of the Morgellons skin specimens studied so far. Other species of Borrelia including Borrelia garinii, Borrelia miyamotoi, and Borrelia hermsii have also been detected in skin specimens taken from MD patients. The optimal treatment for MD remains to be determined.
# Introduction
Morgellons disease (MD) is an emerging dermopathy with worldwide distribution. The name "Morgellons" is derived from a disease recognized in the seventeenth century in French children by Sir Thomas Browne. These children were noted to have "coarse hairs" protruding from their backs. [bib_ref] Sir Thomas Browne and the disease called Morgellons, Kellett [/bib_ref] The distinguishing feature of MD is the appearance of skin lesions with filaments that lie under, are embedded in, or project from skin [fig_ref] Figure 1: A [/fig_ref]. Filaments can be white, black, or brightly colored. [bib_ref] Skin lesions and crawling sensation: disease or delusion?, Savely [/bib_ref] [bib_ref] The mystery of Morgellons disease: infection or delusion?, Savely [/bib_ref] [bib_ref] Morgellons disease: the mystery unfolds, Savely [/bib_ref] [bib_ref] Morgellons disease: analysis of a population with clinically confirmed microscopic subcutaneous fibers..., Savely [/bib_ref] [bib_ref] Filament formation associated with spirochetal infection: a comparative approach to Morgellons disease, Middelveen [/bib_ref] Furthermore, MD patients exhibit a variety of manifestations that resemble symptoms of Lyme disease (LD), such as fatigue, joint pain, and neuropathy. [bib_ref] Skin lesions and crawling sensation: disease or delusion?, Savely [/bib_ref] [bib_ref] The mystery of Morgellons disease: infection or delusion?, Savely [/bib_ref] [bib_ref] Morgellons disease: the mystery unfolds, Savely [/bib_ref] [bib_ref] Morgellons disease: analysis of a population with clinically confirmed microscopic subcutaneous fibers..., Savely [/bib_ref] [bib_ref] Filament formation associated with spirochetal infection: a comparative approach to Morgellons disease, Middelveen [/bib_ref] A study found that 98% of MD subjects had positive LD serology and/or a tickborne disease diagnosis, [bib_ref] Morgellons disease: analysis of a population with clinically confirmed microscopic subcutaneous fibers..., Savely [/bib_ref] confirming the clinical association between MD and spirochetal infection. Conversely, 6% of LD patients in an Australian study were found to have MD. [bib_ref] Clinical determinants of Lyme borreliosis, babesiosis, bartonellosis, anaplasmosis, and ehrlichiosis in an..., Mayne [/bib_ref] The similarity between MD and an animal disease, bovine digital dermatitis (BDD), an acknowledged spirochetal infection that is associated with ulcerative lesions exhibiting keratin projections, was previously explored. [bib_ref] Filament formation associated with spirochetal infection: a comparative approach to Morgellons disease, Middelveen [/bib_ref] Treponemal spirochetes are the primary etiologic agents of BDD. [bib_ref] Long-term observations on the dynamics of bovine digital dermatitis lesions on a..., Berry [/bib_ref] [bib_ref] An experimental infection model to induce digital dermatitis infection in cattle, Gomez [/bib_ref] A causal relationship between spirochetal infection and filament formation was confirmed by duplication of the clinical disease via experimental infection with pure cultured treponemes. [bib_ref] Long-term observations on the dynamics of bovine digital dermatitis lesions on a..., Berry [/bib_ref] [bib_ref] An experimental infection model to induce digital dermatitis infection in cattle, Gomez [/bib_ref] This prompted further investigation into the possibility of a spirochetal etiology for MD to discover if a similar disease process occurred at the cellular level. [bib_ref] Morgellons disease: a chemical and light microscopic study, Middelveen [/bib_ref] [bib_ref] Characterization and evolution of dermal filaments from patients with Morgellons disease, Middelveen [/bib_ref] [bib_ref] Exploring the association between Morgellons disease and Lyme disease: identification of Borrelia..., Middelveen [/bib_ref] Histopathology of MD Histological studies on MD tissue show that MD filaments are not textile fibers. They are biofilaments of human cellular origin produced by epithelial cells and stem from deeper layers of the epidermis, the upper layers of the dermis, and the root sheath of hair follicles. [bib_ref] Filament formation associated with spirochetal infection: a comparative approach to Morgellons disease, Middelveen [/bib_ref] [bib_ref] Morgellons disease: a chemical and light microscopic study, Middelveen [/bib_ref] [bib_ref] Characterization and evolution of dermal filaments from patients with Morgellons disease, Middelveen [/bib_ref] Histological studies established that these filaments are predominantly composed of collagen and keratin, [bib_ref] Morgellons disease: a chemical and light microscopic study, Middelveen [/bib_ref] [bib_ref] Characterization and evolution of dermal filaments from patients with Morgellons disease, Middelveen [/bib_ref] and are nucleated at the base of attachment to epithelial cells, [bib_ref] Characterization and evolution of dermal filaments from patients with Morgellons disease, Middelveen [/bib_ref] thus demonstrating human cellular origin. Staining of embedded filaments with Congo red resulted in apple-green birefringence suggestive of an amyloid component, although this remains to be confirmed by more specific studies (unpublished data). Staining of embedded filaments with calcofluor-white produced negative results, demonstrating that filaments are not cellulose as found in cotton, linen, or other plant-based textile fibers, or chitin as found in fungal cells and insect exoskeletons (unpublished data).
A preliminary study using scanning electron microscopy (SEM) showed hairlike scales on a blue filament, suggesting that at least some MD fibers are hairs. [bib_ref] Morgellons disease: a chemical and light microscopic study, Middelveen [/bib_ref] The blue coloration of some fibers was first determined to be the result of melanin pigmentation as shown by positive Fontana Masson staining. [bib_ref] Characterization and evolution of dermal filaments from patients with Morgellons disease, Middelveen [/bib_ref] An independent study concurred that embedded blue fibers in an MD specimen (supplied by the authors of this paper) were not textile fibers. SEM revealed that the blue fibers were microscopic hairs with cuticular scaling, and transmission electron microscopy (TEM) revealed darkly stained melanosomes that were not organized, a finding consistent with human hairs (Shawkey MD, unpublished data, 2013).
Microspectrophotometry reflectance data on fibers were consistent with patterns of pigmented tissues. Raman spectroscopy 14 on two separate blue fibers showed relevant peaks that were indicative of carbamate compounds and melanin aromatic rings (Shawkey MD, unpublished observation, 2016). Hence, independent studies using different methodologies provided evidence that Morgellons fibers are hairlike extrusions and that the blue coloration is the result of melanin pigmentation. Although the mechanism for coloration of red fibers is not yet understood, there are no known textile fibers colored by blue melanin pigmentation. 11
## Association of md with borrelia infection
Borrelia spirochetes have repeatedly been detected in MD skin and tissue samples [fig_ref] Figure 3: Spirochete detected with Dieterle silver stain in culture of skin sample from... [/fig_ref]. Initial studies confirmed the presence of Borrelia burgdorferi sensu stricto (Bb ss) spirochetes within dermatological tissue removed from MD lesions of four North American patients. [bib_ref] Characterization and evolution of dermal filaments from patients with Morgellons disease, Middelveen [/bib_ref] A subsequent study reported the detection and identification of Borrelia garinii in Morgellons skin samples obtained from an Australian patient. [bib_ref] Morgellons: a novel dermatological perspective as the multisystem infective disease borreliosis, Mayne [/bib_ref] A larger study subsequently reported the detection of Borrelia spirochetes in 25 MD subjects. [bib_ref] Exploring the association between Morgellons disease and Lyme disease: identification of Borrelia..., Middelveen [/bib_ref] Detection of Borrelia DNA by polymerase chain reaction (PCR) followed by Sanger sequencing in two independent laboratories determined that the Borrelia spirochetes detected in these studies were predominantly Bb ss, but B. garinii and Borrelia miyamotoi were also reported. More recently, studies of MD specimens in two additional laboratories have detected Borrelia DNA of three Borrelia spp., Bb ss, B. garinii, and Borrelia hermsii.The fact that four different laboratories have been able to detect Borrelia DNA in Morgellons specimens shows that these findings are reproducible. Motile spirochetes identified as Borrelia spp. have been cultured in Barbour-Stoenner-Kelly (BSK)-H medium inoculated with MD dermatological tissue. This demonstrated that the spirochetes present in MD tissue are alive and viable. [bib_ref] Exploring the association between Morgellons disease and Lyme disease: identification of Borrelia..., Middelveen [/bib_ref] The culture of Borrelia spirochetes can be challenging because of fastidious growth requirements and pleomorphism. [bib_ref] Formation and cultivation of Borrelia burgdorferi spheroplast L-form variants, Mursic [/bib_ref] [bib_ref] An in vitro study of the susceptibility of mobile and cystic forms..., Brorson [/bib_ref] Therefore, PCR amplification of cultured spirochetes has been used to confirm the presence and provide molecular identification of live Borrelia spirochetes in MD dermatological tissue. [bib_ref] Exploring the association between Morgellons disease and Lyme disease: identification of Borrelia..., Middelveen [/bib_ref] The combination of Borrelia culture and PCR lends strong support to the clinical association of spirochetal infection with MD.
Although the common denominator in the evolution of MD lesions seems to be infection with Borrelia spp., the etiology of MD is presumed to be multifactorial. Secondary etiologic factors such as genetic background, hormonal influences, immune status, and the presence of other coinvolved infections appear to play a role in the development of this phenomenon. [bib_ref] Skin lesions and crawling sensation: disease or delusion?, Savely [/bib_ref] [bib_ref] The mystery of Morgellons disease: infection or delusion?, Savely [/bib_ref] [bib_ref] Morgellons disease: the mystery unfolds, Savely [/bib_ref] [bib_ref] Morgellons disease: analysis of a population with clinically confirmed microscopic subcutaneous fibers..., Savely [/bib_ref] [bib_ref] Filament formation associated with spirochetal infection: a comparative approach to Morgellons disease, Middelveen [/bib_ref] Other pathogens have been detected in Morgellons tissue samples. Strains of Helicobacter pylori and closely related bacteria were detected along with Borrelia spp. in tissue samples from MD lesions.One study detected Treponema denticola along with Borrelia spp. in some Morgellons specimens, 12 while another study found H. pylori, T. denticola, and Bartonella henselae in Morgellons specimens.A putative role of Agrobacterium in MD has not been confirmed.
The precise mechanism of Morgellons filament formation has not been elucidated. In MD lesions, collagen and keratin filaments arise from proliferative keratinocytes and fibroblasts in human epithelial tissue. [bib_ref] Morgellons disease: a chemical and light microscopic study, Middelveen [/bib_ref] [bib_ref] Characterization and evolution of dermal filaments from patients with Morgellons disease, Middelveen [/bib_ref] Borrelia has the capability to invade fibroblasts and keratinocytes and to replicate inside these cells. [bib_ref] Fibroblasts protect the Lyme disease spirochete, Borrelia burgdorferi, from ceftriaxone in vitro, Georgilis [/bib_ref] [bib_ref] Invasion of human skin fibroblasts by the Lyme disease spirochete Borrelia burgdorferi, Klempner [/bib_ref] [bib_ref] Interactions between Borrelia burgdorferi and mouse fibroblasts, Chmielewski [/bib_ref] We speculate that infection and replication of Borrelia within keratinocytes and fibroblasts alter keratin and collagen gene regulation. Furthermore, intracellular sequestration of Borrelia may be a factor contributing to the development of refractory infection and MD. Borrelia spirochetes have been isolated in vitro from monolayers of keratinocytes and fibroblasts that were treated with antibiotics. [bib_ref] Fibroblasts protect the Lyme disease spirochete, Borrelia burgdorferi, from ceftriaxone in vitro, Georgilis [/bib_ref] [bib_ref] Invasion of human skin fibroblasts by the Lyme disease spirochete Borrelia burgdorferi, Klempner [/bib_ref] Spirochetes were detected in MD dermatological tissue taken from patients who had been given aggressive antibiotic therapy. [bib_ref] Exploring the association between Morgellons disease and Lyme disease: identification of Borrelia..., Middelveen [/bib_ref] Persistent infection and resistance to antibiotic treatment may therefore result from sequestration of Borrelia spirochetes within keratinocytes and fibroblasts in MD patients.
The diagnosis of LD is controversial, and research showing a connection between MD and LD has been challenged. A study from the Centers for Disease Control and Prevention (CDC) concluded that MD was not caused by a pathogen.However, in that study the search for spirochetal pathogens was limited to nonspecific silver nitrate staining on a small number of tissue samples and commercial two-tiered serological LD testing as recommended by the CDC.Although Borrelia spirochetes can be readily detected in MD tissue, the use of sensitive and specific molecular methods is required. [bib_ref] Exploring the association between Morgellons disease and Lyme disease: identification of Borrelia..., Middelveen [/bib_ref] Two-tiered serological testing for LD lacks sensitivity for a number of reasons, and seronegativity in positive cases
## 352
Middelveen and Stricker occurs. [bib_ref] Seronegative Lyme disease, Dattwyler [/bib_ref] [bib_ref] Serologic tests for Lyme disease: more smoke and mirrors, Stricker [/bib_ref] Detection of LD is complicated by the genetic diversity of Borrelia spp. [bib_ref] Genetic diversity of Borrelia burgdorferi and detection of B. bissettii-like DNA in..., Girard [/bib_ref] [bib_ref] Investigation of genotypes of Borrelia burgdorferi in Ixodes scapularis ticks collected during..., Ogden [/bib_ref] [bib_ref] Occurrence of different genospecies of Borrelia burgdorferi sensu lato in Ixodid ticks..., Péter [/bib_ref] Other methods such as PCR detection of Borrelia DNA and antigen detection are not standardized and vary in sensitivity and specificity. [bib_ref] Evidence of a Lyme borreliosis infection from the viewpoint of laboratory medicine, Lange [/bib_ref] [bib_ref] PCR in laboratory diagnosis of human Borrelia burgdorferi infections, Schmidt [/bib_ref] Although these tests can be effective means of detecting Borrelia spirochetes in human fluids and tissues, they are not accepted as diagnostic by the CDC. [bib_ref] Exploring the association between Morgellons disease and Lyme disease: identification of Borrelia..., Middelveen [/bib_ref] [bib_ref] PCR in laboratory diagnosis of human Borrelia burgdorferi infections, Schmidt [/bib_ref] Thus, the relationship between MD and LD shown in clinical, histopathological and molecular studies remains controversial.
## Historical view of md
While the first reports of LD in the USA date back to the 1970s, [bib_ref] Lyme arthritis: an epidemic of oligoarticular arthritis in children and adults in..., Steere [/bib_ref] filamentous dermatological manifestations were not reported to be associated with Borrelia infection until 2005. [bib_ref] Skin lesions and crawling sensation: disease or delusion?, Savely [/bib_ref] Although this temporal disparity may seem anomalous to some, there are explanations. Sir Thomas Browne's description of dermal hairlike extrusions coupled with movement sensations dates back to 1674. [bib_ref] Sir Thomas Browne and the disease called Morgellons, Kellett [/bib_ref] In 1938, Ekbom published a seminal account of delusional parasitosis with sensations of insects crawling on or under skin, and it is notable that many of Ekbom's study subjects had syphilis, a known spirochetal infection. [bib_ref] Praeseniler Dermat-zooenwahn, Ekbom [/bib_ref] LD is not a new disease: a 5,300-year-old mummy is the oldest known case. [bib_ref] New insights into the Tyrolean Iceman's origin and phenotype as inferred by..., Keller [/bib_ref] Thus, Morgellons cases may have occurred sporadically in parallel with LD, and the association between LD and MD may have gone unrecognized and unreported.
Misdiagnosis of MD is likely to be common as the filaments are microscopic and invisible without sufficient magnification or, if observed under magnification, may be miscategorized as textile fibers. One could also argue that because serological testing for LD lacks sensitivity, [bib_ref] Seronegative Lyme disease, Dattwyler [/bib_ref] [bib_ref] Serologic tests for Lyme disease: more smoke and mirrors, Stricker [/bib_ref] patients frequently go undiagnosed. The few patients who do get a positive two-tiered test for LD are likely to be treated early for Bb infection and therefore may fail to develop MD. We see many MD patients who are diagnosed with other dermatological conditions (prurigo nodularis, eczema, etc) or mental illnesses, and thus were not tested for LD. Although MD is usually associated with chronic symptoms of LD, some patients develop MD lesions in conjunction with a documented tick bite and/or erythema migrans rash.
## Md and psychiatric diagnoses
There are over 250 peer-reviewed articles linking LD and associated tickborne diseases to mental illness. [bib_ref] Late-stage neuropsychiatric Lyme borreliosis. Differential diagnosis and treatment, Fallon [/bib_ref] [bib_ref] The psychoimmunology of Lyme/tick-borne diseases and its association with neuropsychatric symptoms, Bransfield [/bib_ref] [bib_ref] Why neurodegenerative diseases are progressive: uncontrolled inflammation drives disease progression, Gao [/bib_ref] A spectrum of psychiatric illnesses occurs in patients with neuroborreliosis, including paranoia, schizophrenia, bipolar disorder, delusions, sensory hallucinations, major depression, and mania. [bib_ref] Association between Lyme disease and schizoaffective disorder, bipolar type: is it inflammation..., Mattinley [/bib_ref] Neurological symptoms associated with LD range in severity from marginal to critical, and can include cognitive impairment, dementia, insomnia, irritability, anxiety, depression, personality disorders, psychosis, and suicidal or homicidal tendencies. [bib_ref] The psychoimmunology of Lyme/tick-borne diseases and its association with neuropsychatric symptoms, Bransfield [/bib_ref] The effects of microbial infection can impact neuronal functioning, and as a result of infection a prolonged inflammatory process can drive the chronic progression of neurodegenerative diseases. [bib_ref] The psychoimmunology of Lyme/tick-borne diseases and its association with neuropsychatric symptoms, Bransfield [/bib_ref] [bib_ref] Microbes' roadmap to neurons, Kristensson [/bib_ref] Levels of certain proinflammatory cytokines have been reported to increase in the cerebrospinal fluid of patients with neurological manifestations of LD. [bib_ref] Inflammation and central nervous system Lyme disease, Fallon [/bib_ref] Similarly, elevated inflammatory markers that correlate with excess of immune cytokines have been described in MD patients. [bib_ref] Morgellons disease, illuminating an undefined illness: a case series, Harvey [/bib_ref] The outer membrane lipoproteins of Borrelia are proinflammatory and can mimic host antigens, theoretically resulting in autoimmune reactivity. [bib_ref] The psychoimmunology of Lyme/tick-borne diseases and its association with neuropsychatric symptoms, Bransfield [/bib_ref] The fact that elevated anti-nerve antibody reactivity can be found in pretreatment LD patients and in posttreatment patients with persistent symptoms supports this theory. [bib_ref] The psychoimmunology of Lyme/tick-borne diseases and its association with neuropsychatric symptoms, Bransfield [/bib_ref] Although MD may result from an infectious process, there may be a psychiatric component as well, and some (but not all) MD patients exhibit neuropsychiatric symptoms. In a study of 25 Morgellons patients, 23 had prior psychiatric diagnoses including bipolar disorder, attention deficit disorder, obsessive-compulsive disorder (OCD), and schizophrenia. [bib_ref] Morgellons disease, illuminating an undefined illness: a case series, Harvey [/bib_ref] The fact that MD patients may show neuropsychiatric symptoms complicates the diagnosis and explains why some health care providers consider MD to be a delusional disorder. To further complicate matters, some patients with MD who do not exhibit psychiatric abnormalities have been misdiagnosed with other conditions such as lichen sclerosus or prurigo nodularis.
Furthermore, lack of scientific knowledge has led patients to interpret the physical presence of dermal filaments and symptoms of formication as parasitic infection. In these cases, the false belief of parasitic infestation is not genuinely delusional because patients are in fact misinterpreting symptoms caused by aberrant production of human biofibers. Furthermore, we find that electrostatic energy and mechanical energy can cause movement of filaments, interpreted by some patients as the movement of a living organism, and small insects such as fruit flies can adhere to open lesions, leading some patients to believe they are infested. If such patients do have a psychiatric condition such as OCD, then the false belief can be intensified or reinforced.
Hypothetically, some patients with delusional disorders could mistakenly believe they have MD. In addition, in our experience some LD patients exhibit crawling or stinging sensations, or can have ulcerative lesions -without developing associated dermal filaments. The overlap of mental illness, LD, and MD highlights the explicit need for a
## 353
Morgellons Disease universally accepted clinical definition of MD. The following case definition for MD has been proposed: a somatic Lyme-like illness associated with spontaneously appearing, slowly healing, filamentous and ulcerative skin lesions. The key diagnostic criterion is colored, white, or black filaments protruding from or embedded in skin. However, because there is an overlap of LD, MD, and mental illness with a spectrum of different symptoms, we recommend that LD be considered in the differential diagnosis of patients with Lyme-like symptoms in conjunction with formication with or without ulcerations.
## Clinical classification of md
In a recent publication, [bib_ref] Exploring the association between Morgellons disease and Lyme disease: identification of Borrelia..., Middelveen [/bib_ref] we proposed a clinical classification scheme that reflects the duration and location of MD lesions:
1. Early localized: lesions/fibers present for less than three months and localized to one area of the body (head, trunk, extremities). 2. Early disseminated: lesions/fibers present for less than three months and involving more than one area of the body (head, trunk, extremities). 3. Late localized: lesions/fibers present for more than six months and localized to one area of the body (head, trunk, extremities). 4. Late disseminated: lesions/fibers present for more than six months and involving more than one area of the body (head, trunk, extremities).
As noted in that article, the classification scheme provides a medical framework that should help to validate and standardize the diagnosis of MD. Further studies are needed to determine whether this classification will have therapeutic and prognostic significance for MD patients. 13
## Treatment of md
Since a clinical classification of MD has not been universally accepted, optimal treatment for the disease remains unsettled. Nevertheless, several therapeutic principles have emerged: 1) the earlier the treatment is initiated in the course of MD, the better the outcome appears to be; 2) treatment should be aimed at the underlying tickborne disease; 3) prolonged combination antibiotic therapy may be necessary to eradicate dermopathy; and 4) antiparasitic therapy may be useful in some patients with MD. At this point, the most logical treatment is supported by the guidelines of the International Lyme and Associated Diseases Society. [bib_ref] Evidence assessments and guideline recommendations in Lyme disease: the clinical management of..., Cameron [/bib_ref] Although treatment with antipsychotic agents has been proposed for patients with neuropsychiatric symptoms of MD, this treatment generally fails without concomitant therapy of the underlying tickborne disease. [bib_ref] Morgellons: a novel dermatological perspective as the multisystem infective disease borreliosis, Mayne [/bib_ref] Additional approaches with agents such as dapsone merit further study. 44
# Conclusion
In summary, MD is an emerging dermopathy that is associated with Borrelia infection, and the growing number of MD cases reflects the increase in tickborne diseases around the world. Although some medical practitioners erroneously consider MD to be caused by a delusional disorder, studies have shown that MD is a somatic illness that appears to be triggered by Borrelia infection. The optimal treatment for MD remains to be determined.
[fig] Figure 1: A) Skin lesions on the hand of a Morgellons disease patient. (B) Skin lesions on the buttocks of a Morgellons disease patient. Note: Figures courtesy of the Charles e. Holman Morgellons Disease Foundation, Austin, TX, USA, used with permission. [/fig]
[fig] Figure 2: Multicolored fibers in Morgellons disease skin lesion. Notes: 400× original magnification. Figure courtesy of the Charles E. Holman Morgellons Disease Foundation, Austin, TX, USA, used with permission. International Journal of General Medicine 2016:9 submit your manuscript | www.dovepress. [/fig]
[fig] Figure 3: Spirochete detected with Dieterle silver stain in culture of skin sample from Morgellons disease patient. Notes: 1000× original magnification. Figure courtesy of Marianne J Middelveen. [/fig]
[fig] Figure 4: Borrelia spirochetes immunostained with anti-Borrelia antibody in culture of skin sample from Morgellons disease patient. Notes: 1000× original magnification. Figure courtesy of Marianne J Middelveen. International Journal of General Medicine 2016:9 submit your manuscript | www.dovepress.com Dovepress Dovepress [/fig]
|
Surgical Versus Non-Surgical Treatment for Vertebral Compression Fracture with Osteopenia: A Systematic Review and Meta-Analysis
BackgroundSurgical and non-surgical interventions are the two categories for treatment of vertebral compression fractures (VCFs). However, there is clinical uncertainty over optimal management. This study aimed to examine the safety and effectiveness of surgical management for treatment of VCFs with osteopenia compared with non-surgical treatment.MethodsWe conducted a systematic search through electronic databases from inception to June 2014, with no limits on study data or language. Randomized controlled trials (RCTs) evaluating surgical versus non-surgical interventions for treatment of patients with VCFs due to osteopenia were considered. Primary outcomes were pain and adverse effects. A randomeffects model was used to calculate the pooled mean difference (MD) or risk ratios with 95% confidence interval (CI).ResultsSixteen reports (11 studies) met the inclusion criteria, and provided data for the meta-analysis with a total of 1,401 participants. Compared with conservative treatment, surgical treatment was more effective in reducing pain (short-term: MD -2.05, 95% CI -3.55 to -0.56, P=0.007; mid-term: MD -1.70, 95% CI -2.78 to -0.62, P=0.002; long-term: MD -1.24, 95% CI -2.20 to -0.29, P=0.01) and disability on the Roland-Morris Disability score (short-term: MD -4.97, 95% CI -8.71 to -1.23, P=0.009), as well as improving quality of life on the Short-Form 36 Physical Component Summary score (short-term: MD 5.53, 95% CI 1.45 to 9.61, P=0.008) and the Quality of Life Questionnaire of the European Foundation for Osteoporosis score (short-term: MD -5.01, 95% CI -8.11 to -1.91, P=0.002). Indirect comparisons PLOS ONE | between vertebroplasty and kyphoplasty found no evidence that the treatment effect differed across the two interventions for any outcomes assessed. Compared with the sham procedure, surgical treatment showed no evidence of improvement in pain relief and physical function. Based on these two comparisons, no significant difference between groups was noted in the pooled results for adverse events.ConclusionCompared to conservative treatment, surgical treatment was more effective in decreasing pain in the short,mid and long terms. However, no significant mid-and long-term differences in physical function and quality of life was observed. Little good evidence is available for surgical treatment compared with that for sham procedure. PV and BK are currently used to treat VCFs with osteopenia, with little difference in treatment effects. Evidence of better quality and from a larger sample size is required before a recommendation can be made.
# Introduction
An estimated 1.4 million vertebral compression fractures (VCFs) are reported worldwide every year [bib_ref] An estimate of the worldwide prevalence and disability associated with osteoporotic fractures, Johnell [/bib_ref]. Patients with VCF may present with severe back pain, functional disability, and a decrease in quality of life [bib_ref] Costs and quality of life associated with osteoporosis-related fractures in Sweden, Borgström [/bib_ref] [bib_ref] The relationship of health-related quality of life to prevalent and incident vertebral..., Silverman [/bib_ref]. The most common etiology of VCFs is osteoporosis, although trauma, infection, and tumour can also lead to VCFs. Bone metastases and multiple myeloma can lead to generalised osteoporosis or weakening of bone at specific sites [bib_ref] Skeletal complications of malignancy, Coleman [/bib_ref]. As a result, fractures were caused, especially painful VCFs. Aging is associated with a decrease in bone mineral density [bib_ref] Women's health in exercise and aging: What do we know?, Li [/bib_ref] [bib_ref] Women and exercise in aging, Kendall [/bib_ref] [bib_ref] Effects of small-volume soccer and vibration training on body composition, aerobic fitness,..., Connolly [/bib_ref]. VCFs secondary to osteopenia arising from osteoporosis, fall or tumours are a cause of morbidity in older adults. These fractures will increase with the increasing aging of the population [bib_ref] Does this woman have osteoporosis?, Green [/bib_ref] [bib_ref] An 8-week reactive balance training program in older healthy adults: A preliminary..., Paquette [/bib_ref].
Surgical and non-surgical methods are used to treat VCFs. Standard non-surgical treatment includes bed rest, analgesia, use of a back brace or corset, and physical support. Minimally invasive techniques, such as percutaneous vertebroplasty (PV) and balloon kyphoplasty (BK), are popular for treating painful VCFs [bib_ref] New technologies in spine: kyphoplasty and vertebroplasty for the treatment of painful..., Garfin [/bib_ref] [bib_ref] Prospective evaluation of pain relief in 100 patients undergoing percutaneous vertebroplasty: results..., Mcgraw [/bib_ref] [bib_ref] Balloon Kyphoplasty Outcomes Group: Balloon kyphoplasty for symptomatic vertebral body compression fractures..., Garfin [/bib_ref] [bib_ref] Reduction of pain and fracture incidence after kyphoplasty: 1-year outcomes of a..., Grafe [/bib_ref]. PV involves the percutaneous injection of bone cement into the affected vertebra to stabilize the fractured vertebral body, and results in immediate pain relief. It was initially devised in France by , and first performed to treat hemangioma in the cervical spine [bib_ref] Temperature elevation caused by bone cement polymerization during vertebroplasty, Deramond [/bib_ref]. Since then, the operation has become widely used and proved valuable in treating osteoporotic VCFs. BK is a type of PV in which inflatable bone tamps are used to restore vertebral height. After removing the balloons, the resulting intravertebral cavity is filled with bone cement to stabilize the vertebral body. No consensus has been reached over which is the best management strategy for patients with VCFs.
Several published systematic reviews have assessed the effects of surgical management for the treatment of VCFs [bib_ref] Percutaneous vertebroplasty versus conservative treatment of osteoporotic vertebral compression fractures: a systematic..., Tian [/bib_ref] [bib_ref] Vertebroplasty versus non-operative therapy for the treatment of osteoporotic vertebral compressive fractures:..., Cao [/bib_ref] [bib_ref] Comparing pain reduction following vertebroplasty and conservative treatment for osteoporotic vertebral compression..., Liu [/bib_ref]. However, previous reviews focused on one type of surgical management, such as PV or BK. To synthesise the latest trial reports, we have performed this systematic review and meta-analysis of randomized controlled trials (RCTs) of surgical management for treatment of VCFs with osteopenia compared with non-surgical treatment. This review includes trials evaluating a variety of different surgical methods as used to treat people with VCFs due to osteopenia, which provide an overall assessment on the use of surgical managements in this patients population. To our best knowledge, this study is the first comprehensive meta-analysis of surgical management versus non-surgical management in this issue.
# Methods
## Search strategy
We did a systematic review in accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines [bib_ref] Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement, Moher [/bib_ref]. For a completed PRISMA checklist, see S1 Checklist. The Cochrane Central Register of Controlled Trials, PubMed, EMBASE, Web of Science, China Biology Medicine disc, China National Knowledge Infrastructure, and Wanfang Database were searched from the earliest available date to June 2014. We used "Vertebroplasty", "Fractures, Bone" and "RCTs" as search terms (full details of the search in S1 File). Reference lists of selected articles were reviewed for other potentially relevant citations.
## Inclusion criteria
Types of studies. Only RCTs investigating surgical versus non-surgical interventions for the treatment of patients with VCFs were included. No limits on publication dates or any language restrictions were placed.
Types of participants. Patients with VCFs due to osteopenia arising from osteoporosis or tumours were included. We excluded patients who had severe cardiopulmonary comorbidity, untreatable coagulopathy, systemic or local spine infection, suspected underlying malignant disease, radicular syndrome, spinal cord compression syndrome, or contraindications to magnetic resonance imaging.
Types of interventions. We included an experimental group that received surgical treatment, and a control condition in which either conservative treatment or a sham procedure was administered. Surgical treatments included vertebroplasty, kyphoplasty, pedicle screw fixation, anterior reconstruction or fusion, and posterior reconstruction or fusion. Patients in the sham surgical intervention group underwent the same procedure as those in the experimental group up to the insertion of the needle. Conservative interventions consisted of bed rest, medication (e.g., analgesics), use of a corset, brace, or cast, and rehabilitation (e.g., physiotherapy).
Types of outcome measures. The primary outcomes were self-reported pain and adverse effects. All reported symptomatic and asymptomatic adverse events were recorded. Secondary outcomes were function or disability, health-related quality of life, and hospitalization costs. The costs included procedural costs, rehabilitation costs, and primary costs. All outcome measures were pre-specified before review commencement. We categorized outcomes as shortterm (not longer than three months), mid-term (six months), and long-term (twelve months or more) follow-up. All outcome measures were pre-specified before review commencement.
## Data extraction
Two reviewers (Guo JB and Xie B) independently extracted and cross-checked data on trials. The decision to include studies was made initially on the basis of the study title and abstract. When a study could not be excluded with certainty at this stage, the full-text was obtained for evaluation. Disagreements were resolved by discussion and, where necessary, in consultation with a third reviewer (Zhang WY). Extracted information included the first author, publication year, study design (e.g., intervention, follow-up length, outcome measures), characteristics of participants (e.g., age, gender, inclusion/exclusion criteria, sample size, duration of complaint), and information to assess the risk of bias.
## Quality assessment
The risk of bias of the included studies was independently assessed by two authors (Guo JB and Chen BL) using the Cochrane Collaboration's "Risk of Bias". We assessed the following domains: sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective outcome reporting, and other sources of bias. Risk of bias was categorized as low, unclear, or high for each of the included studies. Disagreements were resolved by third party adjudication.
# Statistical analysis
Quantitative data were entered into Review Manager (RevMan 5.1). When appropriate, results of comparable groups of studies were pooled in a meta-analysis using the random-effects model. For dichotomous outcomes, risk ratios (RRs) with 95% confidence intervals (CIs) were calculated. For continuous outcomes, mean difference (MDs) with 95% CIs or standardized mean differences with 95% CI were calculated. P values were 2-tailed, and a value of P<0.05 was considered statistically significant. Heterogeneity between studies was measured by Chi 2 statistic (P<0.1) and quantified with I² statistic [bib_ref] Measuring inconsistency in meta-analyses, Higgins [/bib_ref]. We judged the I 2 values of less than 25% to be minimal heterogeneity, less than50% to be moderate heterogeneity, and 50% or greater to be substantial heterogeneity. In attempting to dissipate any heterogeneity, subgroup analyses were performed on studies. Since the different trials implemented various types of surgical management, trials were divided according to the type of intervention (e.g., PV, BK). If there were two or more surgical methods in the same comparison, then the subgroup analysis was conducted. Potential publication bias was evaluated using funnel plots.
When studies required for the meta-analysis did not provide means and standard deviations at follow-up, we contacted corresponding authors to obtain the unreported data. In the few cases in which a response was unavailable, we estimated the standard deviations using the formula suggested in the Cochrane Handbook for Systematic Reviews of Interventions [bib_ref] The Cochrane Collaboration's tool for assessing risk of bias in randomised trials, Higgins [/bib_ref].
# Results
Our search identified a total of 10,374 records. After removing duplicates, 6,251 records were screened. A further 74 were determined to be potentially relevant for full-text review. Among these studies, 11 studies (16 reports) met the inclusion criteria [bib_ref] A randomized trial of vertebroplasty for painful osteoporotic vertebral fractures, Buchbinder [/bib_ref] [bib_ref] A randomized trial of vertebroplasty for osteoporotic spinal fractures, Kallmes [/bib_ref] [bib_ref] Percutaneous vertebroplasty compared with optimal pain medication treatment: short-term clinical outcome of..., Voormolen [/bib_ref] [bib_ref] Efficacy and safety of balloon kyphoplasty compared with non-surgical care for vertebral..., Wardlaw [/bib_ref] [bib_ref] Balloon kyphoplasty for the treatment of acute vertebral compression fractures: 2-year results..., Boonen [/bib_ref] [bib_ref] Investigational vertebroplasty safety and efficacy trial (INVEST): patient-reported outcomes through 1 year, Comstock [/bib_ref] [bib_ref] Percutaneous vertebroplasty compared to conservative treatment in patients with osteoporotic vertebral compression..., Chen [/bib_ref] [bib_ref] Vertebroplasty versus conservative treatment in acute osteoporotic vertebral compression fractures (Vertos II):..., Klazen [/bib_ref] [bib_ref] Percutaneous vertebroplasty is not a risk factor for new osteoporotic compression fractures:..., Klazen [/bib_ref] [bib_ref] Twelve-months follow-up in forty-nine patients with acute/semiacute osteoporotic vertebral fractures treated conservatively..., Rousing [/bib_ref] [bib_ref] Percutaneous kyphoplasty versus conservative treatment of acute and subacute osteoporotic vertebral compression..., Xie [/bib_ref] [bib_ref] Balloon kyphoplasty versus non-surgical fracture management for treatment of painful vertebral body..., Berenson [/bib_ref] [bib_ref] Randomized controlled trial of percutaneous vertebroplasty versus optimal medical management for the..., Farrokhi [/bib_ref] [bib_ref] Effect of vertebroplasty on pain relief, quality of life, and the incidence..., Blasco [/bib_ref] , and provided data for the meta-analysis. A manual search of reference lists from these studies did not yield any new eligible study [fig_ref] Fig 1: Flow chart of study selection [/fig_ref].
## Study characteristics
Eleven studies (16 reports) were included, and a summary of their characteristics is shown in [fig_ref] Table 1: Characteristics of included studies [/fig_ref]. Six studies were small single-center trials, and five [bib_ref] A randomized trial of vertebroplasty for painful osteoporotic vertebral fractures, Buchbinder [/bib_ref] [bib_ref] A randomized trial of vertebroplasty for osteoporotic spinal fractures, Kallmes [/bib_ref] [bib_ref] Efficacy and safety of balloon kyphoplasty compared with non-surgical care for vertebral..., Wardlaw [/bib_ref] [bib_ref] Balloon kyphoplasty for the treatment of acute vertebral compression fractures: 2-year results..., Boonen [/bib_ref] [bib_ref] Investigational vertebroplasty safety and efficacy trial (INVEST): patient-reported outcomes through 1 year, Comstock [/bib_ref] [bib_ref] Percutaneous vertebroplasty compared to conservative treatment in patients with osteoporotic vertebral compression..., Chen [/bib_ref] [bib_ref] Vertebroplasty versus conservative treatment in acute osteoporotic vertebral compression fractures (Vertos II):..., Klazen [/bib_ref] [bib_ref] Percutaneous vertebroplasty is not a risk factor for new osteoporotic compression fractures:..., Klazen [/bib_ref] [bib_ref] Balloon kyphoplasty versus non-surgical fracture management for treatment of painful vertebral body..., Berenson [/bib_ref] were multicenter trials. All the studies used surgical treatment as the intervention. Two types of interventions were used; eight studies used PV, three studies used BK. Regarding controls, ten studies [bib_ref] Percutaneous vertebroplasty compared with optimal pain medication treatment: short-term clinical outcome of..., Voormolen [/bib_ref] [bib_ref] Efficacy and safety of balloon kyphoplasty compared with non-surgical care for vertebral..., Wardlaw [/bib_ref] [bib_ref] Balloon kyphoplasty for the treatment of acute vertebral compression fractures: 2-year results..., Boonen [/bib_ref] [bib_ref] Percutaneous vertebroplasty compared to conservative treatment in patients with osteoporotic vertebral compression..., Chen [/bib_ref] [bib_ref] Vertebroplasty versus conservative treatment in acute osteoporotic vertebral compression fractures (Vertos II):..., Klazen [/bib_ref] [bib_ref] Percutaneous vertebroplasty is not a risk factor for new osteoporotic compression fractures:..., Klazen [/bib_ref] [bib_ref] Twelve-months follow-up in forty-nine patients with acute/semiacute osteoporotic vertebral fractures treated conservatively..., Rousing [/bib_ref] [bib_ref] Percutaneous kyphoplasty versus conservative treatment of acute and subacute osteoporotic vertebral compression..., Xie [/bib_ref] [bib_ref] Balloon kyphoplasty versus non-surgical fracture management for treatment of painful vertebral body..., Berenson [/bib_ref] [bib_ref] Randomized controlled trial of percutaneous vertebroplasty versus optimal medical management for the..., Farrokhi [/bib_ref] [bib_ref] Effect of vertebroplasty on pain relief, quality of life, and the incidence..., Blasco [/bib_ref] used conservative treatment and two [bib_ref] A randomized trial of vertebroplasty for painful osteoporotic vertebral fractures, Buchbinder [/bib_ref] [bib_ref] A randomized trial of vertebroplasty for osteoporotic spinal fractures, Kallmes [/bib_ref] [bib_ref] Investigational vertebroplasty safety and efficacy trial (INVEST): patient-reported outcomes through 1 year, Comstock [/bib_ref] used sham surgical treatment. All participants had VCFs due to osteopenia. One study [bib_ref] Balloon kyphoplasty versus non-surgical fracture management for treatment of painful vertebral body..., Berenson [/bib_ref] arised from bone metastases and multiple myeloma. Another study [bib_ref] Efficacy and safety of balloon kyphoplasty compared with non-surgical care for vertebral..., Wardlaw [/bib_ref] was caused by either osteoporosis or tumours, and the other participants all had osteoporotic VCFs. Data from 1,401 participants were extracted, of whom 701 received surgical treatment.
## Qualitative assessment
Methodological details are reported in S1 and S2 Figs. Three (18.75%) trials [bib_ref] Percutaneous vertebroplasty compared to conservative treatment in patients with osteoporotic vertebral compression..., Chen [/bib_ref] [bib_ref] Efficacy and safety of balloon kyphoplasty compared with non-surgical care for vertebral..., Wardlaw [/bib_ref] [bib_ref] Balloon kyphoplasty for the treatment of acute vertebral compression fractures: 2-year results..., Boonen [/bib_ref] [bib_ref] Investigational vertebroplasty safety and efficacy trial (INVEST): patient-reported outcomes through 1 year, Comstock [/bib_ref] [bib_ref] Percutaneous vertebroplasty compared to conservative treatment in patients with osteoporotic vertebral compression..., Chen [/bib_ref] [bib_ref] Vertebroplasty versus conservative treatment in acute osteoporotic vertebral compression fractures (Vertos II):..., Klazen [/bib_ref] [bib_ref] Percutaneous vertebroplasty is not a risk factor for new osteoporotic compression fractures:..., Klazen [/bib_ref] [bib_ref] Twelve-months follow-up in forty-nine patients with acute/semiacute osteoporotic vertebral fractures treated conservatively..., Rousing [/bib_ref] [bib_ref] Percutaneous kyphoplasty versus conservative treatment of acute and subacute osteoporotic vertebral compression..., Xie [/bib_ref] [bib_ref] Balloon kyphoplasty versus non-surgical fracture management for treatment of painful vertebral body..., Berenson [/bib_ref] [bib_ref] Effect of vertebroplasty on pain relief, quality of life, and the incidence..., Blasco [/bib_ref] were not provided with allocation concealment, which resulted in unclear risk of selection bias. Thirteen (81.25%) trials [bib_ref] Percutaneous vertebroplasty compared with optimal pain medication treatment: short-term clinical outcome of..., Voormolen [/bib_ref] [bib_ref] Efficacy and safety of balloon kyphoplasty compared with non-surgical care for vertebral..., Wardlaw [/bib_ref] [bib_ref] Balloon kyphoplasty for the treatment of acute vertebral compression fractures: 2-year results..., Boonen [/bib_ref] [bib_ref] Percutaneous vertebroplasty compared to conservative treatment in patients with osteoporotic vertebral compression..., Chen [/bib_ref] [bib_ref] Vertebroplasty versus conservative treatment in acute osteoporotic vertebral compression fractures (Vertos II):..., Klazen [/bib_ref] [bib_ref] Percutaneous vertebroplasty is not a risk factor for new osteoporotic compression fractures:..., Klazen [/bib_ref] [bib_ref] Twelve-months follow-up in forty-nine patients with acute/semiacute osteoporotic vertebral fractures treated conservatively..., Rousing [/bib_ref] [bib_ref] Percutaneous kyphoplasty versus conservative treatment of acute and subacute osteoporotic vertebral compression..., Xie [/bib_ref] [bib_ref] Balloon kyphoplasty versus non-surgical fracture management for treatment of painful vertebral body..., Berenson [/bib_ref] [bib_ref] Randomized controlled trial of percutaneous vertebroplasty versus optimal medical management for the..., Farrokhi [/bib_ref] [bib_ref] Effect of vertebroplasty on pain relief, quality of life, and the incidence..., Blasco [/bib_ref] were at high risk of performance bias, reflecting the difficulty in blinding surgeons. Only one (6.25%) trial [bib_ref] Percutaneous kyphoplasty versus conservative treatment of acute and subacute osteoporotic vertebral compression..., Xie [/bib_ref] did not describe the patients who withdrew from the study. Finally, 11 (68.75%) trials stated per protocol the primary method of analysis, and the remaining trials did not describe the method of analysis.
## Comparison 1: surgical versus conservative interventions for
Treating VCF with Osteopenia Self-related pain In total, ten trials evaluated self-related pain using a numeric rating scale (NRS) and a visual analog scale (VAS). Among these trials, nine (1,014 patients) reported data at short-term follow-up period, four (597 patients) at mid-term follow-up period and five (604 patients) at long-term follow-up term. We found significant improvements with surgical management for pain relief in the short term (MD −2.05, 95% CI, −3.55 to −0.56; P = 0.007), mid term (MD −1.70, 95% CI, −2.78 to −0.62; P = 0.002) and long term (MD −1.24, 95% CI, −2.20 to −0.29; P = 0.01) [fig_ref] Fig 2: Self-related pain for surgical versus conservative treatment [/fig_ref]. In the test for subgroup differences, PV was seen to be more effective than conservative treatment at short-term (MD −1.28, 95% CI, −2.42 to −0.13; P = 0.03) [fig_ref] Fig 2: Self-related pain for surgical versus conservative treatment [/fig_ref] , mid-term (MD −1.79, 95% CI, −3.38 to −0.20; P = 0.03) [fig_ref] Fig 2: Self-related pain for surgical versus conservative treatment [/fig_ref] , and long-term follow-up periods (MD −1.38, 95% CI, −2.64 to −0.12; P = 0.03) [fig_ref] Fig 2: Self-related pain for surgical versus conservative treatment [/fig_ref] in reducing pain. We only found significant improvements with BK for pain relief in the short term (MD −3.60, 95% CI, −5.80 to −1.39; P = 0.001) [fig_ref] Fig 2: Self-related pain for surgical versus conservative treatment [/fig_ref]. In addition, it can be observed that the funnel plot had an asymmetrical distribution regarding self-related pain at short-term follow-up period [fig_ref] Fig 3: Funnel plot of included studies regarding self-related pain in the short term [/fig_ref]. There was a suggestion of publication bias according to the shape of the funnel plot.
## Adverse events
Overall, no significant difference between surgical and conservative treatments was observed for adverse events from seven trials consisting of 855 patients (139/433 versus 124/422; RR 1.10, 95% CI, 0.85 to 1.43; P = 0.46) [fig_ref] Fig 4: Overall adverse events for surgical versus conservative treatment [/fig_ref]. In the test for subgroup differences, five trials reported no significant difference between PV and conservative treatments(39/214 versus 32/ 207; RR 1.29, 95% CI, 0.65 to 2.59; P = 0.47), and two reported no significant difference between BK and conservative treatments(100/219 versus 92/215; RR 1.07, 95% CI, 0.86 to 1.31; P = 0.55). One trial was not entered into this forest plot because of incomplete data. The study did not report adverse events in the control group [bib_ref] Randomized controlled trial of percutaneous vertebroplasty versus optimal medical management for the..., Farrokhi [/bib_ref]. Further analyses by types of adverse events showed that musculoskeletal disorders were the most common adverse events in both groups. Among musculoskeletal disorders, the frequency of new fractures was the highest (97/402 versus 80/373), but no statistically significant difference was observed between groups (P = 0.62) [fig_ref] Fig 5: Incidence of new fractures for surgical versus conservative treatment [/fig_ref]. For individual adverse events, no statistically significant difference between the surgical and conservative treatment groups was observed (S3 One study [bib_ref] Vertebroplasty versus conservative treatment in acute osteoporotic vertebral compression fractures (Vertos II):..., Klazen [/bib_ref] stated that adverse events related to surgical procedure include hematoma and urinary tract infection. Other adverse events that have been reported include infections, cardiovascular and vascular disorders, injury or procedural complications, and blood and lymphatic disorders.
## Function or disability
Four trials collected data for the Roland-Morris Disability (RMD) score, with three (372 patients) reporting a decrease in the RMD score after surgical intervention at short-term followup period (MD −4.97, 95% CI, −8.71 to −1.23; P = 0.009) [fig_ref] Fig 6: RMD score for surgical versus conservative treatment in the short term [/fig_ref]. In the test for subgroup differences, RMD score decreased with BK compared with that with conservative treatment (MD -5.99, 95% CI, -10.60 to -1.38; P = 0.01). Only one trial [bib_ref] Balloon kyphoplasty for the treatment of acute vertebral compression fractures: 2-year results..., Boonen [/bib_ref] at six months follow-up was noted, and surgical treatment was more effective in disability on the RMD score. However, no significant differences were observed between groups at twenty-four months follow-up.
## Health-related quality of life
We used more than one measure of quality of life, and we preferentially included the following measures: the Short-Form 36 (SF-36), the Quality of Life Questionnaire of the European Foundation for Osteoporosis (QUALEFFO), and the European Quality of Life-5 Dimensions (EQ-5D). SF-36: Four trials, with a total of 614 patients reported data for the Short-Form 36 Physical Component Summary (SF-36 PCS) score. We observed a greater increase at short-term followup period after surgical procedure (MD 5.53, 95% CI, 1.45 to 9.61; P = 0.008) [fig_ref] Fig 7: SF-36 PCS score for surgical versus conservative treatment [/fig_ref]. In the test for subgroup differences, we observed significant differences in favor of BK at short-term follow-up period (MD 7.16, 95% CI, 5.72 to 8.60; P < 0.00001). However, no difference was observed between the two groups at long-term follow-up period (MD 1.20, 95% CI, −1.33 to 3.72; P = 0.93) [fig_ref] Fig 7: SF-36 PCS score for surgical versus conservative treatment [/fig_ref]. A single trial [bib_ref] A randomized trial of vertebroplasty for osteoporotic spinal fractures, Kallmes [/bib_ref] reported that surgical treatment was more effective in improving quality of life on the SF-36 PCS score at six months follow-up. Considering the Short-Form 36 Mental Component Summary (SF-36 MCS) score, only small study effects were found in favor of the surgical group at short-term follow-up period (MD 7.38, 95% CI, 1.13 to 13.63; P = 0.02) (S4 QUALEFFO: Two trials (144 patients) collected data for the QUALEFFO score, which reported that surgical treatment was more effective than conservative treatment at short-term follow-up period (MD −5.01, 95% CI, −8.11 to −1.91; P = 0.002) (S5 A single trial [bib_ref] Effect of vertebroplasty on pain relief, quality of life, and the incidence..., Blasco [/bib_ref] reported that significant improvement was obtained at both six and twelve months follow-up in the surgical group.
EQ-5D: Three trials, with a total of 500 patients reported data for EQ-5D. However, no significant differences were observed between surgical and conservative management for the EQ-5D score at short-or long-term follow-up periods. [fig_ref] Fig 6: RMD score for surgical versus conservative treatment in the short term [/fig_ref].
## Hospitalization costs
Two trials [bib_ref] Vertebroplasty versus conservative treatment in acute osteoporotic vertebral compression fractures (Vertos II):..., Klazen [/bib_ref] reported hospitalization costs. Data from Klazen [bib_ref] Vertebroplasty versus conservative treatment in acute osteoporotic vertebral compression fractures (Vertos II):..., Klazen [/bib_ref] showed that the cost of surgery in the first month was higher than the cost of conservative treatment because of procedural costs (P < 0.0001). However, the difference was no longer significant at one year (P = 0.087), which resulted from continued pain and loss of function in the conservative treatment group. The resulting incremental cost-effectiveness suggested that vertebroplasty was warranted for the patients with vertebral fractures treated at a mean of 5.6 weeks after the start of symptoms. Data from Fritzelldemonstrated that BK costs more than standard medical treatment with a significant difference of 7,833€(95% CI, 1671 to 12,511), and from this trial it was not possible to demonstrate that BK was more cost-effective than standard medical treatment in patients treated for an acute or subacute vertebral fracture because of osteoporosis.
## Comparison 2: surgical intervention versus a sham surgical
Procedure for Treating VCF with Osteopenia Self-related pain Three trials collected data for self-related pain, with two (198 patients) reporting pain relief after surgical intervention at short-term follow-up period (MD −4.97, 95% CI, −8.71 to −1.23; P = 0.009) [fig_ref] Fig 8: Self-related pain for surgical versus sham surgical treatment [/fig_ref]. However, no significant differences were found in the mid term. Only one trial [bib_ref] Investigational vertebroplasty safety and efficacy trial (INVEST): patient-reported outcomes through 1 year, Comstock [/bib_ref] at twelve months follow-up period was noted, and there were no significant Surgical Versus Non-Surgical for Vertebral Compression Fracture differences between two groups. Only one surgical method PV was used, so we did not make any subgroup analyses in comparison 2.
## Adverse events
Data on adverse events were available for both studies. No significant difference between surgical intervention and sham surgical intervention was observed in the pooled results for overall adverse events (20/106 versus 16/103; RR 1.32, 95% CI, 0.80 to 2.18; P = 0.28) (S7 Adverse events in the two studies included incident fracture, injury or procedural complications, tachycardia and rigors, osteomyelitis, tightness in the back or rib cage, and pain [bib_ref] A randomized trial of vertebroplasty for painful osteoporotic vertebral fractures, Buchbinder [/bib_ref] [bib_ref] A randomized trial of vertebroplasty for osteoporotic spinal fractures, Kallmes [/bib_ref].
## Function or disability
Three trials collected data for the RMD score, and no significant difference was found among groups at short-term (MD 0.03, 95% CI, −1.77 to 1.83; P = 0.98) and mid-term follow-up periods (MD −0.73, 95% CI, −2.76 to 1.30; P = 0.48 (S8 A single trial [bib_ref] Investigational vertebroplasty safety and efficacy trial (INVEST): patient-reported outcomes through 1 year, Comstock [/bib_ref] reported that there were no significant differences between groups at twelve months follow-up.
## Health-related quality of life
A single trial [bib_ref] A randomized trial of vertebroplasty for osteoporotic spinal fractures, Kallmes [/bib_ref] found no significant difference between the two intervention groups at one month follow-up for the SF-36 PCS score, SF-36 MCS score, and EQ-5D score. Considering the QUALEFFO score, no significant differences were observed between groups at any time points measured, except at one week, which favored the sham procedure group [bib_ref] A randomized trial of vertebroplasty for painful osteoporotic vertebral fractures, Buchbinder [/bib_ref].
# Subgroup analysis
The primary comparison in this review was any surgical intervention versus any conservative intervention. Secondary comparison analyses were any surgical intervention versus a sham surgical procedure. The subgroup analysis included two comparisons as follows: PV versus conservative treatment and BK versus conservative treatment. Only one outcome, the self-related pain at short-term follow-up period, showed significant heterogeneity between the treatment effects of the different classes of intervention. In all other cases, threre was no significant heterogeneity between the treatment effects of the different classes of intervention [fig_ref] Table 2: Summary of results. [/fig_ref]. One outlying trial was the cause of this heterogeneity in the self-related pain at short-term followup period [bib_ref] Balloon kyphoplasty versus non-surgical fracture management for treatment of painful vertebral body..., Berenson [/bib_ref] ; when this trial was excluded from the analysis, the result remained significant (MD −1.85, 95% CI, −3.68 to −0.02; P = 0.05), but the test for between subgroup heterogeneity was no longer significant (P = 0.30).
# Discussion
Previous reviews have examined the results of PV versus conservative treatment or BK versus conservative treatment, but none has combined the surgical methods. BK is a type of PV, and they invovle similar techniques. In addition, both of them have less invasive procedures with similar risk profiles compared to open surgery. Therefore, this review collated all the evidence from a large number of trials, which evaluated vertebroplasty and kyphoplasty into one review to assess the overall effect of surgical versus non-surgical treatment of VCFs with osteopenia. It also allowed an indirect comparison of between PV and BK, and provided no evidence of differences in the treatment effect between the two types of surgical methods. Statistical significance was reported in comparison one for five outcomes: self-related pain, RMD, SF-36 PCS, SF-36 MCS, and QUALEFFO. The most of the significant differences we observed were at short-term follow-up period. Only the improvements of self-related pain were seen in the short, mid, and long terms. This meta-analysis suggested a favorable association of surgical treatment in reducing pain, improving function and quality of life compared with conservative treatment. By contrast, when we focused on the two studies [bib_ref] A randomized trial of vertebroplasty for painful osteoporotic vertebral fractures, Buchbinder [/bib_ref] [bib_ref] A randomized trial of vertebroplasty for osteoporotic spinal fractures, Kallmes [/bib_ref] in which surgical procedure was compared with sham procedure, we saw no evidence of improvement in pain relief, physical function and quality of life. The kind of these two studies is generally speaking difficult to achieve in surgical procedures, and often reaches the limit of ethically acceptable studies. Both studies are multicenter, double-blind RCTs published in the New England Journal of Medicine. The results from the two papers have aroused worldwide attention and intense controversy [bib_ref] Effectiveness of vertebroplasty: a recent controversy, Munk [/bib_ref] [bib_ref] Vertebroplasty:about sense and nonsense of uncontrolled "controlled randomized prospective trials, Aebi [/bib_ref] [bib_ref] Response to "randomized vertebroplasty trials: bad news or sham news?, Kallmes [/bib_ref]. The placebo effect may be explained by several reasons, such as narcotic effects, psychological effects, and neural excitatory effects. However, at long-term follow-up period, pain relief with placebo treatment was similar to that with surgical management. Samuel Butler [bib_ref] Effectiveness of vertebroplasty: a recent controversy, Munk [/bib_ref] stated that these two studies comprised samples fewer than 300 patients in both the treatment and control groups, and many patients were unwilling to accept randomization, especially for patients in severe pain. Thus, the small size of sample and selection bias made these studies untrustworthy. In spite of these limitations, we still observed a statistical trend, which suggested that vertebroplasty was more effective than sham treatment. For example, data from Kallmes [bib_ref] A randomized trial of vertebroplasty for osteoporotic spinal fractures, Kallmes [/bib_ref] showed a higher rate of clinically meaningful improvement in pain relief in the vertebroplasty group than that in the control group (64% versus 48%, P = 0.06). Vertebroplasty may be effective with a larger sample size. Therefore, high-quality, well-designed studies are needed to address the areas of uncertainty.
The incidence of new fractures was similar after surgical and conservative therapy. Thus, the occurrence of new vertebral fractures was due to ongoing osteoporosis and not the type of therapy. About 20% of patients with a history of osteoporotic vertebral fractures will experience a new vertebral fracture within a year, depending on the severity of the prior fracture [bib_ref] Patients with prior fractures have an increased risk of future fractures: a..., Klotzbuecher [/bib_ref]. In addition, no statistically significant difference was observed between groups in other adverse events. In this meta-analysis, we did not believe that the risk of adverse events after surgery should be a contraindication for the operation.
The findings from our review were consistent with findings from previous reviews. Five reviews have been published, and all of them included trials only using vertebroplasty as the experimental group [bib_ref] Percutaneous vertebroplasty versus conservative treatment of osteoporotic vertebral compression fractures: a systematic..., Tian [/bib_ref] [bib_ref] Vertebroplasty versus non-operative therapy for the treatment of osteoporotic vertebral compressive fractures:..., Cao [/bib_ref] [bib_ref] Comparing pain reduction following vertebroplasty and conservative treatment for osteoporotic vertebral compression..., Liu [/bib_ref] [bib_ref] Effectiveness of vertebroplasty using individual patient data from two randomised placebo controlled..., Staples [/bib_ref]. Four of these reviews restricted the results to RCTs only, whereas one [bib_ref] Vertebroplasty versus non-operative therapy for the treatment of osteoporotic vertebral compressive fractures:..., Cao [/bib_ref] included randomized and non-randomized clinical trials. Three of these reviews [bib_ref] Percutaneous vertebroplasty versus conservative treatment of osteoporotic vertebral compression fractures: a systematic..., Tian [/bib_ref] [bib_ref] Comparing pain reduction following vertebroplasty and conservative treatment for osteoporotic vertebral compression..., Liu [/bib_ref] found that vertebroplasty in pain relief is more obvious than that in conservative treatment at short-term, mid-term, and long-term follow-up. Four trials [bib_ref] Percutaneous vertebroplasty versus conservative treatment of osteoporotic vertebral compression fractures: a systematic..., Tian [/bib_ref] [bib_ref] Comparing pain reduction following vertebroplasty and conservative treatment for osteoporotic vertebral compression..., Liu [/bib_ref] [bib_ref] Effectiveness of vertebroplasty using individual patient data from two randomised placebo controlled..., Staples [/bib_ref] compared vertebroplasty with a sham procedure in patients with VCFs, and all failed to show an advantage of vertebroplasty over placebo.
# Strengths
Our study had several strengths. The review highlights the wide range of surgical techniques being used in the treatment of VCFs with osteopenia, and previous reviews focused on one type of surgical management. We made subgroup analyses for vertebroplasty and kyphoplasty, and it allowed an indirect comparison between them. Our review methods were systematic and exhaustive. We employed a broad search strategy without any publication dates or language restrictions. Relevant citations in selected articles were also examined. All trials included in our review were RCTs. Study selection, quality assessment, and data extraction were performed by at least two independent reviewers. We mapped all possible treatment comparisons, and six outcomes were reported in this meta-analysis. Some heterogeneity was observed in the metaanalysis, so we generated random-effects models that accounted for between-study variability.
# Limitations
First, all studies included in this review used pain scales. However, different studies used different pain surveys as outcomes, which possibly created some of the heterogeneity we observed among trials. Second, we noted minimal focus on hospitalization cost analysis of the surgical intervention; therefore, little is known about the cost-effectiveness and economic value of these surgical managements. Third, we estimated the standard deviations for several studies using the formula suggested in the Cochrane Handbook for Systematic Reviews of Interventions because a response was unavailable from certain authors.
## Implications for research
Evidence of better quality and from a larger sample size is required before a recommendation can be made. Further RCTs are needed until we can establish whether surgical management is more effective than non-surgical management in patients with VCFs due to osteopenia. Future studies should identify which type of surgical treatment provides more advantages. The shortcomings of the present evidence in this review include small study size, selection bias, performance bias, detection bias, and attrition bias. Future trials should be reported following the CONSORT guidelines [bib_ref] statement: updated guidelines for reporting parallel group randomised trials, Schulz [/bib_ref] , and data should be analyzed according to intention to treat principles. This review illustrated the need for the universal use of relevant outcome measures, especially adverse events and hospitalization cost outcomes. Moreover, systematic data collection at short, mid-, and long-term follow-up is essential. Short term: not longer than 3 months; Markers represent point estimates of mean difference, marker size represents study weight in random-effects meta-analysis. Horizontal bars indicate 95% confidence intervals. CI, confidence interval; IV, inverse variance. (TIF) S6 [fig_ref] Fig 5: Incidence of new fractures for surgical versus conservative treatment [/fig_ref] score for surgical versus conservative treatment in the short and long terms. Short term: not longer than 3 months; Long term: 12 months or more; Markers represent point estimates of mean difference, marker size represents study weight in random-effects meta-analysis. Horizontal bars indicate 95% confidence intervals. CI, confidence interval; IV, inverse variance. Short term: not longer than 3 months; Mid term: 6 months; Markers represent point estimates of mean difference, marker size represents study weight in random-effects meta-analysis. Horizontal bars indicate 95% confidence intervals. CI, confidence interval; IV, inverse variance. (TIF)
## Supporting information
[fig] Fig 1: Flow chart of study selection. For details of study identification. doi:10.1371/journal.pone.0127145.g001 [/fig]
[fig] Fig 2: Self-related pain for surgical versus conservative treatment. A: mean difference (MD) at the end of the intervention (not longer than 3 months). B: MD at six months. C: MD at long-term follow-up period (12 months or more). PV, percutaneous vertebroplasty; BK, balloon kyphoplasty; CI, confidence interval; IV, inverse variance. [/fig]
[fig] Fig 3: Funnel plot of included studies regarding self-related pain in the short term. Short term: not longer than 3 months. Log of mean difference were plotted against the standard error of mean difference of each study to identify asymmetry in the distribution of trials.doi:10.1371/journal.pone.0127145.g003 Surgical Versus Non-Surgical for Vertebral Compression Fracture PLOS ONE | DOI:10.1371/journal.pone.0127145 May 28, 2015 [/fig]
[fig] Fig 4: Overall adverse events for surgical versus conservative treatment. Markers represent point estimates of risk ratios, marker size represents study weight in random-effects meta-analysis. Horizontal bars indicate 95% confidence intervals. CI, confidence interval; M-H, mantel-haenszel. doi:10.1371/journal.pone.0127145.g004 Surgical Versus Non-Surgical for Vertebral Compression Fracture [/fig]
[fig] Fig 5: Incidence of new fractures for surgical versus conservative treatment. doi:10.1371/journal.pone.0127145.g005 [/fig]
[fig] Fig 6: RMD score for surgical versus conservative treatment in the short term. Short term: not longer than 3 months. doi:10.1371/journal.pone.0127145.g006 [/fig]
[fig] Fig 7: SF-36 PCS score for surgical versus conservative treatment. A: mean difference (MD) at the end of the intervention (not longer than 3 months). B: MD at long-term follow-up period (12 months or more). doi:10.1371/journal.pone.0127145.g007 Surgical Versus Non-Surgical for Vertebral Compression Fracture [/fig]
[fig] Fig 8: Self-related pain for surgical versus sham surgical treatment. Short term: not longer than 3 months; Mid term: 6 months. [/fig]
[fig] S1: Checklist. PRISMA Checklist of this meta-analysis. (DOC) File. Search strategies for all databases. (DOC) Fig. Risk of bias graph using the Cochrane Risk of Bias tool. (TIF) S2 Fig. Risk of bias summary using the Cochrane Risk of Bias tool. (TIF) S3 Fig. Individual adverse events for surgical versus conservative treatment. Markers represent point estimates of risk ratios, marker size represents study weight in random-effects metaanalysis. Horizontal bars indicate 95% confidence intervals. CI, confidence interval; M-H, mantel-haenszel. (TIF) S4 Fig. SF-36 MCS score for surgical versus conservative treatment in the short term. Short term: not longer than 3 months; Markers represent point estimates of mean difference, marker size represents study weight in random-effects meta-analysis. Horizontal bars indicate 95% confidence intervals. CI, confidence interval; IV, inverse variance. (TIF) S5 Fig. QUALEFFO score for surgical versus conservative treatment in the short term. [/fig]
[table] Table 1: Characteristics of included studies. [/table]
[table] Table 2: Summary of results. [/table]
|
Formation in Interprofessional Education in Nursing and Medical Students Globally. Scoping review
Objective. This work sought to know the state of the art related to the theme of Interprofessional Education (IPE) in the training of Nursing and Medical students and the level of evidence developed thus far. Methods. and Development of professional skills. The highest level of evidence is presented by the articles dealing with didactics; on the contrary, no articles were found that dealt with topics related with early inclusion of IPE in the medical and nursing curricula, which are currently necessary to complement the focus of patient-centered care. Conclusion. The thematic nuclei show that the level of evidence in the literature is varied, although mostly descriptive in scope, highlighting the development of professional skills as a result of interprofessional education.La formación en educación interprofesional en estudiantes de enfermería y medicina en el mundo. Scoping reviewResumenObjetivo. Conocer el estado del arte, relacionado con la temática de Educación Interprofesional (EIP) en la formación de estudiantes de enfermería y medicina, y el nivel de evidencia desarrollado hasta el momento. Métodos. Revisión Sistemática exploratoria, declarada como Scoping review por la Joanna Briggs Institute -JBI-, en la que se realizó una búsqueda en bases de datos de Embase, Science direct, Pubmed-Medline, Academic search complete, BVS, Scopus y ERIC, entre 2009-2019 utilizando los términos DeCS y MeSH de Interprofessional education, education research, healthcare professionals, nursing y medicine. Se seleccionaron 39 artículos originales luego del proceso de crítica con los criterios de JBI. Resultados. Emergieron cuatro núcleos temáticos: Experiencias y percepciones del aprendizaje interprofesional, Didácticas vinculadas a la EIP, Indicadores empíricos relacionados con la EIP y Desarrollo de competencias profesionales. El mayor nivel de evidencia lo presentan los artículos sobre empleo de didácticas; por el contrario, no se encontraron artículos que trataran temáticas relacionadas con la inclusión temprana de EIP en los currículos de enfermería y medicina, pues son actualmente necesarios para complementar el enfoque de atención centrada en el paciente. Conclusión. Los núcleos temáticos muestran que el nivel de evidencia de la literatura es variado, aunque en su mayoría de alcance descriptivo, lo que Descritores: educação interprofissional; educação em enfermagem; educação médica; estudantes de enfermagem; estudantes de medicina; currículo; revisão.Invest Educ Enferm. 2020; 38(2): e06 Formation in Interprofessional Education in Nursing and Medical Students Globally. Scoping review
resalta el desarrollo de competencias profesionales como resultado de la educación interprofesional.
Descriptores: educación interprofesional; educación en enfermería; educación médica; estudiantes de enfermería; estudiantes de medicina; curriculum; revisión. Resultados. Emergiram quatro núcleos temáticos: Experiências e percepções da aprendizagem interprofissional, Didáticas vinculadas à EIP, Indicadores empíricos relacionados com a EIP e Desenvolvimento de competências profissionais. O maior nível de evidência o apresentam os artigos sobre emprego de didáticas; pelo contrário, não se encontraram artigos que trataram temáticas relacionadas com a inclusão precoce de EIP nos currículos de enfermagem e medicina, sendo atualmente necessários para complementar o enfoque de atenção centrada no paciente. Conclusão. Os núcleos temáticos mostram que o nível de evidência da literatura é variado, embora na sua maioria de alcance descritivo, ressaltando o desenvolvimento de competências profissionais como resultado da educação interprofissional.
# Introduction
O ne of the ways to respond to the Millennium Development Goals with respect to universal health is related with guaranteeing the formation of competent professionals, who are focused on responding to the needs of the people through collaborative work among the different health disciplines; however, this conception tends to be removed from reality in global health systems, where confrontations among professionals play a harmful role for users. This is how interprofessional education (IPE) emerges as a response to this problem, as a teaching and learning approach that joins students from two or more professions to learn about their roles,in which they learn with, learn of, and learn about each other to improve collaboration and the quality of health care.As presented, one of the principal problems to which interprofessional education responds is the definition of roles among health professionals, especially between physicians and nurses; this has been considered a constant problem in the clinical setting in which a thin line exists between what each one can do and can cause confusion in the roles. According with the Theory of Roles,the role is a set of prescriptions that define the behavior of a group member in a given position within the group; however, when students are not in contact during their formation with other disciplines with which they will work, it is difficult for them to manage to identify the extent of their tasks or define which these are.
According with a systematic review by Reeves et al.,IPE seeks to improve collaboration among distinct types of health professionals and from social care, premise that is based on 15 studies reviewed, which assessed the effectiveness of IPE interventions, evidencing positive results regarding patient satisfaction, conduct of the collaborating team, and diminished rates of clinical errors in the service teams. The evidence from the reviews foundare aimed exclusively at the impact of interventions, added to their focus being given within the context of collaborative practices, making it necessary to inquire if during the period of university formation, and within the curricula of Nursing and Medical programs, these incorporate strategies to promote IPE.Consequently, the review proposed develops more broadly the topic of this theme, not only focusing on interventions but on the experiences provided by this education and the skills that can be improved with it.
To develop IPE, we must consider methodologies that permit students to be active, interactive, reflexive, and be centered on the patient. These methods may be used to create opportunities, to compare and contrast the functions and responsibilities, power and authority, ethics and codes of practices, knowledge and skills; to establish effective relations and develop and reinforce aptitudes for collaborative practice.Having described its importance in professional formation, it is imperative to know the state of the art related with the theme of interprofessional education (IPE) on the formation of students from the Nursing and Medical careers and the level of evidence developed until now, thus, determining the contributions and knowledge gaps present in the literature.
# Methods
Descriptive Scoping Reviewwhich sought to answer the questions: What is the current state of scientific knowledge of the IPE phenomenon in Nursing and Medicine? What is the evidence of its applicability, benefits, and limitations? What are the research gaps on this theme? The search was performed in the Embase, Science direct, Pubmed, Medline, Academic search complete, BVS, Scopus, and ERIC databases in English, Spanish, and Portuguese, limited to between 2009 and 2019, with this time range being adequate to encompass the evolution and current status of the theme, which is relatively novel in Latin America. The study used DeCS and MeSH terms of interprofessional education, education research, healthcare professionals, nursing and medicine, performing the search with equations using AND and OR Boolean operators. Although the context of this review is found globally, for effects of the search, in some of the databases the information was filtered by Latin America, North America, Europe, Asia, Africa and Oceania, which permits recognizing the places in the world where the IPE theme is more or less developed.
The inclusion criteria for this review considered original articles, which present data from research results available in full text, which included subjects like professors and/or Nursing and Medical students. The study excluded texts and theses due to their length and because they address the theme of education without being specific on the interprofessional. It also excluded reflexion articles because they do not contribute concrete and specific results, which could be classified within the levels of evidence provided by the JBI (11) and review articles because contribute secondary sources of information and contain interpretations by the authors about the primary sources, which for this review, are not relevant. The search and selection strategy of articles is detailed in, which has been adapted from the PRISMA flowchart for Scoping Review.
## Identification
## Search
## Eligibility included
For information analysis, the articles were collected in the Microsoft Excel program, where a matrix was created with their critique and analysis. To conduct the critique of the quality of the articles, the rigor criteria were used in the research, bearing in mind their approach. For qualitative studies, the review applied criteria by Gubaof credibility, transferability, dependence, and confirmability. Quantitative studies used criteria of internal, external, construct, and statistical validity.Regarding the analysis of theme addressed in the articles, development of the matrix defined as variables to consider the research approach, study design, data collection manner, analysis used, their principal results, and the conclusions; then, with this information condensed, the work followed the steps exposed by Peters et al.,which include question, objective, inclusion criteria, participants, principal concept, context, search and selection of articles, extraction and classification of results and discussion.
# Results
The bibliographic search led to the selection of 120 original articles, which after the critique excluded 81 of them, 45 quantitative studies which mostly had faults in their internal validity; 25 qualitative studies that did not clarify the confirmability of the data or the reflexivity in the research process and 11 mixed studies due to declaring this methodology without evidencing its integration in the results. In the end, 39 original articles were analyzed, having as criteria their thematic and/or methodological affinity, thereby, deriving into four thematic groups: Experiences and perceptions of interprofessional learning, Didactics linked to IPE, Empirical indicators related with IPE, and Development of professional skills. With respect to the methodological distribution of the articles after their integration in the principal concepts, shows that the concept most approached was the development of professional skills by 36%. The predominant methodological approach was qualitative research by 28.2% of the total analyzed and development of psychometric research with 12.8%. With respect to their geographic location, 61.5% of the articles are from North America, comprising the United States and Canada. Two articles were identified in Latin America and none in Africa. shows, not only how limited IPE research is in Latin America, but the progress in design and implementation of interprofessional formation programs in Nursing and Medicine taking place in the United States and Europe. Oceania Australia (n = 2) 2 (5.1)
## Concepts
## Table 2. articles located by geographic location
Besides the contributions found in each of the articles reviewed, which will be presented in the discussion, the work presents the levels of evidence according with the JBI (11) and the research gaps according with each thematic category.
# Discussion
The following presents the discussion among the relevant findings of the emerging themes in this review, levels of evidence, and explicit gaps.
## Development of professional skills (n=14)
# Contributions found
Research on IPE has demonstrated that it favors the development of professional skills, framed within knowledge, skills, and attitudes that permit optimizing interpersonal relationships with colleagues, peers, and users, besides recognizing their shared and autonomous roles and of supporting the capacity to make shared decisions. Initially, on addressing the preparation of students upon IPE, Judge et al.,proposed a quasi-experimental study to evaluate the response capacity against IPE through cases that had to be solved by students from odontology, medicine, nursing, pharmacy, dietetics, and physical therapy. These authors concluded that the average score in the Scale of preparation for interprofessional learning in nursing, after the intervention, was higher compared with other disciplines. Now, on approaching the professional skills developed with IPE, we approach the perception of students and its measurement with valid instruments; thus, from this qualitative approach diverse authors gathered from the students that the knowledge and skills they developed most were related with the capacity for reflexion, communication, identification of their role and that of the other, construction of team relationships, collaboration, and mediation to make decisions, to provide patient-centered quality care.Other studies have described that when students from the health sciences learn and work in interprofessional teams, they obtain greater clarity of their role and ability to plan tasks than when they are in intra-professional teams, added to the positive impact this has on the sensitization on the role of other professions.Moreover, the interventions conducted based on didactics around IPE evidenced promotion in the development of skills related with communication, teamwork, conflict resolution, decision making, roles and attitudes on care by the interprofessional team.For Baker and Durham,elaborating an IPE course permitted improving collaboration, conflict resolution, and communication among students from medicine, nursing, and pharmacy; aspects also indicated by Castillo Parra et al.,emphasizing on the construction of new knowledge with collective learning. Other authors who used simulation for interprofessional learning described that their participants changed their preconceived ideas on physician, nursing, and pharmaceutical professionals.In a more practical level, a program was designed consisting of interviews of patients, followed by a team meeting to develop an integral care plan. Statistically, no significant changes were noted on the attitudes, but the comments from the focal groups highlighted the value of collaboration among health professions and the potential benefit of participating more than once to obtain longer lasting experience with a patient and have additional exposure to work as interprofessional team.It is also highlighted that IPE on the care of individuals with chronic diseases, like diabetes, where physiopathological knowledge is transversal in medicine and nursing, permits development of clinical skills.From the perspective of bioethics in the practice of health sciences, diverse studies foundmanaged to support that under IPE development is achieved of this vision of ethics and values in students added to the practice and interdisciplinary teamwork. Thereby, it is recognized that at curricular level the programs of health sciences have tried to wage on IPE as a way of enhancing professional collaboration, identifying the role to perform, and developing transversal values for the care of patients, in addition to the joint commitment to maintain knowledge updated to benefit the users, that is, application of evidence-based practice but interprofessional.In this same sense, the study by Harper Boland et al.,determined that with IPE students improved their skills for teamwork (p value of the pre and post difference <0.001) and their level of trust on themselves and on others, added to the students expressing greater comprehension of the differences in the values and ethics of the multiple professions, which generated increased recognition and respect for the differences within the work carried out by each profession.
Finally, the study by González et al.,evaluates the impact of IPE in developing skills inherent to this learning, such as communication, role definition, teamwork, and decision making, finding a positive effect in the students' self-perceptions, especially with interprofessional communication, in the dimensions of oral expression, active listening, and conflict management (p = 0.018; p = 0.018; p = 0.036 and p<0.001, respectively), revealing that this type of education helps future health professionals to center on caring for people and not on the exclusivity of their roles.
The evidence collected shows that the skills developed most were role definition and teamwork aimed at the care of patients being safe and humanized, supporting the idea that the implementation of IPE in undergraduate health sciences curricula prepares students for collaborative practice, becoming the opportunity to learn from each other; however, consensus must exist and the political will of the Faculties and Deanships to include the work with other professions within their study programs.
## Levels of evidence
Of the articles analyzed in this thematic, six are in a level of significance of which three are in level 2because they are studies of mixed methodology and the rest in level 3because they are qualitative studies. With respect to the level of effectiveness, eight articles are presented, of which four correspond to level 2Cbecause of their quasi-experimental designs; one article in level 3Edue to being a study without control group and three articles with level 4Bbecause of being investigations with cross-sectional design.
## Explicit gaps
Of the articles analyzed on this thematic, some gaps are derived that guide to research in professional skills within IPE, such as: -Describe the development of professional skills framed within interpersonal aspects and of critical judgment in the practice. -Identify the influence of collaborative work on the development of specific professional skills in care and community areas. -Develop educational programs to promote learning of interpersonal skills and collaborative work.
## Didactics linked to ipe (n=11)
# Contributions found
With respect to the concept of didactics, diverse results are noted on the description and development of pedagogic strategies; thus, predominating the use of standardized simulation and role play, which are evaluated positively, given that they permit acquiring skills to develop necessary skills to provide better care. Also, the use of other specific collaboration strategies with students and professionals in clinical settings, like using TeamStepps and Aspire model have permitted improving the quality and safety of care, strengthening health teams. Currently, evidence exists of new strategies, like using elearning that has had favorable results. The active methodologies used in IPE generate a positive learning experience of soft skills perceived by students in health careers. They have reported that it helps them to develop skills to clarify roles, use democratic and horizontal models in decision making, and interprofessional communication based on respect/trust, generating greater awareness as a result of the interactive and dialogical nature of the didactics and by having an academic from communications.Interprofessional education in complex scenarios of high-fidelity simulation and problem-based learning with experience practice permits students to improve not only communication, but also the trust from patients in an interprofessional team, achieving that established in the study plans.Among the basic recommendations to apply simulation, there is the suggestion of selecting adequate criteria for self-appraisal and coordination in administrative aspects; planning must include the development of teaching materials and supply of efficient technical equipment.From a real professional scenario in clinical units, continuous training of the health staff is quite common, focused on generating quality and safety environments, and using diverse didactics. Two experiences, use of TeamStepps and ASPIRE model, experienced by professionals and graduate nursing students, medicine and other health careers, has evidenced positive results after sessions of interprofessional collaboration through workshops, discussions with experts, and simulations. The participants presented improvement in team structure, communication skills, leadership, situation monitoring, and mutual support.Among the strategies most often used in IPE, there is simulation; findings of the experiences with this type of didactic have resulted favorable. Upon evaluating the perception in the socialization and interprofessional assessment (91 students from nursing, medicine, and pharmacy), statistically significant improvements were evidenced; 92% enjoyed the interaction opportunity and 81% report that it sharpens their awareness of the roles other disciplines perform in the delivery of care.Similarly, in another experience, 329 students from health professions, under the same simulation didactic, 90% was very much in agreement with the advantages and benefits of this and 60% stated that the sessions would change their professional behavior.Another experience that evaluated communication skills, exchange of information, and interaction of teamwork with 166 students from health careers, through a pre-and post-test of the interprofessional simulation, evidenced no statistically significant differences, but the participants considered the experience valuable and declared that the observations by professors and standardized patients were very useful in their professional growth.However, simulation is not foreign to some limitations and barriers in the experiences, having difficulties, like in the programming schedules, funding, and staffing.Regarding ICT, IPE learning strategies were compared on knowledge, skills, and teamwork attitudes; for mixed learning (classroom plus e-learning) versus virtual learning (e-learning). Both groups reported significant increase in teamwork skills, but not in communication and conflict resolution; nevertheless, the mixed-learning cohort reported improvement in the domains of attitude, while the virtual-learning cohort reported improved leadership.Simulation is an excellent learning strategy because it provokes a positive impact by being a strategy that generates benefits in the development of soft skills. It is important to recognize the benefits offered by this didactic, which is why it may be worth to incorporate it to curricular plans in Nursing and Medicine careers. Other didactics that support and permit overcoming barriers must be considered, like using virtual sessions of which more research is warranted.
## Levels of evidence
As per the level of evidence in this thematic, eight articles have a level of effectiveness, represented by six quasi-experimental studies classified in level 2Cand two articles with cross-sectional design placing them in level 4B.Regarding level of significance, there are three articles of which one is in level 2due to its mixed methodology and the other two in level 3for being qualitative studies.
## Explicit gaps
The gaps derived from the analysis of articles about this theme are aimed at aspects of IPE, such as:
-Implement and test different didactic strategies that have been developed and evaluated as effective. -Describe the experiences of professors and students on the use of didactics, like standardized simulation, role play, and elearning in IPE. -Evaluate the impact of implementing these didactics within formation curricula of health programs.
## Experiences and perceptions of interprofessional learning (n=9)
# Contributions found
With respect to the experiences of students from diverse health professions regarding IPE, satisfaction exists on their participation in collaborative activities, added to the fact of knowing the emotional experience and roles of each one in the health team is important to achieve practices focused on teamwork. Additionally, these instances permit identifying benefits, facilitating aspects, as well as barriers and/or factors affecting their development.
In interprofessional learning, individuals cannot simply come together in situations to learn to work together, rapport must be developed through interpersonal skills, given that said relationship has the potential to break down some of the stereotypical perceptions professionals have among themselves, making it easier and natural for them to work together.To achieve the aforementioned, recognition of emotions gains relevance; thus, interprofessional activities in diverse settings generated similar emotions in students, who considered that identifying blind spots of their own role and borrowing the coworker's lens to comprehend what they do and experience, promotes convergence into a shared vision on caring for the users.Identifying the professional role becomes a central element as a result of IPE, given that by reflecting on their collaborative learning students recognize a change in preconceived authoritarian ideas about themselves and on the recognition of the roles of other professions.However, students also reported that collaboration in IPE was seen as something secondary to principal learning.From the aforementioned, the presence of aspects that affect the IPE can be deduced. In this regard, leadership, working together, and a common setting for work groups are recognized as facilitating aspects, achieving as benefits better care for the user; development of optimal interprofessional relationships, and work satisfaction. Learning was hindered by inadequate staffing, productivity pressure, and programming challenges, considering this last element the principal negative factor. Barriers included leadership with hierarchical approaches, poor communication, and lack of knowledge of roles.In an IPE experience with 704 medical and nursing students in primary health care (PHC), perception regarding interprofessional work and interdependence evidences that 98% consider that interprofessional work is important, with women being most critical on the level of importance on teamwork (p <0.01), along with being more demanding in considering it essential in PHC (p = 0.035).In relation with the attitudes of educators with respect to an IPE experience, professors who had participated in IPE reported higher intention to participate or continue participating than professors without experience in such. The combination of the disposition perceived by administrators and attitudes toward IPE was the best predictor of the intention to participate in IPE, although no significant differences were detected among the groups of professors with respect to these attitudes.The importance of knowing these experiences and perceptions on IPE for the curricula of health programs implies incorporating these types of instances in Nursing and Medicine, given that student-centered educational programs may be favored by using IPE, where students are able to channel their emotions together, favoring their regulation among peers. Finally, today, IPE opens undoubtedly a gap that can be analyzed from the gender perspective, of how women visualize comprehensiveness and teamwork as a transcendental factor in their formation.
## Levels of evidence
The level of evidence in this thematic is predominantly significant, with seven articles, of which one is in level 2for being of mixed methodology and the other six in level 3due to being qualitative studies. Only two articles had a level of effectiveness in 4Bdue to being cross-sectional correlational descriptive studies.
## Explicit gaps
## Empirical indicators related with ipe (n=5)
# Contributions found
Empirical indicators refer to processes measuring the IPE phenomenon, hence, the measurement instruments will be described, considering that the literature reports that these have been aimed at evaluating preparation processes in interprofessional learning, interprofessional collaboration, attitudes toward IPE, and the results this can have on the practice. Beginning with the preparation Scale for interprofessional learning, Serbian authors made their process of adaptation and validation.This instrument was designed in 1999 by Parsell and Blighto evaluate results expected from collaborative learning in medical students; it is comprised of three factors with 19 items measuring teamwork and collaboration, professional identity, and professional roles. For the Serbian adaptation and validation of this scale, it was determined that the exploratory factor analysis with 19 items revealed two factors representing 51.1% of the total variance with internal reliability α = 0.90.Moreover, Iverson et al.,designed and validated an instrument to assess interprofessional learning and the results it can have on the practice. Thus, these authors, through the literature and a panel of experts identified four basic aspects to consider within the instrument: values and ethics; roles and responsibilities; interprofessional communication; and teamwork. From there, 26 items were conformed, evaluating the skill with dichotomy scale, which do not have construct validity, but in its face validity, the instrument had inter-evaluator scores ≥ 0.78 and content validity was = 0.93, considered acceptable. Retaking the thematic of ethics and professional values mentioned, a scale was designed to identify bioethical aspects shared by diverse deontological codes of the health professions. With the values identified, a Likert-type survey of attitudes was designed, which requested your giving value to your estimated ethics of each of the values in relation to your profession. Reliability was tested through Cronbach's alpha, obtaining a score of 0.905; however, the authors do not report other statistical tests to validate this instrument.Continuing with that referring to the disposition for interprofessional learning, a 27-item instrument was created from the literature to measure attitudes toward interprofessional collaboration in students and health professionals, denominated Jefferson Scale of Attitudes toward Interprofessional Collaboration (JeffSATIC). From the factor analysis performed for the construct validity, two factors emerge: working relationships and responsibility, which through Bartlett's sphericity test represents 51% of the total variance; against confidentiality, Cronbach's alpha was 0.80.Lastly, an instrument was identified to measure skills derived from interprofessional education; this was developed by authors from the United States to evaluate results related with the collaborative practice at undergraduate level. Based on a document published on the theme, they created a 42-item questionnaire applied to 481 participants for its validation. Four factors were defined: teamwork; values and ethics; interprofessional communication; and roles and responsibilities, thus, the factor analysis demonstrated that the instrument explains 79% of the variance. Each component showed a high degree of internal consistency with Cronbach's alpha ranging from 0.96 to 0.98.From these empirical indicators described, it may be concluded that greater progress is still needed to measure aspects or dimensions related with IPE, especially on the thematic of professional skills shared by health professionals, added to patients' perceptions of care from interdisciplinary teams. The instruments described are tools that must be adapted and validated to the context so they can be used in suitable manner.
## Levels of evidence
The construction and validation of scales and instruments represent the progress of knowledge toward empirical indicators that permit measuring the IPE phenomenon in concrete and tangible manner; however, due to the psychometric approach, where validity and reliability results are obtained, no evidence is produced as such that should be classified in the JBI levels.
## Explicit gaps
For this theme, the research gap is more methodological tan thematic, which is why it is necessary to conduct psychometric research that construct and validate instruments that measure skills in integral and scale manner regarding skills, knowledge, attitudes, and organizational environment to measure the impact of IPE on health professionals.
Finally, knowing the role of another health professional facilitates performance in relation with the limits of action and that influences undoubtedly on the leadership future nursing or medicine professionals will assume in a given situation, where recognition of what others do favors communication; thus, IPE experiences are an opportunity for students to learn. It is suggested to include this methodology in the Nursing and Medical curricula if a true effect is expected on the change of stereotypes from the other professions.
Among the limitations of this review, we have that due to this being an exploratory review, it was not possible to fully guarantee that it was systematic and was not focused on a methodology of specific studies, added to the selection bias, which excluded articles beyond the range of time established for the search and documents, like theses or degree works that could have contributed more information to the IPE thematic.
It can be concluded that the state of the art related with the interprofessional theme on the formation of nursing and medical students globally proved varied, although most of the productions were of descriptive scope. Four thematic groups are highlighted therein: Experiences and perceptions del interprofessional learning, Didactics linked to IPE, Empirical indicators related with IPE, and Development of professional skills, with the thematic of didactics having the greatest progress, with presence of experimental studies that demonstrated the effectiveness of simulation activities and e-learning to develop knowledge and skills that lead to interprofessional learning, with these coming mostly from North America and Europe, where the highest scientific production has been reached in this theme.
Research gaps primarily concentrate in demonstrating the importance of early integration of IPE in the curricular syllabus of the Nursing and Medicine careers; evaluating the different types of IPE didactics within the formation curricula of health programs and identifying the influence of collaborative work on the development of professional skills, especially interpersonal skills. It is relevant to know the perception of patients on the impact of IPE on their care and identify the influence of collaborative work on the development of professional skills, as well as on the perception of the faculty body involved in teaching interpersonal skills through interprofessional education.
Information on aids and funding. Research project "Promotion of interpersonal skills in students from Nursing and Medicine through the use of interprofessional simulation" funded by the Vice-Rectory of Research and Development through the Direction of Scientific and Technological Research DICYT No. 031902MI -Universidad de Santiago de Chile. |
Modifications to residential neighbourhood characteristics and risk of 79 common health conditions: a prospective cohort study
## Description of study population
Men and women participating in HeSSup were from a stratified random sample of the Finnish population based on four age groups (20-24, 30-34, 40-44, and 50-54). The eligible population (N=64,797) was identified from the Finnish population register and an invitation to participate was posted for them, along with a baseline questionnaire. 1 Between June 7, 1998 and May 23, 1999 and January 7 and August 12, 2003, a total of 23,655 responded had data on residential neighbourhoods and were successfully linked to electronic health records of national registers until December 31, 2012. The Turku University Central Hospital Ethics Committee approved the study.
The FPS sample comprised the entire public sector personnel of 10 cities and 21 hospitals in the same geographical areas. The sample included in the present analysis comprised 91,131 men and women aged 17 to 77 who responded to the survey, had data on residential neighbourhoods, and were successfully linked to electronic health records from national registries up to December 31, 2018. Corresponding to the sex distribution of Finnish public sector employees, almost 80% of the participants were women. eFigure 2 shows timeline for data collection for study of neighbourhood characteristics, morbidity and mortality and that of neighbourhood characteristics and lifestyle changes by type of data in both cohort studies. The 5-year time lag is commonly used in prognostic studies and it is also the frequency for health checks in the general population that aim to prevent disease and premature mortality.
## Further details on the assessment of green space
Green space was defined as any open land surface that was partly or completely covered with grass, trees, shrubs, or other vegetation, and included, for instance, parks, forests and community gardens. 4 A national wide normalized difference vegetation index (NDVI) map was generated, as the residential places used in the analyses were located in different part of Finland. NDVI dataset was computed using 30mx30m resolution Landsat 5 Thematic Mapper images available in Google Earth Engine.We selected images with maximum 30 percent of cloud cover between June and August (the maximum vegetation greenness period of the year in Finland). To minimise NDVI value variation due to atmospheric conditions, we used atmospherically corrected surface reflectance images. Water bodies and clouds were masked out in each images using the Quality Assessment band of Landsat 5. NDVI was estimate using the visible Red and Near Infrared (NIR) bands as NIR-Red/NIR+Red.As the water bodies has been masked out, the NDVI values ranged from 0 to 1, where 0 denotes the lowest vegetation greenness or non-vegetation cover and 1 the highest cover. As a single year collection of Landsat 5 TM images with a maximum of 30 percent of cloud cover during the maximum vegetation greenness period did not cover the whole study region, we used images of consequent three years period to generate a cloud-free composite maps of NDVI median values for the whole Finland in the study period. We assumed that urban green areas, mostly forest and grass, do not exhibit substantial inter-annual NDVI changes in short period, unless the vegetation cover has changed into other cover type, e.g. built-up area. The created NDVI maps were exported from Google Earth Engine as geotiff files. Statistical metrics of NDVI maps in 250x250m grid, used to assess the degree of residential greenspace, were estimated using R software in the Finnish CSC's high-performance supercomputer.
We validated NDVI data by comparisons with green space in high-resolution Aerial Photograph data in major Finnish cities. Indeed, NDVI is a widely applied technique for vegetation mapping 6-8 in epidemiological studies. 9,10
## Further details about lifestyle risk factors
We repeatedly assessed the following lifestyle factors at years 1 and 4 or 5 using identical standard survey instruments in both cohorts:
## Further details about morbidity assessment
Participants were linked by their unique identification number to national registries of hospital discharge information (recorded by the National Institute for Health and Welfare) and mortality (recorded by Statistics Finland). These electronic health records include cause and date of hospitalisation or mortality and their coverage -all hospital types, including private hospitals, and records cover emergencies -reflects the comprehensive nature of Finland's public health-care system. Additional information on site-specific cancers, diabetes, cardiovascular disease (including hypertension), psychotic disorders, dementia, Parkinson's disease, multiple sclerosis, epilepsy, asthma, chronic obstructive bronchitis, inflammatory bowel disease, liver disease, rheumatoid arthritis, gout, and renal failure was available via record linkage to the Drug Reimbursement Register of the Social Insurance Institution of Finland. The diagnosis for incident disease was coded according to the World Health Organization's International Classification of Diseases Tenth Revision (ICD-10). 14 As previously,we focused on fifteen ICD-10 disease chapters that concern infectious and parasitic diseases (A00-B99), neoplasms (C00-D48), diseases of the blood (D50-D89), endocrine, nutritional, and metabolic diseases (E00-E90), mental and behavioural disorders (F00-F99), diseases of the nervous system (G00-G99), the eye (H00-H59), the ear (H60-H95), the circulatory system (I00-I99), the respiratory system (J00-J99), the digestive system (K00-K93), the skin (L00-L99), the musculoskeletal system (M00-M99), and the genitourinary system (N00-N99), injuries and poisoning (S00-T98), and external causes (V01-Y98). 14 Participants with a history of a chronic disease were excluded as shown in eTable 1. W00-W19 Self-harm X60-X84 *Participants with a history of the disease at baseline were excluded from the analysis of neighbourhood characteristics and new-onset disease.
## Etable 1. included icd-10 chapters and codes and exclusions for baseline cases
## Further details about correction for multiple testing
Adjustment for multiple testing in outcome-wide studies is used to avoid false negative findings and several techniques are available for this, including adjustment for false discovery rate (FDR) of 5% and the more conservative Bonferroni correction.However, a policy of not correcting for multiple testing has also been recommended because reducing the type I error (i.e. false positives or "chance findings") increases the type II error (false negatives) for associations that are not null.A further suggestion is to abandon statistical significance (p-value) and instead focus on the point estimate and its confidence intervals considering all the values between the interval's limits reasonably compatible with the data.In this study, we corrected all step 1 analyses using Bonferroni approach for 79 tests -the number of health outcomes tested in relation to each neighbourhood characteristic. The threshold for statistical significance was 0.0006 yielding significant associations with 12 health outcomes for neighbourhood education, 16 for neighbourhood income, 20 for neighbourhood unemployment, and 9 for neighbourhood green space. The corresponding numbers were 30, 36, 40 and 29 without Bonferroni correction (p < 0.05) and 12, 15, 17 and 6 with a more stringent Bonferroni correction for 4 x 79 tests (threshold for statistical significance 0.00016). For detailed results, see eTable 6. All main analyses in steps 2-4 were based on a conventional alpha level (p < 0.05) although, in sensitivity analyses of smaller sample sizes, deviation from main results were interpreted by focusing on point estimates and 95% confidence intervals.
In an effort to be as transparent as possible, by providing point estimates, their 95% confidence intervals and precise p-values (4 decimal places) in appendices, we leave opportunity for interested and statistically-minded readers to apply a method of multiple test adjustment of their choice. In appendix 2, we also report all associations of the four neighbourhood characteristics and their changes with 79 health outcomes based on continuous exposure measures to avoid any arbitrary thresholds for categorisation.
## Summary of supplementary results
Steps of analysis, number of tests conducted, and number of statistically significant findings are summarised in eTable 2.
Proportional hazards assumption: eTable 3 shows results from tests of the proportionality assumption. Of the 79 associations between neighbourhood education and health outcomes, 74 met the assumption whereas a statistically significant (p<0.05) interaction between exposure and length of follow-up was observed for 5 health outcomes. The corresponding figures were 73 and 6 for neighbourhood income, 72 and 7 for neighbourhood unemployment, and 74 and 5 for neighbourhood green space. As the log-log plots in eFigure 4 indicate, deviation from the proportional hazards assumption was most cases caused by the weakening of the association over time Characteristics of the study population: Baseline characteristics of the participants for analyses of health outcomes (Sample A) and lifestyle change (Sample B) are shown in eTable 4. In eTable 5 we show comparison of the two participating cohort studies, the population-based HeSSup and the FPS occupational cohort. Supplementary results by step of analysis (table 1) are as follows:
The first analysis step: eTable 6 shows hazard ratios for the associations between each of the four dichotomised (advantaged vs disadvantaged) neighbourhood characteristics and each of the 79 health outcomes. Hazard ratios are adjusted for age, sex, education and cohort. Tests of interactions between neighbourhood characteristics and cohort on health outcomes suggest that only two associations differed between cohorts after adjusting for multiple testing: The associations of neighbourhood income with mental disorders (p interaction 0.0002) and substance abuse (p interaction 0.0001) were stronger in the populationbased HeSSup study than in the FPS occupational cohort (eTable 7).
The second analysis step: In eTable 8, we report hazard ratios for the associations of change in each neighbourhood characteristic and those 30 health outcomes that were associated with neighbourhood characteristics after Bonferroni correction. These analyses were stratified by baseline level of the neighbourhood characteristic and hazard ratios are adjusted for age, sex, education and cohort. eTable 9 shows a comparison of prevalent and new-onset disease rates between participants who moved and did not move residential address during the observation period. Of all 79 health outcomes before follow-up, the prevalence was higher among movers than non-movers for 23 and lower in four of these endpoints. The corresponding numbers for the rate of new-onset disease during the follow-up were 29 and three. Results from stratified analyses for non-movers and movers are provided in eTables 10 and 11. The following 14 health outcomes were associated with both favourable and unfavourable changes in neighbourhood characteristics either in the total cohort or those with stable residential address or both: death, endocrine diseases, diabetes, mental and behavioural disorders, disorders due to substance abuse, psychotic disorders, diseases of the nervous system, stroke, chronic obstructive bronchitis, diseases of the digestive system, digestive and abdominal symptoms, diseases of the skin, osteoarthritis and soft tissue disorders. Of these, robust associations with substance abuse disorders were seen in movers only.
The third step of analysis: eTable 12 shows baseline characteristics of the study population in analysis of 8 non-randomised neighbourhood modifications in participants who did not move. While differences in age and sex between the groups were small, those exposed to favourable neighbourhood modifications or being exclusively based in advantaged neighbourhoods tend to have higher education than those exposed to unfavourable neighbourhood modification or remaining in disadvantaged neighbourhoods. eFigure 5 shows that the main results on four health outcomes (diabetes, stroke, diseases of the skin and osteoarthritis), adjusted for age, sex, education and cohort, remained after excluding those who did not have stable employment during the observation period. Results for men and women, participants with more than one lifestyle risk factors and those living in urban areas are provided in eTables 13 and 14. The results show that the associations were directionally consistent with the main findings although statistical significance was not always reached perhaps due to smaller sample size (for analyses of subgroups with no risk factors or rural residential address, statistical power was diminished). The results were also robust to multivariable adjustments (eTable 15). eTableshows results from a sensitivity analysis using 750m x 750m instead of 250m x 250m spatial units to define neighbourhood characteristics. Averaging over larger areas means lower data resolution; that is, for areas that have heterogenous characteristics, an average over a wider area will result in a failure to capture with precision the local conditions that might influence health. Associations with health are therefore expected to be weaker for the larger areas. Confirming this, we found an attenuation in the magnitude of the effect estimates when using a cruder 750m x 750m spatial unit.
The fourth step of analysis: eTable 17 shows that the main findings on parallel changes in neighbourhood characteristics and lifestyle were little changed in analyses of participants with stable employment. This was also the case when analyses were restricted to participants who lived in urban areas (eTable 18).
The fifth step of analysis: These results are available as an excel file in appendix 2.
## Supplementary analyses on skin diseases:
In detailed analysis of diagnoses, eTable 19 shows that the association between change in neighbourhood unemployment and risk of skin diseases may be driven by health conditions related to hygiene. eTable 2. Steps of analysis, number of tests conducted, and statistically significant findings
## Efigure 5. association of change in neighbourhood characteristics with health outcomes among participants with no change in residential address or employment
Hazard ratios are adjusted for age, sex, education and cohort. eTable 13. Association of change in neighbourhood characteristics with subsequent health outcomes in non-movers by sex eTable 14. Association of change in neighbourhood characteristics with subsequent health outcomes in non-movers who had more than one lifestyle risk factors and lived in urban areas eTable 15. Association of change in neighbourhood characteristics with subsequent health outcomes in non-movers after multivariable adjustments. |
Early ultrasound diagnosis of conjoined twins at eight weeks of pregnancy: A case report
BACKGROUNDConjoined twins are a rare occurrence, and the majority of these malformations are detected during second trimester screening.CASE SUMMARYHerein we report a case of conjoined twins, which was diagnosed by ultrasound at 8 wk gestation and was normal at 7 wk gestation. The two fetuses shared one heart and were diagnosed as thoracopagus twins. This is the first report of conjoined twins diagnosed at 8 wk gestation. The pregnancy was terminated electively at 9 wk gestation. Because some congenital malformations can be diagnosed earlier, a prenatal ultrasound examination at an early gestational stage cannot be dismissed.CONCLUSIONThis case demonstrates that a 7-8 wk gestation might be the earliest period when conjoined twins can be diagnosed by ultrasound.
## Open-access: this article is an
open-access article that was selected by an in-house editor and fully peer-reviewed by external reviewers. It is distributed in accordance with the Creative Commons Attribution NonCommercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See
# Introduction
Conjoined twins are extremely rare fetal abnormalities with a reported prevalence of 1:50000-1:100000 . Conjoined twins result from postzygotic splitting 13 d after fertilization. Approximately 70% of conjoined twins die within 24-48 h after birth or have a fatal congenital disease; thus, early diagnosis and treatment are desirable. In a previous study, conjoined twins were diagnosed by ultrasound between 18 and 22 wk gestation. More recent research has reported the diagnosis of fetal defects in twin pregnancies at 11-13 wk gestation . Termination of pregnancy as early as possible continues to be the best treatment. Herein we report a case of conjoined twins suspected of cardiac fusion diagnosed by prenatal ultrasound at 8 wk gestation and confirmed by surgery.
## Case presentation
## Chief complaints
At 8 wk gestation, a pregnant woman presented with low progesterone level.
## History of present illness
A 30-year-old woman had a normal prenatal ultrasound examination at 7 wk gestation. The patient did not have any other discomfort.
## History of past illness
The patient is G4P1. She delivered her first child vaginally.
## Personal and family history
The patient has no special family history.
## Physical examination
The patient's physical examination was normal.
## Laboratory examinations
Laboratory testing had revealed a low progesterone level.
## Imaging examinations
The first prenatal transvaginal ultrasound was performed by CD-6 (Mindray, Shenzhen, China) at 7 wk gestation. Two-dimensional ultrasound images showed that the embryo was normal with crown-rump length of 12 mm.
After 7 d, the prenatal ultrasound was performed again using a Volusion E8 (GE Healthcare, Milwaukee, WI, United States). Two-dimensional ultrasound images showed a single yolk and two embryos in the gestational sac. The lengths of the two embryos were 18 mm and 20 mm. The embryos were connected without distinct boundaries, and there was no fluid between them. The two embryos shared one heart. Color Doppler flow imaging showed that pulsating blood flow signals existed at the junction of the two embryos. November 6, 2020 Volume 8 Issue 21
## Final diagnosis
Conjoined twins (thoracopagus twins).
## Treatment
Terminated the pregnancy through abortion.
## Outcome and follow-up
Discharge after a period of observation. November 6, 2020 Volume 8 Issue 21
# Discussion
Conjoined twins are very rare monozygotic variants, most of which are monochorionic monoamniotic twins . Although the basis of conjoined twins has not been established, there are two main theories. The first theory is that incomplete division of an early embryo produces identical twins with the same anatomic structures. The other theory is that conjoined twins are the result of two separate embryos reuniting again. Conjoined twins are classified based on the conjoined body parts. The most common type of conjoined twins is an abdominal union, which accounts for 87% with a mortality rate of 51%, while dorsal unions occur in up to 13% of conjoined twins. Cardiac fusion usually occurs in thoracopagus twins, but rarely occurs in ischiopagus, craniopagus, or pygopagus twins. Along with the improvements in surgical technique, survival rates for separation procedures have increased. Although cardiac fusion is relatively common in conjoined twins, the success rate of surgical separation and diagnostic accuracy for this type is extremely low.
In most cases of conjoined twins, a severely deformed connection is fatal for the surviving fetus. However, severe malformations are always caused by the untimely diagnosis, which delays optimal surgical treatment. Chen et alstudied the outcomes of postnatal surgical separation of conjoined twins diagnosed before 15 wk gestation. The survival rate of emergency surgery (10%-29%) was reported to be much lower than nonemergent, planned surgery (81%-83%). Recently, Heuer et alsuccessfully separated total-fusion craniopagus twins at 10 mo of age. Dunawayconcluded that there are many unknowns regarding the research of separating craniopagus twins in stages. Moreover, in couples who opt to continue the pregnancy, a multidisciplinary team for prenatal and postnatal management is necessary because multiple malformations are often found .
It was also demonstrated that early diagnosis of conjoined twins is necessary. Early diagnosis of conjoined twins has the following advantages. First, through the early evaluation of the degree of fusion, the optimal treatment time and strategy can be selected. In addition, early diagnosis will minimize maternal and fetal mortality. Early evaluation will also help parents to choose early termination, thereby reducing the risk of trauma during vaginal delivery.
# Conclusion
In conclusion, the incidence of conjoined twins is low, but the rate of mortality and malformations (such as severe congenital lung disease) is high. Termination of pregnancy may be a wise choice, especially when the vital organs of the body are connected. This study suggests that active and regular antenatal ultrasonic examinations provide a great opportunity for good prognosis. Interdisciplinary integration also plays an important role in the diagnosis and treatment of conjoined twins . We suppose that an ultrasound examination at 7-8 wk of pregnancy plays an important role in the early diagnosis of congenital malformations in twins. Therefore, apart from the heartbeat and calculation of gestational age, sonographers should also notice the fetal structure in early examinations. |
Dr. Edward Jenner on the Influence of Artificial Eruptions in Various Diseases
## 1822]
JDtr. Jenner on Artificial Eruptions. 91 symptoms of which are too well known. In addition to the usual phenomena, however, " there was a perceptible enlargement about the centre of the left side of the thorax." To engage his mind, and without any hopes of ultimate recovery, the antimonial ointment was rubbed on the protuberant part until pustules were produced, which was effected in a few days. In a week there was some amendment in regard to the cough and other distressing symptoms. The application was continued with the view of keeping the pustules in full activity. At the expiration of a fortnight the general bad symptoms had considerably abated, and the patient's looks were much improved. From this time the convalescence was so rapid that, in six weeks more, no vestige of disease remained, and the youth began to renew his ordinary avocation, that of a stone-cutter. Dr. Jenner very properly wishes that it may not be inferred from this case that lie supposes genuine confirmed phthisis is to be cured by artificial eruptions. The fourth case was a woman, 54 years of age, subject to spasmodic asthma. She used the ointment on the nape of the neck, after which the complaint was more slight, and the intervals longer.
Case V. A seaman, 47 years of age, was suddenly seized with inflammation of the right eye, but not so violent as to prevent him pursuing his avocations for nearly three weeks, when a chill came on daily between three and four o'clock. " About half an hour after each attack, a pain seized the right side of the head, principally about the orbit of the eye, extending in the course of the temporal muscle. It continued to return every afternoon at a certain hour, and at length became 80 violent as to deprive him of sight and intellect. In one of the paroxysms he grew enraged with his wife, because he supposed that she had not lit him a candle, although one was burning before him. These periodical attacks became marked in the end with raving madness. In a paroxysm, with extreme severity of pain, he was at the point of destroying one of his children. He was bled from the arm, leeched, and took purgatives up to drastics, but they took no effect upon the malady. Seeing the impression which tartar emetic had made on affections connected with a disordered state of the brain and nerves, I did not hesitate to direct the application of the ointment, and it was applied to the left arm. Pimples followed in twenty-four hour's, end as soon as they became acuminated, and contained a little limpid fluid, the patient found ease; the pain continued to abate, and at the end of three days it was quite gone. He continues well. This man, like many others who ply as mariners on our river, (the Severn,) was a hard drinker." P. 9.
[June Case VI. The sixth case was a boy, 12 years of age, who had been ill six months, with what is called chronic hepatitis. His liver felt hard, enlarged, and very sensible to the touch. The cachectic appearance and general emaciation were such as indicated the probability of a fatal termination.
At this period the ointment was applied, with the usual cutaneous effect, " after which he recovered with astonishing rapidity, and no vestige of the induration or enlargement within the abdomen remains." Mild aperients were used during his indisposition. Cases VII. VIII. IX. The seventh and eighth cases arc equivocal, or unsatisfactory. The ninth case was that of a young woman with mania, the patient of Mr. Fewster of Thornbury. Her ravings were violent. Leeches were applied about the head, and the ointment upon the leech-bites. "As soon as vesicles appeared, she was well." Having neglected to continue the ointment, she experienced a relapse, and became decidedly maniacal. The ointment was reapplied, and she soon recovered.
CaseX. (A patient of Mr. Fry of Dursley,) Mrs. B. a;tat 23, was seized with puerperal mania the second day after parturition, and became totally unmanageable, refusing food and medicines. The antimonial ointment was rubbed along the inner surface of the forearm, from the joint to the wrist; but a fortnight elapsed before a vesiculated eruption was brought out. " As soon as the eruptions appeared, amelioration of her symptoms was evident." She was kept under the influence of the medicine nearly three weeks, during which time she progressively improved. During the external applications she took occasional purgatives, and a solution of tartarized antimony internally in nauseating doses.
Considering that this was puerperal mania, and that the other means were pretty efficient, we can hardly agree with our worthy author, that u her restoration may be chiefly imputed to the ointment."
Case XI. The eleventh case was insanity unconnected with pregnancy. The disease was removed by the ointment; but the patient relapsed, and when the work went to press, her friends were endeavouring to get her into Saint Luke's. Case XVI. "We shall pass on to the 16th case, the intervening ones not being particularly interesting.
Mr. F. S. aitat. 37, a hard drinker, caught cold in an open carriage, which was succcedcd by hffimoptoc. IIis rcspira-1822] ration, at the time of report, was very quick and difficult, with cough, and expectoration of viscid phlegm. Has been bled repeatedly during his illness, now ot thirteen weeks duration, and with some relief. Last of all lie used the ointment, which excited a very irritative crop of pustules on the chest. When they were fully formed, and had discharged, the patient found great relief of respiration, but not so decidedly of the cough. In another month the patient was farther improved, but the account breaks off, when he was very far from well. Case XVII. This was a man, 40 years of age, apparently phthisical for 25 years past, and subject to attacks of hacmoptoe. " After being severely indisposed with affections of the chest, viz. cough, and impeded respiration; he was seized, last summer, with a dangerous recurrence of the haimorrliage. It was concluded that he had pulmonary consumption in the last stage, and the case afforded every indication of terminating fatally, but the hacmoptoe, which became more and more violent, was at last arrested by superacetate of lead, and the tart. emet. ointment was applied to the chest. As soon as the eruptions vesiculated, he got better; when they died away, he began again to feel uncomfortable about the chest, complaining that he felt " plugged up in breathing." He finds immediate ease by a renewal of the eruptions, and has, in consequence, continued under their influence for nearly twelve months past. His skin is so irritable, that pimples almost immediately follow the application, though in some individuals three days will elapse before they will be excited. The ointment gives him most relief when applied to the opposite side of the chest to that most affected.'' 23.
Case XVIIT. We marvel much that our able and intelligent author did not place the following cure to its real cause ?the removal of obstruction from the hepatic duct, rather than the application of antimonial ointment. We shall give the case entire, that our readers may judge for themselves. >? " John Gay, 39 years of age,?Was taken ill about four months ago, with feelings of languor and nervous debility, accompanied with dyspepsia, bilious and acid eructations: he had also pain in the right hypochondrium, dry sore throat, and general feverishness. With these symptoms he had pyrosis. His complaint continued to grow worse, especially a dull pain which had been goina; forwards in the region ot the liver, till he was seized with symptoms of complete obstruction of the common duct of the gall-bladder. His skin became tinted with a deep blackish yellow colour. The food and medicines which he took, for some time after the symptoms of of obstruction, regurgitated in an unaltered state, from the stomach. 9-1 Analytical Reviews.
[June He found great difficulty in procuring medicines, from different medical men, that would act on his bowels. Tn this emergency, scanty evacuations of slime were procured by the administration of clysters ; but he passed no solid stercoraceous stools ; pills of soap and rhubarb, combined with an aromatic, and also the ointment, were now prescribed for him. Pimples appeared within twentyfour hours. These suppurated, and discharged pus with unusual freedom, and disposition to continue to discharge. About the time at which the pustules appeared, a sudden burst of evacuations took place, consisting first of bilious coloured fluid, next of slime of a green hue, and an abundance of shreds of a skinny appearance. In his first stools, at this time, my patient observed a mass, which he conceived to be food and medicines which he had taken previously, and had remained unaltered in tha alimentary canal. The pain in his right shoulder and right hypochondrium abated rapidly, and his stools came away more solid, but still enveloped in slime. He is now convalescent, but tender under the margins of the right ribs, and possessing a mitigated degree of unhealthy action about the liver. As his health improved, the eruptions evinced a disposition to dry up, but they have been continued. He had, four years ago, a constitutional sore, which has occasionally healed and re-appeared: its final suppression may have had some connexion with his present complaint. His recovery went on and was perfect in far less time than I could reasonably expect, considering the extreme state of debility to which he was brought by his long and severe sufferings. Within six weeks he resumed his laborious occupation, which was that of a sawyer." 24.
"VYe were lately in attendance on a remarkable case of jaundice from simple obstruction of the ducts, some particulars of which may not be uninteresting. A young medical gentleman in the city liad been attending a bad labour?at first protracted, then with retention of the placenta, and afterward, we believe, uterine haemorrhage. The fatigue or anxiety?or both combined, produced jaundice in about 48 hours after the accouchement. Dr. B. became quite yellow, but continued to go about his professional avocations, complaining of lassitude and all the usual symptoms of jaundice, with very yellow urine and white stools. The jaundice continuing, and the tint deepening for ten or twelve days, with loss of strength and appetite, the patient and friends got alarmed, and we were requested to see the gentleman, with Dr. Farre, and occasionally Dr. Babington. There was at this time no pain in the side or at the pit of the stomach; but the skin was intensely yellow, the stomach irritable, the stools white, the mind desponding, and the physical powers all depressed.
The usual remedies were tried, as the warm-bath?mercurials, aperients, local bleeding from the region of the liver, &c. &c. but all without the slightest advantage. In short, the com-
## 1822]
Dr. Jenner on Artificial Eruptions. 05 plaint went on nearly seven weeks, the patient getting weaker and more emaciated every day, with great gastric irritability. Some of the attendant physicians considered the case as very dangerous, the obstruction being looked upon as of some fixed or organic nature, which would terminate by effusion in some of the cavities. About the middle of the seventh week, however, when great despondency prevailed on all sides, Dr. B. complained of pain, for th< first time, at the pit of the stomach, which gradually but rapidly increased, till it ended in the tortures of gall-stones. In about 24 or SO hours from the commencement of the pain, vomiting, and spasms, a burst of bilious fasces came awav, with cessation of pain, and, of course, an end of the jaundice. This took place in the night, and the servant threw away the first two or three motions, so that the biliary concretions were not detected, but no doubt could exist as to their having been past by stool.
The above case is highly important in many respects. In the first place it shows how quickly mental anxiety will produce jaundice by obstructing the biliary ducts, probably from spasms of their mouths, whereby the bile stagnates in the passages, and becomes inspissated there. In the second place it teaches that complete obstruction to the natural course of the bile?in other words, simple jaundice, may continue many weeks without being fatal. We have seen but one other case where jaundice from gall-stones continued so long, and ultimately did well. In the case here alluded to, which was also that of a medical man, the obstruction remained thirteen weeks, and then terminated by painful expulsion of gallstones. In general, however, simple jaundice lasting longer than two or three weeks, becomes a dangerous disease, and requiring a very guarded prognosis. In old people it is still more dangerous than in young. In organic diseases of the liver, as where a tubercle presses on and obstructs the ductus communis, jaundice will last many months without proving fatal. We are at this time (April 1822) attending with Mr. Brien, an intelligent surgeon of Spencer Street, a man who has been intensely jaundiced for more than ten months. He has organic disease of the liver, probably tubercles and is greatly emaciated, but not yet dropsical, To return to the case of our young medical friend. It appears either that the obstructing cause lay dormant in the duct for so many weeks, like a foetus in utero, until its size or other circumstances induced expulsive efforts in the duct; or else the cause was gradually and slowly advancing to the extremity of the tube, and did not cause pain till it came to its mouth, which is known to be much more sensible than Analytical llevieivs.
[June any oilier portion of the canal. In cither case, we apprehend that Ave have very little power ovnr the mechanical obstruction. In the case under consideration, every mean was tried in vain?including impregnation of the system with mercury?emetics?electricity?warm baths?soap and alkalies ?purgatives?frictions of mercurial ointment, &c. &c. &c.
After we had ceased to administer remedies, the expulsive nisus in the duct came on spontaneously, and the concretion was dislodged. But we may here mention that we have seen some instances of jaundice from mental affections, where the cause was very quickly removed, (at a very early period of the complaint,) by administering live or six grains of calomel and two grains of opium over-night, and next morning giving the patient a brisk emetic. We pursued this plan on the supposition that spasmodic constriction of the biliary duct is the first link in the production of biliary concretions, and consequently jaundice; and that the opium and calomel tended to relax this constriction and increase the vis a tergo, or secretion of bile. The mechanical action of vomiting was employed for very obvious reasons. We have to apologise for this digression ; but hope it is not entirely devoid of interest.
We have now presented our readers with an analysis of those cases which Dr. Jenner has laid before the profession.
The trials, as he acknowledges, which have been made of exciting vesiculo-pustular eruptions, areas yet limited, u but he trusts the general favourable results of the experiments made, will apologize for his hazarding a few. physiological hints for the consideration of those who may til ink a wider basis to work upon desirable." These hints are chiefly speculative, and founded on his "favourite pedestal analogy."
It appears that about 40 years ago two papers were published on the subject of the external application of tartrite of antimony?the one by Mr. Gaitskill, an able and experienced surgeon at Rotherhithe;?the other by Dr. Bradley. The latter gentleman relates some interesting observations respecting the use of the remedy in rheumatic affections. In every instance it appeared to Dr. Bradley to be a remedy of great cfiicacy; but the disagreeable symptoms produced by it caused many to desert its use. In recent cases, he observes, the first or second application often removed the complaint;
but where the cases were of long standing, it was necessary to persevere in the frictions for three or four weeks. In several cases the eruption appeared on distant parts of the body.
Dr. Robinson has lately written (in the Medical Repository, January 1821) on the good cflects of the antimonial 1822]
Dr. Joiner on Artificial Eruptions.
disorder, and may we not from thence infer what appears to me to be a pretty general law of Nature, that she often gets rid of diseased action affecting vital organs, by exciting eruptions in other parts not vital?" 31.
Among the illustrations of this doctrine brought forward by Dr. Jenner, he mentions a remark made to him by his able and excellent friend, John Fleming, Esq. now a member of the British Senate, and formerly at the head of the medical department at Calcutta?namely, that when attacked with intermittents, and when he took the bark frequently, he seldom could get well until herpetic eruptions appeared on liis lips. That illustrious observer, Dr. Ferriar, in his medical histories and reflections, has occasionally alluded, in the most pointed terms, to the subject now before us.# " Cutaneous eruptions (he observes) oiten extinguish dangerous diseases." Madness and melancholy, epilepsy, delirium protracted after fever, dyspepsia, and various pulmonary affections, arc all observed to be mitigated on the appearance of cutaneous disorders especially on the return of those that, after becoming habitual, had been suddenly suppressed.
Iluxham has also related some general facts that bear on this subject?and it is particularly worthy of notice that lie describes the sympathy between the skin and lun?-s which elucidates the favourable effect of artificial eruptions in some pulmonary disorders.
Here Dr. Jenner regrets, and with reason, that Dr. Parry's * It is a doctrine of very old standing indeed; but in the humoral pathology, physicians considered what we now look upon as metastases of diseased action, to be translations of matter, and eruptions, abscesses &c. were believed to be so many drains for carrying the peccant humour out of the constitution.?Rev.
Vol. 111. No. 9. O [June " Elements of Pathology" is not sufficiently studied and appreciated by the profession. We hope \vc shall prove instrumental, in the present number of onr journal, in obviating this unwarrantable apathy of our brethren to such a valuable work. At all events, we have sent forth an abstract of Dr. Parry's volume, that will be read and reflected on, where the original coukl never hope to travel?perhaps where it has travelled hitherto in vain.
In order to display, in a still stronger point of view, the analogy between the phenomena of artificial and spontaneous eruptions, Dr. Jenner draws an interesting sketch of the rise, progress, and termination of what is called an exanthematous fever?the small-pox, for example. " Moibid animal matter, generated by this disease, is applied to the body either by what is termed the natural or artificial mode. After a given space of time, in either case, diseased action is manifested by that constitutional derangement which is designated fever. This goes on for a limited period, when eruptions appear on the skin, which soon shew on their apices vesiculated specks. Here the disease, as far as it depended on the primary adion of the infectious matter which called it into existence, terminates. But now a new train of symptoms comes on, consequent to the diseased action excited on the skin by the pustules, the influence of which is felt in proportion to their numbers, their malignancy, the disposition of the constitution, and the extent to which they penetrate the skin. The fever, in the first and second instances, has two distinct origins. In the first instance, it arises from the influence of the morbid matter inhaled, or intentionally applied ; in the second, from diseased action going forward on the skin, and, in many instances, also on the mucous membranes of the fauces, trachea, and ramifications of the bronchi. The rapidity with which, in some instances, the secondary diseased action follows the primary, often obscures the distinction. Of this the ordinary phenomena of confluent small-pox and scarlatina exhibit familiar instances. In the first of these the skin is often'so quickly and universally assailed, that there is, in most instances, no interval of cessation. Nature is in a hurry to call out her guards." 40.
Here our esteemed author introduces some practical remarks on the benefits which may be derived from sedative applications, where the pustules are formed so thick upon the cutis as to augment, in a high degree, the secondary fever. An experiment was made by Mr. Fry, in the case of a young woman, who had a full burthen of distinct smallpox, and whose countenance was loaded with pustules. " In this state one cheek was sopped with liq. lythargyri somewhat more diluted than I intended, while the other was suffered to take its couvse for the sake of comparison. The consequence was, 1822] that although, from excessive occupation, this process was not repeated by Mr, F. the effect was nevertheless very manifest, for the pustules were so much checked in their progress to maturation, that they could be scarcely said to have maturated at all." 40.
As Dr. Jenner rests this proposition of checking the exuberance of eruptions by sedative applications, on a single case, and that not under his own eye, we arc justified in considering it as a mere speculation ; and we would hazard another speculation, namely, that the practice would prove a dangerous one. We grant, indeed, that cool air is always advantageous in eruptive diseases, as moderating the eruptive fever, and refreshing the patient; but we would be very loth to apply sedative and repellent applications to the eruptions themselves, for a similar purpose. We arc sure Dr. Jenner will excuse the freedom of this remark, the propriety of which, we think, is not a little strengthened by the following observation of Dr. Jenner himself.
" Even in small-pox, though the disease itself cannot possibly disappear wholly, the eruptions, when in a vivid stato of maturation, may so lose their prominent appearance, as on a sudden to become flattened and excite distress in the constitution, which is often followed by fatal consequences." 44.
Dr. Jenner remarks that, when eruptions appear early, and without the proper vesicular character, the prognosis is unfavourable. But we must pass over several pages occupied with hints, suggestions, and speculations respecting several diseases, as our limits are already exceeded.
The formula of antimonial ointment used by Dr. Jenner is the following. Ant. tart. 5ij. ungt. cetacei 5ix. sacchari albi 5}. hydrarg. sulphur, rub. v gr. m. ft. unguentum. The tartrite of antimony should be finely levigated. The sugar preserves the ointment from rancidity. The ointment sometimes brings out crops of pustules in one day?sometimes it requires several. " In the case of a lady, where two parts of tartar emetic and one of simple cerate were used, eruptions appeared in a few hours. A communication from Mr. J. Fosbrokc is inserted, in which several interesting particulars are stated relative to the action of the antimonial ointment in some cases of hysteria. A case is also introduced from the Rev. G. G. Jenner, ill titrating the beneficial effects of the said ointment. We shall give it in the author's own "words, as the case is short. " William Ilolloway, ajtat. 26,?Very tall, and of rather a spare habit, about the end of December, or beginning of January last, was attacked with violent pain in the left side, and considerable swelling about the region of the liver, with most of the usual symptoms that attend hepatitis, together with others indicative of pulmonary affection. He wan bled, blistered, and took various medicines, under the direction of several medical gentlemen in the neighbourhood. These remedies afforded him a temporary relief; but he soon grew worse, and his malady continued to increase for six weeks, when I mentioned his case to you. By your advice, I furnished him with some of the ointment of tartarized antimony, and directed him to rub it on the chest. In twenty-four hours, eruptions appeared. The enlargement about the liver soon began to subside, the pain abated, and at the expiration of a month he was able to follow his usual occupation of a mason and bricklayer. " In September last, during the unsettled weather, he went to assist a neighbour in securing his corn; when, after using great bodily exertion, and drinking freely of cider while he was very warm, he felt himself much indisposed, and in two days afterwards was seized with chills. The pain in the side returned, attended with pain on the top of the shoulder and in the chest, shortness of breath, cough, quick pulse, (I never found it under 120,) and great debility. He now used the ointment, unassisted by any internal remedy. He received it from me with the most enthusiastic rapture, and used it moro profusely than I intended, not only on the chest, but on the shoulder, and wherever he felt pain. A large crop of pustules was the result, which maturated, and continued to discharge plentifully for nine days, when ho was able to resume his work, and is now free from all complaint." 63.
We must now take leave of our worthy author, with the expression of our high respect for ltis talents,zeal, and usefulness; and with ardent wishes tluit the remedial process pointed out in these paires may answer the expectations of himself, and prove equally advantageous in the hands of others.
ointment rubbed on the region of the stomach in hoopingcough. By tartrite of antimony, Dr. Jenner remarks, we can not only crcate vesicles, but at the same time produce diseased action in the skin itself, by deeply deranging its structure beneath the surface?one cause, perhaps, why the sympathetic affection excited by the lytta, and those changes induced by antimony, are very different. Our author observes that, in small pox, a different degree of secondary fever will follow, when the skin is simply affected on the surface, and when it is partially destroyed beneath. " Whoever has observed the deranged state of health where vesiculated eruptions have been called into action by an effort of Nature, must have seen how often they arrest the progress of the original |
Functional connectivity of the human amygdala in health and in depression
To analyse the functioning of the amygdala in depression, we performed the first voxel-level resting state functionalconnectivity neuroimaging analysis of depression of voxels in the amygdala with all other voxels in the brain, with 336 patients with major depressive disorder and 350 controls. Amygdala voxels had decreased functional connectivity (FC) with the orbitofrontal cortex, temporal lobe areas, including the temporal pole, inferior temporal gyrus and the parahippocampal gyrus. The reductions in the strengths of the FC of the amygdala voxels with the medial orbitofrontal cortex and temporal lobe voxels were correlated with increases in the Beck Depression Inventory score and in the duration of illness measures of depression. Parcellation analysis in 350 healthy controls based on voxel-level FC showed that the basal division of the amygdala has high FC with medial orbitofrontal cortex areas, and the dorsolateral amygdala has strong FC with the lateral orbitofrontal cortex and related ventral parts of the inferior frontal gyrus. In depression, the basal amygdala division had especially reduced FC with the medial orbitofrontal cortex, which is involved in reward; and the dorsolateral amygdala subdivision had relatively reduced FC with the lateral orbitofrontal cortex, which is involved in non-reward.
# Introduction
There is considerable evidence that the amygdala is involved in emotion [bib_ref] Rethinking the emotional brain, Ledoux [/bib_ref]. Moreover, resting state functional connectivity (FC) between brain areas, which reflects correlations of activity, is a fundamental tool in helping to understand the brain regions with altered connectivity and function in mental disorders [bib_ref] Great expectations: using whole-brain computational connectomics for understanding neuropsychiatric disorders, Deco [/bib_ref] , and changes in amygdala FC have been related to depression [bib_ref] Neural mechanisms of the cognitive model of depression, Disner [/bib_ref] [bib_ref] Identifying predictors, moderators, and mediators of antidepressant response in major depressive disorder:..., Phillips [/bib_ref] [bib_ref] Circuits regulating pleasure and happiness -mechanisms of depression, Loonen [/bib_ref] [bib_ref] Fronto-limbic effective connectivity as possible predictor of antidepressant response to SSRI administration, Vai [/bib_ref] [bib_ref] Resting-state functional connectivity of the amygdala and longitudinal changes in depression severity..., Connolly [/bib_ref] [bib_ref] Successful group psychotherapy of depression in adolescents alters fronto-limbic resting-state connectivity, Straub [/bib_ref]. Furthermore, the amygdala responses to faces [bib_ref] Neurons in the amygdala of the monkey with responses selective for faces, Leonard [/bib_ref] may be influenced by antidepressant drugs [bib_ref] Neural processing of reward and punishment in young people at increased familial..., Mccabe [/bib_ref] [bib_ref] It's the way that you look at it'-a cognitive neuropsychological account of..., Harmer [/bib_ref].
The aim of the present paper was to examine the FC of the amygdala in depression at the voxel level. We analysed every amygdala voxel for significantly different FC with voxels throughout the rest of the brain in depressed people vs controls, and tested whether any differences of FC found are correlated with the symptoms of depression, in order to advance understanding of the amygdala and depression. The advantage of voxel-level FC is that we can show exactly which amygdala voxels have altered FC with the exact (voxel-level) parts of other brain areas. In order to perform this voxel-level FC analysis, we utilized and required a uniquely large sample of 336 patients with major depressive disorder and 350 controls. In a previous study with 70 participants, lower FC between the amygdala and hippocampus and parahippocampus was reported [bib_ref] Abnormal amygdala resting-state functional connectivity in adolescent depression, Cullen [/bib_ref].
This investigation is very different from a previous analysis of FC in depression [bib_ref] Medial reward and lateral non-reward orbitofrontal cortex circuits change in opposite directions..., Cheng [/bib_ref] , as follows. First, we focused on the FC of the amygdala, not the whole brain, in order to analyse the FC of the amygdala in depression. The methodology used here is quite different from, and more statistically powerful than, a whole brain voxel-to voxel FC analysis, which, because there are so many voxels pairs in the whole brain, is rather insensitive, for it carries a huge burden to correct for the multiple comparisons [1 133 760 771 voxel pairs requiring normally P < 10 À8 for any effect to be significant [bib_ref] Medial reward and lateral non-reward orbitofrontal cortex circuits change in opposite directions..., Cheng [/bib_ref] ]. The patients in the present study included some of those in our earlier study for whom we had the data needed for the present study. Second, we performed a parcellation of the amygdala based on its FC, showed which parts the brain each amygdala subdivision was related to, and showed how the FC of each amygdala subdivision was different in depression. Third, we describe here how amygdala connectivity was correlated with the depression severity and duration, which was not performed in the previous study. Part of the reason for these differences is that in the previous investigation we focused on voxel-to-voxel whole brain connectivity, which limits the results that can be established, whereas here we focus on the amygdala, and are able to report significant differences in its FC in depression, and even of its subdivisions.
A new theory of depression is that the lateral orbitofrontal cortex has increased sensitivity of a non-reward attractor in and that the reciprocally related medial orbitofrontal cortex reward system is underactive in depression [bib_ref] A non-reward attractor theory of depression, Rolls [/bib_ref] , and there is evidence consistent with this [bib_ref] Reward and punishment processing in depression, Eshel [/bib_ref] [bib_ref] The brain reward circuitry in mood disorders, Russo [/bib_ref] [bib_ref] Reward processing dysfunction in major depression, bipolar disorder and schizophrenia, Whitton [/bib_ref] [bib_ref] Medial reward and lateral non-reward orbitofrontal cortex circuits change in opposite directions..., Cheng [/bib_ref] [bib_ref] A non-reward attractor theory of depression, Rolls [/bib_ref]. It was therefore of interest in the present investigation whether the amygdala had altered connectivity with these other brain systems already implicated in depression. The new theory is that the lateral orbitofrontal cortex, which is activated when expected reward is not obtained (which can cause sadness), can enter and maintain an ongoing mood state in a recurrent 'attractor' network more readily in depression [bib_ref] A non-reward attractor theory of depression, Rolls [/bib_ref] , as described more fully in the 'Discussion' section.
# Methods
## Participants
There were 336 patients with a diagnosis of major depression and 350 controls. The data available for this study were from Xinan (First Affiliated Hospital of Chongqing Medical School in Chongqing, China) and Taiwan . This is a subset of participants in a previous study in which very different analyses were performed [bib_ref] Medial reward and lateral non-reward orbitofrontal cortex circuits change in opposite directions..., Cheng [/bib_ref] , as is made clear in the Introduction. All participants were diagnosed according to the Diagnostic and Statistical Manual of Mental Disorder-IV criteria for major depressive disorder. Depression severity and symptomatology were evaluated by the Hamilton Depression Rating Scale (HAMD, 17 items) [bib_ref] A rating scale for depression, Hamilton [/bib_ref] and the Beck Depression Inventory (BDI) [bib_ref] Assessment of depression: the depression inventory, Beck [/bib_ref]. One hundred and twenty-five of the patients were not receiving medication at the time of the neuroimaging. [fig_ref] Table 1: A summary of the demographic information and the psychiatric diagnosis in the... [/fig_ref] provides a summary of the demographic information and the psychiatric diagnosis of the participants, with further details in the Supplementary Material.
## Image acquisition and preprocessing
Data for resting state FC analysis were collected in 3T MRI scanners in an 8 min period in which the participants were awake in the scanner not performing a task using standard protocols described in the Supplementary Material.
Data preprocessing was performed using DPARSF (Chao-Gan and Yu-Feng, 2010) (http://restfmri.net) that is a toolbox based on the SPM8 software package. The first 10 EPI scans were discarded to suppress equilibration effects. The remaining scans of each subject underwent slice timing correction by sinc interpolating volume slices, motion correction for volume to volume displacement, spatial normalization to standard Montreal Neurological Institute (MNI) space using affine transformation and nonlinear deformation with a voxel size of 3 Â 3 Â 3 mm, followed by spatial smoothing (8 mm Full Width Half Maximum FWHM). To remove the sources of spurious correlations present in resting-state BOLD data, all fMRI time-series underwent band-pass temporal filtering (0.01-0.1 Hz), nuisance signal removal from the ventricles, and deep white matter, and regressing out any effects of head motion using the Friston et al. 24 head motion parameters procedure [bib_ref] Movement-related effects in fMRI time-series, Friston [/bib_ref]. Finally, we implemented additional careful volume censoring (scrubbing) movement correction as reported by [bib_ref] Methods to detect, characterize, and remove motion artifact in resting state fMRI, Power [/bib_ref] to ensure that head-motion artefacts are not driving observed effects. The mean framewise displacement (FD) was computed with FD threshold for displacement being 0.5 mm. In addition to the frame corresponding to the displaced time point, 1 preceding and 2 succeeding time points were also deleted to reduce the spill-over effect of head movements. Subjects with >10% displaced frames flagged were completely excluded from the analysis as it is likely that such high-level of movement would have had an influence on several volumes. Global signals were not regressed out, for reasons described elsewhere [bib_ref] Medial reward and lateral non-reward orbitofrontal cortex circuits change in opposite directions..., Cheng [/bib_ref]. Because we do not regress out the global signal, most of the functional connectivities are positive, and the interpretation of a decrease of FC is just that the correlation between the activities of the pair of nodes has decreased. Considering the potential effect of gender [bib_ref] Laterality patterns of brain functional connectivity: gender effects, Tomasi [/bib_ref] , age [bib_ref] A brain-wide study of age-related changes in functional connectivity, Geerligs [/bib_ref] and head motion [bib_ref] Spurious but systematic correlations in functional connectivity MRI networks arise from subject..., Power [/bib_ref] [bib_ref] Methods to detect, characterize, and remove motion artifact in resting state fMRI, Power [/bib_ref] on FC, any effects of gender ratio, years of education, age and head motion between the patient and control groups were regressed out in all analyses. There were no differences in the gender ratios, age and mean FD (P > 0.05 in all cases), though the number of years of education was lower in the patients than controls. However, none of the FC link differences found between patients and controls was correlated significantly (FDR, P < 0.05) with the number of years of education. We also note that the Taiwanese sample included patients with depression in remission while under antidepressant treatment, and thus their scores on the HAMD assessment were in the low range.
## Hypothesis based voxel-wise association studies
In this paper we utilize what we term 'hypothesis-based voxellevel FC analysis' in which we select a brain region of interest, but then calculate for every voxel in that region whether it has FC with individual voxels in every other brain region. In the present paper, we select the amygdala as the region of interest, given the research on it described above implicating it in depression, and then we show exactly which amygdala voxels have altered FC in depression with which individual voxels in every other brain area. Given that the amygdala has 149 voxels, and that there are 47 619 3 Â 3 Â 3 mm voxels in the automated anatomical atlas (AAL2) brain, the number of voxel pairs in this study was approximately (149 Â 47 619). As noted in the section 'Introduction', this methodology is quite different from, and more statistically powerful than, a whole brain voxel-to voxel FC analysis, which, because there are so many voxels pairs in the whole brain, is rather insensitive, for it carries a huge burden to correct for the multiple comparisons [1 133 760 771 voxel pairs requiring normally P < 10 À8 for any effect to be significant [bib_ref] Medial reward and lateral non-reward orbitofrontal cortex circuits change in opposite directions..., Cheng [/bib_ref] ].
In the present study, each resting-state fMRI volume included 47,619 voxels, and the amygdala region of interest had 149 voxels in the AAL2 atlas. For each pair of voxels in the amygdala and voxels in all other brain areas, the time series were extracted and their Pearson correlation was calculated for each subject, to provide the measure of FC, followed by z-transformation. Two-tailed, two-sample t-tests were performed on Fisher's z-transformed correlation coefficients to identify significantly altered FC links in patients with depression compared with controls within each imaging centre. The effects of age, gender ratios, head motion (mean FDs) and education were regressed out within each dataset in this step by a generalized linear model [bib_ref] Head size, age and gender adjustment in MRI studies: a necessary nuisance?, Barnes [/bib_ref] [bib_ref] The autism brain imaging data exchange: towards a large-scale evaluation of the..., Martino [/bib_ref]. After obtaining the t-test results (P value for each FC) for each centre, the Liptak-Stouffer z score method [bib_ref] On the combination of independent tests, Liptak [/bib_ref] [described in detail in previous studies [bib_ref] Voxel-based, brain-wide association study of aberrant functional connectivity in schizophrenia implicates thalamocortical..., Cheng [/bib_ref] ] was then used to combine the results from the individual datasets. Specifically, the P-value of each FC resulting from the twosample t-test in the previous step was converted to its corresponding z score. This was calculated firstly as in equation: z ¼ U À1 ð1 À p k Þ, where U is the standard normal cumulative distribution function and k represents the k centre. Next, a combined z score for a FC was calculated using the Liptak-Stouffer formula: Z ¼ P w k z k = P w 2 k , where w k ¼ square root of sample size is the weight of the kth dataset. Finally, the Z is transformed into its corresponding P-value, and an FDR procedure was used to correct for multiple comparisons across the 149 Â 47 619 voxel pairs. In the present study, FDR correction for the FC between any pair of voxels was used, and results are presented based on this statistical test with FDR P < 0.05, corresponding to a P threshold of 6:89 Â 10 À4 in the Z-tests.
## Visualization of the differences in fc for each voxel
To illustrate in some of the figures the extent to which voxels in different brain areas had differences of FC between patients and controls, we used a measure for the association (MA) between a voxel i and the brain disorder. This was defined as: MA ¼ N a , where N is the number of links between voxel i and every other voxel in the brain that have a P-value of less than a (which in the present study with FDR correction was P < 0.05, corresponding to a P threshold of 6:89 Â 10 À4 ) in t-tests comparing patients with controls. A larger value of MA implies a more significant difference in FC. To ensure clarity, for a given voxel, the MA is the number of altered FC after FDR correction involving this voxel, so a higher MA indicates more significantly different FC links to other brain areas for that voxel.
## Clinical correlates
We also investigated whether the differences in FC between patients and controls were correlated with clinical variables [the HAMD [bib_ref] A rating scale for depression, Hamilton [/bib_ref] , BDI [bib_ref] Assessment of depression: the depression inventory, Beck [/bib_ref] , and illness duration [bib_ref] Brain morphometric abnormalities in geriatric depression: long-term neurobiological effects of illness duration, Bell-Mcginty [/bib_ref] [bib_ref] Microstructural white-matter abnormalities associated with treatment resistance, severity and duration of illness..., De Diego-Adelino [/bib_ref] ]. Since the Taiwan dataset had only 54 patients and their HAMD scores are low due to the effects of medication, this correlation analysis was performed on the Xinan dataset. The samples that were more than 3 standard deviations away from the sample's mean were removed from this analysis. Specifically, for each brain region identified in the hypothesis based voxel-wise association studies, we first calculated the partial correlation between the clinical scores and the voxel-wise FCs between the significant voxels in that brain region and the amygdala, with head motion, education, sex and age as covariates so that they did not contribute to the correlation. Then the mean correlation between the clinical scores and voxel-wise FCs was defined as the overall correlation between the significant voxels in that brain region and the amygdala. Finally, a permutation test with 5000 randomizations of the patient labels was used to assess the statistical significance of the mean correlation.
# Results
The fMRI resting state FC analyses were performed with 336 patients with a diagnosis of major depression, and 350 controls, and this large population was sufficient to allow voxel-level analysis with FDR corrected statistics of the differences of FC of amygdala voxels with all other voxels in the brain (excluding the cerebellum) in patients vs controls.
## A voxel-level association study (vas) of amygdala voxels with different fc in depressed patients
As shown in [fig_ref] Figure 1: Anatomical location of consistently different FC in depression obtained from the voxel-based... [/fig_ref] and [fig_ref] Table 2: Numbers of voxels in different AAL2 areas with significantly different FC with... [/fig_ref] , there were a number of amygdala voxels with different FC in patients with depression compared to controls. In all cases, a reduction in FC was found in the depression group. The largest clusters of voxels with altered (reduced) FC with the amygdala were in the medial orbitofrontal cortex (634 voxels, [fig_ref] Table 2: Numbers of voxels in different AAL2 areas with significantly different FC with... [/fig_ref]. (These voxel numbers are those with altered FC with amygdala voxels with P < 0.05 FDR corrected.) Additional areas with voxels with reduced FC with the amygdala in depression included the lateral orbitofrontal cortex, parahippocampal and fusiform gyri; inferior and middle temporal gyri; the temporal pole; the insula; occipital visual areas; the mid-cingulate cortex; the striatum (including parts of the caudate and putamen); and the precentral and postcentral gyrus [fig_ref] Table 2: Numbers of voxels in different AAL2 areas with significantly different FC with... [/fig_ref] ; [fig_ref] Figure 1: Anatomical location of consistently different FC in depression obtained from the voxel-based... [/fig_ref].
Supplementary [fig_ref] Figure 3: Voxel-level parcellation of the amygdala [/fig_ref] confirms that the results with the combined dataset shown in [fig_ref] Figure 1: Anatomical location of consistently different FC in depression obtained from the voxel-based... [/fig_ref] are consistent with those obtained in a single dataset. Supplementary [fig_ref] Figure 4: Summary of amygdala FC differences in depression [/fig_ref] shows the correlation between the mean t value corresponding to the Xinan dataset and the mean t value corresponding to the Taiwan dataset for all the voxel-wise functional connectivities involving the amygdala. This provides confirmation that the results from the two datasets are consistent.
## Analysis of the fc links that were different in patients with depression
To investigate the brain areas between which there was different FC in depression, and whether it was increased or decreased, the FC of the voxels with significant differences of FC (after FDR correction at P < 0.05, and within the voxel clusters shown in [fig_ref] Table 2: Numbers of voxels in different AAL2 areas with significantly different FC with... [/fig_ref] were measured for each of the AAL2 regions within which the voxels were located. (A list of abbreviations of the AAL2 areas are provided in [fig_ref] Table 1: A summary of the demographic information and the psychiatric diagnosis in the... [/fig_ref] The FC differences are shown in [fig_ref] Figure 2: The voxel-level FC for amygdala voxels that are significantly different in the... [/fig_ref] at the voxel level, with the voxels shown arranged by the AAL2 areas in which they are found. [fig_ref] Figure 2: The voxel-level FC for amygdala voxels that are significantly different in the... [/fig_ref] shows that the amygdala voxels with altered FC with other brain areas tend to be in different parts of the amygdala. This is a new level of precision attained in this investigation in which the location in the amygdala of voxels with functional connectivities with the location of voxels in other brain areas are revealed, and how these differ in depression.
First, the different FC of amygdala voxels with the medial orbitofrontal cortex, area 13, brain region is considered. This region is involved in reward and subjective pleasure [bib_ref] Value, pleasure, and choice in the ventral prefrontal cortex, Grabenhorst [/bib_ref]. The relevant voxels are in AAL2 regions such as OFCmed, OFCant, OFCpost, Olfactory and Rectus [fig_ref] Figure 1: Anatomical location of consistently different FC in depression obtained from the voxel-based... [/fig_ref] ; [fig_ref] Table 2: Numbers of voxels in different AAL2 areas with significantly different FC with... [/fig_ref]. The voxels within this area 13 cluster have high positive correlations between them, and have generally the same pattern of altered FC in depression [bib_ref] Medial reward and lateral non-reward orbitofrontal cortex circuits change in opposite directions..., Cheng [/bib_ref] , so this cluster is described in the remainder of this paper as OFC13. [fig_ref] Table 2: Numbers of voxels in different AAL2 areas with significantly different FC with... [/fig_ref]. For the Lateral OFC the AAL2 areas included were OFClat and IFG_Orb. For the Temporal lobe the AAL2 areas included were TPO, ITG and MTG. These AAL2 areasare listed in [fig_ref] Table 1: A summary of the demographic information and the psychiatric diagnosis in the... [/fig_ref]. The peak MA value shown is the number of links between the peak voxel in the brain regions listed under 'Areas' and the amygdala voxels, where these links are significantly different FDR corrected at P < 0.05. The MNI coordinates are the peak of the cluster in the brain regions listed under 'Areas'. In all cases in this table, the FC links are weaker in the depressed group; thus these MA values show the number of decreased FC links in the depression group. There were a total of 97 740 links between the voxels of amygdala and other brain areas which showed a significant difference between controls and patients with depression (FDR correction, P < 0.05).
Second, some voxels in the amygdala have reduced FC with a more lateral to anterior part of the orbitofrontal cortex, in AAL2 areas OFClat and Frontal_Inf_Orb_2 [fig_ref] Figure 1: Anatomical location of consistently different FC in depression obtained from the voxel-based... [/fig_ref] , and this is an area that is BA 47/12 , is termed here OFC47/12, and is involved in non-reward and unpleasant events [bib_ref] Value, pleasure, and choice in the ventral prefrontal cortex, Grabenhorst [/bib_ref] [bib_ref] A non-reward attractor theory of depression, Rolls [/bib_ref]. There is some overlap of the amygdala voxels with reduced connectivity in depression in the medial and lateral orbitofrontal cortex areas, but some other amygdala voxels have reduced FC only with the medial orbitofrontal cortex, perhaps because more amygdala voxels have significant FC with the more extensive medial than with the lateral orbitofrontal cortex.
Third, some voxels in the amygdala have decreased FC with some temporal cortex areas including the inferior temporal gyrus, fusiform gyrus and temporal pole, areas known to be involved in visual and multimodal processing [bib_ref] Invariant visual object and face recognition: neural and computational bases, and a..., Rolls [/bib_ref] [fig_ref] Figure 1: Anatomical location of consistently different FC in depression obtained from the voxel-based... [/fig_ref]. This reduced FC also extended posteriorly into earlier visual areas including the occipital cortex and lingual gyrus, areas to which the amygdala sends backprojections [bib_ref] Anatomical organization of the primate amygdaloid complex, Amaral [/bib_ref].
Fourth, some amygdala voxels had reduced FC in depression with medial temporal lobe areas such as the parahippocampal, perirhinal and entorhinal cortex implicated in memory [fig_ref] Figure 1: Anatomical location of consistently different FC in depression obtained from the voxel-based... [/fig_ref] ; [fig_ref] Table 2: Numbers of voxels in different AAL2 areas with significantly different FC with... [/fig_ref].
Fifth, some amygdala voxels were found to have reduced FC with the insula [fig_ref] Figure 1: Anatomical location of consistently different FC in depression obtained from the voxel-based... [/fig_ref] , in an area implicated in autonomic output [bib_ref] Functions of the anterior insula in taste, autonomic, and related functions, Rolls [/bib_ref].
Sixth, some amygdala voxels were found to have reduced FC with more ventral parts of the striatum including parts of the caudate and putamen [fig_ref] Figure 1: Anatomical location of consistently different FC in depression obtained from the voxel-based... [/fig_ref].
Seventh, some amygdala voxels were found to have reduced FC with some somatosensory/motor areas including the middle cingulate cortex and pre-and post-central gyrus [fig_ref] Figure 1: Anatomical location of consistently different FC in depression obtained from the voxel-based... [/fig_ref] ; [fig_ref] Table 2: Numbers of voxels in different AAL2 areas with significantly different FC with... [/fig_ref].
## Amygdala voxel-level fc in healthy participants, using parcellation
To analyse whether some parts of the amygdala had changes especially related to depression, we first performed a parcellation of the amygdala in healthy controls, and then examined the differences of FC of each division of the amygdala between controls and participants with depression. The parcellation analysis was performed on the healthy controls, so that we could investigate how the strengths of the connectivities in different parts of the amygdala in healthy individuals might be different in depression. For healthy control participants the voxel-wise FC pattern of the amygdala with voxels in other brain areas is shown in [fig_ref] Figure 3: Voxel-level parcellation of the amygdala [/fig_ref] -C (k-means was used to perform this clustering. The number of clusters was selected to be the maximum number in which each cluster was spatially discrete). Three subdivisions were found in each hemisphere [fig_ref] Figure 3: Voxel-level parcellation of the amygdala [/fig_ref] , a dorsal amygdala subdivision close to the central nucleus of the amygdala (1, blue); a dorsolateral amygdala subdivision (2, yellow), and a ventral amygdala subdivision (3, red). Similar parcellation was found, if the FC of each amygdala voxel with other amygdala voxels was used to perform the clustering. The pattern of FC for the different amygdala subdivisions with areas of the orbitofrontal cortex is different, as shown in the diagrams in [fig_ref] Figure 3: Voxel-level parcellation of the amygdala [/fig_ref] and C. The basolateral part of the amygdala (subdivision 3, red) has high FC with medial orbitofrontal cortex areas (including AAL2 areas REC and OFCmed). The dorsolateral part of the amygdala (subdivision 2, yellow) has especially strong FC with the lateral orbitofrontal cortex and its related ventral parts of the inferior frontal gyrus (including AAL2 areas OFClat, IFGorb, IFGtriang and IFG operc). The dorsal part of the amygdala (subdivision 1, blue) has relatively strong FC with the posterior orbitofrontal cortex (AAL2 area OFCpost and the area immediately posterior to this, OLF which includes the olfactory tubercle).
## Functional connectivities for different amygdala subregions in depression
The differences of FC in depression of these three amygdala subdivisions are shown in [fig_ref] Figure 3: Voxel-level parcellation of the amygdala [/fig_ref] and E. The contrast is (healthy controls-depressed group) for the three subdivisions. (A negative value for z in [fig_ref] Figure 3: Voxel-level parcellation of the amygdala [/fig_ref] and E thus represents a weaker FC in patients with depression.) [fig_ref] Figure 3: Voxel-level parcellation of the amygdala [/fig_ref] and E shows that all three amygdala subdivisions show reduced FC with the AAL2 areas listed in [fig_ref] Table 2: Numbers of voxels in different AAL2 areas with significantly different FC with... [/fig_ref]. A difference that was consistent across both hemispheres is that the basal amygdala division (subdivision 3, red) had especially reduced FC with some medial orbitofrontal cortex areas (including AAL2 areas OFCmed, REC, PFCventmed and OLF). In contrast, the dorsolateral amygdala subdivision (2, yellow) had relatively reduced FC with lateral orbitofrontal cortex and related areas (including AAL2 areas OFClat, IFGorb, IFGtriang and IFGoperc) (relative to subdivision 3). These differences are of interest, for the medial orbitofrontal cortex is involved in reward, and the lateral orbitofrontal cortex in non-reward and punishment [bib_ref] Value, pleasure, and choice in the ventral prefrontal cortex, Grabenhorst [/bib_ref] [bib_ref] The orbitofrontal cortex and emotion in health and disease, including depression, Rolls [/bib_ref].
## Clinical symptom correlates of the reduced amygdala functional connectivities in depression
The hypothesis that some of the symptom scores or the illness duration was related to the reduced FC identified between the medial orbitofrontal cortex and amygdala in depression was tested by grouping together all the medial orbitofrontal cortex areas. The results [fig_ref] Table 3: Correlations between the FC links and the depression symptom severity scores In... [/fig_ref] , top row) show that the reduction in FC between right amygdala voxels, and the medial orbitofrontal cortex areas (OFCmed, OFCpost, OFCant, rectus and OLF) were significantly correlated with the BDI score; and that the reduction of FC between both sides of the amygdala and the medial orbitofrontal cortex areas was related to the duration of the illness. These correlations of the FC of amygdala voxels with the medial orbitofrontal cortex areas are significant FDR corrected for multiple comparisons.
In the remainder of [fig_ref] Table 3: Correlations between the FC links and the depression symptom severity scores In... [/fig_ref] and in Supplementary [fig_ref] Figure 1: Anatomical location of consistently different FC in depression obtained from the voxel-based... [/fig_ref] , we show for illustration that for individual AAL2 areas there were significant correlations (P < 0.05 uncorrected) between some of the AAL2-based region of interest-wise FC links and the symptom severity scores and illness duration. The correlations are in the direction that the weaker an FC link is, the higher is the clinical score for the depression. For [fig_ref] Table 3: Correlations between the FC links and the depression symptom severity scores In... [/fig_ref] , the correlation is that between the average of the FC of the FDR corrected significant voxels in an AAL2 region with an amygdala voxel, and the clinical measures. The BDI score was correlated with decreased FC between the amygdala and the OFC_med, and OFC_post, with the cuneus, with the fusiform cortex, with the temporal pole, with the mid-cingulate cortex and with occipital areas and the lingual gyrus [fig_ref] Table 3: Correlations between the FC links and the depression symptom severity scores In... [/fig_ref]. The illness duration was correlated with decreased FC between the amygdala and the medial orbitofrontal cortex areas including OFCmed, OFCpost, OFCant and Gyrus Rectus; with the fusiform gyrus; and with the inferior temporal gyrus. For Supplementary [fig_ref] Figure 1: Anatomical location of consistently different FC in depression obtained from the voxel-based... [/fig_ref] , the correlations shown are those between every voxel-based link between all voxels with FDR corrected differences in depression between the amygdala and an AAL2 region, and the clinical measures. The statistical significance of these effects is shown in [fig_ref] Table 3: Correlations between the FC links and the depression symptom severity scores In... [/fig_ref]. Supplementary [fig_ref] Figure 1: Anatomical location of consistently different FC in depression obtained from the voxel-based... [/fig_ref] shows, extremely interestingly, that a decrease of FCs of the medial orbitofrontal cortex with the amygdala is correlated with more severe depression (measured by the BDI and illness duration). To address any possible effect of the medication on the correlation between the clinical variables and functional connectivities, we also performed a partial correlation as described in the section 'Methods' by adding medication, head motion, education, sex and age as covariates. The results were consistent with the correlations shown in [fig_ref] Table 3: Correlations between the FC links and the depression symptom severity scores In... [/fig_ref] , which indicates that the results in [fig_ref] Table 3: Correlations between the FC links and the depression symptom severity scores In... [/fig_ref] hold independently of the medication. The HAMD [bib_ref] A rating scale for depression, Hamilton [/bib_ref] score was not significantly correlated with the functional connectivities of the amygdala.
## Fc in unmedicated patients with depression
Within the depressed group from the Xinan dataset, 125 were not receiving medication, and 157 patients were receiving medication. We were able to confirm that the main findings shown here in [fig_ref] Figure 1: Anatomical location of consistently different FC in depression obtained from the voxel-based... [/fig_ref] for all the depressed patients were not due just to the effects of the medication, in that in the 125 unmedicated [fig_ref] Table 2: Numbers of voxels in different AAL2 areas with significantly different FC with... [/fig_ref] for the left AMYG. (E) The same as (D) for the right AMYG. patients, differences from controls in the same brain regions were similar, as shown in Supplementary [fig_ref] Figure 2: The voxel-level FC for amygdala voxels that are significantly different in the... [/fig_ref]. However, it was of interest that in the unmedicated patients, the precuneus had higher FC with the amygdala (see Supplementary [fig_ref] Figure 2: The voxel-level FC for amygdala voxels that are significantly different in the... [/fig_ref]. Consistently, in further analyses, in which the unmedicated and medicated depressed patients were compared, it was found that in the medicated patients there was much less of an increase of FC between the precuneus and the amygdala. It thus may be that the medication decreases the FC between the precuneus and amygdala. (The medication consisted in most cases of selective serotonin reuptake inhibitors including fluoxetine, paroxetine, sertraline, citalopram and escitalopram; or serotonin-norepinephrine reuptake inhibitors such as venflaxine, or a tetracyclic antidepressant such as mirtazepine.)
In this study, any effects of the medication were difficult to analyse, because those on medication were more likely to be longterm than first episode patients, so we limit the analysis to what has just been described. In summary, the reductions in FC shown in [fig_ref] Figure 1: Anatomical location of consistently different FC in depression obtained from the voxel-based... [/fig_ref] and [fig_ref] Table 2: Numbers of voxels in different AAL2 areas with significantly different FC with... [/fig_ref] in patients with depression were also found in unmedicated patients (Supplementary [fig_ref] Figure 2: The voxel-level FC for amygdala voxels that are significantly different in the... [/fig_ref] , and are not due just to the effects of the medication.
# Discussion
One main finding is that the amygdala has reduced FC with a major region with altered FC in depression, the medial orbitofrontal cortex BA 13 [bib_ref] Medial reward and lateral non-reward orbitofrontal cortex circuits change in opposite directions..., Cheng [/bib_ref] , which is implicated in reward [bib_ref] Value, pleasure, and choice in the ventral prefrontal cortex, Grabenhorst [/bib_ref] [fig_ref] Table 2: Numbers of voxels in different AAL2 areas with significantly different FC with... [/fig_ref] ; [fig_ref] Figure 1: Anatomical location of consistently different FC in depression obtained from the voxel-based... [/fig_ref] shows for illustrative purposes the locations of the voxels in the amygdala in brain slices that have significant correlations (P < 0.05) with the BDI scores or with illness duration. [fig_ref] Figure 1: Anatomical location of consistently different FC in depression obtained from the voxel-based... [/fig_ref]. The reduced FC of the amygdala with the medial orbitofrontal cortex was correlated with the increase in the measures from the BDI [fig_ref] Table 3: Correlations between the FC links and the depression symptom severity scores In... [/fig_ref] ; Supplementary Figure S1), and with the illness duration, making it likely that this FC link between the amygdala and the medial orbitofrontal cortex is related to the depression. This finding is consistent with the non-reward attractor theory of depression, which includes the hypothesis of less activity in medial orbitofrontal cortex reward-related areas in depression [bib_ref] A non-reward attractor theory of depression, Rolls [/bib_ref] , as well as increased activity in the reciprocally related lateral orbitofrontal cortex. The new theory is that the lateral orbitofrontal frontal cortex, which is activated when expected reward is not obtained (which can cause sadness), can enter and maintain an ongoing mood state in a recurrent 'attractor' network more readily in , This attractor or short-term memory state may be triggered into increased activity by a strong nonreward event in the environment, or may be more sensitive in some depressed people, and more easily therefore triggered into a high firing rate state that is associated with a sad mood. Given that the amygdala has some roles in emotion [bib_ref] Rethinking the emotional brain, Ledoux [/bib_ref] , its reduced FC with the medial orbitofrontal cortex which is involved in reward and positive mood, may contribute to the lowering of mood by being somewhat disconnected from the orbitofrontal cortex in depression. A possible clinical implication is that altering the functioning of the lateral orbitofrontal cortex may release the medial orbitofrontal cortex. There is already evidence that repetitive transcranial stimulation (rTMS) of the lateral orbitofrontal cortex may be helpful in the treatment of some patients with depression [bib_ref] Neural correlates of successful orbitofrontal 1 Hz rTMS following unsuccessful dorsolateral and..., Fettes [/bib_ref] [bib_ref] 1Hz rTMS of the right orbitofrontal cortex for major depression: safety, tolerability..., Feffer [/bib_ref]. An additional part of the theory of depression is that, given that the lateral and medial orbitofrontal cortex tend to have reciprocal activations [bib_ref] Abstract reward and punishment representations in the human orbitofrontal cortex, O'doherty [/bib_ref] , decreased activity of the reward-related medial orbitofrontal cortex may also be related to depression [bib_ref] A non-reward attractor theory of depression, Rolls [/bib_ref] [bib_ref] The orbitofrontal cortex and emotion in health and disease, including depression, Rolls [/bib_ref]. Consistent with this, the medial orbitofrontal cortex has reduced FC with systems implicated in memory, including the medial temporal lobe [bib_ref] Medial reward and lateral non-reward orbitofrontal cortex circuits change in opposite directions..., Cheng [/bib_ref] , and posterior cingulate cortex [bib_ref] Increased functional connectivity of the posterior cingulate cortex with the lateral orbitofrontal..., Cheng [/bib_ref] [which can be considered a gateway to the hippocampus [bib_ref] The storage and recall of memories in the hippocampo-cortical system, Rolls [/bib_ref] ]. This reduced FC may reduce happy memories. It is suggested in this context that the reduced FC in depression of the amygdala with the medial orbitofrontal cortex described here is related to the reduced hedonia in depression.
It is noted that the great development of the orbitofrontal cortex in primates and especially in humans in evolution, may enable the human orbitofrontal cortex to make a correspondingly greater contribution to emotion in humans than the amygdala, and to depression. Indeed, the FC changes of the medial and lateral orbitofrontal cortex were significant for large numbers of voxels in the whole brain FC analysis [bib_ref] Medial reward and lateral non-reward orbitofrontal cortex circuits change in opposite directions..., Cheng [/bib_ref].
Second, some FC reductions of the amygdala with medial temporal lobe areas such as the parahippocampal, perirhinal and entorhinal cortex implicated in memory were found [fig_ref] Figure 1: Anatomical location of consistently different FC in depression obtained from the voxel-based... [/fig_ref] ; [fig_ref] Table 2: Numbers of voxels in different AAL2 areas with significantly different FC with... [/fig_ref]. This decrease in FC of the amygdala with medial temporal lobe memory-related areas is similar to that of the medial orbitofrontal cortex, which has greatly reduced FC with the medial temporal lobe memory system [bib_ref] Medial reward and lateral non-reward orbitofrontal cortex circuits change in opposite directions..., Cheng [/bib_ref] , and which with its role in reward may be related to the reduced processing of happy memories, and therefore an imbalance towards unhappy memories, in depression [bib_ref] Medial reward and lateral non-reward orbitofrontal cortex circuits change in opposite directions..., Cheng [/bib_ref] [bib_ref] A non-reward attractor theory of depression, Rolls [/bib_ref].
Third, some voxels in the amygdala had decreased FC with some temporal cortex areas including the inferior temporal gyrus, temporal pole, and fusiform gyrus, areas known to be involved in visual and multimodal processing [bib_ref] Invariant visual object and face recognition: neural and computational bases, and a..., Rolls [/bib_ref] [fig_ref] Figure 1: Anatomical location of consistently different FC in depression obtained from the voxel-based... [/fig_ref]. These decreases of these functional connectivities were correlated with the severity of the symptoms and the illness duration [fig_ref] Table 3: Correlations between the FC links and the depression symptom severity scores In... [/fig_ref] ; Supplementary [fig_ref] Figure 1: Anatomical location of consistently different FC in depression obtained from the voxel-based... [/fig_ref] , and these functional connectivities were higher in medicated than unmedicated patients (Supplementary [fig_ref] Figure 2: The voxel-level FC for amygdala voxels that are significantly different in the... [/fig_ref]. This is thus strong evidence that the reduction of connectivity between temporal cortex areas and the amygdala is important in depression. These temporal cortex areas may introduce inputs relevant to emotion to the amygdala, in that neurons in the primate inferior temporal visual cortex respond to faces [bib_ref] Visual neurons responsive to faces in the monkey temporal cortex, Perrett [/bib_ref] [bib_ref] Invariant visual object and face recognition: neural and computational bases, and a..., Rolls [/bib_ref] , and similar neurons are found in the amygdala [bib_ref] Neurons in the amygdala of the monkey with responses selective for faces, Leonard [/bib_ref] , linking these regions to emotional responses to faces. The hypothesis is that these temporal cortical areas provide important inputs to the amygdala, and that backprojections from the amygdala reach these areas and also earlier cortical including occipital visual areas [bib_ref] Amygdalo-cortical projections in the monkey (Macaca fascicularis), Amaral [/bib_ref] [bib_ref] Anatomical organization of the primate amygdaloid complex, Amaral [/bib_ref].
Fourth, some amygdala voxels had reduced FC with the middle cingulate cortex, involved in motor function, in depression. This pathway has been identified in macaques, and it has been suggested is involved in influences of amygdala face processing subsystems [bib_ref] The representation of information about faces in the temporal and frontal lobes, Rolls [/bib_ref] on emotional face expressions visual, some occipital areas (see [fig_ref] Table 1: A summary of the demographic information and the psychiatric diagnosis in the... [/fig_ref]. Voxels with different FC from controls are shown by the blue shading, with decreases evident in depression. associated with social communication and emotional constructs such as fear, anger, happiness and sadness [bib_ref] Amygdala interconnections with the cingulate motor cortex in the rhesus monkey, Morecraft [/bib_ref]. Interestingly, no effects were found relating the amygdala in depression to a different cingulate area involved in reward and pleasure, the anterior cingulate cortex [bib_ref] Value, pleasure, and choice in the ventral prefrontal cortex, Grabenhorst [/bib_ref]. This again emphasizes the importance of the orbitofrontal cortex and the regions connected to it in depression [bib_ref] A non-reward attractor theory of depression, Rolls [/bib_ref].
Fifth, the results of the parcellation of the amygdala based on its voxel-level FC in 350 healthy controls are of great interest. The basal division of the amygdala (subdivision 3, red in [fig_ref] Figure 3: Voxel-level parcellation of the amygdala [/fig_ref] has high FC with medial orbitofrontal cortex areas, and the dorsolateral part of the amygdala (subdivision 2, yellow) has especially strong FC with the lateral orbitofrontal cortex and its related ventral parts of the inferior frontal gyrus. In depression, the basal amygdala division had especially reduced FC with medial orbitofrontal cortex areas; and the dorsolateral amygdala subdivision (2, yellow) had relatively reduced FC with lateral orbitofrontal cortex and related areas. These differences are of interest, for the medial orbitofrontal cortex is involved in reward, and the lateral orbitofrontal cortex in non-reward and punishment [bib_ref] Value, pleasure, and choice in the ventral prefrontal cortex, Grabenhorst [/bib_ref] [bib_ref] The orbitofrontal cortex and emotion in health and disease, including depression, Rolls [/bib_ref]. At the cytoarchitectonic level, three main divisions have been described [bib_ref] Cytoarchitectonic mapping of the human amygdala, hippocampal region and entorhinal cortex: intersubject..., Amunts [/bib_ref] , a centromedial group (the central nucleus and medial nucleus) which may correspond to our subdivision 1; a superficial group (which may correspond to our functional subdivision 2); and a laterobasal group (which may correspond to our functional subdivision 3). However, as a cytoarchitectonic study, that did not reveal evidence about the connectivity of the three subdivisions of the amygdala, which is provided by the present results.
A strength of this study is the large number of participants, which enabled robust voxel-level FC to be analysed, enabling identification of precisely defined parts of brain areas that had altered connectivity that was related to the depression, such as the lateral part of the lateral orbitofrontal cortex. A limitation is that it would be useful to extend this investigation to activations of the amygdala; and also to effective (that is directed) connectivity [bib_ref] Estimation of directed effective connectivity from fMRI functional connectivity hints at asymmetries..., Gilson [/bib_ref]. From the present investigation, it appears that the main routes via which the amygdala has altered connectivity with other brain regions that may contribute to depression are via the medial orbitofrontal cortex, the lateral orbitofrontal cortex, the medial temporal lobe memoryrelated areas, and the temporal cortex. Another possible route is via the basal ganglia (see [fig_ref] Table 2: Numbers of voxels in different AAL2 areas with significantly different FC with... [/fig_ref] , which have connections with the habenula, which in turn provides a route for cortical areas and the amygdala to influence serotonergic and dopaminergic neurons [bib_ref] The roles of the orbitofrontal cortex via the habenula in non-reward and..., Rolls [/bib_ref]. The relatively few differences in the amygdala FC in depression with some other areas implicated in depression, including the anterior and subcallosal cingulate cortex [bib_ref] Positron emission tomography imaging in depression: a neural systems perspective, Mayberg [/bib_ref] [bib_ref] The subcallosal cingulate gyrus in the context of major depression, Hamani [/bib_ref] [bib_ref] Neuronal coding of implicit emotion categories in the subcallosal cortex in patients..., Laxton [/bib_ref] was notable, and in contrast to the orbitofrontal cortex [bib_ref] Medial reward and lateral non-reward orbitofrontal cortex circuits change in opposite directions..., Cheng [/bib_ref]. Moreover, and consistently, the changes of FC of the amygdala in depression were less significant than the changes in other areas such as the medial orbitofrontal cortex, lateral orbitofrontal cortex and parahippocampal gyrus [bib_ref] Medial reward and lateral non-reward orbitofrontal cortex circuits change in opposite directions..., Cheng [/bib_ref]. A possible limitation of the study was that acutely depressed and remitted depressed patients were included, and there might be differences. Further, although some of the patients were on medication, a highlight of the study was that we were able to examine the differences in a much larger group of unmedicated patients than has ever been studied previously. The importance of the present study is that by focusing on the amygdala, and using very large neuroimaging datasets of patients with depression and controls, we were able to characterize the altered FC in depression of the amygdala with other brain regions.
[fig] Figure 1: Anatomical location of consistently different FC in depression obtained from the voxel-based Association Study (vAS). Voxels showing the largest number of voxel-level FC differences with the amygdala in patients with depression. The colour bar represents the measure of association (MA) given by the number of significantly different FC links relating to each voxel. Voxels with MA larger than 10 are indicated here and elsewhere to show where the main differences are between patients with depression and controls. The right of the brain is on the right of each slice. The Y values are in MNI coordinates. [/fig]
[fig] Figure 2: The voxel-level FC for amygdala voxels that are significantly different in the depressed and the control group, separated by the AAL2 region in which the significant voxels were located. For each AAL2 area illustrated, the left six slices through that area at the MNI Y level indicated show the locations of the voxels with different FC with the amygdala. The right four slices at Y¼À2, 1, 4 and 7 show the amygdala voxels with different FC in depressed patients compared to controls for that brain area. MA values are shown. Voxels with decreased FC are shown in blue, and with increased FC in red/yellow. Voxels are indicated where the FC with the paired region is P < 0.05 (FDR corrected). For the Medial OFC (orbitofrontal cortex) subdiagram the AAL2 areas included were OFCmed, OFCant, OFCpost, Rectus and OLF (see [/fig]
[fig] Figure 3: Voxel-level parcellation of the amygdala. Voxel-level parcellation in controls. (A) The three subdivisions shown on coronal slices at Y¼À4 (top) and Y¼À1. The left side of the brain is on the left of the images. (B) Voxel-level parcellation of the left amygdala (AMYG) based on its FC in healthy controls with other brain areas. The polar plot shows the correlations of the voxels in each subdivision of the amygdala with the significantly different voxels in orbitofrontal cortex AAL2 areas. FC (r) values are indicated by the distance from the centre of the polar plot, with the scale shown indicating the r value. A two-way repeated measures analysis of variance (ANOVA) showed by the interaction term (P < 0.0001) that the three amygdala subdivisions had different FC with these orbitofrontal cortex areas. (C) Voxel-level parcellation of the right amygdala based on its FC in healthy controls with other brain areas. The polar plot shows the correlations of the voxels in each subdivision of the AMYG with the significantly different voxels in orbitofrontal cortex AAL2 areas. The interaction term in the ANOVA was again significant. Differences in FC for these three divisions of the amygdala in depression. (D) The mean z value for the difference in FC (patients with depression-healthy controls) of the links between voxels in each subdivision and the significant ROIs showed in [/fig]
[fig] Figure 4: Summary of amygdala FC differences in depression. The amygdala networks that show different FC in patients with depression. Left: ventral view. Ventral view of the brain. A decrease in FC is shown in blue, and an increase in red, at the voxel level, with the scale shown on the right calibrated using the measure of Association of each voxel (MA, see text). AMYG, amygdala; HIP, hippocampus; ITG, inferior temporal gyrus; MCC, mid-cingulate cortex; motor, pre-and post-central gyrus and Rolandic operculum; OFC, orbitofrontal cortex; PHG, parahippocampal area; FFG, fusiform gyrus; TPO, temporal pole; [/fig]
[table] Table 1: A summary of the demographic information and the psychiatric diagnosis in the present study HAMD, Hamilton Depression Rating Scale; BDI, Beck Depression Inventory; Mean FD ¼ mean framewise displacements. [/table]
[table] Table 2: Numbers of voxels in different AAL2 areas with significantly different FC with amygdala voxels in patients with depression [/table]
[table] Table 3: Correlations between the FC links and the depression symptom severity scores In the top row of the table, the correlations are shown for the average of all FDR significant amygdala voxels with the average of all FDR significant medial orbitofrontal voxels from the medial OFC AAL2 areas, and these statistics are significant when considering the correlations with the symptoms of just the amygdala and medial orbitofrontal cortex voxels (FDR, P < 0.05). The rest of the table provides, for illustrative purposes, correlations of the symptoms with the average strength of the FC between significant voxels in the amygdala with those in other AAL2 brain areas. The values in the lower part of the Table in bold font are significant at P < 0.05 (permuta- [/table]
|
Access to the NHS by telephone and Internet during an influenza pandemic: an observational study
Objectives: To examine use of a novel telephone and Internet service-the National Pandemic Flu Service (NPFS)-by the population of England during the 2009-2010 influenza pandemic. Setting: National telephone and Internet-based service.Participants: Service available to population of England (n=51.8 million).Primary and secondary outcome measures:Primary: service use rate, by week. Numbers and agespecific and sex-specific rates of population who: accessed service; were authorised to collect antiviral medication; collected antiviral medication; were advised to seek further face-to-face assessment. Secondary: daily mean contacts by hour; proportion using service by telephone/Internet.Results:The NPFS was activated on 23 July 2009, operated for 204 days and assessed 2.7 million patients (5200 consultations/100 000 population). This was six times the number of people who consulted their general practitioner with influenza-like illness during the same period (823 consultations/100 000 population, rate ratio (RR)=6.30, 95% CI 6.28 to 6.32). Women used the service more than men (52.6 vs 43.4 assessments/1000 population, RR1 21, 95% CI 1.21 to 1.22). Among adults, use of the service declined with age (16-29 years: 74.4 vs 65 years+: 9.9 assessments/1000 population (RR 7.46 95% CI 7.41 to 7.52). Almost three-quarters of those assessed met the criteria to receive antiviral medication (1 807 866/ 2 488 510; 72.6%). Most of the people subsequently collected this medication, although more than onethird did not (n=646 709; 35.8%). Just over one-third of those assessed were advised to seek further face-toface assessment with a practitioner (951 332/ 2 488 504; 38.2%).Conclusions: This innovative healthcare service operated at large scale and achieved its aim of relieving considerable pressure from mainstream health services, while providing appropriate initial assessment and management for patients. This offers proof-of-concept for such a service that, with further refinement, England can use in future pandemics. Other countries may wish to adopt a similar system as part of their pandemic emergency planning.Rutter P, Mytton O, Ellis B, et al.
# Introduction
England started planning for the emergence of a new influenza virus with pandemic potential following a warning by the Chief Medical Officer in 2002 about the risks of being unprepared.The plan foresaw that widespread infection would cause a surge in demand for primary care and hospital services that could quickly overwhelm them. Planning, therefore, focused on a new way of providing first response to patients with influenza-like illness (ILI) during a pandemic. The concept involved telephone call centres and the Internet and evaluation of symptoms, provision of advice and authorisation of antiviral medication in appropriate clinical circumstances. This was to be the first time in the UK that patients would be assessed and given a prescription medication without being seen by a doctor or nurse.The system was intended to keep general practitioners free to see patients who developed serious illness as a result of influenza infection, but to continue also managing their normal caseload.
On 11 June 2009, the WHO formally declared the first influenza pandemic for 40 years.The virus had first emerged in Mexico in March 2009. By May of that year it had spread worldwide. The first case in the UK occurred in Scotland in late April, and was followed shortly by the first cases in England. The UK response was characterised by two phases-an initial
Strengths and limitations of this study ▪ Complete data ascertainment from the 2.7 million people who used the National Pandemic Flu Service. ▪ Data were self-reported, without clinical validation or follow-up. ▪ Comparisons with GP consultation data illustrate how healthcare-seeking behaviour during the pandemic differed by age and sex.
containment response from May to June 2009 and a subsequent mitigation (or 'treatment-only') phase.The containment phase involved a proactive response, with the aim of limiting and delaying the virus' spread. It included school closures, early treatment of suspected cases based on clinical criteria and contact tracing to provide prophylaxis to close contacts of cases. The incidence of influenza began to rise in mid-June. By July 2009, ILI was putting great pressure on the country's National Health Service (NHS), particularly general practice (GP), which in England provides the majority of front-line primary care. In response, the government activated its emergency plan to establish this national telephone and Internet-based service. We report on the experience with this novel service for assessment and treatment of pandemic influenza, and the way that it was used by the public.
## Context: the national pandemic flu service
The service, first activated on 23 July 2009 during the mitigation phase of the public health response, was named the National Pandemic Flu Service (NPFS). Any member of the public with 'flu-like' symptoms could access it by telephone or on the Internet (Internet users answering a series of questions on-screen). Telephone users were asked the same questions by a call centre operative.
Each assessment consisted of a set of yes/no questions. Via a standard algorithm, the responses to these questions determined the advice given to each patient and whether or not a course of antiviral medication was authorised for the patient. Each patient passed through an assessment of up to three stages (see [fig_ref] Figure 1: Algorithm for telephone/Internet-based assessment [/fig_ref].
The first stage was a preassessment screen, to redirect patients in need of urgent personal medical attention. Patients reporting features such as unconsciousness, seizures or meningococcal-like rash were told to call an emergency ambulance rather than continue using the service. Patients who had recently returned from a malarial region were directed to see their general practitioner urgently instead of using the service, as were all children under 1 year and pregnant women with a history of respiratory disease. All other patients entered the second stage of the assessment.
The second stage identified whether the patient had the classic symptoms of influenza. The question asked was: "Does the patient have a high temperature and at least two of the following symptoms?-Widespread muscle and joint aches, a cough, headache, blocked or runny nose, sore throat, vomiting, watery diarrhea, cannot stop crying (only children)". Those who answered 'yes' progressed to the third stage of the assessment. The remainder were informed that they were unlikely to have influenza and their contact with the service ended, with advice being given.
Patients arriving at the third assessment stage were, therefore, those identified as having ILI. In this third stage, further information was collected on four factors: how long they had been symptomatic; whether antivirals had already been taken; risk factors predisposing to severe illness (such as chronic lung disease and immunocompromise) and symptoms suggestive of severe illness (cough productive of yellow, green, brown or bloody phlegm; uncharacteristic behaviour change; pleuritic chest pain; respiratory distress; drowsiness). Those who had been ill for more than a week were also asked whether they had non-'flu-like' symptoms of acute illness (eg, persistent vomiting or severe earache).
The answers given in the third stage determined whether or not the patient was given an authorisation number to obtain a course of antiviral medication. The answers also determined what advice was given to the patient, including whether or not they were also advised to contact another healthcare service-usually their general practitioner, but sometimes a midwife (if pregnant) or a renal specialist (those with pre-existing renal disease).
A network of antiviral distribution centres was established around England, in locations such as community centres, sports centres and pharmacies. Any assessed patient who required antiviral medication was asked to find a 'flu friend' (a relative, friend or neighbour) to collect this on their behalf from a distribution centre, using the authorisation number provided by the service. Collection by a third party was intended to reduce the spread of infection at crowded distribution points.
The algorithm had been developed as part of pandemic planning, and was refined when clinical information about the influenza A (H1N1) pdm09 virus was available. The algorithm was developed and refined by a group involving clinical experts from a range of organisations, including the Health Protection Agency and the Royal College of General Practitioners (RCGP), based on their clinical experience and review of the available evidence. Legally classified as a 'medical device', the algorithm underwent fast-track regulatory approval by the Medicines and Healthcare Regulatory Authority. The Department of Health commissioned the RCGP to assure the quality of the system. General practitioner liaison advisers, coordinated by the RCGP, were based in the service's call centres.
Throughout the service's operation, the algorithm was kept under review by the clinical subgroup of the Department of Health's Pandemic Influenza Clinical and Operational Group. Refinements to the algorithm were introduced to improve safety and to minimise disruption to other services. These refinements were made based on the service's evolving experience and emerging information about the pandemic. During the service's operation, the Department of Health monitored reports of clinical incidents from a number of sources including the National Patient Safety Agency (NPSA); the call centre-based GP advisers; Strategic Health Authority and Primary Care Trust Flu Leads; the Health Protection Agency; clinical networks; and the public.
The decision to activate the service was taken on the basis of rising rates of ILI, coupled with information provided through close liaison with local public health officials, who were monitoring the pressure being experienced by primary care services in each area of the country. The decision to deactivate the service was taken when ILI rates were approaching background levels, and use of the service had diminished very substantially.
# Methods
Data from the telephone and Internet arms of the service were captured on a single database. For every patient who accessed the service, this recorded the date and time of accessing the service; basic demographic information (date of birth, sex); responses to each question and, therefore, the outputs of the algorithm. These data provided surveillance information during the pandemic in real time. Every collection of a course of antiviral medication was also recorded.
For comparison purposes, we also obtained data about GP consultations for ILI. The RCGP has undertaken ILI surveillance in GP for over 40 years. The system uses around 100 sentinel GPs across England, covering a population of approximately 800 000. It extracts summary information (based on Read codes 7 ) from GP electronic records. This is used to estimate the rate of ILI consultations in the population of England as a whole. [bib_ref] The representativeness of sentinel practice networks, Fleming [/bib_ref] These estimates, by week of consultation, were supplied by the RCGP Research and Surveillance Centre. Based on these estimated rates, we estimated the number of GP ILI consultations in the whole of England by week and compared this to the number of NPFS consultations by week.
Population sizes by age group were obtained from the Office for National Statistics estimates for mid-2009. Comparisons between groups were made using two-sided t tests. The Exact method was used to calculate 95% CIs for rate ratios (RR). Data were analysed in Excel 2011 and STATA V.12.1.
This analysis was part of an enhanced public health surveillance programme during the pandemic. No explicit ethical approval was required for data collection because it had a statutory basis under the Health Service (Control of Patient Information) Regulations SI1438/2002.
# Results
The NPFS was activated across England on 23 July 2009. It remained operational for 204 days, until 10 February Contacts occurred in two clear waves (figure 2). The first was short, covering June and July 2009. The second was more prolonged between September and December 2009. The rate of NPFS contacts at the peak of the first wave was more than three times higher than in the second wave (135 vs 38 consultations/100 000 population/day, RR 3.59, 95% CI 3.53 to 3.65).
Over the same period, there were just 429 000 GP consultations related to ILI in England (827 consultations/ 100 000 population). These also occurred in two waves (figure 2). NPFS contacts exceeded GP consultations by a factor of almost six (RR=5.76, 95% CI 5.72 to 5.80). The difference was most marked in the service's second week of operation (748 vs 41.1 consultations/100 000 population/week, RR=18.2, 95% CI 17.6 to 18.9) and subsequently diminished over time.
The number of GP consultations related to influenzalike illness had been rising sharply in June and early July 2009, before the NPFS was activated. In the week beginning, 13 July 2009, there were 154 GP consultations/ 100 000 population/week (80 000 consultations in total). During the subsequent week, the NPFS was activated and the number of GP consultations fell to 133/100 000 people/week (69 000 consultations in total).
Just over half of patients (1 305 902/2 488 504; 53%) used the Internet service arm; the remainder used the telephone. Each method had distinct daily usage patterns (figure 3). Usage of both peaked between 9:00 and 10:00. Just under one-quarter (22%; 279 150/1 250 184) of telephone contacts happened at other times. Usage of the Internet-based service was more evenly spread throughout the day. Just under half of its usage was outside '9-5' h (43%, 639 684/1 482 404). In total, an average of 13 500 assessments were completed per day. The mean daily volume of assessments was higher during the week than at the weekend (14 600 vs 10 600, p<0.0001).
Of the 2.73 million contacts with the NPFS, 9% (n=244 078, figure 4) were immediately diverted because they related to a child aged under 1 year, a pregnant woman with respiratory disease or because the patient described features of a potential medical emergency. The remaining patients entered the full assessment process. Almost three-quarters of these (1 807 866/2 488 510; 73%) received authorisation to collect a course of antiviral medication. Most did subsequently collect this, although a substantial minority did not (n=646 709; 36%). The algorithm determined that almost two-thirds of patients (n=1 537 178; 62%) did not require further assessment by a clinician. These patients received selfcare advice. The remainder were advised to seek in-person medical assessment (n=951 332; 38%), the form of which depended on its indication (figure 4).
Use of the NPFS varied demographically [fig_ref] Table 1: National Pandemic Flu Service [/fig_ref]. Women used it more than men (52.6 vs 43.4 assessments/1000 population, RR 1.21, 95% CI 1.21 to 1.22), but men appeared to show a greater preference than women for the NPFS relative to visiting their GP (ratio of NPFS:RCGP 6.1 for men vs 5.6 for women, table 2). Use declined with age: the assessment rate was seven times greater among 16-29-year-olds than among the over-65s (74.4 vs 9.9 assessments/1000 population, RR 7.46, 95% CI 7.41 to 7.52). Preferential use of the NPFS rather than GP was particularly high among the young and middle-aged [fig_ref] Table 2: Comparison of GP consultation and NPFS assessment rates [/fig_ref].
Men were offered antiviral medication slightly more often than women (74% for men vs 72% for women, χ 2 =1812, p<0.0001). Antiviral medication authorisations were highest for the age group between 1 and 15 years (78%, table 1), generally reduced with age (χ 2 =6.1×10 5 , p<0.0001) and were lowest for the over-65s (59%).
Of those eligible to collect a course of antiviral medication, just under two-thirds did so (1 161 156/1 807 866; 64%). Those aged 16-29 years were the least likely to collect their medication (317 155/528 211; 60%) and those aged 65 years and over the most likely (36 182/ 49 152; 74%). When a child aged between 1 and 15 years was deemed eligible for antiviral medication, this medication was collected in 64% of cases.
Men were slightly more likely than women to access the service via the Internet rather than telephone (53% vs 52%, χ 2 =288.61, p<0.0001) but variation between age groups was greater (figure 5). Young adults were more likely to opt for Internet-based access (25-44 years, 59%) and the oldest group least likely to do so (75 years and over, 28%, RR 2.1 95% CI 2.0 to 2.1).
# Discussion
The National Pandemic Flu Service was designed, in advance of the pandemic, to prevent the NHS being overwhelmed by people seeking medical advice and treatment of 'flu-like' symptoms due to heightened public awareness and concern. By providing antiviral medication to 'flu friends' at distribution centres, it was intended to enable the sick to stay at home and thereby reduce the spread of infection. It was planned to quickly provide antiviral medication to those who might benefit from it.
When the pandemic came and this new service was activated, the population proved very willing to use it. The service dealt with a high volume of contacts, assessing 2.7 million patients in less than 7 months. It distributed antiviral medication to over one million people. We suggest that its activation and use in the 2009-2010 pandemic provides successful proof-of-concept for this approach.
A strength of this article is that it uses data on all patients who accessed the service. A weakness, because of the nature of the service, is that the information was reported by patients and was not clinically validated. Apart from the incident reporting systems described later, there are no patient follow-up data to assess outcomes or confirm underlying diagnoses. Results of microbiological sampling of patients who used the service have been published elsewhere.Also, our estimates for GP consultations are based on the RCGP sentinel surveillance system. Owing to the nature of sampling this may not accurately reflect the true picture across England.
## Service operation
The cost of developing and maintaining the NPFS was the main infrastructure-related expenditure associated with the country's pandemic preparedness and response. The total cost of infrastructure in the UK (including some non-NPFS costs) was £27.7 million during the preparedness phase and £65.7 million during the response phase.Although the telephone service was established at relatively short notice, a great deal of planning had occurred during an extended prepandemic period. The government contracted with an existing telephone triage service that was experienced in training temporary staff to use health-related algorithms and was able to scale service provision up and down according to demand.
The UK had a regular direct telephone and email service for health consultations called NHS Direct. The telephone service handled around five million calls each year and around five million people used the online symptom checker. Prior to the opening of the NPFS, some patients were contacting NHS Direct. The proportion of calls to NHS Direct that were triaged using the cold/flu algorithm rose steadily during June and July, matching the rise in general practitioner consultations. Outside the 'flu season' the proportion is typically less than 2%. This proportion peaked at 35% on 14 July, likely equivalent to around 4000-6000 daily calls (or 8-12 calls/100 000). This is substantially lower than peak NPFS usage when it first opened (over 130 assessments/ 100 000/day, around half of which were by phone). The greater propensity to contact the NPFS may be because of its ability to provide access to antiviral medication (which NHS Direct was not), or because it was specifically advertised in relation to the pandemic.
## Impact on primary care workload
The NPFS was activated after the GP consultation rate for ILI had risen for 7 weeks, and had doubled over the previous week. The weekly rate had risen to over 150 consultations/100 000 population. Although not as high as in previous influenza pandemics, this was the highest level experienced in England in a decade of 'flu seasons'. Moreover, the consultation rate was 5-10-fold higher than the national average in some areas of the country,causing huge pressure on primary care services in some areas.
The GP consultation rate fell in the week that the NPFS was activated (week beginning 20 July 2009), and again in its second week. We cannot conclusively attribute this to the introduction of the NPFS. Other measures suggest that influenza activity may have peaked at around the same time: hospital admissions (week beginning 20 July), critical care admissions (week beginning 27 July), deaths (week beginning 27 July) and Google searches for influenza (week beginning 27 July). [bib_ref] Influenza A (H1N1) pdm09 in England, Mytton [/bib_ref] However, the fall in GP consultations was far more pronounced than the fall in these other measures. The clear impression from general practitioners is that the service played an important role in reducing pressure on GP at a very busy time. COMPARISON WITH OTHER UK COUNTRIES Wales, Scotland and Northern Ireland chose not to be part of the NPFS. Comparisons with these countries may give an insight into the effect of the NPFS in reducing GP consultations. While there are important geographical differences between the countries their health systems, although separate, are broadly comparable.
Wales and Northern Ireland had a similar experience to England. Wales had a summer peak (92.8 consultations/ 100 000 population) in the same week as England (week beginning 20 July), followed by a sustained second wave in the autumn (peaking at 66.2 consultations/100 000 in the week beginning 12 October).In terms of GP consultations, the ratio between the summer and autumn peaks was 1.4 in Wales compared with 3.8 in England. This may suggest that the NPFS succeeded, in England, in reducing GP consultations during the second wave.
Northern Ireland also experienced two waves. The peak of the first wave was later than in England and Wales (week beginning 27 July). Unlike in England and Wales, the second wave had a higher peak than the first wave. This too might suggest that the NPFS in England played a role in keeping GP consultations low in the second wave. Wales is arguably a better epidemiological comparator than Northern Ireland, because Wales and England adjoin whereas Northern Ireland and England are on separate islands. The Scottish experience was very different. In Scotland, there was no summer peak in GP consultations. As in the other UK countries, there was a sustained autumn wave that peaked in October. The absence of a summer peak in Scotland may partly be explained by the earlier closure of the schools for summer holidays.Because the trajectory of the 'flu' pandemic appeared very different in Scotland, it is not appropriate to make comparisons with England.
Differences between the four nations in the pressure created by the pandemic were apparent from an early stage. In England, there was marked pressure on primary care, particularly in London and the West Midlands. By contrast, the pressure on primary care in Wales, Scotland and Northern Ireland was less intense. It was for these reasons that England chose to use the NPFS, while the other nations continued to rely purely on their usual primary care systems. 4
## Telephone and internet use in the nhs
In the UK, most households have Internet access, and more than two-thirds seek health information online. [bib_ref] The characteristics and motivations of online health information seekers: cross-sectional survey and..., Powell [/bib_ref] Telephone consultations have been available throughout England since the launch of NHS Direct in 1998. Staffed by nurses and health advisors, this system handled over four million calls per year, and its website received 10 million hits. NHS Direct was shown to reduce demand on out-of-hours GP services. 14 Its scale and electronic medical records also made it useful for surveillance, including influenza surveillance. Subsequently, in 2013, NHS Direct was replaced by NHS 111, another telephone and Internet-based service.
The NPFS was launched to a public that has become increasingly familiar with seeking medical help and information by phone and through Internet, as it does in many other aspects of life. But the NPFS was the first system in England-and, to our knowledge, the first national system in the world-in which non-medically trained lay people acted as the primary interface with patients, working to a standard algorithm and providing a drug that normally requires medical prescription, as deemed appropriate.
## The risks of algorithm-based care
During the service's operation, concerns were raised about the possibility of misdiagnoses, particularly that led to serious illness being missed. The number of such incidents that were reported was small. The NPSA was informed of 72 incidents. By comparison, it was informed of 2410 incidents pertaining to pandemic-related healthcare as a whole. The Agency reviewed the most serious 186 of these 2410 incidents in detail. Only one serious incident was found to be related to a delay in diagnosis following use of the NPFS-that is one patient in 2.7 million who accessed the service. Diagnostic error, including delayed diagnosis, is a common problem within conventional systems of assessment and management. The RCGP concluded that "the first mass application of non-clinical-based triage appears to have been a qualified success story", and recommended amendments for future pandemics, based on the 2009 experience.A Department of Health evaluation set out specific details of problems encountered, and also laid out modifications to planned NPFS use in future pandemic plans.
## Patterns of usage
The highest NPFS usage rate was seen among those aged 16-29 years. The rate then decreased with age, and was also lower among children. The decrease with age among adults is consistent with the pattern of GP consultations and hospital admissions. [bib_ref] Influenza A (H1N1) pdm09 in England, Mytton [/bib_ref] However, the magnitude of this relationship was greater for the NPFS than it was for GP. The NPFS usage rate fell by a factor of five between the ages of 44 and 65 years. The GP consultation rate fell by a factor of 3. This suggests that younger people preferred newer, more immediate ways of accessing the NHS while older people preferred the familiarity of their general practitioner. It may also be because older people have comorbid conditions, and therefore felt that they needed a face-to-face consultation with a doctor. The preferential use of the NPFS by young and middle-aged people may suggest that the system has value as a surveillance tool in those age groups.
The relatively low NPFS usage rate for children despite the high incidence of influenza among children 6 may reflect a lack of awareness that parents could use the NPFS on behalf of their children or a parental preference to access usual medical care for their children.
One-third of those authorised to receive antiviral drugs through the NPFS did not subsequently collect this prescription. Collection rates for drugs prescribed by a general practitioner in England are far higher than this (97% of drugs prescribed for infections are collected, although other authors have also shown that around 30-50% of patients with a long-term condition are non-compliant with medication. [bib_ref] Suitability of measures of self-reported medication adherence for routine clinical use: a..., Garfield [/bib_ref] However, in the case of the NPFS it is perhaps inevitable that a proportion of users of the service chose found that a selflimiting illness was running its course and so did not follow through with treatment.
There is a notable difference between the number of people who were advised by the NPFS to seek further medical attention and the number of people who did so. The service advised 1.2 million people to see their general practitioner. As we have reported elsewhere, [bib_ref] Influenza A (H1N1) pdm09 in England, Mytton [/bib_ref] there were a total of 580 000 GP consultations relating to ILI during the pandemic. Although some patients may have attended hospital emergency departments, this strongly suggests that a large number of the 1.2 million patients advised to seek further medical attention did not do so. Further research is required to understand the full scale of this, the reasons for this and the implications of it.
## Widespread use of antiviral medication
The NPFS made antiviral medication available to a large number of people with ILI, without the need to consult a medical practitioner. The number of such prescriptions made through the NPFS (1 161 157) dwarfed the number dispensed by community pharmacists on the prescription of a general practitioner (10 610 from June 2009 to February 2010). [bib_ref] Influenza A (H1N1) pdm09 in England, Mytton [/bib_ref] It seems unlikely that such high availability of prescription antiviral medication could have been achieved using the general practitioner system alone without this causing very severe disruption.
Criticism of heavy use of antiviral medication centres on concerns about antiviral resistance from their 'unnecessary' use in what was perceived as a relatively mild influenza syndrome. While a small number (45) of antiviral-resistant cases were reported,there was no evidence of sustained person-to-person transmission of antiviral-resistant strains in the UK despite extensive testing and monitoring. And while it is correct that case fatality rates were lower than in previous influenza pandemics, [bib_ref] Mortality from pandemic A/H1N1 2009 influenza in England: public health surveillance study, Donaldson [/bib_ref] there were many cases of severe illness requiring intensive care [bib_ref] Hospitalization in two waves of pandemic influenza A (H1N1) in England, Campbell [/bib_ref] and 70 children in England died. [bib_ref] Paediatric mortality related to pandemic influenza A H1N1 infection in England: an..., Sachedina [/bib_ref] It has also been noted that the positivity rate among those authorised to collect antivirals by the NPFS were just over 30% in adults and just over 50% in children (although children were only sampled after the peak of the second wave). However, these rates were very similar (after age-adjustment) to the positivity rates among GP ILI consultations.In addition, the NPFS sampling scheme relied on nasal swabs taken by the patient without clinical supervision, rather than sputum or nasopharyngeal aspirates which have higher positivity rates. [bib_ref] Review of rapid diagnostic tests for influenza, Gavin [/bib_ref] The stated NPFS positivity rates may, therefore, underestimate the true incidence of infection.
In the first pandemic of modern times when a prevaccine protective option was available in the early stages of the pandemic, we would argue that it was right to allow wide access to it, and that more deaths of young people and higher levels of serious complications were potentially prevented.
The NPFS gave patients a rapid access to antiviral medication, facilitating early use. Early use is thought to be critical for reducing illness severity and intensity in the individual. [bib_ref] Early versus late oseltamivir treatment in severely ill patients with 2009 pandemic..., Kumar [/bib_ref] Also, the widespread use of antiviral medication in the community, alongside other measures, may have contributed to limiting the burden of influenza, particularly in the second wave. 34 Antiviral medication may reduce transmission of the virus, [bib_ref] Use of antiviral drugs to reduce household transmission of pandemic (H1N1) 2009,..., Pebody [/bib_ref] particularly among children, [bib_ref] Neuraminidase inhibitors for preventing and treating influenza in children, Wang [/bib_ref] and widespread use at a population level might reduce the overall attack rate. [bib_ref] Analysis of the effectiveness of interventions used during the 2009 A/H1N1 influenza..., Halder [/bib_ref] In facilitating widespread early treatment, the NPFS may have contributed to preventing adverse outcomes, such as hospitalisation, critical care admission and death. CONCLUSION A total of 2.7 million people accessed the NPFS by telephone or Internet. They received advice and, in two-thirds of cases, authorisation to collect antiviral medication. The service seemed to appeal particularly to younger patients. Pressure on primary care fell dramatically after its introduction, although this cannot wholly be attributed to the service. Together with the formal evaluation and review of safety data, the experiences presented here offer evidence of the feasibility of rapid, national implementation of a telephone-based and Internet-based triage and antiviral authorisation service, based on non-clinically trained operators following defined algorithms. With further refinement, England can use this in future pandemics (which could be of much greater severity). Other countries may wish to adopt a similar system as part of their pandemic emergency planning.
[fig] Figure 1: Algorithm for telephone/Internet-based assessment. Rutter P, Mytton O, Ellis B, et al. BMJ Open 2014;4:e004174. doi:10.1136/bmjopen-2013-004174 2010. During this time, it received 2.73 million unique patient contacts (5200 contacts/100 000 population). [/fig]
[fig] Figure 2: Weekly rates of general practice (GP) consultations with influenza-like illness and National Pandemic Flu Service consultations, by week, in England, April 2009-February 2010. [/fig]
[fig] Figure 3: Daily mean contacts per hour to each arm of the National Pandemic Flu Service. [/fig]
[fig] Figure 4: Patient assessment by the National Pandemic Flu Service (204 days, July 2009-February 2010). Rutter P, Mytton O, Ellis B, et al. BMJ Open 2014;4:e004174. doi:10.1136/bmjopen-2013-004174 [/fig]
[fig] Figure 5: Use of the Internet-based arm of the National Pandemic Flu Service as a percentage of Internet-based and telephone-based use, by population group. [/fig]
[table] Table 1: National Pandemic Flu Service (NPFS) assessments, authorisations for antiviral medication and actual collections of antiviral medication [/table]
[table] Table 2: Comparison of GP consultation and NPFS assessment rates [/table]
|
Responding to the deaf in disasters: establishing the need for systematic training for state-level emergency management agencies and community organizations
Background: Deaf and hard-of-hearing (Deaf/HH) individuals have been underserved before and during emergencies. This paper will assess Deaf/HH related emergency preparedness training needs for state emergency management agencies and deaf-serving community-based organizations (CBOs).Methods: Four approaches were used: 1) a literature review; 2) results from 50 key informant (KI) interviews from state and territorial-level emergency management and public health agencies; 3) results from 14 KI interviews with deaf-serving CBOs in the San Francisco Bay Area; and 4) a pilot program evaluation of an emergency responder training serving the Deaf/HH in one urban community. Results: Results from literature review and state and territorial level KIs indicate that there is a substantive gap in emergency preparedness training on serving Deaf/HH provided by state agencies. In addition, local KI interviews with 14 deaf-serving CBOs found gaps in training within deaf-serving CBOs. These gaps have implications for preparing for and responding to all-hazards emergencies including weather-related or earthquake-related natural disasters, terrorist attacks, and nuclear-chemical disasters.Conclusion: Emergency preparedness trainings specific to responding to or promoting preparedness of the Deaf/ HH is rare, even for state agency personnel, and frequently lack standardization, evaluation, or institutionalization in emergency management infrastructure. This has significant policy and research implications. Similarly, CBOs are not adequately trained to serve the needs of their constituents.
## Significance
Almost no information exists in the peer-reviewed literature about the emergency preparedness training standards and current trainings provided for Deaf and hard-of-hearing (Deaf/HH) populations. However, recent national and international disasters that required humanitarian efforts illustrate the fact that Deaf and hard-of-hearing (Deaf/HH) individuals are often not only underserved, but also particularly vulnerable, in preparing for, responding to, and recovering from emergencies. In the United States, according to a landmark 2004 report by Stout, "a failing grade" was given to U.S. public warning and emergency communications systems serving the Deaf/HH post-9/11. According to the Office of Homeland Security's 2006 Nationwide Plan Review of 2,800 state and local emergency operations plans (EOPs) and related documents which included interviews with over 1,000 public safety and homeland security officials across the US, the word Deaf appeared only 8 times in their entire report.
However, FEMA's Office of Disability Integration and Coordination's list of key concepts provide a starting point at the national level to provide guidance for ensuring people with disabilities are included in, and not left out of, the emergency management infrastructure across the US. These concepts include self-determination, no "One-Size-Fits-All," equal opportunity, inclusion, integration, physical access, equal access, effective communication, and program modifications.
# Background
The Deaf/HH population There are 48 million Deaf, deaf-blind, and hard-of-hearing (HH) people living in the United States [bib_ref] Hearing loss prevalence in the United States, Lin [/bib_ref]. Communication needs vary depending on level of hearing loss and cultural orientation. The diverse communication modalities in this population include American Sign Language (ASL), Signed Exact English (SEE), Pidgin Signed English (PSE), Cued Speech, lip-reading and spoken English. Recent research indicates that this population faces serious health disparities due to communication barriers and low literacy rates, including a higher risk for obesity, depression, and interpersonal violence. In particular, these communication barriers contribute to increased vulnerability in an emergency situation and present unique considerations for emergency responders [bib_ref] Obesity among adults with disabling conditions, Weil [/bib_ref] [bib_ref] Suicide in deaf populations: a literature review, Turner [/bib_ref] [bib_ref] Situating Research on Safety Promoting Behaviors Among Disabled and Deaf Victims of..., Brownridge [/bib_ref]. Members of the Deaf community do not see themselves as disabled but rather as members of a linguistic minority group centered on the use of sign language, which must be taken into account when designing training programs for both emergency responders and to enhance preparedness efforts by community-based organizations (CBOs) serving the Deaf community. Cultural competence is an important consideration. It is defined by the Office of Minority Health's National Standards for Culturally and Linguistically Appropriate Services (CLAS) in Health Care as "a set of congruent behaviors, attitudes, and policies that come together in a system, agency, or among professionals that enables effective work in crosscultural situations."
## Rationale for evaluation research
There is relatively little research or evaluation on the efficacy of preparedness efforts, particularly for those with disabilities. Given the relatively few Deaf/HH specific training programs available and the lack of evaluation of such programs, there is a need to develop best practices for emergency preparedness training for four audiences: 1) state-level emergency management and public health agencies; 2) local emergency responders; 3) Deaf-serving CBOs; and 4) Deaf community members. To date, there have been no published evaluations of trainings for emergency responders or trainings that target Deaf/HH people. In 2011, a pilot mixed methods evaluation was conducted of a training workshop for law enforcement as emergency responders for the purpose of increasing officers' cultural competency in working with Deaf and hard-of-hearing people (Deaf/HH) during domestic and sexual violence (DV/SV) emergencies.
## History of emergency preparedness training
To date, we have identified only fifteen training programs and disaster simulation exercises for the Deaf/ HH across the country targeting emergency responders (5), certified ASL interpreters (2), deaf-blind people (1), government agencies (4) and the Deaf community [bib_ref] Suicide in deaf populations: a literature review, Turner [/bib_ref]. Some training programs target multiple audiences. These training programs range from onetime workshops to more extensive training modules [fig_ref] Table 1: Existing deaf preparedness training workshops and disaster simulation exercises* [/fig_ref].
These types of targeted training efforts for the Deaf/ HH began for the first time during the decade post-9/ 11. For example, in 2006, the Helen Keller National Center provided a training workshop on emergency and disaster preparation for deaf-blind people. The 12-hour multi-session curriculum focused on: disasters that affect different areas of the country; how to set up a personal support network; emergency bag and disaster kit preparation; possibilities to consider in an evacuation situation; communication with emergency responders; use of personal and emergency alert systems; rental and homeowner's insurance; and different aspects of water and food safety.
In addition, in 2010, the first training of its kind in the country began for certified ASL interpreters to work as emergency responders through a partnership between California Emergency Management Agency (CalEMA) and NorCal Center on Deafness. Three courses were offered for approximately 100 Certified ASL interpreters. According to Jordan Scott at CalEMA, when the program was being developed: "it was determined that, because of the chaotic nature of disasters, and the need for accurate and timely communication to the public in a shelter environment or during a press conference, it was important that prerequisites for the participants be established. In order to be eligible for Cal EMA's program, the interpreter must possess a valid certification (CDI; NAD Level 3, 4 or 5; RID CI/CT; NIC Generalist, Advanced or Master Level.) In addition they must have a minimum of 10 years of community sign language interpreting experience, with 5 years of medical, law enforcement or mental health emergency interpreting experience" (Scott J, California Emergency Management Agency, Personal Communication, January 1, 2012). There has not been an opportunity to actually deploy any interpreters during a disaster and Cal EMA is in the process of developing an on-line refresher course so the interpreters can renew their credential status in preparation for future disasters. In general, the scope of these twelve trainings varies considerably according to target audience and geographic location. While the effectiveness and impact of these trainings still need to be evaluated, save one, the variance between them demonstrates a need for more standardization on a national level.
# Methods
We conducted a literature review, a state agency-level assessment of trainings for the Deaf/HH, a local assessment of CBO capacity, as well as a first-responder training evaluation. All interviews were conducted by trained interviewers. Informed consent was obtained from all participants for publication of this report and any accompanying images.
## Literature review
Inclusion criteria for the literature search in peer-reviewed databases included all literature or reports from the United States in English, 1990-2012. The peer-reviewed literature was searched using PubMed and Google Scholar, and grey literature on training evaluation was searched using Google, as well as by examining Community Emergency Response Team (CERT) newsletters. The following search terms were used: "Deaf CERT training," "Deaf emergency training," "Deaf disaster training," and "Deaf hurricane training," "Deaf earthquake," "Deaf flood," "Deaf fire," and "Deaf all-hazards."
## State-level agency training needs assessment
University of California, Berkeley (UCB) researchers interviewed key informants (KIs) from state and territorial level emergency management or public health agencies in order to assess emergency preparedness information and capacity to respond to the Deaf/HH during an emergency. Fifty-nine KIs (all US states, DC, and territories) were sampled, 50 KI telephone interviews were completed and 55 basic State Emergency Operation Plans (EOPs) were obtained and analyzed from agencies. The Office of Public Health Preparedness and Response (OPHPR) Extramural Research Program Offices (ERPO) and the Division of State and Local Readiness (DSLR) assisted in identifying contacts to help us obtain EOPs. State-level participants were assured of full confidentiality of their information prior to participation. For the purposes of this paper, we report analysis of three training-related interview items about departmental or other trainings attended by the KI interviewees.
Other manuscripts available and in preparation will review results from remaining survey items asked of state personnel [bib_ref] Availability and Readability of Emergency Preparedness Materials for Deaf and Hard-of-Hearing and..., Neuhauser [/bib_ref].
## Local-level training needs assessment sampling frame
A sample of 14 deaf-serving CBOs were selected for KI interviews based on expert opinion from the head of a deaf-serving organization in the San Francisco Bay Area in consultation with the project's National Advisory Board (NAB) of leaders from the Deaf/HH community, a deaf graduate student researcher, and a project consultant with certification in ASL interpreting and expert knowledge of the Deaf community in Northern California. These 14 deaf-serving CBOs have on average 5,970 clients (median: 475; range: 75-60,000 clients) and of the clients served, 20-100% are Deaf or HH. For the purposes of this paper, analysis is reported for trainingrelated survey questions only. Questions included emergency preparedness education for clients and staff, as well as client characteristics, CBO capacity to reach their constituents in an emergency, barriers and supports for the development and dissemination of preparedness materials, and partnerships with other organizations. Participants were assured of full confidentiality of their information prior to participating.
## Instrument
We developed a semi-structured interview guide in written English for nine Deaf/HH KIs and translated the guide into ASL Gloss (a written format that approximates ASL grammar, morphology, prosody, and syntax) for five Deaf/HH KIs. A similar interview guide for Deaf women by [bib_ref] Deaf Women: Experiences and Perceptions of Healthcare System Access, Steinberg [/bib_ref] also used ASL Gloss and defines it as standard written format used to represent ASL [bib_ref] Deaf Women: Experiences and Perceptions of Healthcare System Access, Steinberg [/bib_ref]. Before the interview, all KIs were asked how they preferred to communicate, whether it should be in spoken English, Signed Exact English (SEE), or ASL. An ASL Gloss was created for the interview guide in order to ensure that every KI interview was administered by the interviewers in a consistent format, whether in spoken English or in ASL. The ASL Gloss, developed by a Deaf graduate student researcher in conjunction with a certified ASL interpreter, was created to ensure that the instrument was linguistically and culturally appropriate.
## Local law enforcement training evaluation
We also conducted an evaluation of a law enforcement training in Oakland, CA, to promote better response to domestic violence emergencies involving the Deaf/HH. Participants in the workshop included police officers and other law enforcement personnel, including police dispatchers. Data were collected through (1) a pre-and post-test survey [n=34] administered immediately before and after the training, and (2) two semi-structured focus groups [n=6 and n=13] with the same participants. Focus group activities occurred on the same day after two 2-hour educational outreach/training certification workshops for law enforcement personnel in the San Francisco Bay Area. A trained focus group facilitator conducted the focus group. Survey items included a measurement of attitudes, including perceived capabilities of Deaf people, with six items such as "Deaf people can make their own life decisions" and "Deaf people can have normal one-on-one interactions on a daily basis," and perceived self-efficacy when working with the Deaf/ HH, with ten items such as "I feel confident I could figure out a way to communicate with Deaf people in an emergency." Due to the dearth of instruments on Deaf preparedness, survey items were adapted from several extant instruments [bib_ref] Psychometric Analysis of a Scale Assessing Self-Efficacy for Cultural Competence in Patient..., Assemi [/bib_ref].
# Analytical methods
For all of these research activities, close-ended items were entered and analyzed in SPSS. We developed descriptive statistics and conducted bivariate analysis. For openended items, we applied qualitative coding and content analysis in Excel for identifying themes across state respondents or CBOs that were reported for contextual understanding.
# Results
The key findings include (1)
## Literature review
There is almost no literature about broader emergency preparedness communication issues for and by the Deaf/ HH across various domains. We could find no peerreviewed literature specifically on emergency preparedness training and evaluation for the Deaf/HH. Zero peerreviewed articles were found using databases including PubMed and Google Scholar.
In the "grey" unpublished literature, we were able to find and review only two documents through a Google search. They included the U.S. Fire Administration's (2002) tip sheet for assisting Deaf/HH individuals in the event of an evacuation, and the U.S. Department of Justice's (2009) report entitled "Victims with Disabilities: Collaborative, Multidisciplinary First Response Techniques for First Responders Called to Help Crime Victims Who Have Disabilities Trainer's Guide" [bib_ref] Disaster Preparedness for Vulnerable Populations: Determining Effective Strategies for Communicating Risk, Warning,..., Sullivan [/bib_ref]. The information in these documents aimed to help first responders communicate with the Deaf/HH in an emergency. We found no other grey literature that addressed emergency preparedness training for the Deaf/HH.
## State-level agency training needs assessment
Results indicated a significant association between a statelevel KI's familiarity with communication issues faced by the Deaf/HH and: 1) whether or not the KI's department provides any trainings to him/her or other staff regarding how to serve the Deaf/HH populations during emergencies or disasters (p=0.02); and 2) whether or not the KI or other staff have attended any other trainings outside of their departments on serving the Deaf/HH populations during emergencies or disasters (p=0.02).
In addition, there was a significant association between a KI's familiarity with how to make and accept relay phone calls (a critical mechanism for 2-way communication with Deaf/HH individuals during an emergency) and: 1) whether or not the KI's department provides any trainings to him/her or other staff regarding how to serve the Deaf/HH populations during emergencies or disasters (p=0.008); or 2) whether or not the KI or other staff have attended any other trainings outside their department on serving the Deaf/HH populations during emergencies or disasters (p=0.004).
Approximately half of state KIs reported that their own department had not provided any training, however 67.3% (n=33) had attended other trainings outside of their departments on serving the Deaf population during emergencies [fig_ref] Table 2: State-level Deaf/HH trainings providedHave you or other staff attended any other trainings... [/fig_ref].
# Cbo data analysis
According to 14 KI interviews at local deaf-serving CBO's in SF Bay Area, only 36.4% of CBOs provided specific information on emergency preparedness to their clients and only 35.7% had attended trainings about serving Deaf/HH populations during emergencies or disasters at other organizations. Additionally, only half (50%) of the CBO's provided classes or trainings about emergency preparedness to clients or caregivers, and only 43% provided emergency preparedness educational materials to clients [fig_ref] Figure 1: CBO KIs providing emergency preparedness training for Deaf/HH [/fig_ref].
Despite the fact that 50% of the deaf-serving organizations reported providing medical services, chronic disease management, and skilled nursing, only 14.3% reported providing emergency preparedness services. Therefore, in addition to a lack of training, Deaf/HH clients from these organizations are underserved in terms of emergency preparedness, evacuation services in a disaster response phase, and recovery assistance following an event. Only 21.4% of CBOs provided evacuation services and only 14.3% provided recovery assistance to their clients. In addition, despite the fact that deaf-serving organizations are expected to be uniquely attuned to the literacy and functional needs of their respective constituents, written emergency preparedness materials designed for dissemination to Deaf audiences appear to be lacking. The readability of 100% (5/5) of the sample of materials collected from deaf-serving CBO's tested above the average literacy level for Deaf/HH populations [bib_ref] Availability and Readability of Emergency Preparedness Materials for Deaf and Hard-of-Hearing and..., Neuhauser [/bib_ref].
On average, KIs from deaf-serving CBOs reported a median of 12 months since attending specific training on emergency preparedness (range: 1-72 months ago). In terms of frequency of attendance, 28.6% of respondents said it was a one-time training, 28.6% said it was an annual training, and 42.9% said they attended training at some other interval [fig_ref] Table 4: Deaf-serving CBO training capacity [/fig_ref]. In addition, KIs from deaf-serving CBOs had various levels of familiarity with communication issues facing their constituents in an emergency [fig_ref] Figure 2: Deaf/HH-serving CBO familiarity with communication issues [/fig_ref].
Half of the KIs (50%) in deaf/HH-serving CBOs reported that the CBO provided any classes or training for clients or caregivers about emergency preparedness [fig_ref] Figure 1: CBO KIs providing emergency preparedness training for Deaf/HH [/fig_ref]. However, only a little over a third of the KIs (36.4%) reported that the CBO provided training specific to the issues that Deaf/HH people face in an emergency [fig_ref] Table 4: Deaf-serving CBO training capacity [/fig_ref].
## Evaluation of local law enforcement training
Results from both the survey (N=34) and two focus groups (total N=19) demonstrated that participants gained cultural competency skills post-training as indicated by items measuring attitudes towards the Deaf/HH, perceived better self-efficacy when working with the Deaf/ HH both in a DV emergency and in a large-scale emergency, as well as by demonstrating knowledge of communication and translation needs for interacting with Deaf/ HH individuals during emergencies. The attitudes subscale of the survey showed that the training had a positive impact on general attitudes towards the Deaf/HH, including perceived self-efficacy when working with the Deaf/ HH (t(33) =−5.02, p<0.01), which, in this case, is partly a reflection of cultural competence, but not on their perception of the capabilities of the Deaf/HH (t(33) =−0.34, p=0.74).
However, survey participants demonstrated a lack of knowledge about federal and state-level policy and the law, which can have serious implications at the time of an emergency. For example, emergency responders encountering a Deaf/HH person at the time of an emergency may not consider securing an American Sign Language interpreter, which is required by law. Post-training, few participants were able to mention the two federal laws which protect the rights of Deaf Americans by name: The Americans with Disabilities Act (ADA) and Section 504 of the Rehabilitation Act of 1973.
# Discussion
In this study, data from multiple sources: a literature review, State KI interviews, deaf-serving CBO KI interviews, and the training evaluation with first responders, all showed a critical lack of training about Deaf/HH emergency preparedness for Deaf/HH clients, CBOs, and emergency responders. Given the lack of evaluation of these types of trainings in the literature, a much more rigorous approach to assessing efficacy of the trainings is needed. The literature review indicates that there is a critical need for evaluation of trainings specific to Deaf/HH individuals in order to ensure effective training procedures and emergency preparedness programs for these populations. We also need further research on factors that promote or hinder preparedness communication capacity of agencies that serve Deaf/HH populations in disasters, deaf/HHserving CBOs, and Deaf/HH individuals [bib_ref] Availability and Readability of Emergency Preparedness Materials for Deaf and Hard-of-Hearing and..., Neuhauser [/bib_ref]. On a system-wide level, little information exists about national/state guidelines for emergency communication for Deaf/HH groups. However, according to FEMA's recent conference on Promising Practices in Inclusive Emergency Management (2011), one promising practice for training state emergency management personnel can be found in Hawaii, where people with disabilities are training emergency responders [bib_ref] Hearing loss prevalence in the United States, Lin [/bib_ref].
## State kis
Despite the lack of training described by state officials interviewed, several of them brainstormed solutions and cited barriers such as cultural issues, identifying the Deaf before and during emergencies, including Deaf/HH individuals in mass notification systems, and tailoring messages to this population. They also identified systemic challenges such as a lack of training about these issues, lack of accessible infrastructure, a need for more compliance with existing legislation for mass communication, and the need for improved collaboration and partnerships with local and state-level deaf-serving organizations.
Despite the lack of training noted by state officials, several respondents discussed technological initiatives to improve access for the Deaf/HH. One respondent, referring to a question about state-level strategies for improving emergency communication for the Deaf/HH, discussed a initiative to implement a new (State) Alert and Warning System with pilot tests in two counties in which people would register their information into the system and choose how they would like to be contacted in the event of an emergency.
## Deaf-serving cbos
Fourteen CBOs reported serving Deaf/HH clients who are minority, low-income, limited English proficiency, and low-literacy. In addition, most respondents reported serving clients who are homeless, homebound, or have limited or no transportation. These demographics indicate that training is especially important for these organizations because deaf-serving CBOs are working with individuals who have multiple risk factors before and during a disaster that may impact their safety more than many other individuals and these CBOs may have unique access to these individuals. In addition, training for deaf-serving CBOs is much needed due to the fact that these CBOs are non-profit organizations operating on very limited budgets. Especially in this economic climate, CBOs often lack resources for capacity building, which is problematic in light of the fact that Deaf/HH communities depend on local CBOs in the event of a disaster.
## Policy implications
Significant gaps in emergency preparedness training for serving the Deaf/HH exist despite the fact that in 2011 President Obama recommitted to enforcing and protecting the civil rights of people with disabilities on the 21st anniversary of the Americans with Disabilities Act (1990): "The promise of the ADA was that all Americans should have equal access and equal opportunity, including Americans with disabilities. The ADA was about independence and the freedom to make of our lives what we will. We celebrate that today, and we recommit ourselves to ending discrimination in all its forms," said President Obama.
This current research, coupled with more stringent enforcement of existing federal legislation, provides the kind of data and guidance that can impact national policy regarding the implementation of emergency response and planning tools that are tailor-made for the Deaf and hard-of-hearing. The timing of this research is fortuitous. Over a decade of research indicates that input from the Deaf/HH community and from people with disabilities is increasingly being acknowledged as important and actionable, both within the research community and also in emergency management. Recent trends in emergency management include incorporating a whole community planning approach, which is defined byas an emerging community-oriented approach to models for practice of emergency management, and that recognizes:
". . .in large scale disasters, the needs of survivors outweigh collective resources and capabilities of government . . . it really looks beyond traditional governments approach and all thinking government can solve disaster management challenges on its own. And it's really acknowledging that even small and medium sized events can be helped when government expands its reach and delivers services more efficiently by partnering with the community. " Acknowledging resource limitations in a large-scale disaster, whole community planning approaches involve: 1) non-governmental organizations, 2) businesses, and 3) government.
The more prepared emergency responders are in meeting the needs of Deaf/HH individuals in emergencies, the better prepared they will be in serving other vulnerable populations such as linguistic minorities. Lessons learned by addressing communication barriers faced by the Deaf/ HH can be applied to 120 million Americans who also face communication barriers: low health-literate members of society (90 million) and linguistic minorities (30 million). Targeted preparedness training materials can be disseminated to community-based organizations, communication strategies can be targeted for unique needs of diverse communities, and training can increase the capacity of both organizations and individuals to help serve the overwhelming needs seen in many recent disasters, potentially increasing resilience in diverse communities.
# Limitations
Limitations of the study include the small sample size inherent in US state-level analysis, the small number of Deaf-serving CBOs in our sample, and the single evaluation of a training program. However, there are only 50 states, 1 DC, and 8 territories, so this limitation is common to all state-level analysis. In addition given that deafness is a relatively low incidence disability, there is a small universe of Deaf-serving CBOs from which to choose. For the evaluation study, only one of the instruments that we adapted for use was peer reviewed due to the fact that there was a lack of available instruments. Our instruments were also newly created and pilottested with experts in, and affiliated with, the Deaf community and were created out of necessity because there were none available. Future researchers could attempt to replicate findings. Findings may have applicability to other populations with communication barriers such as limited-English proficiency populations.
# Conclusion
## Future directions
Our research indicates the need for: 1) increased accessibility and involvement of the Deaf/HH in training and exercises with guidance from state personnel; 2) required cultural competence training for first responders in order for them to better understand the diversity within this population (Deaf/deaf-blind/hard-of-hearing/late deafened) and improved emergency responder communication approaches; 3) standardized guidelines for CBOs to participate in local emergency and preparedness planning and exercises as fully as possible; and 4) development and dissemination of national guidelines about functional and access needs specific to the Deaf/HH population to emergency preparedness and response trainers across the nation. Establishing a national taskforce is a critical first step in the development of appropriate guidelines.
Since 9/11, important first steps have been taken to train emergency responders, Deaf/HH individuals, and sign-language interpreters to prepare for emergencies or disasters. However despite the limitations of our predominantly qualitative study, our research indicates that trainings specific for Deaf/HH are rare, even for state agency personnel, frequently lack standardization or evaluation, have not been institutionalized in the emergency management infrastructure, and thus there is no consistent curriculum across agencies. Programs on which we found information had few quality control or quality improvement measures and demonstrate a lack of coordination of efforts across agencies and organizations which are providing the trainings. On a policy level, there needs to be more mainstreaming of the needs of the Deaf/HH in emergency training and operations. In the long term, increasing awareness among first responders, emergency management agencies, and deaf-serving CBOs about the needs of Deaf before, during, and after disasters may lead to higher-level policy changes and improved outcomes for Deaf individuals.
Given the enormous diversity within the Deaf/HH population, developing, implementing, and evaluating emergency preparedness training cannot be achieved without forming alliances between agencies charged with emergency response and the Deaf/HH community, including getting input from experts in the emergency preparedness field who are Deaf/HH. Our fundamental recommendation regarding accessible emergency preparedness communication design is to ensure close participation of its intended beneficiaries as well as those involved in the communication dissemination and evaluation. All-hazards emergency preparedness communication challenges can be mitigated by a participatory design process, or a co-production of knowledge between the lived experience of Deaf/HH people, EMS system practitioners, and emergency preparedness researchers [bib_ref] Availability and Readability of Emergency Preparedness Materials for Deaf and Hard-of-Hearing and..., Neuhauser [/bib_ref] [bib_ref] Community knowledge in environmental health science: coproducing policy expertise, Corburn [/bib_ref] [bib_ref] Participatory Design of Mass Health Communication in Three Languages for Seniors and..., Neuhauser [/bib_ref].
[fig] Figure 1: CBO KIs providing emergency preparedness training for Deaf/HH. Clients or Caregivers (N=7). [/fig]
[fig] Figure 2: Deaf/HH-serving CBO familiarity with communication issues (N=14). [/fig]
[table] Table 1: Existing deaf preparedness training workshops and disaster simulation exercises* [/table]
[table] Table 2: State-level Deaf/HH trainings providedHave you or other staff attended any other trainings on serving the Deaf and Hard of Hearing populations during emergencies or disasters? (N=49) [/table]
[table] Table 3: Frequency of Deaf/HH training attendance by state agencies [/table]
[table] Table 4: Deaf-serving CBO training capacity [/table]
|
COVID-19 and the intensive care unit: vaccines to the rescue
# Declarations
## Conflicts of interest
The authors declare that they no conflict of interest.
## Open access
This article is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License, which permits any non-commercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by-nc/4. 0/.
## Publisher's note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. |
New method for detection of complex 3D fracture motion - Verification of an optical motion analysis system for biomechanical studies
Background: Fracture-healing depends on interfragmentary motion. For improved osteosynthesis and fracturehealing, the micromotion between fracture fragments is undergoing intensive research. The detection of 3D micromotions at the fracture gap still presents a challenge for conventional tactile measurement systems. Optical measurement systems may be easier to use than conventional systems, but, as yet, cannot guarantee accuracy. The purpose of this study was to validate the optical measurement system PONTOS 5M for use in biomechanical research, including measurement of micromotion. Methods: A standardized transverse fracture model was created to detect interfragmentary motions under axial loadings of up to 200 N. Measurements were performed using the optical measurement system and compared with a conventional high-accuracy tactile system consisting of 3 standard digital dial indicators (1 μm resolution; 5 μm error limit).Results: We found that the deviation in the mean average motion detection between the systems was at most 5.3 μm, indicating that detection of micromotion was possible with the optical measurement system. Furthermore, we could show two considerable advantages while using the optical measurement system. Only with the optical system interfragmentary motion could be analyzed directly at the fracture gap. Furthermore, the calibration of the optical system could be performed faster, safer and easier than that of the tactile system.Conclusion:The PONTOS 5 M optical measurement system appears to be a favorable alternative to previously used tactile measurement systems for biomechanical applications. Easy handling, combined with a high accuracy for 3D detection of micromotions (≤ 5 μm), suggests the likelihood of high user acceptance. This study was performed in the context of the deployment of a new implant (dynamic locking screw; Synthes, Oberdorf, Switzerland).
# Background
Various conditions are important for sufficient fracturehealing. In addition to adequate blood supply and a reduction in fracture size, axial interfragmentary motion is one of the most important factors for indirect (secondary) bone healing [bib_ref] The Dynamic locking screw (DLS) can increase interfragmentary motion on the near..., Doebele [/bib_ref] [bib_ref] Far cortical locking can reduce stiffness of locked plating constructs while retaining..., Bottlang [/bib_ref] [bib_ref] Interfragmentary motion in tibial osteotomies stabilized with ring fixators, Duda [/bib_ref] [bib_ref] Stiffness modulation of locking plate constructs using near cortical slotted holes: a..., Gardner [/bib_ref] [bib_ref] The influence of induced micromovement upon the healing of experimental tibial fractures, Goodship [/bib_ref] [bib_ref] Elastic plate osteosynthesis, biomechanics, indications and technique in comparison with rigid osteosynthesis, Sturmer [/bib_ref]. A deficiency in callus formation and delayed or non-union have been reported in diverse studies as the result of inadequate interfragmentary movements [bib_ref] The Dynamic locking screw (DLS) can increase interfragmentary motion on the near..., Doebele [/bib_ref] [bib_ref] The role of rigidity in fracture fixation. an overview, Uhthoff [/bib_ref] [bib_ref] The effects of external mechanical stimulation on the healing of diaphyseal osteotomies..., Wolf [/bib_ref] [bib_ref] Biomechanics of locked plates and screws, Egol [/bib_ref] [bib_ref] Locking compression plates for osteoporotic nonunions of the diaphyseal Humerus, Ring [/bib_ref] [bib_ref] Combination of interfragmentary screws and locking plates in distal meta-diaphyseal fractures of..., Horn [/bib_ref]. The optimal range for this micromotion seems to be 400 μm [bib_ref] The effects of external mechanical stimulation on the healing of diaphyseal osteotomies..., Wolf [/bib_ref]. Therefore, current biomechanical analyses have focused on the development of osteosynthetic implants for optimal interfragmentary motion [bib_ref] The Dynamic locking screw (DLS) can increase interfragmentary motion on the near..., Doebele [/bib_ref] [bib_ref] Far cortical locking can reduce stiffness of locked plating constructs while retaining..., Bottlang [/bib_ref] [bib_ref] Biomechanics of locked plates and screws, Egol [/bib_ref] [bib_ref] Interfragmentary movements in the early phase of healing in distraction and correction..., Duda [/bib_ref]. In the context of the development of new implants biomechanical tests are highly important. The purpose of this study was to find a motion analysis system for biomechanical tests, which allows the detection of three-dimensional interfragmentary motion with a high accuracy directly at the fracture gap of biomechanical specimens (osteosyntheses). Conventional tactile measurement systems are highly accurate (up to1 μm), but the test is time-consuming, laborious, and at times defective. A further disadvantage of tactile systems is the capture of movement direction in only 1 dimension per tactile unit in most cases. The integration of tactile measurement systems in an existing biomechanical set up could be exceedingly difficult. The integration of optical measurement systems in established biomechanical set ups is easy [bib_ref] Software techniques for two-and three-dimensional kinematic measurements of biological and biomimetic systems, Tyson [/bib_ref]. Using optical systems there is no interaction between test set up and measurement system. The detection of three-dimensional motion is also possible. But is it possible to detect interfragmentary motion in a range of about 400 μm [bib_ref] The effects of external mechanical stimulation on the healing of diaphyseal osteotomies..., Wolf [/bib_ref] [bib_ref] Accuracy of an optical active-marker system to track the relative motion of..., Maletsky [/bib_ref] ? The optical measurement system PONTOS 5 M (GOM -Optical Measuring Techniques, Braunschweig, Germany) is an established system for motion analysis in the automotive and aerospace industry (used for crash-tests and vibration-analysis of airplane wings). In this study we evaluate PONTOS 5 M for the detection of 3D interfragmentary micromotion in a standardized fracture model. For reference, we used a tactile measurement system consisting of 3 dial indicators designed at the German aerospace centre DLR (Oberpfaffenhofen, Germany). We hypothesized that the optical measurement system would provide the same or higher accuracy for detecting fracture gap movements as the conventional system.
# Results
Indirect measurement of the fracture movement μm. Dial indicator 2 showed a slight tendency for dependency of mean difference and size of the measurement. The regression approach in [fig_ref] Figure 1: shows the agreement between the three dial indicators and PONOTS 5 M [/fig_ref] indicates this relationship. There is a bias, which tends to be zero for small interfragmentary motion and is the greater the greater the motion is: with an average motion of 0 μm the bias was 2.5 μm, but at an average of -1.8 μm the bias was -4.5 μm. The 95% limits of agreement were of 5 μm around the bias.
## Direct measurement of the fracture movement
Regarding the y-displacement of the interfragmentary motion there was a clear dependency of the magnitude of the motion and the difference between the dial indicator and PONTOS 5 M: the higher the load the higher the differences between the two methods [fig_ref] Figure 2: Bland Altman Plot [/fig_ref]. For an average interfragmentary motion of 0 mm the bias between the dial indicator and PONTOS 5 M was 2 μm. For an averaged motion of 120 μm in contrast, the bias was -140 μm. The limits of agreement were of 40 μm around the bias [fig_ref] Table 1: Mean differences and 95% limits of agreement [/fig_ref].
The agreement of PONTOS 5 M with the dial indicator concerning the z-displacement of the interfragmentary motion is shown in [fig_ref] Figure 3: Bland Altman Plot [/fig_ref]. There was a wider spread of the differences with a higher force and therefore a larger interfragmentary motion. For an average interfragmentary motion of -1.31 mm the bias was 30 μm with limits of agreement of -110 and 170 μm. For an average of 0 mm interfragmentary motion the bias was -40 μm with limits of agreement of -80 μm and 10 μm [fig_ref] Table 1: Mean differences and 95% limits of agreement [/fig_ref].
Regarding the α angle (pitch-angle) there was a bias between the dial indicator and PONTOS 5 M between -0.05 and 0.13 degrees. The difference and especially the variability of the difference clearly depend on the applied load . With a force of 100 N the average angle is -4.1 degree. The bias of the two methods at 100 N is 0.13 degree with limits of agreement of -0.38 [fig_ref] Table 1: Mean differences and 95% limits of agreement [/fig_ref].
# Discussion
The aim of the present study was to validate the efficacy of the optical measurement system PONTOS 5 M compared with a conventional tactile measurement system. The results show that both measurement systems were capable of analyzing interfragmentary movements with high accuracy (resolution of about ≤5 μm). However, the optical measurement system was able to analyze 3D motions whereas the tactile system used in this study only performed 2D measurements of the fracture motion. For this reason it was only possible to compare the 2D data obtained using the optical measurement system with the corresponding data of the tactile measurement system. This is one limitation of the present study.
In order to assess the accuracy of the optical measurement system we attached passive markers of the optical measurement system to the spindle of the dial indicators, analysed the collapse of the measuring spindle and compared the values of the dial indicators with those of the optical measurement system.
The resolution of the dial indicators is known to be 1 μm. The accuracy between the tactile measurement system and optical measurement system showed a mean difference of -5.3 μm (dial indicator 1), 0.4 μm (dial indicator 2) and -1.8 μm (dial indicator 3). The variability of the differences between the two methods was between 5 μm and 9 μm. This was only little wider than the error limit of the dial indicators. The difference between the two methods was comparable with the precision within the dial indicators. With the optical measurement system interfragmentary movement can be detected directly at the fracture gap as the passive markers are very small and can be attached nearly anywhere. Contrariwise using the digital indicators the measurement of the interfragmentary motion was only possible indirectly by using the cranks. The average deviation in interfragmentary motion measured directly at the fracture gap (PONTOS) and indirectly using the tactile system and the cranks was 120 μm (y-displacement), -1310 μm (z-displacement) and -4.1 degree (α angle). These results are rather surprising. The large difference of the measured interfragmentary motion can be explained by the set up of the tactile system using the cranks. The cranks become deformed by the mechanical load during the test which was shown by attaching the passive markers on the cranks [fig_ref] Figure 6: Test set-up [/fig_ref]. Using the cranks tends to result in a decreasing of accuracy. This effect was more pronounced the higher the axial load was. In addition to the high accuracy of the optical measurement system, this system is much easier to use in comparison to the dial indicator method, because the passive markers are self-adhesive and, as mentioned, can be attached nearly anywhere. The dial indicators, in contrast, each requires a specially produced crank for attachment. This setup is not only time-consuming, but also expensive, and the accuracy of the measurement depends directly on the cranks. In contrast, each passive marker of the optical measurement system functions like a 6-DOF-sensor, so that several different data sets for the object can be obtained. Due to its high accuracy, PONTOS 5 M is regularly used in the automobile and airplane industry for car crash-tests or vibration-analysis of airplane wings. Therefore, large amounts of data are available from different testing setups. In biomechanical setups for musculoskeletal research, diverse types of fracture models have been validated. We decided to use a simple model with a transverse fracture gap in our study in order to exclude measurement deviations as far as possible. There are some publications that deal with the application of optical measuring systems in biomechanics. Here also the accuracy of the systems was part of the research. A common system in the biomechanical field is the Vicon system [bib_ref] Effects of camera switching on fine accuracy in a motion capture system, Kuxhaus [/bib_ref] [bib_ref] Systematic accuracy and precision analysis of video motion capturing systems-exemplified on the..., Windolf [/bib_ref]. A study by Windolf et al showed an accuracy of 64 ± 5 microns using the Vicon-460 system [bib_ref] Systematic accuracy and precision analysis of video motion capturing systems-exemplified on the..., Windolf [/bib_ref]. Arbitrary changes in camera arrangement revealed variations in mean accuracy between 76 and 129 μm. This is less accurate than measuring with the PONTOS System.
# Conclusions
In this study, we attempted to expand the applicability of the PONTOS 5 M optical measurement system for biomechanical assessments. As the need for implant improvement, especially with regard to plate stiffness, is acute, the need for an accurate, validated, easy-to-handle 3D measurement system is also high. Within the framework of the presented data and the limitation of only 2D evaluation, we can say that the PONTOS 5 M optical measurement system appears to be a favourable alternative to previously used tactile measurement systems for biomechanical applications. Easy handling combined with a high accuracy (≤ 5 μm) and 3D detection of motions suggests the likelihood of high user acceptance. The use of simple passive markers suggest an easier handling at cadaver models as compared to mechanical devices, which need accurate disinfection after usage.
# Methods
In 6 surrogate specimens, a standardized transverse fracture model with a fracture gap of 3 mm was created, as described previously [bib_ref] The Dynamic locking screw (DLS) can increase interfragmentary motion on the near..., Doebele [/bib_ref] [bib_ref] Far cortical locking can reduce stiffness of locked plating constructs while retaining..., Bottlang [/bib_ref]. The cylindrical bone surrogates were manufactured using polyoxymethylen-copolymerisat with a Young's modulus of 3.1 GPa (diameter 30 mm, wall thickness 7 mm, cylinder length 120 mm). Osteosynthesis was performed using a bridge-plating configuration with a standard 11-hole, 3.5-mm locking compression plate (Synthes, Oberdorf, Switzerland). Plates were fixed with 3 locking screws (55 mm) on each fragment placed in the second, third, and fourth hole from the fracture site . All screws were tightened to 1.5 Nm, with the plate elevated 2 mm from the surrogate surface. Specimens were inserted into a special testing frame mounted on a testing machine (Zwick 2.5 KN; Ulm, Germany) and an axial load of up to 100 N with a constant rate of 10 mm/min was applied. Measurements of the micromotion of the fragments were performed using a self-made high-accuracy tactile measurement system made of 3 digital dial indicators and an optical measurement system.
## Optical measurement system
For measuring the 3D fracture motion we used the optical analysis system PONTOS 5 M (GOM -Optical Measuring Techniques, Braunschweig, Germany). The system is offered in 4 different configurations (5 M, 4 M, 12 M or High Speed) with several camera resolutions (up to 4096 × 3072 pixel) and frame rates (up to 5000 Hz) to cope for diverse applications. The individual system consists of two CCD cameras. For the detection of the motion passive markers are required. These passive markers are simple white dots. The size is adapted to the object size and camera resolution. The points are recorded and tracked by the PONTOS software . The accuracy of the measurements is directly addicted by the resolution of the cameras. Each passive marker is detected in the single image as ellipses in the size of several pixels. The center of the marker can therefore be determined in the sub-pixel space by evaluation of a best fit at the contour of the ellipses, which is done automatically by the software. The PONTOS 5 M system was set up and calibrated for a measurement volume of 350 × 280 × 280 mm according to the manufacturers documentation. The geometrical setup as well as the optical distorsion factors of lenses are considered in the calibration procedure. The frame rate was 4 Hz. We used white selfadhesive dots with a diameter of 2 mm. At least 3 points per object are needed. For a higher accuracy we used all in all about 200 passive markers. It is not necessary to add the points in a special pattern. But doing that, the system could learn to identify the objects with the help of the different patterns. A group of points functions like a six degrees of freedom (6DoF)-sensor. The high number of points allowed fitting two cylindrical geometry elements which represent the physical cylindrical fracture fragments. The 6DoF motion could therefore be directly represented at the location in the centre of the fracture gap for each fragment. Relative motion in all six degrees of freedom where analysedrespectively.
## Reference measurement system
For reference measurements we used digital indicators from the company MAHR (Marcator 1086; Mahr, Goettingen, Germany). To date, dial indicators represent the gold standard in measuring motion in the range of micrometers. We used 3 digital indicators. Each dial gauge can specify one degree of freedom. The resolution of the individual gauge was 1 μm (Span of error 5 μm, Repeatability 2 μm). This accuracy is ensured by the company. For positioning the 3 dial indicators to the osteosynthesis two cranks were constructed and made for measure using a cnc milling machine [fig_ref] Figure 6: Test set-up [/fig_ref]. The measurement set up was developed in cooperation with the German Aerospace Center (DLR, Oberpfaffenhofen, Germany). The dial indicators were connected to a computer system by an usb interface. The data recording was carried out using the supplied software from Mahr. Due to the placement of the indicators far from the fracture gap, the interfragmentary motion could not be detected directly. Instead, interfragmentary motions were indirectly measured by analyzing the motion of the measuring spindle of the dial indicators. The motion at the fracture gap has to be calculated using the geometry of the cranks and the values of the dial indicators. Using the values of digital indicator 1 and 2 it was possible to calculate the z deviation (Δz) and the angle alpha (Δa) between the both cylinder axis (fracture fragments) [fig_ref] Figure 6: Test set-up [/fig_ref]. Using digital indicator 3 it was possible to Osteosynthesis. Osteosynthesis with standard 11-hole, 3.5-mm locking compression plate (Synthes, Oberdorf, Switzerland). Plates were fixed with 3 locking screws (55 mm) on each fragment placed in the second, third, and fourth hole from the fracture site. calculate the y deviation (Δy). To compare the accuracy of the two measurement systems (PONTOS 5 M vs tactile measurement system), we conducted an additional measurement. For this purpose, the passive markers of the optical measurement system were attached directly to the spindles of each of the 3 indicators . With this set up the motion of the spindles could be detected with the optical measurement system, by measuring the point to point motion. Both values (digital indicator and PONTOS) could be compared. Simultaneous data collection was achieved by trigger points. A total of 3 data-sets, 200 values per set, were collected, 1 for each dial indicator.
# Statistical analysis
For analyzing the agreement between the optical measurement system PONTOS 5 M and the indirect measuring by the dial indicators Bland-Altman-Plots were calculated. For each comparison we checked the assumption of uniform differences and uniform variability. If these assumptions were violated a regression approach was applied. We used a Generalized Estimation Equation Model (GEE) to account for the different measurements made in one specimen (the force was varied between 0 and 100 N in increments of 10 N for each specimen, which results in 11 measurements per specimen). The results of this model were used to calculate mean differences and 95% limits of agreement as described by Bland and Altman [bib_ref] Measuring agreement in method comparison studies, Bland [/bib_ref] [bib_ref] Agreement between methods of measurement with multiple observations per individual, Bland [/bib_ref].
[fig] Figure 1: shows the agreement between the three dial indicators and PONOTS 5 M. For dial indicators 1 and 3 there was no obvious dependence of the amount of measuring spindle motion and the mean differences between the direct (PONTOS 5 M) and indirect measurement (dial indicator). For dial indicator 1 there was a bias of -5.3 μm with limits of agreement of -14.3 μm and 3.6 μm. For dial indicator 3 the mean difference was 0.4 μm with limits of agreement of -5.3 μm and 4.6 [/fig]
[fig] Figure 2: Bland Altman Plot: Interfragmentary motion Δy. The differences of interfragmentary motion (Δy) between the dial indicator and PONOTS 5 M are plotted against the average of the two methods. The solid line indicates the systematic bias between the two methods. The upper and lower limits show the 95% limits of agreement. [/fig]
[fig] Figure 3: Bland Altman Plot: Interfragmentary motion Δz. The differences of interfragmentary motion (Δz) between the dial indicator and PONOTS 5 M are plotted against the average of the two methods. The solid line indicates the systematic bias between the two methods. The upper and lower limits show the 95% limits of agreement. [/fig]
[fig] Figure 4, Figure 5: Bland Altman Plot: Interfragmentary motion Δa. The differences of interfragmentary motion (α angle) between the dial indicator and PONOTS 5 M are plotted against the average of the two methods. The solid line indicates the systematic bias between the two methods. The upper and lower limits show the 95% limits of agreement. Additional measurement: Direct comparison of both Systemes. For this purpose, the passive markers of the optical measurement system were attached to the spindles of the 3 dial indicators. [/fig]
[fig] Figure 6: Test set-up. Two cranks were fixed to position the 3 dial indicators. Passive markers were fixed directly to the fracture gap. [/fig]
[table] Table 1: Mean differences and 95% limits of agreement [/table]
|
Getting under the skin of the primary care consultation using video stimulated recall: a systematic review
Background: Video stimulated recall (VSR) is a method of enhancing participants' accounts of the consultation using a video recording of the event to encourage and prompt recall in a post consultation interview. VSR is used in education and education research, and to a lesser extent in medical and nursing research. Little is known about the sort of research questions that lend themselves best to the use of VSR or the impact of the specific VSR procedure on study quality. This systematic review describes studies in primary care that have used the method and aims to identify the strengths, weaknesses and role of VSR. Methods: A systematic literature search has been conducted to identify primary care consultation research using VSR. Two authors undertook data extraction and quality appraisal of identified papers and a narrative synthesis has been conducted to draw together the findings. In addition, theory on classifying VSR procedures derived from other disciplines is used as a lens through which to assess the relevance of VSR technique.Results: Twenty eight publications were identified that reported VSR in primary care doctor-patient consultation research. VSR was identified as a useful method to explore specific events within the consultation, mundane or routine occurrences, non-spoken events and appears to particularly add value to doctor's post consultation accounts. However, studies frequently had insufficient description of methods to properly evaluate both the quality of the study, and the influence of VSR technique on findings.Conclusions: VSR is particularly useful for study of specific consultation events when a 'within case' approach is used in analysis, comparing and contrasting findings from the consultation and post-consultation interview. Alignment of the choice of VSR procedure and sampling to the study research question was established as particularly important in the quality of studies. Future researchers may consider the role of process evaluation to understand further the impact of research design on data yielded and the acceptability of the method to participants.
# Background
The consultation has been long a subject of interest for researchers seeking to gain further understanding of the doctor-patient relationship and interaction. In 1969, Byrne and Long audio recorded over 2500 consultations to research verbal behaviours between doctors and patients; since then, there has been increasing use of video recordings to facilitate observational consultation research [bib_ref] Using video-recorded consultations for research in primary care: advantages and limitations, Coleman [/bib_ref]. An alternative method to indirect observation of the consultation is to seek participants' accounts of events by interview, focus group or workshops and these methods have been used in a recent publication by The Health Foundation 'When doctors and patients talk: making sense of the consultation'.
Participant accounts are retrospective and limited to that which is remembered and reported; however, recall accuracy and completeness may be enhanced by playing back the video-recorded consultation within the interview context: 'video stimulated recall' (VSR). VSR may be useful for improving recall, for uncovering cognitive processes and as a tool to facilitate reflections on elements of many different social interactions. The method of VSR has been used extensively in educational and counselling researchand to a lesser extent in medicine and nursing. When data derived from participant accounts using VSR is combined with consultation analysis an in-depth exploration of consultation events may be achieved.
Stimulated recall can also be achieved with the use of audio recordings, in place of video. However, the advantage of using video recordings is that the visual stimulus may be a stronger stimulus for recall and the participant may also comment or reflect on their non-verbal behaviours.
VSR is described as useful for the study of patientprofessional interactional components of the consultation and complex, context dependent occurrences, in addition to permitting more accurate recall of events that may have been forgotten [bib_ref] Video elicitation interviews: a qualitative research method for investigating physician-patient interactions, Henry [/bib_ref]. The technique is also complex, costly and time consuming and it is suggested it should be reserved for research questions that cannot be answered with consultation analysis or participant interviews alone [bib_ref] Video elicitation interviews: a qualitative research method for investigating physician-patient interactions, Henry [/bib_ref] [bib_ref] Combining qualitative interviews with videorecorded consultations: Gaining insight into GPs' decision-making, Coleman [/bib_ref]. Henry at al [bib_ref] Video elicitation interviews: a qualitative research method for investigating physician-patient interactions, Henry [/bib_ref] conducted a literature review of studies using the method; however in the absence of quality appraisal of the studies, no empirical evidence was presented to guide future researchers in the most appropriate use of VSR or to illuminate the methodological strengths and weaknesses particular to VSR. The question remains as to which types of research question lend themselves best to this method.
VSR may be conducted in a number of different ways. For example, the video may be shown in entirety prior to a semi-structured interview or the participant may be asked to comment during playback on specific areas of interest. The nuances of VSR procedure are considered important in the design of research although they have not been previously described in medical literature.
In summary, VSR appears to be an important methodology for researching the consultation but what is missing from the existing literature is an understanding of the strengths and weaknesses particular to the method, the way in which VSR procedure relates to study quality and the research questions that may be best suited to the method. This systematic review aims to address these gaps, and was conducted as preliminary work for a study that planned to use the method of VSR in exploring the content of osteoarthritis consultations in primary care. In this systematic review, we describe studies in primary care consultation research that have used VSR, in order to describe the utility of the method in consultation research.
## Specific objectives of this review
This systematic review aims to further understanding of the role of VSR in doctor-patient consultation research to describe:
a. The research questions that have been addressed using VSR b. The methodological strengths and weaknesses particular to VSR, including its acceptability to participants c. The procedure of VSR (using the theoretical framework in and how the choice of procedure influences overall considerations of study quality and utility d. The areas of research where VSR adds value Components of SR procedure and theoretical effect on outcome (adapted from Gass and MackeyExample/comments Time between video recorded event and SR Participant recall of events will be greater immediately after the interview.
# Strength of stimulus
Video is an example of a strong stimulus, but the strength of stimulus may be increased still further by additional stimulus for recall e.g. transcripts of consultation. The greater strength of stimulus, the more enhanced the recall will be.
## Procedural structure of accompanying interview
A structured interview is an example of high procedural structure and will result in more specific information relative to the research question.
A low structure approach would involve minimal questioning and the use only of neutral prompts during playback e.g. "what were you thinking then?". This method may be more suitable where the research question concerns cognitive processes at the time of the interview and is less likely to result in researcher contamination.
## Initiation of recall event
The researcher may lead recall by asking the participant to comment on areas of interest to the researcher, or the participant may be asked to comment on aspects of their choice. Again, researcher initiated events may encourage more reflection than recall alone.
## Relationship between video recorded event and line of inquiry
During a VSR interview, a participant may be questioned only on events that occurred during the video, described as a 'concrete relationship to action'. However, they may be asked to abstract to other general events, an example of a 'non-specific relationship to action'. In this instance, their recall may not be as great.
## Participant training
Participants may need training and practice if asked to comment on stimulus in an unstructured way. Training may enhance a participant's ability to reflect on observed events.
# Methods
## Theoretical framework: vsr procedures
Gass and Mackeyhave previously described a classification of SR techniques; this was used as a theoretical framework to inform analysis. One inherent limitation of the technique of VSR is that the feelings and thoughts expressed in the context of a post-consultation interview may not reflect the thoughts at the time of the consultation, and are subject to researcher influence [bib_ref] Video elicitation interviews: a qualitative research method for investigating physician-patient interactions, Henry [/bib_ref]. Careful attention to the procedure of VSR may reduce this effect. Techniques of VSR vary widely and different methods may be more suited to capturing recall, reliving or reflection [bib_ref] Video elicitation interviews: a qualitative research method for investigating physician-patient interactions, Henry [/bib_ref]. Gass and Mackey have reviewed the literature across different disciplines in the techniques of stimulated recall (SR), although not restricted to video, in their text relating to second language research. Their methodological theoretical framework has applications beyond language research and is considered a useful starting point for researchers considering the method. In , the techniques of SR are listed, as adapted from Gass and Mackey's classification. In theory, the recall accuracy will be greatest if the interview takes place immediately after the consultation event, with the highest strength of stimulus and if the stimulus has a concrete relation to the area of questioning. Concern is reported in the psychology literature about the types of memory accessed with delayed recall which is reported to affect validity of responses. However, as short term memory decays within a few hours, it is suggested that there may be not much difference in recall performed at 3 hours, compared with 3 days.
Lyleargues research questions concerned with decision making or cognitive processes during the video recorded event (in this case, the consultation) are most likely to be subject to reinterpretation of proceedings; for this reason, the choice of structure of the post consultation interview and the individual initiating recall are key to reduce the likelihood of reflection. The wording of questions would therefore seem to be of great importance in reducing researcher contamination. There is some empirical evidence for this from a number of studies in psychology around 'think aloud' protocols. Although these do not strictly represent SR, a participant is asked to verbalise thoughts while completing a task. Ericsson and Simon have conducted many reviews on this subject and their consistent finding is that verbalisation during a task does not change performance unless participants are asked to verbalise motives or reasons for their behaviour; in this case, participants are observed to change behaviour. This finding is attributed to participants speculating or theorising about higher cognitive processes that may be automatic [bib_ref] Verbal reports as data, Ericsson [/bib_ref].
## Literature search
Based on the assumption that in primary care the consultation may differ in character and structure from secondary care settings, this review is restricted to studies in primary care. The search was divided into four areas: consultation; primary care; video; and qualitative research. The literature search was conducted in March 2012 and repeated in November 2012 in Medline, Psychinfo, CINAHL, Embase and HMIC, Web of Science and BIOSIS. Additional references were obtained by reference checking, contacting experts, searching conference abstracts and cited reference checking using Web of Science. The search was limited to English language publications.
Given the wide range of terms used for video-elicitation and the possibility that terms exist of which the authors of this review are unaware, the search was left broad and all results relating to video searched for details of stimulated recall. If a post consultation interview was reported in the abstract the full text was reviewed to establish if VSR had been used.
A full list of search terms appears in [fig_ref] Table 2: Search terms used [/fig_ref] and the full Medline Search in Additional file 1. Inclusion and exclusion criteria are listed in [fig_ref] Table 3: Inclusion and exclusion criteria [/fig_ref].
In the first stage of sorting all record titles were screened and exclusions made where possible by the first author (ZP). The remaining records were then viewed as abstracts, by two reviewers independently (ZP and GMcH), and exclusions made where possible. Those titles and abstracts not fulfilling the inclusion criteria at each stage were discounted. The full text of the remaining articles was then requested, including those with no abstracts. Disagreements between the reviewers were resolved by discussion and consensus on inclusion or exclusion reached for both abstract and full text review. All full text articles retrieved were read, decisions made regarding their inclusion, and the reasons for exclusion recorded, again by two reviewers. Exclusions were made serially by each criteria and only one reason recorded for each abstract or full text discounted. An access database containing the data extraction and quality assessment items was designed and piloted by two reviewers and minor amendments made. Thereafter, data extraction and quality appraisal forms were completed for each paper by two reviewers (ZP, and either GMcH or AH) independently. Two papers described methodology only [bib_ref] Combining qualitative interviews with videorecorded consultations: Gaining insight into GPs' decision-making, Coleman [/bib_ref] [bib_ref] A method for study of dilemmas during health care consultations, Timpka [/bib_ref] , with no independent research question; these were not counted in the final sample, but the content of each used to aid quality appraisal of their respective related paper.
## Quality assessment
A list of characteristics for quality assessment was designed, based on the following two sources:
1. Coleman [bib_ref] Using video-recorded consultations for research in primary care: advantages and limitations, Coleman [/bib_ref] cites four aspects of 'bias' of research using video, namely the effect of the video-recorder on the patient and GP (described as internal validity) and the characteristics of patients and GPs who consent to being videotaped, compared with non-consenters (described as external validity). The extent to which authors reported on these aspects was recorded. 2. Papers included used qualitative methodology as a framework for analysis and so questions from the Critical Appraisal Skills Programme (CASP) Qualitative appraisal toolwere incorporated in the checklist. This tool has been used in other qualitative systematic reviews; the eight detailed questions from CASP included are detailed in [fig_ref] Table 4: Quality assessment items derived from CASP checklist [/fig_ref].
## Data extraction
The data extraction elements are shown in [fig_ref] Table 5: Data extraction questions [/fig_ref].
The full data extraction form used by the authors, including the quality assessment is included in Additional file 2. [this has been incorporated into the 'Reviewer's main conclusions'see [fig_ref] Table 5: Data extraction questions [/fig_ref] , and Additional file 2]
## Synthesis
A narrative synthesis approach was used, guided by the aims of the review; this method is ideally suited to combining results from qualitative studies where quantitative synthesis is not possible and easily adaptable to describing process (methods) rather than pooling study results. The outline of SR techniques described by Gass and Mackeywas used as a theoretical framework to inform analysis. Following individual data extraction and quality appraisal, authors met to first discuss and compare findings for each study. Secondly, emergent patterns and themes across studies were discussed. Thereafter, a preliminary synthesis was achieved using tabulation of studies and forming groups and moderator variables used to explore relationships between studies. All authors then contributed to the final report. The study methods and the reporting of results adhere to the guidelines in the PRISMA statement.
# Results and discussion
Identification of studies 2132 papers were identified by the initial search, and 28 ultimately fulfilled inclusion criteria. A flowchart showing the phases of identification, as recommended by PRISMA, is shown in [fig_ref] Figure 1: Phases of identification of papers [/fig_ref] , and [fig_ref] Table 6: Reasons for exclusion [/fig_ref] details the reasons for exclusion.
## Description of included studies
The included articles are described in terms of research question and area in [fig_ref] Table 7: Description of included studies To provide a detailed understanding of the ways... [/fig_ref]. The 28 individual articles refer to 18 sets of video recorded consultation data, and multiple publications from a single data set are listed together in a single row. The areas of research fall into eight categories: decision making; communication (including a subset of cross-cultural communication); doctorpatient relationship; patient experiences; evaluation of the method of VSR; self-management; health promotion and team working. Many of the studies were concerned with generic aspects of the consultation and as such have a relatively unselected sample. However, five studies were focused on specific consultation content: patients' expressed psychological problems [bib_ref] Intercultural communication competence in family medicine: lessons from the field, Rosenberg [/bib_ref] ; discussion of smoking cessation [bib_ref] Factors influencing discussion of smoking between general practitioners and patients who smoke:..., Coleman [/bib_ref] ; HIV risk [bib_ref] Awkward moments in patient-physician communication about HIV risk, Epstein [/bib_ref] ; self-management in long term conditions [bib_ref] Bringing selfmanagement into clinical view: a qualitative study of long-term condition management..., Blakeman [/bib_ref] and health promotion [bib_ref] Prevention in Practice -how do General-Practitioners Discuss Life-Style Issues with their Patients, Arborelius [/bib_ref]. In justifying the choice of method, many sought simply to gain a fuller understanding of participants' experiences. The doctor-patient relationship and communication were the most common areas of inquiry with three studies researching the effect of the computer on the relationship, and three looking specifically at cross-cultural communication. Specific events within the consultation were the focus of the study in studies concerning decisionmaking, or discussions around HIV risk and smoking cessation. Two studies used the method to explore nondeliberate behaviour: unspoken information or non-verbal cues [bib_ref] How do you know what Aunt Martha looks like?' A video elicitation..., Henry [/bib_ref] [bib_ref] The significance for decision-making of information that is not exchanged by patients..., Bugge [/bib_ref].
# General methodological considerations
The results from the quality appraisal are detailed in [fig_ref] Table 8: Findings from quality appraisal [/fig_ref]. Frequently, studies had insufficient detail in their methods section to properly evaluate the quality of the study. Three author groups described their methodology in separate publications [bib_ref] Video elicitation interviews: a qualitative research method for investigating physician-patient interactions, Henry [/bib_ref] [bib_ref] Combining qualitative interviews with videorecorded consultations: Gaining insight into GPs' decision-making, Coleman [/bib_ref] [bib_ref] A method for study of dilemmas during health care consultations, Timpka [/bib_ref] [bib_ref] In what way may videotapes be used to get significant information about..., Arborelius [/bib_ref]. Some authors also reported analysis of different data components in separate publications where there were individual research questions [bib_ref] A method for study of dilemmas during health care consultations, Timpka [/bib_ref] [bib_ref] Factors influencing discussion of smoking between general practitioners and patients who smoke:..., Coleman [/bib_ref] [bib_ref] Bringing selfmanagement into clinical view: a qualitative study of long-term condition management..., Blakeman [/bib_ref] [bib_ref] Prevention in Practice -how do General-Practitioners Discuss Life-Style Issues with their Patients, Arborelius [/bib_ref] [bib_ref] The role of culture in the general practice consultation process, Ali [/bib_ref] [bib_ref] In what way may videotapes be used to get significant information about..., Arborelius [/bib_ref] [bib_ref] Patients comment on video-recorded consultations-the "good" GP and the "bad, Arborelius [/bib_ref] [bib_ref] What is going on when the general practitioner doesn't grasp the situation?, Arborelius [/bib_ref] [bib_ref] Comparison of patients' and doctors' comments on video-recorded consultations, Arborelius [/bib_ref] [bib_ref] General practitioners' comments on video recorded consultations as an aid to understanding..., Arborelius [/bib_ref] [bib_ref] The Quality and Outcomes Framework and self-management dialogue in primary care consultations:..., Blakeman [/bib_ref] [bib_ref] Physicians, patients, and the electronic health record: An ethnographic analysis, Ventres [/bib_ref] [bib_ref] Clinician style and examination room computers: a video ethnography, Ventres [/bib_ref] [bib_ref] Using content analysis of videorecorded consultations to identify smokers' "readiness" and "resistance"..., Coleman [/bib_ref] [bib_ref] Communication between South Asian patients and GPs: comparative study using the Roter..., Neal [/bib_ref]. These associated publications were not always referenced in the included study [bib_ref] The role of culture in the general practice consultation process, Ali [/bib_ref] [bib_ref] Communication between South Asian patients and GPs: comparative study using the Roter..., Neal [/bib_ref]. Multiple publications on the same dataset were generally not felt to be of high methodological quality, predominately due to the lack of alignment between research question and methods, particularly participant sampling. For example, Arborelius et al. [bib_ref] Prevention in Practice -how do General-Practitioners Discuss Life-Style Issues with their Patients, Arborelius [/bib_ref] focused one paper on health promotion advice when only 8 of the original 46 video recorded consultations contained discussion of this nature. Sampling emerged as a particularly important component of quality in research design. For example, Coleman et al. ensured richness of data in their video data about smoking cessation by sampling at every stage of the method; GPs were sampled to represent a range of attitudes to smoking identified on a questionnaire, patients were selected on basis of smoking status and the videos shown to the GPs were chosen to reflect a range of different types of discussion around smoking e.g. smoking cessation discussed in the presence or absence of smoking related problem [bib_ref] Factors influencing discussion of smoking between general practitioners and patients who smoke:..., Coleman [/bib_ref]. Epstein et al. also enhanced sampling by using pre-consultation questionnaires to identify patients for their sample concerned about HIV risk. Although a number of studies described the characteristics of the sample of their study, only one did this with reference to non-consenters enabling the reader to judge the transferability of the results [bib_ref] Factors influencing discussion of smoking between general practitioners and patients who smoke:..., Coleman [/bib_ref].
Five studies only analysed VSR data from either patient or doctor, 10 used VSR data from more than one perspective (patient, doctor or interpreter) and 13 studies analysed both VSR and consultation data together. The research question did not always match the data collected; for example in four studies researching communication [bib_ref] Intercultural communication competence in family medicine: lessons from the field, Rosenberg [/bib_ref] [bib_ref] The role of culture in the general practice consultation process, Ali [/bib_ref] [bib_ref] Doctor-patient communication in primary care with an interpreter: physician perceptions of professional..., Rosenberg [/bib_ref] [bib_ref] Through interpreters' eyes: comparing roles of professional and family interpreters, Rosenberg [/bib_ref] , the VSR interviews were the only data analysed and analysis of the consultation itself may have added value. Furthermore, three of these studies did not study all parties in the consultation.
Conversely, in two studies, the study findings did not appear to represent all the different data sources collected. Gao et al. [bib_ref] Considering Culture in Physician-Patient Communication During Colorectal Cancer Screening, Gao [/bib_ref] researched communication, looking in detail at cross-cultural influences on colorectal screening; in their study only patient VSR and consultation findings are reported despite the methods indicating they also conducted VSR with GPs. Blakeman et al. [bib_ref] The Quality and Outcomes Framework and self-management dialogue in primary care consultations:..., Blakeman [/bib_ref] interviewed both doctors and nurses in their study regarding the influence of the Quality and Outcomes Framework (QOF). The doctor responses appeared to be underrepresented in the results; in this instance this may have been due to the context of the study as nurse consultations may have been more QOF orientated.
In terms of the effect of the video on participants' behaviour, two studies reported that GP behaviour was not affected by the video [bib_ref] Awkward moments in patient-physician communication about HIV risk, Epstein [/bib_ref] [bib_ref] Clinician style and examination room computers: a video ethnography, Ventres [/bib_ref]. Arborelius et al. [bib_ref] In what way may videotapes be used to get significant information about..., Arborelius [/bib_ref] asked GPs if they thought their behaviour was altered on a questionnaire pre and post viewing; 80% reported feeling slightly or not affected, which increased to 90% post viewing of the video. The physicians felt more affected by the presence of the camera than patients. Four other studies mention this as a limitation with no studies giving any empirical evidence to support or refute an effect.
Most studies limited their discussion about ethical implications of the study to a statement about ethics board approval (10 datasets) or that participants consented (14 data sets). In one study, patients were video recorded before their consent was given [bib_ref] The patient and the primary care team: a small-scale critical theory, Timpka [/bib_ref]. Due to the brevity or absence of statements about ethical issues, it was usually unclear what participants had been told was the purpose of the study. In studies where doctor deficiencies were the clear focus of the paper, one wonders if participating GPs knew this in advance, and whether they would have agreed to participate if they had known. In one exception to this, Coleman et al. [bib_ref] Factors influencing discussion of smoking between general practitioners and patients who smoke:..., Coleman [/bib_ref] state that GPs did not know the study was about smoking, presumably to reduce influence of the study on the behaviours and talk of interest. A few studies referred to anonymity and confidentiality, and gave participants the option to withdraw [bib_ref] In what way may videotapes be used to get significant information about..., Arborelius [/bib_ref] [bib_ref] Comparison of patients' and doctors' comments on video-recorded consultations, Arborelius [/bib_ref]. Epstein et al. [bib_ref] Awkward moments in patient-physician communication about HIV risk, Epstein [/bib_ref] disclosed that some GPs were 'visibly upset' when viewing the videos.
The influence of the researcher on the research process was generally under-recognised. Indirectly, this was alluded to in studies using neutral prompts during video playback and participant led recall, to reduce researcher influence. However, beyond this there were no critical reflections whereby authors considered their own role in the research process.
## Acceptability to participants
No studies directly addressed the issue of acceptability of the method to participants. Patient participants have expressed the novelty of watching themselves on screen and directed a number of their comments during playback around this issue. In one dataset, the authors purposely showed the video first in an introductory manner so that participants could become more used to watching themselves on screen, noting that patients 'comment in a neutral and polite way' [bib_ref] Prevention in Practice -how do General-Practitioners Discuss Life-Style Issues with their Patients, Arborelius [/bib_ref] [bib_ref] In what way may videotapes be used to get significant information about..., Arborelius [/bib_ref] [bib_ref] Patients comment on video-recorded consultations-the "good" GP and the "bad, Arborelius [/bib_ref] [bib_ref] What is going on when the general practitioner doesn't grasp the situation?, Arborelius [/bib_ref] [bib_ref] Comparison of patients' and doctors' comments on video-recorded consultations, Arborelius [/bib_ref] [bib_ref] General practitioners' comments on video recorded consultations as an aid to understanding..., Arborelius [/bib_ref]. Acceptability of the method can be inferred to some extent by participant consent rates but only 6 datasets recorded consent rates of patients in any associated paper and none indicated consent rates of GPs. Interestingly, Blakeman et al. [bib_ref] A qualitative study of GPs' attitudes to self-management of chronic disease, Blakeman [/bib_ref] did not incorporate patient VSR into their study design as they anticipated this would be unacceptable to participating GPs. Blakeman has since indicated this assumption was probably unfounded (personal communication).
## Vsr procedure: relationship to research question and study quality
In the Introduction, a classification of six elements of VSR procedure was introduced . This classification comprises: time interval between consultation and VSR; strength of stimulus; structure of interview; who initiates recall; relationship between line of questioning and stimulus and participant training. This classification was used as a lens through which to view the included studies in this review. [fig_ref] Table 9: Techniques of VSR compared with area of research and data used for... [/fig_ref] details the procedures used in each study using this classification. Participant training was not described in any study and similarly the relationship of events on the video to the researchers' line of inquiry in interview was difficult to evaluate in the absence of an interview schedule and so these two elements are not included in the Unfortunately, there is no empirical evidence from this review to comment on the importance of the timing of the VSR event or the strength of the stimulus, due to either a lack of reporting or lack of process evaluation. With regard timing of VSR, 10 papers did not report the length of time between video and VSR event. Of the other 18 studies, the VSR event occurred immediately post consultation in two, and up to two weeks later in the remainder. It was not possible to assess whether the studies with longer intervals had poorer recall. Bugge et al. [bib_ref] The significance for decision-making of information that is not exchanged by patients..., Bugge [/bib_ref] employed more than two post consultation interviews and for some participants, a further telephone interview at six months; it was not clear in this study how the additional post consultation reviews contributed to the results, or how recall differed in each review.
Three author groups enhanced the strength of the stimulus by either showing the video more than once, or by giving the participant a written transcript in addition to the video. Unfortunately, these studies did not evaluate to what extent the additional stimulus elicited additional information from participants.
A number of studies adopted participant-led low structure procedures where the participant was asked to comment on the video with no associated semi-structured interview, and neutral prompts only. As previously suggested, this method would be recommended for exploring decision making; however none of these studies were primarily concerned with decision making. Some studies did not report the nature of the prompts that were given to Ali [bib_ref] The role of culture in the general practice consultation process, Ali [/bib_ref] No mention. States GPs were recorded over a period of time to try and reduce effect Mentions inclusion criteria but doesn't describe these. Not clear in interview if interpreter was used or not, and what questions the patient was asked. Analysis not clearly described. Conclusions appear to be derived from literature review rather than empirical findings.
Characteristics of consenters described in unreferenced related paper only Als [bib_ref] The desk-top computer as a magic box: patterns of behaviour connected with..., Als [/bib_ref] States attempted to recruit a sample of variation, characteristics and consent not described No mention Analysis not described in detail.
Arborelius [bib_ref] Prevention in Practice -how do General-Practitioners Discuss Life-Style Issues with their Patients, Arborelius [/bib_ref] [bib_ref] In what way may videotapes be used to get significant information about..., Arborelius [/bib_ref] [bib_ref] Patients comment on video-recorded consultations-the "good" GP and the "bad, Arborelius [/bib_ref] [bib_ref] What is going on when the general practitioner doesn't grasp the situation?, Arborelius [/bib_ref] [bib_ref] Comparison of patients' and doctors' comments on video-recorded consultations, Arborelius [/bib_ref] [bib_ref] General practitioners' comments on video recorded consultations as an aid to understanding..., Arborelius [/bib_ref] [bib_ref] The GP's dilemmas: a study of knowledge need and use during health..., Timpka [/bib_ref] , Characteristics of consenting patients described but not nonconsenters.
Mentions in 2 papers the influence of the camera was minimal (self-report from participants)
Participant comments during VSR often not aligned to research question as only neutral prompts, therefore small number of comments relevant to study aims [bib_ref] Patients comment on video-recorded consultations-the "good" GP and the "bad, Arborelius [/bib_ref].
Research question not aligned to sampling resulting in small numbers of relevant consultations for some papers [bib_ref] Prevention in Practice -how do General-Practitioners Discuss Life-Style Issues with their Patients, Arborelius [/bib_ref] [bib_ref] What is going on when the general practitioner doesn't grasp the situation?, Arborelius [/bib_ref].
Analysis clearly described in 2 papers in this group [bib_ref] General practitioners' comments on video recorded consultations as an aid to understanding..., Arborelius [/bib_ref].
Possible over-interpretation of participants' comments (particularly assumptions on when GP had failed to 'grasp' situation) [bib_ref] What is going on when the general practitioner doesn't grasp the situation?, Arborelius [/bib_ref] [bib_ref] General practitioners' comments on video recorded consultations as an aid to understanding..., Arborelius [/bib_ref] with limited discussion of implication of findings [bib_ref] Patients comment on video-recorded consultations-the "good" GP and the "bad, Arborelius [/bib_ref] Analysis mostly conducted across case and not within case: within cases analysis and comparison may have enhanced analysis and understanding of cases where difficulties exist in the consultation [bib_ref] Comparison of patients' and doctors' comments on video-recorded consultations, Arborelius [/bib_ref] (where within case approach was used, only 1 minute of consultation analysed.
Blakeman [bib_ref] Bringing selfmanagement into clinical view: a qualitative study of long-term condition management..., Blakeman [/bib_ref] [bib_ref] The Quality and Outcomes Framework and self-management dialogue in primary care consultations:..., Blakeman [/bib_ref] Characteristics of consenting patients and GPs described but not non-consenters.
No mention Data collection, rationale for study and analysis described in detail. Possible limited conclusions to be drawn from the study of one consultation when studying selfmanagement support which may happen longitudinally in the doctor patient relationship.
Only empirical quotes from nurses reported in 2nd paper, yet conclusions refer to doctors and nurses. In 2nd paper, no discussion about how context of nurse or doctor consultation would influence findings in relation to QOF.
Bugge [bib_ref] The significance for decision-making of information that is not exchanged by patients..., Bugge [/bib_ref] Characteristics of consenting patients described but not nonconsenters. Limited characteristics of GPs described
Brief mention as limitation Relative contribution of different post consultation interviews not described (3 per participant).
Analysis well described.
Cegala [bib_ref] A study of doctors' and patients' perceptions of information processing and communication..., Cegala [/bib_ref] Characteristics of consenting patients and GPs described but not non-consenters.
No mention. Effect on behaviour may be more likely as consultation taken out of normal surgery context and separate microphone on table.
Paper based on assumption that participant's spontaneous comments during playback (with no guided prompts) can be used to draw conclusions about patient perceptions of doctor competence in communication exchange.
No information about sampling.
No empirical quotes to support findings.
Coleman [bib_ref] Factors influencing discussion of smoking between general practitioners and patients who smoke:..., Coleman [/bib_ref] Characteristics of consenters and non-consenters presented. GPs sampled to represent a range of attitudes to smoking Discussed as potential limitation. Quantitative methods to support sampling helped gain a maximum variation sample.
Analysis well described.
Author's role as GP and peer to GP participant's not explored. Cromarty [bib_ref] What do patients think about during their consultations? A qualitative study, Cromarty [/bib_ref] No mention of details of video selection or recruitment (videos selected by participating GPs and not researcher)
## No mention
Relative contribution of different phases of post consultation interview not described (unprompted, with video recall and then written transcript).
Analysis not described in depth.
Epstein [bib_ref] Awkward moments in patient-physician communication about HIV risk, Epstein [/bib_ref] Characteristics of consenting patients and GPs described but not non-consenters.
One comment that GPs stated not affected.
Robust analysis strengthened by different approaches including coding of behaviours, attention to conversation flow and classification scheme of the level and depth of discussion of HIV risk.
Discussion of how GPs volunteering to be video recorded may not be representative of GP population.
More than one consultation per GP facilitated robust analysis.
Purposive sampling used to identify patients/ consultations more likely to contain discussion of HIV risk Not clear how video shown or VSR procedure.
Frankel [bib_ref] Patients, doctors, and videotape: a prescription for creating optimal healing environments?, Frankel [/bib_ref] No mention No mention Research question or theoretical framework lacking.
Sample size unclear Participant comments (GP or patient) on video not confidential and revealed to other participant. Consent not mentioned.
Gao [bib_ref] Considering Culture in Physician-Patient Communication During Colorectal Cancer Screening, Gao [/bib_ref] Characteristics of consenting patients described but not nonconsenters. Limited characteristics of GPs described
No mention Recruitment strategy not entirely appropriate: GP interviews not needed to answer research question and weren't utilised.
Three stage analysis clearly described.
Henry [bib_ref] How do you know what Aunt Martha looks like?' A video elicitation..., Henry [/bib_ref] Variation sampling of patients to gain mix of gender, age and race. GPs sampled with respect to years in practice and specialty
No mention Insufficient detail about structure of interview or VSR procedure to judge how appropriate study method was for exploring tacit clues.
No discussion of how context of health maintenance consultations might influence findings.
Rosenburg [bib_ref] Doctor-patient communication in primary care with an interpreter: physician perceptions of professional..., Rosenberg [/bib_ref] [bib_ref] Through interpreters' eyes: comparing roles of professional and family interpreters, Rosenberg [/bib_ref] Characteristics of sample described (patients and interpreters), but not non-consenters
## No mention
Conclusion not supported by results and patient views would have added value and been relevant to research question [bib_ref] Doctor-patient communication in primary care with an interpreter: physician perceptions of professional..., Rosenberg [/bib_ref].
Little information about VSR procedure of format of interview [bib_ref] Through interpreters' eyes: comparing roles of professional and family interpreters, Rosenberg [/bib_ref].
Rosenburg [bib_ref] Intercultural communication competence in family medicine: lessons from the field, Rosenberg [/bib_ref] Recruitment well described. Characteristics of sample described, but unclear how many underwent VSR No mention Method successful in identifying consultations of interest and evidence supports authors' conclusions. No discussions of limitations.
Patients made few comments over video and structure of interview not clear.
Saba [bib_ref] Shared decision making and the experience of partnership in primary care, Saba [/bib_ref] Characteristics of sample described but low consent rate not discussed.
Brief mention of possible effect Robust analysis strengthened by different approaches including analysis within and across cases, contrasting observed and subjective experiences of shared decision making to construct typology of SFM archetypes and using themes from interviews.
Timpka [bib_ref] The patient and the primary care team: a small-scale critical theory, Timpka [/bib_ref] Characteristics of consenting patients described but not nonconsenters.
Brief mention of possible effect Complex study but not clear how much video the participants viewed, the instructions the participants were given when watching the video or the consent arrangements.
Conclusion not supported by results.
participants. Examples of prompts that were reported are listed in [fig_ref] Table 10: Examples of prompts given by researcher during VSR [/fig_ref]. A low structure procedure allows the participant to specify what is discussed but in some cases this method yielded little data. Arborelius et al. [bib_ref] In what way may videotapes be used to get significant information about..., Arborelius [/bib_ref] stated that patients are less likely to comment spontaneously than doctors and Rosenburg et al. [bib_ref] Intercultural communication competence in family medicine: lessons from the field, Rosenberg [/bib_ref] and Epstein et al. [bib_ref] Awkward moments in patient-physician communication about HIV risk, Epstein [/bib_ref] also reported low frequency of comments from patients. In some instances, the small amount of yielded data affected the robustness of the study conclusions, particularly if no additional data was analysed. In a study about the characteristics of a 'human relationship' with a doctor, analysis hinged on 21 of the original 227 patients' spontaneous comments that related to this subject. When doctors were asked to comment on the video with no specific line of inquiry, they usually focused on deficiencies in their behaviour; in one instance the conclusions of the study focused on doctor deficiencies as a result although the study question concerned GP experiences of the consultation [bib_ref] General practitioners' comments on video recorded consultations as an aid to understanding..., Arborelius [/bib_ref].
Conversely, in the studies exploring decision making, there was limited acknowledgement of the possible influence of a semi-structured interview and researcher behaviour in altering participants' accounts of consultation events. However, the use of semi-structured interviews generally elicited more information specific to the research question. Only one study did not use face to face VSR, but instead used a questionnaire to capture GPs' thoughts during video playback in addition to a face to face interview (without VSR); again, the authors did not make clear in the results how the questionnaire results contributed to the findings of the study [bib_ref] Physicians, patients, and the electronic health record: An ethnographic analysis, Ventres [/bib_ref] [bib_ref] Clinician style and examination room computers: a video ethnography, Ventres [/bib_ref].
What does VSR add? The contribution of VSR to findings VSR to explore participants' perceptions VSR was shown to have advantages over a non-stimulated interview approach in three studies with GPs. Firstly, in a study of discussion around smoking cessation, doctor participants showed great surprise at their actions on video; it was apparent from findings presented that the videos had uncovered aspects of behaviour that the GPs had previously not given any thought to, such as the impact of the computer on smoking cessation discussion [bib_ref] Combining qualitative interviews with videorecorded consultations: Gaining insight into GPs' decision-making, Coleman [/bib_ref] [bib_ref] Factors influencing discussion of smoking between general practitioners and patients who smoke:..., Coleman [/bib_ref]. GPs incorporated commentary on the patient's nonverbal response to smoking cessation (viewed on video) to elaborate their accounts. Furthermore, the GPs in this study were asked about the absence of smoking related discussion and without VSR to cue the specific times when smoking could have been discussed, one can hypothesize that un-stimulated recall may not have been as effective. This work showed the importance of the context in which doctors practice in influencing smoking discussions, explaining why few doctors choose to discuss this issue with patients. Coleman et al. attributed the utility of the method to the subject of interest (smoking cessation) being mundane and therefore easily overlooked, and forgotten. In a similar vein, Blakeman et al. reported that VSR was useful for researching 'taken for granted practice'. In their study regarding self-management, a GP expressed annoyance when watching himself weighing a patient revealing insights about the doctor's perceptions of roles, an issue that one can speculate may have been overlooked in a non VSR interview [bib_ref] Bringing selfmanagement into clinical view: a qualitative study of long-term condition management..., Blakeman [/bib_ref]. The third example concerns GPs' reactions to their discussions around HIV risk [bib_ref] Awkward moments in patient-physician communication about HIV risk, Epstein [/bib_ref]. The GPs in this study were 'generally surprised' at their actions and offered unexpected insight into communication barriers, such as the importance of the lack of a simple opening statement in starting HIV risk discussion.
Of the other studies researching patient experiences the added value of VSR was unclear [bib_ref] The role of culture in the general practice consultation process, Ali [/bib_ref] [bib_ref] Patients comment on video-recorded consultations-the "good" GP and the "bad, Arborelius [/bib_ref] [bib_ref] What do patients think about during their consultations? A qualitative study, Cromarty [/bib_ref] [bib_ref] The patient and the primary care team: a small-scale critical theory, Timpka [/bib_ref]. There were no reports of patients showing surprise at the video findings, as has been noted in several VSR interviews with GPs [bib_ref] Combining qualitative interviews with videorecorded consultations: Gaining insight into GPs' decision-making, Coleman [/bib_ref] [bib_ref] Awkward moments in patient-physician communication about HIV risk, Epstein [/bib_ref] [bib_ref] The desk-top computer as a magic box: patterns of behaviour connected with..., Als [/bib_ref]. One interpretation may be that VSR is more useful for enhancing reflection in clinicians; however, the studies with patients had a number of methodological limitations. In general, the lack of detail around methods was accompanied with insufficient detail in results to judge the added value of VSR.
## Vsr to explore non-spoken behaviours
In two studies, non-verbal events were the focus of the research question and the VSR. Bugge et al. [bib_ref] The significance for decision-making of information that is not exchanged by patients..., Bugge [/bib_ref] explored the significance of non-disclosure of information during decision making. In this study the value of VSR was evident; clinicians reported information they typically sought in certain decision making situations, but the video consultations revealed the absence of the reported behaviour. During the VSR interviews the authors were able to unpick the reasons for non-disclosure including assumptions about patient preferences and uncertainty about treatment effectiveness. As clinicians were clearly not aware of some episodes of non-disclosure prior to viewing, a non- [fig_ref] Table 8: Findings from quality appraisal [/fig_ref] Findings from quality appraisal (Continued)
TreichlerCase study of one patient. No mention of sampling.
No mention Limitations associated with the study of one consultation.
Ventres [bib_ref] Physicians, patients, and the electronic health record: An ethnographic analysis, Ventres [/bib_ref] [bib_ref] Clinician style and examination room computers: a video ethnography, Ventres [/bib_ref] Not described Brief mention Analysis well described but no empirical quotes to support findings. More description of consultation context would have increased credibility of findings. Pt VSR comments [bib_ref] Prevention in Practice -how do General-Practitioners Discuss Life-Style Issues with their Patients, Arborelius [/bib_ref] [bib_ref] In what way may videotapes be used to get significant information about..., Arborelius [/bib_ref] [bib_ref] Patients comment on video-recorded consultations-the "good" GP and the "bad, Arborelius [/bib_ref] [bib_ref] Comparison of patients' and doctors' comments on video-recorded consultations, Arborelius [/bib_ref] Difficult consultations [bib_ref] Patients comment on video-recorded consultations-the "good" GP and the "bad, Arborelius [/bib_ref] [bib_ref] What is going on when the general practitioner doesn't grasp the situation?, Arborelius [/bib_ref] [bib_ref] The GP's dilemmas: a study of knowledge need and use during health..., Timpka [/bib_ref] 46Pt GP asked to comment if unsure how to proceed GP VSR comments [bib_ref] Prevention in Practice -how do General-Practitioners Discuss Life-Style Issues with their Patients, Arborelius [/bib_ref] [bib_ref] In what way may videotapes be used to get significant information about..., Arborelius [/bib_ref] [bib_ref] What is going on when the general practitioner doesn't grasp the situation?, Arborelius [/bib_ref] [bib_ref] Comparison of patients' and doctors' comments on video-recorded consultations, Arborelius [/bib_ref] [bib_ref] General practitioners' comments on video recorded consultations as an aid to understanding..., Arborelius [/bib_ref] [bib_ref] The GP's dilemmas: a study of knowledge need and use during health..., Timpka [/bib_ref] Doctor patient relationship [bib_ref] Comparison of patients' and doctors' comments on video-recorded consultations, Arborelius [/bib_ref] [bib_ref] General practitioners' comments on video recorded consultations as an aid to understanding..., Arborelius [/bib_ref] Health promotion [bib_ref] Prevention in Practice -how do General-Practitioners Discuss Life-Style Issues with their Patients, Arborelius [/bib_ref] 12GP Pt and GP questionnaire post viewing (effect of video on behaviour and satisfaction with consultation) [bib_ref] In what way may videotapes be used to get significant information about..., Arborelius [/bib_ref] Consultation [bib_ref] Prevention in Practice -how do General-Practitioners Discuss Life-Style Issues with their Patients, Arborelius [/bib_ref] [bib_ref] What is going on when the general practitioner doesn't grasp the situation?, Arborelius [/bib_ref] [fig_ref] Table 9: Techniques of VSR compared with area of research and data used for... [/fig_ref] Techniques of VSR compared with area of research and data used for triangulation (Continued) [bib_ref] Intercultural communication competence in family medicine: lessons from the field, Rosenberg [/bib_ref] Cross cultural communication 24 (24) stimulated interview could not have reached the same findings. This study also gives further weight to the suggestion that VSR may be particularly useful for doctors. Henry et al. [bib_ref] How do you know what Aunt Martha looks like?' A video elicitation..., Henry [/bib_ref] identified how tacit clues, including non-verbal behaviours, subconsciously inform clinical judgements. In this study, patients were found to be very attuned to doctor body language and doctors often unaware or unable to articulate rationale behind their judgements; however, doctors were found to have a varying sensitivity to tacit clues. Both of these studies have useful implications for our understanding of doctor patient communication and necessitated a VSR approach due to the specific nonverbal or nondisclosure event in the consultation that needed further elucidation.
## Vsr in conjunction with consultation analysis
In this review, the included studies varied in the extent to which different sources of data contributed to the overall analysis, as detailed in [fig_ref] Table 9: Techniques of VSR compared with area of research and data used for... [/fig_ref]. In the studies where the consultation was analysed alongside the VSR interviews, a number of different methods of analysis were used. Analysis was conducted both 'across cases' , and 'within cases'. In across case analysis, VSR interviews were analysed as a whole with no comparison to the relating consultation; in within case analysis, the consultation and VSR transcripts pertaining to one consultation were analysed together.
In the studies using within case analysis, the added value of using VSR was clearly evident. The use of VSR was particularly illuminating in a study exploring shared decision making and the experience of partnership. By comparing and contrasting physician and patient views on episodes of decision making, Saba et al. have been able to shed light on previous work that has identified discordance between satisfaction and shared decision making in consultations [bib_ref] Shared decision making and the experience of partnership in primary care, Saba [/bib_ref]. This study has demonstrated that shared decision making could occur in the presence of mistrust and frustration, and they conclude that both good communication and relationship dynamics are necessary for shared decision making. A further example of the strength of the within case analysis approach comes from Rosenburg at al's study of intra-cultural encounters [bib_ref] Intercultural communication competence in family medicine: lessons from the field, Rosenberg [/bib_ref]. The detailed descriptions in the paper of consultation excerpts alongside patient and doctor responses during interview enabled the authors to draw novel insights about areas for improvement in intracultural encounters, again with important educational implications.
The use of VSR to study specific instances of sensitive talk around HIV risk [bib_ref] Awkward moments in patient-physician communication about HIV risk, Epstein [/bib_ref] was also very successful in identifying the successful elements of HIV risk discussion, with educational implications. Although the VSR component seemed to contribute a small amount to the study findings (compared to consultation analysis), the GP interviews did appear to be useful in eliciting the nature of barriers to effective discussion. In this example, the research participant almost becomes researcher, aiding the interpretation of findings.
In studies that analysed consultations in a silo independent of VSR findings (across case analysis), the analysis was felt to be lacking in depth and rigour with missed opportunities for insight from the data [bib_ref] The desk-top computer as a magic box: patterns of behaviour connected with..., Als [/bib_ref] [bib_ref] Comparison of patients' and doctors' comments on video-recorded consultations, Arborelius [/bib_ref].
# Conclusions
This review highlights that VSR is particularly useful for the study of specific consultation events when analysis adopts both a within and across case approach. For enhancing participant recall, VSR may be particularly relevant for topics which are routine and easily overlooked, for interviewing doctors and for exploring non-spoken and non-verbal behaviour. The method may be particularly useful for exploring clinicians' perceptions, as differences in rhetoric and behaviour can be explored; the use of interviews alone to research doctor perceptions has been criticised [bib_ref] Ticking boxes and changing the social world: Data collection and the new..., Checkland [/bib_ref] [bib_ref] Critical reflections on the rise of qualitative research, Pope [/bib_ref] and VSR may provide a useful alternative. Blakeman et al. [bib_ref] Bringing selfmanagement into clinical view: a qualitative study of long-term condition management..., Blakeman [/bib_ref] state that the method helps to explore interactions that may have remained unremarkable to both participant and researcher, particularly where the researcher has the same professional background as the participant ('shared conceptual blindness').
In reviewing study quality, frequently there was insufficient reporting of methods to properly evaluate this; one contributory factor to this may be that many journals' word limits may not facilitate proper reporting of complex methodology. Ensuring the technique of VSR, the study sampling and the choice of data sources align to the research question have emerged as particularly important elements in the quality of these studies. VSR studies may generate a lot of data, and care needs to be taken to ensure data collected are relevant to the research question, and represented in the study findings. Studies identified in this review have generally not used opportunities to evaluate their methods e.g. by reporting how un-stimulated recall compared to recall, or how different aspects of data contributed to findings. Stop the tape when you felt uncertain as how to go on [bib_ref] The GP's dilemmas: a study of knowledge need and use during health..., Timpka [/bib_ref] Comment on anything new, unusual or different [bib_ref] In what way may videotapes be used to get significant information about..., Arborelius [/bib_ref] What do you think when you look at the videotape? [bib_ref] In what way may videotapes be used to get significant information about..., Arborelius [/bib_ref] Stop the tape when you identify thoughts feelings or behaviours associated with decision making [bib_ref] Shared decision making and the experience of partnership in primary care, Saba [/bib_ref] Stop the tape at moments you feel important or where you wish to comment, describe what you were thinking or feeling (Preceded with reminder of study focus -communication and cultural differences) [bib_ref] Intercultural communication competence in family medicine: lessons from the field, Rosenberg [/bib_ref] Tell me what was happening [bib_ref] A qualitative study of GPs' attitudes to self-management of chronic disease, Blakeman [/bib_ref] As stated in the introduction, there is concern, particularly in the psychology and sociology literature about the method of VSR producing 'a second-order reconstituted account' [bib_ref] Recording social life: Reflexivity and video methodology, Lomax [/bib_ref] , influenced by the degree of researcher 'interference' in the process of VSR. Few authors commented on this limitation, with some [bib_ref] Awkward moments in patient-physician communication about HIV risk, Epstein [/bib_ref] stating the counter argument, that using participants as experts to interpret their own behaviour yielded unexpected issues. To some extent the argument here will be influenced by a researcher's theoretical and epistemological viewpoint; a post-positivist approach would align with the need to maximise validity and reduce researcher interference, whereby an interpretivist approach would sit more comfortably with the need to respect the differences between viewpoints and make sense of findings using the meanings derived from the 'actors' within the consultation. In the papers included, researchers did not make their viewpoint explicit. However, the majority of studies did aim to elucidate participant experience in some way, and as such vigorous attention to validity of recall may be less important than research in other disciplines where the concern is to accurately reflect cognitive processes.
In this review, studies which have tried to reduce researcher interference, for example by using only neutral prompts during VSR, have often resulted in small amounts of data, much of which was unrelated to the research question. This may have been due to lack of participant or researcher training in the method. The findings of this review suggest that although the limitations of moderate to high structure reviews/ post consultation interviews should be acknowledged, that these methods usually resulted in richer data related to the research question than low structure, participant-led approaches. Prompts given by researchers during playback may still remain 'neutral' while providing a context e.g. study aim or orientation for the participant to comment.
VSR is an intrusive methodology and it is likely that ethical issues arise during the conduct of these studies, such as patient distress during video review. Guillemin and Gillam refer to this as 'ethics in practice' as opposed to 'procedural ethics' , concerned with consent processes and formal approval [bib_ref] Ethics, reflexivity, and "ethically important moments" in research, Guillemin [/bib_ref]. No study referred to any ethical issues arising during data collection. Related to this is the issue of acceptability, and how participants react to VSR, which remains unknown.
Lomax [bib_ref] Recording social life: Reflexivity and video methodology, Lomax [/bib_ref] argues a reflexive stance is essential when collecting video data as the entire research process has a distorting effect on 'real life'. Increased reporting of the ethical issues 'in practice' and the influence of the researcher on the process and would increase the quality of reporting of these studies. These issues are common to other qualitative research [bib_ref] No room for reflexivity? Critical reflections following a systematic review of qualitative..., Newton [/bib_ref] , although particularly relevant to VSR, as evidenced by the distress during VSR described in one study [bib_ref] Awkward moments in patient-physician communication about HIV risk, Epstein [/bib_ref]. This review was conducted with a systematic search. Searching all papers containing reference to video for evidence of VSR, instead of restricting the search by identified terms for VSR, has identified more studies than a previous literature review [bib_ref] Video elicitation interviews: a qualitative research method for investigating physician-patient interactions, Henry [/bib_ref] , which also did not quality appraise identified studies. A strength of this review is the use of quality assessment, using the CASP toolto both inform results and underpin conclusions. Furthermore, the use of the classification described by Gass and Mackey as a theoretical framework to inform analysis has resulted in practical conclusions that will hopefully assist researchers considering the use of the method. No study was excluded based on methodological quality and the heterogeneity of studies may limit the robustness of the synthesis. The most striking difference was in the design and reporting of participant consent in older studies, possibly conducted in an era where the use of video was not as widespread as it is today.
In summary, this systematic review furthers understanding of both the role of VSR in understanding the consultation and the methodological strengths and weaknesses of this approach. Future researchers using the method may consider factoring in process evaluation to gain further understanding of how VSR contributes to recall, the acceptability to participants and how changes to methodology influence findings.
## Additional files
Additional file 1: Example Search in Medline. Full search history conducted in Medline database.
[fig] Figure 1: Phases of identification of papers. [/fig]
[table] Table 2: Search terms used [/table]
[table] Table 3: Inclusion and exclusion criteria [/table]
[table] Table 4: Quality assessment items derived from CASP checklist [/table]
[table] Table 5: Data extraction questions [/table]
[table] Table 6: Reasons for exclusion [/table]
[table] Table 7: Description of included studies To provide a detailed understanding of the ways in which white and South Asian patients communicate with white GPs and to explore any similarities and differences in communication To compare the patients' and the doctors' comments on video-recorded consultations in order to increase understanding of shortcomings in patient-doctor relationship To describe and understand the experiences of general practitioners in consultations To describe and understand patients' positive and negative experiences of General Practitioners To describe the specific behaviour in consultations where the patient experiences a satisfying human relationship with the GP [/table]
[table] Table 8: Findings from quality appraisal [/table]
[table] Table 9: Techniques of VSR compared with area of research and data used for triangulation [/table]
[table] Table 10: Examples of prompts given by researcher during VSR [/table]
|
Clinician views on best practice community care for people with complex emotional needs and how it can be achieved: a qualitative study
Background: Individuals with Complex Emotional Needs (CEN) services, a working description to refer to the needs experienced by people who may have been diagnosed with a "personality disorder", face premature mortality, high rates of co-morbidity, service user and treatment costs. Service provision for this population is recurrently identified as needing to be transformed: there are serious concerns about quality, accessibility, fragmentation of the service system and the stigma and therapeutic pessimism service users encounter. Understanding clinician perspectives is vital for service transformation, as their views and experiences shed light on potential barriers to achieving good care, and how these might be overcome. In this study, we aimed to explore these views.Methods:We used a qualitative interview design. A total of fifty participants from a range of professions across specialist and generic community mental health services across England who provide care to people with CEN took part in six focus groups and sixteen one-to-one interviews. We analysed the data using a thematic approach.Findings: Main themes were: 1) Acknowledging the heterogeneity of needs: the need for a person-centred care approach and flexibility when working with CEN, 2) 'Still a diagnosis of exclusion': Exploring the healthcare providerlevel barriers to providing care, and 3) Understanding the exclusionary culture: exploring the system-based barriers to providing care for CEN. Across these themes, staff highlighted in particular the need for care that was person-centred, relational, empathic, and trauma informed. Major barriers to achieving this are stigmatising attitudes and behaviour towards people with CEN, especially in generic mental health services, lack of development of coherent service systems offering clear long-term pathways and ready access to high quality treatment, and lack of well-developed structures for staff training and support.Discussion: Overall, the findings point towards clinician views as generally congruent with those of service users, reinforcing the need for priorities towards systemwide change to ensure that best practice care is provided for people with CEN. Particularly prominent is the need to put in place systemwide training and support for clinicians working with CEN, encompassing generic and specialist services, and to challenge the stigma still experienced throughout the system.
# Introduction
Approximately 4.4% percent of the population have needs that meet the diagnostic criteria for "personality disorder", a diagnosis that is associated with high rates of co-morbidity, premature mortality, high levels of service useand high treatment costs.
There are serious concerns about the quality and accessibility of servicesand there is limited evidence to inform service improvement and the implementation of more acceptable, effective, and cost-effective models of care. It has also been acknowledged that there is a need to make "personality disorders" central to strategies for improving care and ensuring that young people receive timely and appropriate care, similar to the efforts made for other severe mental disorders, such as mood and psychotic disorders.
Despite evidence for structured multimodal treatment models for patients with a "personality disorder" -such as Dialectic Behaviour treatment, Mentalization-based therapy or Transference focused therapy, challenges remain within the research towards understanding the best ways to develop services for people who experience these problemsand to address the barriers and challenges to achieving the best care for them, especially with young people. In considering initiatives to improve the quality and reach of services, co-production with service users will be essential to improve the acceptability and accessibility of services. Understanding clinician perspectives is also vital because they are responsible for delivering care, and therefore their views and experience may directly assist in identifying potential barriers to achieving good care, and how these might be overcome.
The use of the term "personality disorder" raises questions and debates within the field as it has been seen as a problematic label by some staff, patients and carers. Within the literature, service users often report negative consequences as a result of being diagnosed with a "personality disorder", including stigmatisation and exclusion from services. Stigmatising attitudes among mental health professionals and a lack of therapeutic optimism are viewed by service users as significant barriers to the delivery of best practice care. With these issues in mind and through consultation and collaboration with service users and lived experience researchers, in this paper we use the working term Complex Emotional Needs (CEN) to refer to the needs experienced by people who are likely to be diagnosed with a "personality disorder". We use this term CEN in preference to "personality disorder", recognising that view and experience the latter as pejorative and stigmatising.
Our recent qualitative meta-synthesis also conducted as part of the NIHR Mental Health Policy Research Unit's programme to inform NHS policy in this area, summarised literature on clinician perspectives on good practice in community services for people with CEN (e.g., the need for high quality holistic and personalised care, and the importance of having continuous therapeutic relationships and care that is proactive and collaborative). These priorities were mainly congruent with those found in other studies on service user and carer viewsand the needs of policy makers. While there was overall agreement regarding what best practice should be, more research is needed to understand the barriers and facilitators to achieving it.
This study seeks to fill this gap by exploring staff views on best practice provision of community mental health care for people with CEN and identifying the barriers and facilitators to providing such care. In this study, we focus on the views of staff about the delivery of community mental health care, including both specialist and generic services. This study is part of a programme of research delivered by the NIHR Mental Health Policy Research Unit to inform the design, development, and delivery of community mental health services for this population in England. The study objectives were:
1. To explore the experiences of staff working with individuals with CEN in community mental health services, including NHS and voluntary organisations across a range of English service settings. 2. To explore the views of staff working with individuals with CEN about how community services can best meet their needs.
# Method
## Design
Qualitative interview study, using individual and focus group interviews.
# Conclusions:
Staff working with this service user group report that delivering best practice care requires services to be flexible, integrated, and sustainably funded, and for staff to be supported through ongoing training and supervision.
## Research team and perspectives
This study was co-produced from inception, through design and in delivery by six members of the Mental Health Policy Research Unit (UF, RS, KT, SO, AS, SJ b ), a group of four experts by experience (EB, DA, SJ a , TJ) and nine experts by occupation (MC, PM, SMN, JB, OD). The Mental Health Policy Research Unit (MHRPU) helps those involved in making nationwide plans for mental health services make decisions based on good evidence. The MHPRU makes expert views and evidence available to policymakers in a timely way and conducts research that is directly useful for mental health policy. A co-production group was established at the outset of the study and group meetings occurred approximately every quarter between July 2019 and December 2020; work on co-producing the manuscript for publication was conducted via video-calls and emails through early 2021 with all manuscript authors. Co-production work ran throughout the course of the study (e.g., via coproduction of the study protocol, interview topic guide and analyses). Two researchers (UF & KT) led the focus groups and interviews alongside an expert by experience present during focus group sessions (EB, TJ, DA). Alongside two members of the MHPRU research team (UF, RS), experts by experience conducted coding of interview transcripts, with input from experts by occupation for reviewing of the themes (AS, SJ b , JB). This range of inputs was designed to increase the validity of our analyses; capturing content regarding clinical expertise as well as those aspects of care understood by experts by experience that may be lost when using a traditional researcher lead approach. Data analyses were jointly undertaken by all manuscript authors, with key themes established in an iterative process through review, reflection and discussion. By working collaboratively with researchers, experts by experience and experts by occupation we were able to engage in a reflexive process in which we explored the data through these differing lenses and thus were able to reflect more effectively on the effect that these personal perspectives may have on our interpretation. This approach allowed us to move beyond a content analysis of seeking out the 'consensus view' towards a reflective thematic analysis in which we see alternative views to address potential blind spots that each researcher brings with them thereby increasing the validity of the analysis. Researchers in the MHPRU (KT) and an expert qualitative researcher (Dr Nicola Morant) provided training on conducting qualitative interviews (including roleplay exercises and an observation of a live interview) and training on analysing qualitative research with the lived experience members of the team. An MHPRU researcher (Sarah Carr) supported KT in preparing a protocol around promoting the emotional well-being of lived experience researchers who conducted the research interviews. One of the experts by occupation (JB, a clinical psychologist) offered the lived experience researchers individual support sessions about the emotional content and impact of the work throughout the interview period.
## Inclusion and exclusion criteria
Participants were eligible to take part in the study if they worked in community and outpatient settings, inclusive of those working in specialist "personality disorder" or CEN services, or those providing non-specialist services such as generic mental healthcare to people with CEN. Furthermore, criteria were extended to include those working in social services and the third sector who may be working with service users with a "personality disorder" diagnosis or CEN who do not have or do not want access to statutory services. As the focus was on community services, we excluded participants working in forensic, crisis or inpatient settings.
## Recruitment and sampling
Information about the study was circulated via email and social media (e.g., Twitter) to a range of networks, including relevant professional organisations and networks, and workshops/conferences. Interested participants contacted our research team who provided further information about the study, ascertained whether they met the study inclusion criteria, and organised a date for the interview or focus group to take place. Purposive sampling was used to ensure representation of a range of professions and staff who work directly with people with CEN, and of several types of setting, including specialist CEN services, and relevant generic and voluntary sector services. We recruited participants from across England to capture a range of service models and settings.
## Participant characteristics
A total of 50 participants took part in six focus groups and 16 one-to-one interviews. The sample included 27 women and 23 men. Forty-two participants identified as White, four as Asian, two as Black Caribbean, and two as having other or mixed ethnic background. Participants were recruited from across England, including the Midlands (n = 19), North West (n = 10), North East (n = 6), South West (n = 6), London (n = 5), and South East (n = 4).
Twenty-one participants reported working in specialist 'personality disorder' or CEN services, with 29 working in generic community mental health services where individuals with CEN were frequently supported. Most participants were working in NHS services (n = 32), with the remaining working in local authority social care settings (n = 5) or voluntary/third sector services (n = 13).
Participants came from a range of professional backgrounds: psychologists (n = 9) and assistant/trainee psychologists (n = 5), support workers (n = 6), social workers (n = 6), peer workers and experts by experience (n = 5), psychotherapists and counsellors (n = 5), nurses (n = 4), occupational therapists (n = 3), psychiatrists (n = 3), and commissioner and/ or managers (n = 2).
## Data collection
Data were collected between July 2019 and October 2020. We conducted focus groups and one-to-one semi-structured interviews according to participant preference. Focus groups and interviews were conducted in person or via telephone prior to the COVID-19 pandemic and online via Microsoft TEAMs and Zoom following the introduction of social distancing in March 2020. For remote sessions, participants read and completed online consent forms in advance of the interview, and the researcher reconfirmed consent at the start of the session with recorded verbal consent taken from all participants.
Topic guides were co-produced and agreed upon within study meetings with members of the MHPRU research team, the Lived Experience Advisory Group and service user representatives, and experts by occupation. These guides consisted of key questions including a) how services currently operate; how services can address the needs of service users with CEN; b) how services can support staff to deliver best care; c) what best practice looks like; d) what the challenges and facilitators are to providing this care; e) what the important outcomes for services and for service users are; f ) what training and support is needed. Interviews were audio-recorded using either an encrypted digital recorder or the online platform recording option and saved to an encrypted server. Each session lasted between 45 to 90 min. Audio recordings of focus groups and interviews were transcribed by an external company. A member of the research team checked the transcripts for accuracy and pseudonymised all participants. All transcripts were allocated a unique ID number and imported to NVivo ProV12for analysis.
# Analysis
We adopted a phenomenological approach, a form of qualitative enquiry that emphasizes experiential, lived aspects to ensure we were fully capturing participants experiences and perspectives. Specifically, we undertook a reflexive thematic analysis following the phases outlined by Braun and Clarke. Guided by the overarching study aim to explore the experiences of staff in providing care for people with CEN, the first author (UF) read over the transcripts and then re-read them, noting aspects of interest. Following this, members of the research team (EB, DA, AS, RS) who comprised a range of experts by experience and experts by profession, reviewed the transcripts to identify potential codes. Codes were defined as a single idea associated with a segment of data, identifying what is of interest in the data. We used Nvivo12to organise the extracted chunks of data and associated codes. Following initial coding, related codes were sorted, grouped and labelled as preliminary themes, defined as the central concepts that capture and summarize the core point of a coherent and meaningful pattern in the data. Themes were checked and discussed with other members of the team (SJ b , JB, KT) to ensure they were capturing something significant or noteworthy in the data and captured an important aspect relative to the research aims. To improve validity, transparency, and quality in our analysis, we constantly reviewed the themes within the wider study group. Anonymised data extracts illustrate each theme and key analytic points.
## Ethics
Ethical approval was granted by the Psychiatry, Nursing and Midwifery Research Ethics Subcommittee of King's College London (reference HR-18/19-10,795) prior to data collection with a later amendment to move data collection online in response to COVID-19 restrictions and guidance.
## Findings themes
Three overarching themes were developed, each with explanatory subthemes outlined in.
## Acknowledging the heterogeneity of needs: the need for a person-centred care approach and flexibility when working with cen
There was consensus amongst participants that there is a need to acknowledge the heterogeneity of needs of service users and that care therefore needs to have flexibility and be person centred at its core to address the needs of the service user.
## Working with the person not the diagnosis
Several elements of best practice community care for people with CEN were detailed within the study. Underpinning these principles was the importance of person-centred, individualised care:
"It's about the ethos of person-centred care… People just need compassion. They need to be loved. As a nurse, that is what people need to do. That is what people need to understand, is that a lot of people, they just need people to listen. " (ID 27, female, Expert by Experience working in specialist services).
Participants also described the importance of responding flexibly to the fluctuating and often complex and intersectional psychosocial needs of service users.
"Creation of inclusive pathways and overarching views of the person in multiple systems is needed and needs flexibility to ensure that the pathways bend and flex to the chaos that someone is experiencing. " (ID04, female, Occupational Therapist working in specialist services).
It was felt that adapting to the heterogeneous needs and preferences of individual service users necessitated a move away from 'one-size-fits-all' models of service delivery towards more flexible care pathways. Participants also described that responding to the range of service users' needs (e.g., psychosocial, relational, and medical needs) required collaboration not only between specialist and generic mental health services but also between mental health services and primary care, social services and the third sector.
"This just fits with what the Royal College of Psychiatrists is saying: you have a service that spans multiple services and there is something that keeps particular people in mind so that they're not just disposed of by putting them into a different service or disposed of by not being allowed into one service. " (PT32, male, Occupational Therapist).
Coproduction and the involvement of experts by experience was felt by staff to be a key element that could improve services.
"There just needs to be more partnership. It goes back to the fundamental parts of how services should be designed around relationships. There needs to be a relationship between services and this cohort of people so that they are influencing policy and services at commissioner level at the very least. That will ultimately improve people's lives and save money and make clinicians' lives easier as well. It just makes everyone's life easier; I think. " (PT17, male, social worker).
Overall, staff views across the range of professionals agreed that to achieve the best care for those with CEN required a person-centred approach that allowed flexibility to address the service users' needs which may differ vastly from one person to the next. Empathy, compassion, and authenticity -characteristics reported to be key to the provision of best practice -were described as fundamental to the provision of relational care and to the modelling of supportive but boundaried relationships.
"Boundaries are important… it is core to the work that you need to be boundaried as an individual and as an organisation. They [the service user] will violate them and it is the consistency and the way you continue to support them but you do need to have some kind of boundaries and rules but you have to be consistent and explain why so they can learn that abuse isn't okay but you will keep with them to change and support them. " (PT13, female, clinical psychologist).
For this theme it could be seen that best practice was described as holding boundaries and caring enough to have challenging and difficult times together. This was felt to create trust and therapeutic bond necessary for this work. As a result, relationships were regarded as a crucial aspect of care and as an active agent of change for those with CEN.
## The staff characteristics needed to deliver this care
Participants suggested that characteristics such as empathy and compassion enabled staff to see and relate to service users as individuals rather than simply their diagnosis or behaviours, providing care tailored to individuals: A capacity for reflection and for authenticity in relationships were also seen as essential to the provision of best practice care. Participants identified core qualities of empathy, compassion, authenticity, flexibility, supportiveness, and good interpersonal skills as essential in ensuring staff can build and sustain therapeutic relationships with people with CEN. Staff need to have the strength to hold a person who has intense emotional responses to life and provide a 'window of tolerance' so the person can bear the anguish they feel and learn to tolerate it. They also recognised that training and supervision are needed to develop and sustain these skills.
[formula] " [/formula]
## Acknowledging underlying trauma
Participants highlighted the importance of addressing trauma for those with CEN and not focusing solely on presenting symptoms but also engaging with the wider range of difficulties that service users experience and their causes. This included considering how adverse experiences have affected people. Trauma-informed approaches were suggested to provide a way of working with people that was collaborative and empowering. However, participants also conceded that there was a lack of clarity or consensus as to what constitutes a 'trauma-informed approach' .
"I hear a lot of the same things from people across the country. I don't think it's an issue in one area, but it is interesting that you say that everyone says, "Oh, I'm trauma-informed, " but it's not necessarily used right. " (ID26, female, Senior Practitioner in a CEN service).
This acknowledges that best practice often meant that staff feel they have to do things that service users don't want at times. It's about connecting with service users and providing care with integrity, transparency and compassion.
"For trauma-informed, we would want our managers to have the same, to be able to provide the supervision around what's needed. We'd need ongoing training because we've got new staff coming and going all the time. " (PT18, male, counsellor in general mental health services).
By acknowledging the role that trauma may have, staff felt this could inform their work as well as provide insight and understanding into how the service user experiences relationships. By acknowledging this complexity staff felt that it helps keep the individual and their experiences at the centre of care.
## 'still a diagnosis of exclusion': exploring the healthcare provider level barriers to providing care
While there was a consensus on what best practice looks like, significant barriers were identified to achieving it. These focused on the providers: fear of the risks associated with complex emotional needs, staff feelings of hopelessness to help and create change, the challenge of working in a culture of stigma and labelling, and the need to mitigate these barriers through supporting staff.
## Provider fear of the risks associated with complex emotional needs
Participants described that some staff in generic mental health services were fearful about working with people with CEN because of the severe distress, crisis, recurrent self-harm, and suicidal ideation or intent frequently experienced by this group. Managing such crises and high levels of risk were reported by staff members as leaving them distressed and feeling an intense sense of powerlessness which was felt to link to therapeutic nihilism, stigma and exclusion.
## "[staffs] level of fear gets so high that they think, "i don't have the resources to do this. they need to be somewhere else. " [this leads to a] buck-passing phenomenon and it's like when people go from one service to another to another to another. " (id32, male, occupational therapist).
Participants suggested that such avoidant and riskaverse behaviours were exacerbated by staff feeling overworked and by a lack of resources and demands to hit targets rather than provide the most effective care for the individual.
## "when you've got a caseload of 50, you'd only need to have 5 people with emotional difficulties on your caseload and your focus is then drawn to them because of the way that they present at services. it would then have a knock-on effect on the outcomes for the other people that were on that person's caseload, you know, and you've got a manager saying, "well, they've done a 12-week dbt programme. discharge them off now. " well, that's not going to work for this person, but you're stuck by the constraints of the service that you work for and their delivery targets, their kpis. " (id19, female, social worker working in generic services).
Overall, staff felt it important to reflect on the nature and impact that working with crises and high-risk behaviours have on the staff as well as understanding the personal and organisational needs when it comes to these aspects of care for CEN.
## Staff feelings of hopelessness to help and create change
For participants working in generic services and in the third sector described finding it challenging to work with service users with CEN, because they felt they could not provide effective help. As a result, some of these staff felt that their role was not about therapeutic support and more about striving to keep people alive.
"You make a connection with somebody, you come to care for them. You don't want them to die. " (ID18, male, counsellor in general mental health services).
The lack of therapeutic optimism described by some staff was compounded by the lack of resources and services described above. Participants noted that they had left roles in statutory services as a direct result of these issues. Understanding this aspect of care was felt as imperative to understanding the compassion fatigue that was experienced by staff, as well as understanding where stigmatising attitudes may develop. These reflections show the interplay between the personal hopelessness but also the frustrations related to the organisational limits and complexity facing caring for individuals with CEN.
## The challenge of working in a culture of stigma and labelling
Participants highlighted the stigmatisation of individuals with CEN as a systemic problem within healthcare services, with service users excluded from or receiving poorer care from services as a result of their diagnosis. This was particularly felt to be an issue in generic mental health services in which some staff reported having changed the way they employed diagnostic terminology to help service users avoid such exclusionary outcomes.
"Every time I assessed someone; I always say 'Diagnostic uncertainty' because that meant you were allowed to give them a service. So, you gave them a service, but you didn't give them a label, then they could come and get a couple of years of help. " (ID16, male, third sector service lead).
Participants reported having witnessed stigmatisation and its consequences. Stigmatising attitudes were described as engrained in the culture of services and attributed to a lack of knowledge about and misunderstanding of CEN as well as negative experiences with service users with "personality disorder" diagnoses.
## "i think a lot of people just think it is too difficult. it is not as easy to treat as psychosis, depression, all these comparatively easy-to-define presentations. just i think that they find it easier when it fits into their training and mindset and belief structure. so, coming up against that because it is the dominant-i suppose it is the dominant world view in a lot of services. " (id17, male, social worker).
However, participants also suggested that stigmatising and negative attitudes reflected an attempt by staff to avoid the emotional impact of their work, serving as an unhealthy coping mechanism where support, supervision and training was lacking.
## "they often are seeing people in crisis, having selfharmed, on in-patient wards... i think if i was a nurse, i probably would also say things that aren't very nice if i didn't have the containment. " (id28, male, service lead).
Regardless of cause, stigmatisation was understood by participants to be a major personal and system-wide barrier to the provision of best practice care for people with CEN. By understanding the underlying causes of such negative attitudes, it was felt that training and support for staff could be developed to help tackle such cultures that have such a negative impact on service users and their care.
## The need to mitigate these barriers through supporting staff
Participants placed high value on training that was coproduced and co-delivered with people with lived experience of complex emotional needs, which, alongside lived experience involvement in service delivery and design, helped provide an "empathy injection" to an area often at risk of stigma and preconceptions.
"It's the collaborative training, that is absolutely crucial for practitioners to hear. Because people-We have a very narrow view of each other, we have huge judgments. If we don't meet people, we know that any kind of attitude change means that you need to understand from the person's point of view and put yourself in their shoes. " (ID47, male, Clinical Psychologist in generic mental health services).
Participants also described the need for practitioners in both generic and specialist services to be provided with supervision as well as reflective practice to manage the challenges and emotional demands of providing this care.
## "that value of supervision and making sure it happens. i could probably go on about that for a while, but i think the basic theme is get people thinking about people in a way that promotes empathy and
give them enough skills to make them feel confident and competent in working with that client group. " (ID32, male, occupational therapist working in non-specialist services).
Participants valued the role modelling aspect of supervision and training: the relationships built with supervisors could reflect their aspirations for their therapeutic relationships, for example in providing safety, opportunities for reflection, and the sense of being compassionately held:
"I think the training wants to role-model reflection from the trainer, and that they're emotionally containing of the staff who are coming on the training. " (ID14, female, social worker in a homeless service).
While challenges and barriers exist for staff working with CEN it was viewed that staff skill, which was identified as important to the provision of best practice, could be enhanced through training and support. As a result of providing more training and supporting staff more actively, these barriers and challenges could be mitigated and thus reduce the impact of compassion fatigue, bad experiences, and exposure to poor practice.
## Understanding the exclusionary culture: exploring the system-based barriers to providing care for cen
To fully understanding the exclusionary nature of the "personality disorder" diagnosis for those with CEN it was felt by participants that there were systemic and cultural challenges that had a key role to play. These included the inaccessible and limited nature of services, the lack of specialist training to help staff support service users with CEN and the poor resourcing and lack of funding for services.
## The inaccessible and limited availability of services
Participants described the limited availability of services for people with CEN and the exclusionary nature of services for those with a "personality disorder" label. Across a range of areas, a combination of restrictive eligibility criteria and long waiting lists made it hard to access appropriate specialist services. Access to generic mental health services was also reported as challenging for those with CEN, with services reportedly excluding service users because of their "personality disorder" diagnosis or restricting care for this reason.
"I think services still are set up to exclude people with emotional difficulties… I think any type of 'personality disorder' diagnosis or trait is still a diagnosis of exclusion from most places and most services. It firmly does place people in that, kind of, they're choosing to behave this way and it's not. It's not that. " (ID19, female, social worker).
"It seemed to me that you had to be three quarters on your way to recovery before you could even be accepted by a 'personality disorder' service in the NHS. They were quite rigid. ... They are also rather risk averse. " (ID09, male, psychotherapist working in a CEN service).
Exclusions from generic services, e.g., short-term talking therapies such as CBT, was often attributed to service users being "deemed to be too risky". In contrast, in the face of high thresholds and limited services outside generic mental health pathways, participants noted that it often seems as though risk has escalated and the only way into services is via a crisis pathway.
"I suppose the other thing is the difficulty of accessing secondary care now... I had an emergency myself a while back, a crisis, [it was] two weeks before I could see the psychiatrist. That was the urgent care. That was the, "This is a lady who's right on the edge. " That is not good service. That is not a good way to treat somebody. " (ID22, female, peer worker in a CEN service).
Such high thresholds for care were reported to result in an escalation of distress and behavioural disturbance, which services still might not respond to. Seemingly 'extreme' behaviours reinforced stigmatising attitudes and a tendency for staff to label service users with CEN as 'attention-seekers' or 'manipulative' .
"[some staff say] "Oh, well, they are just attentionseeking", or, "They have got to take responsibility", all this kind of stuff… It just sets them up for failure. " (ID08, male, Mental health nurse and psychotherapist in a non-specialist service).
Overall, staff were concerned and frustrated by the lack of available and accessible services, an aspect that fed their feelings of hopelessness. This was intensified by the negative impact related to the difficult pathways to care. Service users not only found themselves excluded from generic mental health services, but found specialist services inaccessible as well.
## The lack of specialist training to help staff support service users with cen
Participants reported a lack of specialist training in supporting people with CEN, which they described as a barrier to the provision of quality care by generic mental health services and the third sector.
"There is a paucity of training but also a paucity of where they are available. There are quite a lot of courses but if you are in one world and not the other you might not get that information. " (ID24, female, third sector support worker).
A lack of information about what courses were available, or a lack of support from managers to attend training, exacerbated this problem.
The importance of supervision and reflective practice has been discussed above. Where this was not embedded staff reported feeling less able to engage in the care of people with CEN and to manage risks. Having specialists within teams was identified as an important way to support staff to provide the best practice care, by providing role modelling opportunities for staff. Without training, supervision and role-modelling, stigmatising attitudes and perceptions were felt to be sustained or born-out of staff burnout and lack of knowledge.
"I've worked in mental health for the past 20 years. I probably spent the first ten years of my career with very negative views about 'personality disorder' . I think, in retrospect, that's because nobody ever taught me anything about it, and the only thing that I ever did learn was from people who had also never had any training. So, you know, very, kind of, jaded, burnt-out staff who were very negative. " (ID04, female, Occupational Therapist working in specialist services).
Without specialist training to help staff understand the nature of CEN as well as the importance of such underpinning concepts as trauma and relational care, staff felt that it would continue to be a challenging environment for staff and service users. These elements were felt as essential to providing good care but were often restricted to specific staff groups or professionals, therefore leaving staff ill-equipped to care for service users.
## Poor resourcing and lack of funding
The under-resourcing of services was reported to directly influence the quality of care for people with CEN, contributing to long waiting lists, delayed discharge, and staff burnout. The importance of sustained investment and funding was repeatedly discussed by participants.
"We need money to fund services; without it we have lots of waitlists so there is no seamless care, and it makes issues then for other services. " (ID12, female, social worker working with CEN women in homelessness services).
At the core of discussions about resourcing were that best practice care was relational and took time to do well.
Some staff suggested that the under-resourcing of services for people with CEN reflected a lack of esteem for this group of service users compared to service users with other mental health problems or with physical health problems.
"I also think that there's a lot of stigma about this client group and that, when you are thinking about what your priorities for spending money on are, it's generally the people who are less popular that get [less] money. " (ID32, male, occupational therapist).
Some participants argued for the commissioning of integrated specialist CEN and generic mental health services.
"I left just disheartened at the fact that actually we're still working tiered services… we need to skill up a whole workforce and really get away from these tiered pyramid models because it's everybody's business. " (ID04, female, Occupational Therapist working in specialist services).
Participants also highlighted a need for early intervention services and pathways that allowed service users to progress smoothly through care and to utilise the services that suited their needs. Important to both joined-up, effective working and to adequate resourcing was having buy-in and commitment from service commissioners and managers. Participants noted that all services are stretched and under-resourced and this has created not only an overworked workforce but also competition across services. Participants reflected that in some cases, services are losing sight of service users' needs because of budget and resource pressures.
Overall, it was reflected by staff members that even with the best training, evidence-based treatments and dedicated and compassionate staff, without appropriate funding and staff resourcing, the care they could deliver would be ultimately compromised and fall short of what service users need.
# Discussion
Analysis of interviews with fifty mental health and social care professionals recruited from across England identified best practice principles for providing community care to people with CEN as well as the barriers and facilitators to their implementation. Staff highlighted the need for care that was person-centred, relational and empathic. Importantly, staff felt that the current provision of care should be informed by an understanding that someone's current presentation might be shaped by previous experiences of neglect and abuse, often from the very individuals who had been entrusted to care for them earlier in life. However, the ongoing stigmatisation of people with CEN and the lack of staff training and support created barriers to best practice care.
Previous research has emphasised the need for services to focus on the long-term needs of people with CEN, the importance of high quality and consistent therapeutic relationships and of a balanced approach to safety issues, and the need for staff support and joint working across services. Findings from this study add to these recommendations, highlighting the need for services for people with CEN to be more flexible, more integrated, and better resourced.
Staff considered a "one-size-fits-all" approach inadequate for people with CEN, given the range and fluctuating nature of these needs. However, a lack of flexibility within services and siloed working across services often meant that many needs remained unmet. Addressing the needs of these individuals is likely to require better integration of services and greater flexibility of pathways across both specialist and generic mental health services. This aligns with The Community Mental Health Framework for Adults and Older Adults, a mental health policy aimed at greater integration of primary and secondary care and of the voluntary sector in England, and at the development of accessible, straightforward and seamless mental health services. The focus in research on CEN has been almost entirely on testing of specialist psychotherapies, so that very little evidence is available on how to design services and systems of care that deliver high quality and holistic care to meet the long-term needs of people with CEN. Similarly, there is a lack of evidence on whether and how trauma-informed approaches -the use of which was recommended by participants in this study -contribute to improved outcomes for service users and how they can be best implemented.
Although some training programmes have been shown to improve attitudes towards people with a "personality disorder", training alone is unlikely to address the systemic stigmatisation of people with CEN. We urgently need evidence about the individual and organisational level factors that contribute to the development and embedding of stigma, including the potential role of fear, risk, and emotional burnout. Also needed are behavioural change perspectives on how to reduce stigma and increase positive attitudes towards "personality disorders" as conditions for which there is ample evidence that people respond to appropriate treatment. While there are barriers and challenges to providing care for those with CEN or a "personality disorder diagnosis", staff need to have understood that this is a malleable condition that with a lot of improvements seen in people who are offered and engage with treatment. It is important for staff to see these outcomes to tackle the hopelessness that participants reported within this study.
There is substantial evidence that in England at least, people with CEN continue to experience considerable inequity within services. This study further supports our understanding of this inequality facing individuals with CEN, with staff noting the impact of under-resourcing of services on accessibility of best practice care, as well as the pressures facing staff where training, support and supervision are not resourced to ensure staff can deliver effective care. This requires service development to not only consider the needs of service users, but also those of staff, in order to tackle the barriers facing effective care and to integrate elements of best practice such as supervision and reflective practice across services and professional approaches. The findings in this paper also add to our knowledge of what best practice should look like and the often-undervalued elements, such as therapeutic relationships, that are not seen as 'evidence-based models of care' but that are imperative to facilitate effective person-centred care.
# Strengths and limitations
Participants were drawn from across England and from a range of settings and occupations. Topic guides were developed in collaboration with experts by profession and by experience, and these experts also contributed to the analysis and interpretation of data. Some limitations should be noted. Participants were self-selected, and the sample was likely to have included those with an interest and investment in this research. While the sample captured a range of experiences from those with negative experiences and who were unhappy with services to those wishing to reflect on good practice, this self-selection method may not capture the views of individuals less motivated and actively engaged in perusing service improvement. As a result, we are less likely to have captured staff voices reflecting stronger negative views regarding service users, so our ability to dig further into the factors impacting on the emergence of stigmatising attitudes may be limited.
# Conclusion
Eighteen years on from the publication of "Personality Disorder: No Longer a Diagnosis of Exclusion" [26], a report which called for the transformation of services for people with a "personality disorder" diagnosis in the NHS, many people with CEN continue to be underserved, with barriers to best practice care at the staff, service, and system level. Staff working with this service user group report that delivering best practice care -described as relational, person-centred, and traumainformed -requires services to be flexible, integrated, and sustainably funded, and for staff across both specialist and generic services to be supported through ongoing training and supervision.
## Lived experience commentary by dawn allen, founder of making mental health positive and carer
Some of the frustrations and barriers to effective professional care with those with CEN may be resolved to some extent with a broader approach to including relatives in care planning and therapeutic approaches. As the paper outlines, core skills and elements of best practice don't need significant financial enhancement but come from elements such as flexibility, being person-centred and seeing the person not the diagnosis. Sadly, these are the 'simple' parts that are often missed or are missing when it comes to service users' experiences of "personality disorder" care.
As the paper highlights, positive relationships are fundamental in terms of not only patient to professional but also how to maintain and build towards comfortable and accepting relationships with close relatives for the best interests of all involved. There are naturally some barriers around patient self-stigma, patient confidentiality and consent to share information to relatives which I feel can be overcome with a positive and trusting relationship with all professionals involved the community care of those with complex emotional needs.
One area overlooked within the study, and in a lot of the "personality disorders" literature is around involving families and relatives in the care and conversations around treatment and therapy for the complex needs of patients. It is often an area that is under-valued or not addressed as part of the way in which people manage complex emotional needs. From a carer's perspective, this is a critical part of beginning an open dialogue between mental health care professionals, social care and patients where there is an agreement that the patient would find it helpful for the relative to be involved. More where the parent is the patient, and the son or daughter is the relative wanting inclusion and validation as part of the process and understanding of the condition.
As an example, the way I was informed of my mothers' diagnosis of EUPD was via a consultation letter that I was shown from my mother.
There was no open dialogue in terms of community care and outpatient psychiatric consultants, leading to wards my perception of a very stigmatised exclusion of relatives within the mental health system and a negative perception of mental health professionals prohibiting relationships with families. This is contradictory in terms of the aims towards best practice and positive relationships.
Investment, be that in training, support or systemic, needs to consider these aspects of care because without good communication and trusting relationships we as service users will still be facing an uphill battle. |
Predicting the Perceived Sound Quality of Frequency-Compressed Speech
The performance of objective speech and audio quality measures for the prediction of the perceived quality of frequencycompressed speech in hearing aids is investigated in this paper. A number of existing quality measures have been applied to speech signals processed by a hearing aid, which compresses speech spectra along frequency in order to make information contained in higher frequencies audible for listeners with severe high-frequency hearing loss. Quality measures were compared with subjective ratings obtained from normal hearing and hearing impaired children and adults in an earlier study. High correlations were achieved with quality measures computed by quality models that are based on the auditory model of Dau et al., namely, the measure PSM, computed by the quality model PEMO-Q; the measure qc, computed by the quality model proposed by Hansen and Kollmeier; and the linear subcomponent of the HASQI. For the prediction of quality ratings by hearing impaired listeners, extensions of some models incorporating hearing loss were implemented and shown to achieve improved prediction accuracy. Results indicate that these objective quality measures can potentially serve as tools for assisting in initial setting of frequency compression parameters.
# Introduction
## Nonlinear frequency compression
Frequency lowering techniques are now common in digital hearing aids as an alternative amplification strategy for hearing impaired listeners with severe to profound high frequency hearing loss. For this group of listeners, conventional amplification strategies may result in less than optimal performance due to a combination of inadequate gain at higher frequencies, limited bandwidth of the hearing instruments, and/or the potential presence of high frequency cochlear dead regions [bib_ref] Dead regions in the cochlea: conceptual foundations, diagnosis, and clinical applications, Moore [/bib_ref] , [bib_ref] Effect of stimulus bandwidth on auditory skills in normal-hearing and hearing-impaired children, Stelmachowicz [/bib_ref] , among other factors. Frequency lowering techniques aim to transfer high frequency information to lower frequency regions. These techniques can improve speech recognition, but may also affect perceived sound quality. In this paper, we investigate a broad range of objective indices of sound quality against a database of sound quality ratings for frequency-lowered speech. This work not only evaluates which models may be effective predictors of the impact of one form of frequency lowering on sound quality, but also develops insights as to the key features of a successful model by varying key modelling parameters and evaluating their impact on successful predictions.
Historically, frequency lowering has been achieved in many ways including slow playback, channel vocoding, frequency transposition, and frequency compression, as well as a combination of these alternative strategies. As described in ([3] -Section 8.3), a frequency transposing hearing aid will shift a portion of high frequency spectrum to lower frequencies by a fixed amount (in Hz). Frequency compression can be linear or nonlinear; in linear frequency compression, all frequencies are proportionally reduced by the same factor (i.e., the output frequency is a fixed fraction of the input frequency), while in nonlinear frequency compression, only frequencies above a certain threshold are compressed.
Currently, frequency lowering is available as an option in commercial hearing aids as frequency transposition (termed ''Audibility Extender'') in Widex hearing aids [bib_ref] Linear frequency transposition: extending the audibility of high-frequency energy, Kuk [/bib_ref] , as nonlinear frequency compression (termed ''SoundRecover'') in Phonak and Unitron hearing aids and as ''frequency compression'' in Siemens hearing aids, as spectral warping (termed ''SpectralIQ'') in Starkey hearing aids, and as ''Frequency Composition'', another type of frequency transposition , in Bernafon hearing aids.
The differences in frequency lowering implementations can be expected to result in considerably different perceptual effects, even when they are fitted for the same audiometric configuration. For example, McDermott [bib_ref] A Technical Comparison of Digital Frequency-Lowering Algorithms Available in Two Current Hearing..., Mcdermott [/bib_ref] conducted an electroacoustic comparison of Widex's Audibility Extender and Phonak's Sound Recover technologies. Spectrographic analyses with different speech and music samples showed that while both schemes were effective in lowering high frequency content, they also introduced distortion that may affect perception. It is therefore imperative to investigate the perceptual effects of frequency lowering technologies, and to develop electroacoustic tools and computational models that can predict these perceptual effects.
Since this paper focuses on a particular frequency lowering strategy, viz. the nonlinear frequency compression (NFC), a brief description of the NFC processing and the evidence surrounding its effectiveness is presented below.
The NFC processing in today's commercial hearing aids uses compression only for frequencies above a cutoff value. Mathematically, the relationship between input and output frequencies is given by
[formula] F out~F in , F in vF c F 1{p c |F p in , F in §F cð1Þ [/formula]
where F in is the input frequency in Hz, F out is the corresponding output frequency, F c is the cutoff frequency, p is the compression exponent, and 1/p is the compression ratio (CR). [fig_ref] Figure 1: Illustration of nonlinear frequency compression [/fig_ref] illustrates the concept showing short time spectra of speech before and after nonlinear frequency compression. The specification of the F c ensures that the lower frequency information is unadulterated while the spectral content beyond the F c is compressed into a narrower bandwidth. The NFC scheme has been evaluated with both adults and children. Simpson et al. [bib_ref] Improvements in speech perception with an experimental nonlinear frequency compression hearing device, Simpson [/bib_ref] investigated the performance of prototype NFC with 17 hearing impaired adults with moderate to profound sensorineural hearing loss and found that the phoneme recognition scores for the group improved by 6%. Glista et al. [bib_ref] Evaluation of nonlinear frequency compression: Clinical outcomes, Glista [/bib_ref] evaluated NFC processing with 13 adults and 11 children with moderately severe to profound sloping high-frequency hearing losses. On average, recognition scores for plurals and consonants improved significantly for the NFC scheme when compared with conventional amplification. Individual variability was present in the results and children derived greater plural recognition benefit from NFC as well as indicated preference for NFC over conventional amplification when compared to the adults [bib_ref] Evaluation of nonlinear frequency compression: Clinical outcomes, Glista [/bib_ref]. Benefit was also related to audiogram: Those with greater high frequency hearing loss were most likely to demonstrate benefits. More recent studies have evaluated outcomes for those with moderate hearing losses [bib_ref] Evaluation of non-linear frequency compression for school-age children with moderate to moderately-severe..., Wolfe [/bib_ref] , the time course of acclimatization [bib_ref] Perceptual Acclimatization Post Nonlinear Frequency Compression Hearing Aid Fitting in Older Children, Glista [/bib_ref] and effects on sound quality [bib_ref] Nonlinear frequency compression: Effects on sound quality ratings of speech and music, Parsa [/bib_ref] , [bib_ref] Exploring the limits of frequency lowering, Souza [/bib_ref]. Previous work indicated that sound quality is correlated with the strength of the frequency compressor, with stronger settings having poorer sound quality. This underscores the need for formal evaluation of sound quality of the NFC processor. In particular, computational metrics which can effectively predict speech quality perception by hearing impaired listeners could be of use in determining an initial set of NFC parameters that have acceptable sound quality, potentially improving the acceptability of initial fittings.
## Objective sound quality evaluation
The perceived quality of frequency-compressed sounds will depend on a number of variables, including the parameter settings, hearing loss, the type of sound, as well as highly individual, subjective factors like personal experiences, expectations and preferences. In order to estimate the expected sound quality for a given algorithm setting, sound, and listener, one could either follow a data-driven, i.e. statistical model approach (e.g, the Emodel [bib_ref] The E-Model, a Computational Model for Use in Transmission Planning, Itu-T Rec [/bib_ref] , or a perception model approach (e.g. PESQ [bib_ref] Perceptual evaluation of speech quality (PESQ), an objective method for end-to-end speech..., Itu-T [/bib_ref]. A data-driven approach would require a large base of empirical data containing subjective quality ratings of different types of sounds, processed with a variety of algorithm settings, obtained from many listeners with different hearing losses. A perception-model-driven approach would possibly require similar amounts of data if it was developed, trained or optimized, and validated particularly for this application.
Alternatively, existing quality models already validated for similar applications could be tested with a smaller amount of data. However, most of the existing quality models were designed, optimized and validated in the context of telecommunication applications and audio coding for normal-hearing listeners (see [bib_ref] Objective Assessment of Speech and Audio Quality -Technology and Applications, Rix [/bib_ref] for an overview). At present, very few sound quality models for hearing impaired listeners with application for hearing aid (algorithm) quality evaluations have been reported [bib_ref] Speech quality measurement for the hearing impaired on the basis of PESQ, Beerends [/bib_ref] [bib_ref] The Hearing-Aid Speech Quality Index (HASQI), Kates [/bib_ref]. Most of the sound quality models, including current ITU standard methods for speech and audio quality, have in common that they follow the concept of comparing ''internal representations'', computed by a psychoacoustic model, of a test and a reference sound signal, [bib_ref] Perceptual evaluation of speech quality (PESQ), an objective method for end-to-end speech..., Itu-T [/bib_ref]. Detected differences between internal representations are interpreted as quality degradations of the test signal with respect to the reference signal. Hence, these comparison-based models depend on the availability of a reference signal that represents the optimum, or desired, sound quality. This requirement is met in the evaluation of lossy signal processing systems, such as low-bitrate speech and audio codecs, where the unprocessed, original signal serves as a reference.
However, there is no known ideal reference for the evaluation of hearing aids (algorithms) in general. In contrast to audio codecs, whose aim is to produce output signals that are perceptually indistinguishable from the original input, hearing aids aim to alter the input sound in a way that it compensates for the listeners' hearing loss. This intentional alternation could result in the processed sound having higher sound quality than the original signal. Therefore, one would need a ''perfect'' hearing aid to produce the reference signal for sound quality evaluations. Unfortunately, the perfect hearing aid does not exist.
However, many hearing aid algorithms can be evaluated with comparison-based quality models. Examples include speech enhancement algorithms that operate on already distorted input signals, like noisy and/or reverberant speech. In these evaluations, the original, clean speech recording can serve as a reference signal, and comparison signals are generated by adding noise or reverberation. Such a comparison-based method may also be applicable in the case of frequency-lowering algorithms. Frequency-lowering will always degrade the naturalness of the input sounds. If subjective sound quality ratings are dominated by the perceived naturalness and not influenced by the possibly improved speech intelligibility or other positive effects of the frequency compression, comparison-based quality models using the unprocessed input signal as a reference appear potentially qualified as predictors of sound quality also for this class of hearing aid algorithms.
## Objective evaluation of frequency-compressed speech
NFC is a viable choice as an amplification solution for hearing impaired listeners with severe to profound high-frequency hearing loss. Since sound quality is one of the most important factors determining the overall satisfaction and acceptance of hearing aid users [bib_ref] Questionnaires on desirable properties of hearing aids, Hagerman [/bib_ref] [bib_ref] Marketrak VI: 10-year customer satisfaction trends in the US hearing instrument market, Kochkin [/bib_ref] [bib_ref] Sound quality judgment during acclimatization of hearing aid, Ovegård [/bib_ref] [bib_ref] Hearing aid satisfaction: what does research from the past 20 years say?, Wong [/bib_ref] , the negative impact on sound quality by NFC must remain small enough to be overshadowed by the positive effects such as improved intelligibility in order to achieve acceptance and overall preference over conventional processing. Objective sound quality models are useful in this context as they can assist in the initial specification of NFC parameters that achieve a balance between intelligibility enhancement and quality degradation. In the present study, a number of established as well as rather new, mostly perceptual speech and audio quality models, including yet unpublished extended versions, have been applied to predict subjective sound quality ratings of nonlinear frequency-compressed speech [bib_ref] Nonlinear frequency compression: Effects on sound quality ratings of speech and music, Parsa [/bib_ref]. It should be noted that none of the applied models was designed for the present application and some of them were particularly optimized for different experimental conditions, like (monaural) headphone presentation instead of loudspeaker presentation as used in the present study. In spite of this caveat, promising predictor candidates will be identified below.
## Subjective speech quality measurement
Speech signals and subjective quality ratings used to test the quality models were obtained from an earlier study [bib_ref] Nonlinear frequency compression: Effects on sound quality ratings of speech and music, Parsa [/bib_ref] and shall be described only briefly here. See [bib_ref] Nonlinear frequency compression: Effects on sound quality ratings of speech and music, Parsa [/bib_ref] for a more detailed description.
Speech quality ratings were obtained from a group of 12 normal hearing adults (ages 21-27 years, mean = 24 years), 12 normal hearing children (ages 8-18 years, mean = 12 years), 12 hearing impaired adults (ages 50-81 years, mean = 69 years), and 9 hearing impaired children (ages 8-17 years, mean = 12 years). The listeners with hearing impairment also participated in a field study with NFC hearing aids [bib_ref] Evaluation of nonlinear frequency compression: Clinical outcomes, Glista [/bib_ref]. All normal hearing listeners had pure tone thresholds of 20 dB HL or better across the audiometric frequency range. [fig_ref] Figure 2: Mean and standard deviations of the pure-tone air conduction thresholds for hearing... [/fig_ref] depicts the mean and standard deviation of the pure tone thresholds for hearing impaired adults and children. While the children had higher absolute thresholds than adults, statistical analysis revealed there was no significant interaction between frequency-specific thresholds and age [bib_ref] Evaluation of nonlinear frequency compression: Clinical outcomes, Glista [/bib_ref] , [bib_ref] Nonlinear frequency compression: Effects on sound quality ratings of speech and music, Parsa [/bib_ref].
A database of frequency-compressed speech samples was created to obtain the subjective quality ratings. The database was constructed by first programming a prototype NFC hearing aid with specific cutoff frequency (F c ) and compression ratio (CR) parameters, placing this hearing aid in a Brüel and Kjaer portable anechoic test box, and recording the output of the hearing aid separately in response to two male speech samples and two female speech samples. A total of five different NFC settings were investigated: (1) F c = 4000 Hz, CR = 2:1; (2) F c = 3000 Hz, CR = 2:1; (3) F c = 3000 Hz, CR = 6:1; (4) F c = 3000 Hz, CR = 10:1; and (5) F c = 2000 Hz, CR = 2:1.
Hearing aid recordings were later played back to the bilaterally aided study participants in a sound booth through a loudspeaker positioned at 0u azimuth and at a distance of 1.5 m. The playback level was initially presented at 65 dB SPL and then individually adjusted, if needed, to the subject's most comfortable level (MCL) [bib_ref] Nonlinear frequency compression: Effects on sound quality ratings of speech and music, Parsa [/bib_ref]. Calibration was performed using the Brüel & Kjaer Sound Level Meter (SLM), Type 2270. Speech quality ratings were obtained using the ITU standard MUltiple Stimulus with Hidden Reference and Anchors (MUSHRA,protocol mediated by custom software. The MUSHRA protocol uses a numerical quality rating scale from 0 to 100 together with five verbal categories from ''bad'' to ''excellent'' as anchors for orientation. In this method, a ''reference stimulus'' was presented, which was the unprocessed male or female speech sample, while eight ''test stimuli'' were randomly associated with the five NFC-processed stimuli, the original signal itself (''hidden reference''), and a lowpass version and a peak-clipped version of the original signal, respectively. The low-pass version was obtained by passing the original signal through a 10 th order Butterworth filter with a cutoff frequency of 2000 Hz, while the peak-clipped version was obtained through hard-clipping the original stimulus at 25% of its peak value. These stimuli served as ''anchor'' stimuli that represent sound with poor quality, as recommended by. In the MUSHRA method, all stimuli are rated simultaneously using graphical slider buttons. Participants were instructed to listen to each of these stimuli in comparison to the reference, by clicking on the corresponding icons of the software's GUI and adjust the rating sliders such that a satisfactory quality rating of all eight stimuli was achieved. Participants were allowed to re-listen to any stimulus and re-adjust their quality ratings until they were satisfied with the final set of ratings.
Parsa et al. [bib_ref] Nonlinear frequency compression: Effects on sound quality ratings of speech and music, Parsa [/bib_ref] have conducted a detailed statistical analysis of the subjective data collected through this procedure. A high degree of intra-and inter-rater reliability was observed with the speech quality ratings across both normal and hearing impaired groups. As such, this database was used to benchmark the performance of different objective speech quality estimators in predicting the quality of frequency-compressed speech.
## Objective speech quality measurement
This paper evaluated twenty-nine objective quality measures including sub-variants in order to benchmark model performance for prediction of sound quality with NFC processing. [fig_ref] Table 1: Quality models and measures used in this study [/fig_ref] provides a summary of the models/measures tested, and technical information on the nature of each model is provided below.
The objective models investigated in this paper are considered to be ''double-ended'' or ''intrusive'' in their operation. Doubleended models compare features extracted from the signal under test to those extracted from an undistorted reference version of the same signal. The comparisons quantify the degree of perceptual overall difference or similarity. In contrast, single-ended models do not require a reference signal. The models in [fig_ref] Table 1: Quality models and measures used in this study [/fig_ref] can be broadly grouped into: (1) metrics based on speech production model parameters, and (2) metrics based on the comparison of ''internal representations'' obtained with computational models of auditory processing. Double-ended models were chosen for this work, because the subjective quality ratings were obtained using reference signals as well, according to the MUSHRA protocol.
Examples of objective models in the speech production model parameters include the Itakura-Saito Distance (ISD) [bib_ref] A Statistical Method for Estimation of Speech Spectral Density and Formant Frequencies, Itakura [/bib_ref] and the Log-Likelihood Ratio (LLR) [bib_ref] Minimum prediction residual principle applied to speech recognition, Itakura [/bib_ref] , which calculate the weighted similarity between the linear prediction coefficients extracted from the distorted and reference speech signals, and the Log-Area Ratio (LAR)measure, which computes the distance between the area ratio coefficients. Software implementations of ISD, LAR, and LLR were taken from a toolbox provided by Hansen and Pellom [bib_ref] An effective quality evaluation protocol for speech enhancement algorithms, Hansen [/bib_ref].
A basic objective measure that uses an auditory model is the Weighted Spectral Slope Distance (WSSD) [bib_ref] Prediction of perceived phonetic distance from critical-band spectra: A first step, Klatt [/bib_ref] measure. It uses a psychoacoustically motivated bank of critical-band filters to decompose the speech signals, and weights the differences between the slopes of the log magnitudes in each band to produce the final quality measure. Again, the software implementation of the WSSD measure was taken from the Hansen and Pellom toolbox.
The method for the Perceptual Evaluation of Speech Quality (PESQ; ITU-T Recommendation P.862 [bib_ref] Perceptual evaluation of speech quality (PESQ), an objective method for end-to-end speech..., Itu-T [/bib_ref] and its wide band extension (PESQ-WB; ITU-T P.862.2are perhaps the most popularly used objective quality assessment methods incorporating an auditory model. Conceptually, the PESQ technique computes the frame-by-frame internal representations of test and reference signals by first computing the power spectra, and then applying frequency and intensity warping functions based on Zwicker's loudness model [bib_ref] Perceptual evaluation of speech quality (PESQ), an objective method for end-to-end speech..., Itu-T [/bib_ref].
Moore & Glasberg have developed a different loudness model [bib_ref] A Model of Loudness Perception Applied to Cochlear Hearing Loss, Moore [/bib_ref] , characterized by the computation of excitation patterns at the output of each auditory filter and the subsequent computation of the specific loudness pattern. Moore et al. [bib_ref] Measuring and predicting the perceived quality of music and speech subjected to..., Moore [/bib_ref] [bib_ref] Predicting the perceived quality of nonlinearly distorted music and speech signals, Tan [/bib_ref] [bib_ref] Development and Validation of a Method for Predicting the Perceived Naturalness of..., Moore [/bib_ref] and nonlinear distortions (D and R nonlin , respectively) are computed separately and combined to an overall subjective score predictor S overall in the end. For the Moore quality model, a custom software implementation was developed. Correlations between model output measures and mean subjective quality ratings will be reported for D, R nonlin and S overall . Chen et al. [bib_ref] An ERB Loudness Pattern Based Objective Speech Quality Measure, Chen [/bib_ref] devised an objective measure based on the Moore & Glasberg loudness pattern differences between the reference and degraded speech samples. These differences were used to estimate sound quality in the degraded sample versus the reference sample. This measure was computed using custom software.
The speech quality model of Hansen & Kollmeier [bib_ref] Objective modelling of speech quality with a psychoacoustically validated auditory model, Hansen [/bib_ref] (referred to as the ''Hansen model'' below) employs the auditory processing model of Dau et al. [bib_ref] A quantitative model of the 'effective' signal processing in the auditory system:..., Dau [/bib_ref] , ''PEMO'' (for ''PErception MOdel''), for the computation of internal representations of test and reference signals. The bandwidth of the PEMO peripheral (gammatone) filterbank was adapted to telephone-bandpass-filtered speech. A ''band importance weighting'' function is applied to the frequency channels of the internal representations to emphasize higher frequency channels. The overall correlation between the weighted internal representations determines the quality measure qc.
The audio quality model PEMO-Q of Huber and Kollmeier [bib_ref] PEMO-Q -A New Method for Objective Audio Quality Assessment Using a Model..., Huber [/bib_ref] extends the Hansen model to include quality assessment of general, wideband audio signals (including speech) with low to very high audio qualities. By default, it uses the more recent version of the Dau model [bib_ref] Modeling auditory processing of amplitude modulation: I. Modulation Detection and masking with..., Dau [/bib_ref] , with the option of selecting the earlier (much faster) version. The two model versions differ with respect to the processing of amplitude modulations: In [bib_ref] A quantitative model of the 'effective' signal processing in the auditory system:..., Dau [/bib_ref] , the final step of auditory preprocessing (i.e. before the decision stage) is an 8 Hz-lowpass filter, whereas in [bib_ref] Modeling auditory processing of amplitude modulation: I. Modulation Detection and masking with..., Dau [/bib_ref] , this lowpass filter is replaced by a filter bank. The band importance weighting function used in the Hansen model was not adopted for PEMO-Q.
PEMO-Q outputs two quality measures: The Perceptual Similarity Measures PSM and PSM t . The second variant evaluates the temporal course of the instantaneous audio quality, derived from a frame-wise correlation of internal representations, nonlinearly mapped to the overall quality estimator PSM t . In contrast, the PSM is the overall correlation of the complete internal representations. The PSM t is more sensitive to small distortions and more independent of the type of input audio signal [bib_ref] PEMO-Q -A New Method for Objective Audio Quality Assessment Using a Model..., Huber [/bib_ref]. On the other hand, the ''simpler'' PSM seems to be more ''robust'' and more generally applicable, especially for low-to intermediate audio quality. For example, speech enhancement algorithms have been evaluated with this measure and showed good correlations with subjective quality ratings [bib_ref] Objective Measures for the Evaluation of Noise Reduction Schemes, Rohdenburg [/bib_ref] [bib_ref] Quality Assessment for Listening-Room Compensation Algorithms, Goetze [/bib_ref] [bib_ref] Objective measures of speech quality in hearing aids: Prediction of listening effort..., Huber [/bib_ref]. Since speech samples with rather strong quality differences are considered in this study, only the PSM was included here.
By default, PESQ and PEMO-Q apply an asymmetric weighting of differences between internal representations, putting more weight on ''added'' or increased elements of the internal representation of the distorted signal than on ''missing'' or attenuated ones. This step is handled as an option for the Hansen model and PEMO-Q and has been switched on and off in this study. The second binary setting option of PEMO-Q that has been varied is the manner of amplitude processing in the auditory processing model (modulation lowpass vs. filterbank). This leads to four PEMO-Q output measure variants that have been tested in total. Since a benefit of the band importance function used in the Hansen model could only be found in the prediction of speech quality in the context of telecommunication, the Hansen model was also tested with the weighting function deactivated in this study. Hence, four versions of the Hansen model have been tested in total: with and without asymmetry weighting of internal representations, and with and without applying the band importance weighting function.
Finally, the Hearing Aid Speech Quality Index (HASQI) by Kates and Arehart [bib_ref] The Hearing-Aid Speech Quality Index (HASQI), Kates [/bib_ref] was employed. In the computation of the HASQI, a relatively simple auditory processing model is applied to provide internal representations in terms of neural firing rates. Effects of possible sensorineural hearing losses are modeled by estimating the contributions of inner hair cell (IHC) and outer hair cell (OHC) losses to a total hearing loss given by the audiogram. Auditory filter bandwidths are increased and compression ratios are reduced depending on the OHC loss estimate. The effect of IHC loss is modeled by linear signal attenuation in each auditory filter before compression. From the comparison of internal representations of test and reference signals, two intermediate quality metrics are computed, quantifying the amount of linear and non-linear distortions, respectively. The final overall quality index is the product of the linear and the nonlinear distortion metrics. Again, correlations with subjective quality ratings will be reported for linear and nonlinear metrics and overall quality index, referred to as HASQI lin , HASQI nonlin and HASQI comb , respectively.
In summary, the models and measures for speech and audio quality listed in [fig_ref] Table 1: Quality models and measures used in this study [/fig_ref] have been applied to the signals of the database.
The versions of PEMO-Q and the Hansen model with extensions for hearing impaired data have not been published before and will therefore be discussed in the next section.
## Extended versions of pemo-q and the hansen model for hearing impaired
Modifications of the underlying auditory model of the audio quality prediction methods PEMO-Q and its predecessor, the speech quality model of Hansen and Kollmeier, for describing sensorineural hearing losses have been suggested by Derleth et al. [bib_ref] Modelling temporal and compressive properties of the normal and impaired auditory system, Derleth [/bib_ref]. One of their suggestions was to insert an instantaneous expansion and attenuation stage before the adaptation and compression stage in the PEMO [fig_ref] Figure 3: Schematic plot of the auditory processing model PEMO, modified to simulate normal... [/fig_ref] , to account for a loss of sensitivity and reduced dynamic compression. This suggestion has been adopted and implemented in the extended version of PEMO-Q and the Hansen model. Given an audiogram as the input, the hearing thresholds in dB HL at audiometric frequencies are interpolated internally to the center frequencies of the model's peripheral filterbank. The inserted processing stage operates on the output of each of these filters, after envelope extraction. The total hearing loss in the respective channel is decomposed into contributions attributed to losses of inner (IHCL) and outer hair cells (OHCL), respectively, as shown in the equation below [bib_ref] A Model of Loudness Perception Applied to Cochlear Hearing Loss, Moore [/bib_ref] :
## Ihcl~thl-ohcl
In the model stage, the signal amplitude is attenuated by the amount of IHCL. Then, it is raised to the power k, which itself is a function of the input signal amplitude I and OHCL. The attenuation and expansion operations are performed instantaneously (i.e. sample-by-sample). For simulating normal hearing, k equals 1 for any I. For simulating impaired hearing, k(I) becomes larger than 1 and increases with OHCL. The input/output function of this stage is shown in [fig_ref] Figure 4: Input/output functions of the expansion and attenuation model stage for normal hearing [/fig_ref] for normal hearing and two hearing losses. The parameters of the input/function were chosen such that its shape matches the inverse function of the assumed input/output function of the unimpaired cochlea proposed by Moore [bib_ref] Psychoacoustic consequences of compression in the peripheral auditory system, Moore [/bib_ref]. Again, model parameters were not adjusted in the present study.
# Results
The performance of the objective measures was characterized by their correlations with averaged subjective quality ratings of all processing conditions (including hidden reference and anchors), quantified by Pearson's linear correlation coefficient r and Spearman's rank correlation coefficient rs. Quality measure values were taken directly as output by the models and not transformed by, e.g., nonlinear regression functions.
As there were no systematic differences between the quality ratings of normal hearing adults and children, quality ratings were averaged across the complete group of normal hearing subjects. Consequently, results are stated for three groups of subjects separately: (1) normal hearing (adults and children), (2) hearing impaired children, and (3) hearing impaired adults. In addition to averaging across subjects, quality ratings were also averaged across different talkers per processing condition. Results will be stated for both levels of averaging. For clarity, results per subject group will be presented separately in succession and only for a selection of measures. The complete results are reported in the Appendix S1. The relation between subjective ratings and corresponding model predictions will be illustrated by scatter plots for the best performing quality measures. [fig_ref] Table 2: Results of the quality predictions [/fig_ref] shows linear and rank correlations between selected model predictions and mean subjective ratings obtained from normal hearing subjects. (The complete correlation table is given in the Appendix S1) In this table, the suffixes of the PSM and qc measures have the following meaning: fb/lp indicate the modulation filterbank/lowpass version of PEMO-Q respectively, +/2W indicate the activation/deactivation of the frequency-band weighting, and +/2B denote activation/deactivation of the asymmetric weighting of internal representation differences (''Beerends weighting''). When frequency weighting is deactivated in the Hansen model, and the modulation lowpass version is used in PEMO-Q, these models become very similar. They mainly differ in the bandwidth of the peripheral filterbank. It must be noted that because the results obtained with the -B option tended to be better than with option +B, the latter are only reported in the complete set of results given in the Appendix S1.
## Results for normal hearing
It can be seen from [fig_ref] Table 2: Results of the quality predictions [/fig_ref] that the correlation coefficients are reasonable for several of the objective quality metrics, especially when averaged across different talkers per processing condition (right column ''av.'') Four of the five measures with highest correlations are PEMO-Q and related measures (PSM…, qc…), complemented by the LAR (M27 of [fig_ref] Table 1: Quality models and measures used in this study [/fig_ref] measure. The latter, however, shows only a rather poor rank correlation with the subjective ratings. Good correlations, especially rank correlations, are achieved when ratings are averaged across different talkers per processing condition because averaging reduces the statistical noise of the ratings and thus increases the reliability of the data. On the other hand, the significance of the correlation coefficients gets lower as the number of data points is reduced to only eight. shows a scatter plot of the results obtained with the measure PSM_fb-B (M10). It must be noted that the results for two of the four talkers processed by the NFC condition cr2_fc3k were not available due to a labelling error and a processing error of the concerned test files. It can be observed from this figure that the quality ratings for the NFC test conditions are bordered by the ratings of the high quality (hidden) reference and the low quality anchor stimuli (lowpass-filtered and clipped speech, respectively). The scatter plot reveals that the model underestimates the perceived quality of the clipped speech samples whose predicted quality was lower than that of the lowpass-filtered samples, whereas the subjects ranked the qualities of these conditions in an opposite manner. Another noteworthy difference between model predictions and the subjective ratings is the larger predicted quality difference between the reference condition and the processed conditions compared to the differences among processed conditions. Both subjects and model rate the frequency-compressed signals with the lowest cutoff frequency 2 kHz (condition cr2-fc2k) very poorly compared to the other frequency compression conditions. A possible cause for this non-linear relation between compression cutoff frequency and perceptual effect could be the alteration of speech formant frequencies and/or the interference of downshifted higher frequency contents with these formants. Vowel formants have significant energy in the second and third formants in approximately the 1000-4000 Hz region for female speech and approximately 750-3000 Hz for male speech, depending on the specific vowel and talker. Recalling that both male and female speech samples were used, it is reasonable to conclude that the cr2-fc2k setting was strong enough to lower significantly the upper formants of many vowels for both male and female speech and was therefore much more noticeable than settings that likely did not disrupt the harmonic relationships within vowels, such as those that used a 4000 Hz cutoff frequency. shows that in general, the influence of the talker on the measured and predicted quality ratings is very small, except for the
## Results for hearing impaired children
The results obtained for the hearing impaired children are shown in [fig_ref] Table 3: Quality prediction results [/fig_ref] and [fig_ref] Figure 6: Same as Figure 5, but for hearing impaired children [/fig_ref]. According to the correlations with subjective ratings [fig_ref] Table 3: Quality prediction results [/fig_ref] , the prediction accuracy of the models for hearing impaired listeners for separate talkers was very good (r up to 0.94), and even higher when averaged across talkers (r up to 0.99). The slightly better performance of the modulation filterbank version of PEMO-Q in the prediction of normal hearing subjects' ratings [fig_ref] Table 2: Results of the quality predictions [/fig_ref] was not confirmed by the results obtained with hearing impaired children [fig_ref] Table 3: Quality prediction results [/fig_ref]. [fig_ref] Figure 6: Same as Figure 5, but for hearing impaired children [/fig_ref] depicts a scatter plot for PSM_lp-B, hearing impaired (HI) version (M16) and subjective ratings from hearing impaired children. In contrast to , a greater variation of quality ratings and predictions for different talkers can be seen in this figure. Note that the predicted quality of the original, unprocessed condition is always maximal, since this condition is used as the reference by the comparison-based quality models. Hence, possible influences of different talkers cannot be modeled for this condition. Another difference from the normal hearing data concerns the rank order of the three processing conditions with the lowest ratings, i.e. lowpass filtering (LP), clipping (CL) and frequency compression with the lowest cutoff frequency, 2 kHz (cr2-fc2k). Normal hearing subjects ranked LP, CL < cr2-fc2k. The model, in contrast, ranked CL,LP, cr2-fc2k. Hearing impaired children, however, ranked these conditions clearly differently: cr2-fc2k,CL,LP. This rank order was correctly predicted by the quality measure PSM_lp-B (HI) (M16). As a consequence, a very high linear correlation (r = 0.99) and a perfect rank correlation (rs = 1) between PSM_lp-B and subjective ratings were achieved when averaged across talkers. Finally, the range of PSM_lp-B (HI) (M16) values is smaller compared to PSM_fb-B (M10) values used for normal hearing subjects, while the range of subjective ratings is about the same in both cases. This difference is due to the greater sensitivity of the modulation filterbank version of the auditory model PEMO used for the computation of
## Results for hearing impaired adults
Hearing impaired adults exhibited a different rating behavior than hearing impaired children on average, although there were no significant differences between the distributions of hearing losses in these two groups (children's hearing losses were somewhat higher than adults' hearing losses on average). Adults rated NFC conditions generally higher, except for condition cr2-fc2k, whereas their ratings of the anchor condition (LP, CL) were similar to children's ratings (compare [fig_ref] Figure 7: Same as Figure 5, but for hearing impaired adults [/fig_ref] with [fig_ref] Figure 6: Same as Figure 5, but for hearing impaired children [/fig_ref]. As a result, adults' average ratings of cr2-fc2k, LP and CL conditions were very similar, whereas the quality measure PSM_lp-B (HI) (M16) clearly rated cr2-fc2k,LP, CL, which led to a lower correlation between measured and predicted mean quality ratings (r = 0.88) compared to the results obtained for the children (r = 0.99). In contrast, a very high correlation was achieved by the related quality measure qc-W-B (HI) (M8) (r = 0.96). As described earlier, qc-W-B (HI) (M4, M8) mainly differs from PSM_lp-B (HI) (M12, M16) in the lower cutoff frequency of the peripheral filterbank. Looking at [fig_ref] OHCL~m in 65 dB, 0 8: THL f g , f w2000 Hz min 55 dB, 0 [/fig_ref] , the higher correlation obtained with qc-W-B (HI) (M8) is mainly due to a shift of quality ratings of NFC conditions towards higher values compared to the PSM_lp-B (HI) (M16) values, whereas ratings of the anchor conditions LP and CL basically correspond to the PSM_lp-B (HI) (M16) values. This agrees well with the observed difference in subjective ratings between hearing impaired children and adults. summarizes the prediction performances of the different quality measures. Amongst others, it clearly shows the advantage of using models that account for hearing impairment, as evidenced by the higher correlations with subjective quality ratings for most of the evaluations.
# Discussion
The obtained results show that predicting the perceived sound quality of nonlinearly frequency-compressed speech by means of objective, perceptual speech/audio quality models is possible. From the models tested, the PEMO-Q (related) measures, in . Quality prediction results (linear and rank correlation (italic) coefficients) for the hearing impaired adults group. particular, the Perceptual Similarity Measure (PSM) and the speech quality measure qc of the Hansen model (with deactivated frequency weighting function), showed the highest correlations with mean subjective ratings. These measures represent the most ''basic'' PEMO-based quality measures with the lowest degree of optimization for a specific task and hence the highest generality and robustness. This is in line with results of other studies where different quality measures have been applied for tasks other than those for which they were originally developed (e.g. [bib_ref] Objective Measures for the Evaluation of Noise Reduction Schemes, Rohdenburg [/bib_ref] [bib_ref] Quality Assessment for Listening-Room Compensation Algorithms, Goetze [/bib_ref] [bib_ref] Objective measures of speech quality in hearing aids: Prediction of listening effort..., Huber [/bib_ref]. Through the results of the present study, the generality of abovementioned measures is further proven given that the predicted quality degradations caused by very different kinds of distortions (i.e. frequency compression, amplitude clipping and lowpass filtering) correlated highly with subjective ratings. A number of objective measures that were previously either applied to speech or audio quality estimation were evaluated in this paper. Of the objective measures based on speech production model parameters (viz. ISD (M26), LAR (M27), LLR (M28)), LAR (M27) performed the best. In the past, these metrics were primarily used for estimating narrowband (0-4 kHz bandwidth) speech quality (e.g., [bib_ref] An effective quality evaluation protocol for speech enhancement algorithms, Hansen [/bib_ref]. Since some of the PEMO-based quality measures employed a wider bandwidth, experiments were conducted with implementations of ISD (M26), LAR (M27) and LLR (M28) metrics using an extended 8 kHz bandwidth. Results, however, showed a degradation in the correlation coefficients with subjective ratings for LAR (M27) and LLR (M28) metrics compared to the original versions with 4 kHz bandwidth. Similarly, the LPD (M22) metric was developed and validated for narrowband speech applications [bib_ref] An ERB Loudness Pattern Based Objective Speech Quality Measure, Chen [/bib_ref] , and its performance with an extended bandwidth is not investigated in this work and is a worthwhile topic for future study.
Speech quality measures (M22, M25) based on the Moore & Glasberg model [bib_ref] Measuring and predicting the perceived quality of music and speech subjected to..., Moore [/bib_ref] [bib_ref] Predicting the perceived quality of nonlinearly distorted music and speech signals, Tan [/bib_ref] [bib_ref] Development and Validation of a Method for Predicting the Perceived Naturalness of..., Moore [/bib_ref] [bib_ref] An ERB Loudness Pattern Based Objective Speech Quality Measure, Chen [/bib_ref] performed decently with normal hearing data, but were inferior with the hearing impaired data. A revised model of loudness perception incorporating cochlear hearing loss was proposed by Moore and Glasberg [bib_ref] A revised model of loudness perception applied to cochlear hearing loss, Moore [/bib_ref] , and further investigation would be interesting to determine whether the incorporation of this revised model would lead to an improvement in correlations with hearing impaired subjective data.
The Hearing Aid Speech Quality Index (HASQI, M17-M19) showed good rank correlations and moderate linear correlations with hearing impaired data, but only low correlations with normal hearing data. This is in clear contrast to the results reported in a recent publication on an updated version of the HASQI [bib_ref] The hearing-aid speech quality index (HASQI) version 2, Kates [/bib_ref] , which was also tested successfully on frequency-compressed speech. The main reason for the lower correlations observed in the present study was found to be a significant quality overestimation of the lowpass condition. The clipping condition was overestimated as well, although to a lesser extent. Without these anchor conditions, linear correlations of the HASQI comb (M19) with subjective ratings increase to 0.83 for normal hearing data (0.85 when ratings are averaged across different talkers per condition), 0.75 (0.78) for hearing impaired adults' data, and 0.78 (0.82) for hearing impaired children's data. The HASQI submeasure for linear distortions (M17) achieves very high correlations in this case: 0.93 (0.95), 0.89 (0.95) and 0.95 (0.996) for normal hearing subjects', hearing impaired adults' and hearing impaired children's data, respectively.
In the present application, PESQ (M20, M21) surprisingly achieved poor correlations with subjective ratings from normal hearing listeners, but moderate and even high correlations with ratings from hearing impaired adults and children, respectively. The main reason for the poor results regarding normal hearing listeners was found to be a quality overestimation of the lowpassfiltered items by PESQ. Without this processing condition, the linear correlation obtained with the narrow-band version of PESQ increases to 0.89 (0.92) for normal hearing subjects. Correlations obtained with the wideband version of PESQ are lower in this case.
The successful ''basic'' PEMO-Q measures required minimal optimization during development, and use a very simple back-end of the underlying quality model. This result is interpreted as a reflection of the validity of the auditory model originally derived by Dau et al. [bib_ref] A quantitative model of the 'effective' signal processing in the auditory system:..., Dau [/bib_ref] , [bib_ref] Modeling auditory processing of amplitude modulation: I. Modulation Detection and masking with..., Dau [/bib_ref]. It represents a well-founded model of the ''effective'' signal processing in the auditory system and has been validated in a wide variety of psychoacoustical experiments [bib_ref] Modeling auditory processing of amplitude modulation: I. Modulation Detection and masking with..., Dau [/bib_ref] , [bib_ref] A psychoacoustical model for the perception of nonstationary sounds, Münkner [/bib_ref] [bib_ref] A quantitative model of the 'effective' signal processing in the auditory system:..., Dau [/bib_ref] [bib_ref] Modeling auditory processing of amplitude modulation. II. Spectral and temporal integration in..., Dau [/bib_ref] [bib_ref] Psychoacoustics of spectro-temporal effects in masking and loudness patterns, Verhey [/bib_ref] [bib_ref] On the role of envelope fluctuation processing in spectral masking, Derleth [/bib_ref]. Its free parameters were adopted from psychoacoustical modeling and kept fixed in the development of the Hansen model and PEMO-Q as well as in the present study. When applied as a front-end for speech and audio quality prediction, only the bandwidth of the peripheral filterbank (i.e. the range of center frequencies) and the modulation filterbank were set appropriately. The bandwidth used in PEMO-Q appears to have been appropriate for use with the NFC data modeled in this study, too, although we did not evaluate impacts of any adjustments to these original settings. In the hearing-impaired adults group, however, better results were obtained with the quality measure qc with its smaller filterbank bandwidth compared to PSM. This suggests that the bandwidth could still be optimized for the present task and possibly lead to a further improvement of the quality prediction accuracy.
As already pointed out before, the higher correlation with the adults' quality ratings obtained with the measure qc-W-B (HI) (M8) compared to PSM_lp-B (HI) (M16) is mainly due to an upward shift of quality predictions for the 2 kHz-lowpass condition and the frequency-compressed condition with the lowest cutoff frequency of 2 kHz, which is in line with the adults' rating behavior. The upper cutoff frequency of the peripheral filterbank used in the computation of qc-W-B (HI) (M8) is 4 kHz, whereas it is 15.3 kHz for PSM. The two processing conditions of concern cancel (most of) the energy at frequencies higher than 4 kHz. The Quality Prediction of Frequency-Compressed Speech measure qc-W-B (HI) (M8) does not ''notice'' the missing energy above 4 kHz, in contrast to PSM_lp-B (HI) (M16). As a consequence, PSM_lp-B (HI) (M16) detects a larger overall difference between original and processed signals which leads to a lower quality estimate. A possible reason for the better correspondence between hearing impaired adults' ratings and the qc-W-B (HI) (M8) predictions could be a reduced sensitivity of these subjects for missing energy at frequencies above 4 kHz, or/ and a lower perceptual weighting of this degradation component when rating the overall quality. Apparently, the hearing impaired children of this study were more sensitive or less tolerant towards the quality degrading effects of frequency compression than the hearing impaired adults. They rated the quality of frequencycompressed speech samples clearly lower than the adults did, although the distributions of hearing losses as described by audiograms are similar in these two groups; the children's hearing losses are in fact somewhat higher on average than the adults' hearing losses. There are two factors that may be related to this difference between age groups. First, age-related auditory and/or cognitive declines not revealed by the audiogram, but influencing quality perception might be factors related to this result. Moreover, age-dependent different experiences with and expectations on the quality of sound reproducing systems might play a role and have been speculated before. However, to the knowledge of the authors, there is no data yet that would support such hypotheses, so these considerations remain speculative. Second, as is clinically typical, we provided the children with a higher level of gain and output in their hearing aids than was provided to adults [bib_ref] The desired sensation level multistage input/output algorithm, Scollie [/bib_ref]. For this reason, the children in this study may have had enhanced access to low-level effects in the processed signals and may have responded accordingly in their perceptual ratings. Again, this is a speculation and further study would be required to clarify whether this is a factor.
The fact that similar quality rating scores were obtained from normal hearing adults and children does not contradict the hypothesis of an age-related effect, because normal hearing adults in this study were much younger (mean = 24 years) than hearing impaired adults (mean = 69 years). However, in the present data, not all kinds of distortions were rated differently by older listeners, but mainly the frequency-compressed items. Lowpass-filtered and clipped samples received similar quality ratings from children and adults. The same is observed with the quality measure qc-W-B (HI) (M8), in contrast to PSM_lp-B (HI) (M16) for reasons explained above. As a result, qc-W-B (HI) (M8) correlates better with hearing impaired adults' quality rating scores, whereas PSM_lp-B (HI) (M16) is better suited to predict the hearing impaired children's ratings. However, the results obtained from hearing impaired adults and corresponding quality predictions obtained with the measure qc-W-B (HI) (M8) might represent a rather special data subset. In contrast, the measure PSM_lp-B (HI) (M16) appears to have a higher degree of generality and is thus recommended for general application.
## Effect of pemo-q options
Modulation processing. A clear, consistent effect of the kind of modulation processing stage in PEMO-Q on the prediction accuracy could not be found in the results of the present study. Correlations achieved with the modulation lowpass and modulation filterbank versions of the PEMO-Q quality measure PSM, i.e. PSM_lp and PSM_fb, respectively, were mostly comparable and not consistently ranked with regard to subject group, individual vs. averaged talkers, and linear vs. rank correlation coefficients. Since the computational effort required for PSM_fb is about ten times higher than for PSM_lp, the latter might be preferred for practical applications.
Asymmetric weighting of internal representation differences. The first application for which PEMO-Q was originally developed and optimized was the evaluation of audio codecs. For that purpose, a partial, asymmetric assimilation of internal representations of test and reference signals, equivalent to Beerends' asymmetry weighting of internal representation differences, was introduced into the quality model (option +B). In this procedure, negative deviations of the internal representation of the test signal from the reference are partially compensated (differences are halved), whereas positive deviations remain unchanged. This approach follows the hypothesis that ''missing'' components in a distorted signal are perceptually less disturbing than ''additional'' ones. This processing step was found to improve the quality prediction accuracy of PEMO-Q for audio codecs. In the present application, however, better results were obtained without this processing step. The reason for this result is that the negative effect of strong frequency compression and lowpass filtering on sound quality by cancelling energy at medium to high frequencies (.3 kHz) is over-compensated by the assimilation procedure. Consequently, quality estimates are shifted towards higher values disproportionately with increasing frequency compression (i.e., increasing compression ratio and decreasing cutoff frequency) and for the 2 kHz-lowpass filter condition. illustrates this effect with the example of quality prediction results obtained with the quality measure PSM_fb (M9, M10) for the normal hearing subject group with and without assimilation of internal representations (+/2B).
# Summary and conclusions
In this paper, we presented results from a study on quality prediction of frequency-compressed speech in hearing aids by different objective speech and audio quality measures. Objective measures based on the audio quality model PEMO-Q and the related speech quality measure qc of the Hansen model achieved . Effect of PEMO-Q model option +B (partial compensation of differences between test and reference internal representations) on quality estimates for the normal hearing subject group. Gray, small symbols: Results obtained with option -B (i.e. without partial compensation); black, large symbols: results obtained with option +B. doi:10.1371/journal.pone.0110260.g009
Quality Prediction of Frequency-Compressed Speech good correlations with averaged subjective ratings. The extension of these models incorporated hearing impairment, and improved prediction accuracy for quality ratings from hearing impaired subjects. Hearing impaired adults and children rated the quality of frequency-compressed speech differently. Best prediction results were achieved with different quality measures per subject group; adults' ratings were predicted better with a model that analyzes a smaller frequency bandwidth. Optimizations of bandwidth and possibly further model parameters were not carried out in the present study, but are considered to bear significant potential for further improvements. The HASQI model was also evaluated, and revealed consistently high results across subject groups for the linear submeasure of this index. Again, further evaluation could reveal optimization to the main HASQI model to further improve its performance.
In summary, the present study reports first-of-a-kind results on a broad range of objective sound quality evaluation models of frequency-compressed speech. Further studies are needed to validate the performance of the PEMO-Q quality measures for this application with a larger group of hearing impaired listeners and across more NFC parameter settings. Moreover, the predictability of perceived quality of other sound stimuli than speech, such as music, should be investigated. In addition, the objective modeling of sound quality for frequency lowering schemes other than the NFC scheme evaluated here would require further investigation. In addition, possible influences of age and acclimatization to frequency-compressed sound will have to be investigated and modeled. Research on this topic is driven by the fact that a successful objective speech quality model will have the potential to serve as a valuable supplemental tool for the fitting and evaluation of frequency lowering algorithms in hearing instruments.
## Supporting information
Appendix S1.
(DOCX)
[fig] Figure 1: Illustration of nonlinear frequency compression. Dotted, grey line: Power spectrum of unprocessed speech. Solid, black line: spectrum after nonlinear frequency compression. Above a cutoff frequency F c , the spectrum is compressed by a ratio CR = Df orig /Df comp . doi:10.1371/journal.pone.0110260.g001 Quality Prediction of Frequency-Compressed Speech PLOS ONE | www.plosone.org [/fig]
[fig] Figure 2: Mean and standard deviations of the pure-tone air conduction thresholds for hearing impaired adults and children. doi:10.1371/journal.pone.0110260.g002 Quality Prediction of Frequency-Compressed Speech PLOS ONE | www.plosone.org [/fig]
[fig] OHCL~m in 65 dB, 0 8: THL f g , f w2000 Hz min 55 dB, 0:8 : THL f g , f ƒ2000 Hz [/fig]
[fig] Figure 3: Schematic plot of the auditory processing model PEMO, modified to simulate normal and impaired hearing (after[40] and[45]). doi:10.1371/journal.pone.0110260.g003 [/fig]
[fig] Figure 4: Input/output functions of the expansion and attenuation model stage for normal hearing (solid line) and impaired hearing with an assumed total hearing loss of 50 dB (dotted line) and 80 dB (dashed line). doi:10.1371/journal.pone. [/fig]
[fig] Figure 6: Same as Figure 5, but for hearing impaired children. doi:10.1371/journal.pone. [/fig]
[fig] Figure 7: Same as Figure 5, but for hearing impaired adults. doi:10.1371/journal.pone.0110260.g007 [/fig]
[fig] Figure 8: Quality prediction results for the hearing impaired adults group, obtained with quality measure qc-W-B (version for hearing impaired). r: linear correlation, rs: rank correlation. doi:10.1371/journal.pone.0110260.g008 [/fig]
[table] Table 1: Quality models and measures used in this study.(''ref''.: reference; ''HI?'': Does model account for hearing impairment? Yes: 3; no: 2) Quality measure suffixes: fb/lp: modulation filterbank/lowpass model version; +/2 W: with/without frequency band weighting; +/2B: with/without asymmetric weighting of differences (''Beerends weighting''). doi:10.1371/journal.pone.0110260.t001 [/table]
[table] Table 2: Results of the quality predictions (including anchor conditions) for the normal hearing subjects, expressed by linear correlation coefficients and rank correlation coefficients (italic).Subjective ratings were averaged across subjects. Middle column: correlations based on ratings for individual talkers (ind.); right column: ratings averaged across talkers (av.). Quality measure suffixes: fb/lp: modulation filterbank/lowpass model version; +/2W: with/without frequency band weighting; +/2B: with/without asymmetric weighting of differences (''Beerends weighting''). doi:10.1371/journal.pone.0110260.t002 [/table]
[table] Table 3: Quality prediction results (linear and rank correlation (italic) coefficients) for the hearing impaired children group. [/table]
|
Quantifying the role of steric constraints in nucleosome positioning
Statistical positioning describes how the combination of (1) barriers confining nucleosome movement and (2) inter-nucleosomal steric hindrance affects the distribution of densely packed nucleosomes. In this framework, the nucleosomes are modeled as one-dimensional gas, with N non-overlapping finite-sized particles distributed at random in a finite region of length w, where all valid configurations are given equal probability. The effective size d of the particles, determining how tightly they can be packed, parametrizes the effect of steric constraints. Because of the discrete nature of the genomic sequence, all quantities take integer values. Let X k denote the center location of the k-th nucleosome (numbered sequentially from left to right; see figure below). The primary goal of this section is to calculate E[X k ] and Var[X k ].
[formula] 1 2 · · · k · · · L 1 L 2 L 3 L k L N L N +1 X k &% '$ N [/formula]
An equivalent problem that can be easily solved is to examine the distribution of the gaps between nucleosomes. Let L k denote the length of the gap to the left of the k-th nucleosome, and let L N +1 be the length of the gap between the N -th nucleosome and the right barrier. Any configuration can be uniquely specified in terms of these gaps, and the position of the k-th nucleosome is
[formula] X k = (k − 1/2)d + Y k ,(S1) [/formula]
where Y k = k i=1 L i is the partial sum of gap lengths to the left of the k-th nucleosome. Because all configurations have the same total gap length, i.e.
[formula] L 1 + L 2 + ... + L N +1 = w − N d ≡ L, [/formula]
there is a one-to-one map between the set of valid configurations and the set of weak compositions of L into N + 1 ordered non-negative integers. Thus, there are L+N N distinct configurations in total. Noting that the number of configuration with Y k = y is equal to the number weak compositions of y into k parts times the number of weak compositions of (L − y) into N − k + 1 parts, the probability of Y k is
[formula] P (Y k = y) = y+k−1 k−1 L−y+N −k N −k L+N N . [/formula]
(S2)
## S.1.1 statistical positioning with a single barrier
Before calculating Var[X k ] for double barriers, we give a simplified derivation of the most important result: Var[X k ] increases linearly in k at the rate ( + 1) away from a single barrier, where is the mean gap length between nucleosomes. This relation is true even if the barrier is not firm, but merely imposes a partial restriction on a nucleosome. The linear scaling further generalizes to a broader class of models where nucleosomes interact through some interaction potential V , although how fast variance changes between nucleosomes may be different. (In the case of statistical positioning, this potential vanishes unless the nucleosomes overlap, in which case it is infinite.) To prove the linear scaling, note that
[formula] X k = X k−1 + L k + d, [/formula]
where L k is the length of the gap between the two nucleosomes. We now assume that a restriction is imposed on the 0'th nucleosome, causing Var[X 0 ] to be finite. We further assume that all subsequent nucleosomes are allowed to move freely, except for steric interactions with immediate neighbors. If the array of nucleosomes continues indefinitely, the gap length L k becomes independent of the position X k−1 of the nucleosome to the left; note that this independence does not hold for a finite nucleosome array confined to a finite region delimited by two barriers. Using that X k is the sum of two independent random variables then gives
[formula] Var[X k ] = Var[X k−1 ] + Var[L k ]. (S3) Furthermore, for k ≥ 1, E[L k ] is independent of k for a semi-infinite array, giving E[X k ] = E[X 0 ] + kE[L k ] + kd, Var[X k ] = Var[X 0 ] + kVar[L k ]. [/formula]
Thus, the variance increases linearly in this class of models, and the variance gradient, defined as Var[X k+1 ]− Var[X k ], corresponds to the "wiggle room" that the interaction allows.
In the case of statistical positioning, Var[L k ] can be evaluated by noting that P (L k = y) = P (Y 1 = y). In the limit L → ∞ with the ratio L/N = kept fixed, P (Y 1 = y) in Eq. S2 becomes
[formula] P (L k = y) = P (Y 1 = y) → y ( + 1) y+1 .(S4) [/formula]
It is then straightforward to show that E[L k ] = and Var[L k ] = ( + 1), yielding
[formula] Var[X k ] = Var[X 0 ] + k ( + 1) E[X k ] = E[X 0 ] + k( + d) .(S5S a = L y=0 y a y + k − 1 k − 1 L − y + N − k N − k . [/formula]
Using Eq. S2, we can express E[Y k ] and Var[Y k ] as
[formula] E[Y k ] = S 1 S 0 , and Var[Y k ] = E[Y 2 k ] − E[Y k ] 2 = S 2 S 0 − S 2 1 S 2 0 . [/formula]
One practical method for evaluating these sums is to use the generalized binomial expansion
[formula] 1 (1 − z) k = ∞ y=0 y + k − 1 k − 1 z y .(S6) [/formula]
By acting on this equation with (z d dz ) a and multiplying by
[formula] 1 (1−z) N −k+1 , it is straightforward to show that ∞ y=0 y a y + k − 1 k − 1 z y × ∞ j=0 j + N − k N − k z j = z d dz a 1 (1 − z) k × 1 (1 − z) N −k+1 . [/formula]
Because the moment S a is the L'th order term in the z-expansion of left-hand side, it can be evaluated by algebraically simplifying the right-hand side and extracting the L'th order term using Eq. S6. For example, to evaluate S 0 we set a = 0 and expand the right hand side using Eq. S6:
[formula] 1 (1 − z) k 1 (1 − z) N −k+1 = ∞ j=0 j + N N z j . [/formula]
Extracting the L'th order term gives S 0 = L+N N , proving that the probability in Eq. S2 is correctly normalized.
Next, to calculate S 1 , we extract the L'th order term from
[formula] z d dz 1 (1 − z) k × 1 (1 − z) N −k+1 = ∞ j=0 k j + N N + 1 z j , yielding S 1 = k N +L N +1 . We thus get E[Y k ] = S 1 S 0 = k N +L N +1 N +L N = kL N + 1 = k ,(S7) [/formula]
where = L N +1 is the average gap length. Using Eq. S1 and Eq. S7, the expectation value of X k is
[formula] E[X k ] = k( + d) − d/2 . [/formula]
Similarly, S 2 is evaluated by extracting the L'th order term from
[formula] z d dz z d dz 1 (1 − z) k × 1 (1 − z) N −k+1 = kz(1 + kz) (1 − z) N +3 = ∞ j=0 k j + N + 1 N + 2 + k 2 j + N N + 2 z j , giving E[Y 2 k ] = S 2 S 0 = k L+N +1 N +2 + k 2 L+N N +2 N +L N = kL(1 + k(L − 1) + L + N ) (N + 1)(N + 2) . [/formula]
The variance of Y k is then
[formula] Var[Y k ] = E[Y 2 k ] − E[Y k ] 2 = kL(N + 1 − k)(L + N + 1) (N + 1) 2 (N + 2) . [/formula]
Since X k and Y k differ only by a constant, we have
[formula] Var[X k ] = Var[Y k ] = ( + 1) k(N + 1 − k) N + 2 . (S8) [/formula]
In the limit of large N with kept fixed, this expression reduces to
[formula] Var[X k ] → k ( + 1) ,(S9) [/formula]
which is Eq. S5 with Var[X 0 ] = 0. Note that this linear scaling also holds near the boundaries of a long confining interval.
## S.1.3 single barrier limit of the grand canonical ensemble
It is well known in statistical physics that the Grand Canonical Ensemble (GCE) and the Canonical Ensemble (CE) are equivalent in the thermodynamic limit of large N with fixed density w/N . The Nnucleosome term in the GCE partition function is
[formula] Z N = w−N d+N N e N µ , [/formula]
where w is the length of the interval bounding the nucleosomes and µ the chemical potential. To see the equivalence of the two distributions in the thermodynamic limit, initially assume that w is fixed and find N that maximizes Z N , or equivalently,
[formula] log Z N . For large N , log Z N ≈ N µ − N log N w − N d + N − (w − N d) log w − N d w − N d + N . Thus, d log Z N dN ≈ µ − log N w − N d + N + d log w − N d w − N d + N , [/formula]
and N * that maximizes Z N thus satisfies
[formula] µ ≈ log N * w − N * d + N * − d log w − N * d w − N * d + N * . [/formula]
Then, in the thermodynamic limit of N * 1 and w = (N * + 1) + N * d, we get
[formula] µ → log ( + 1) d−1 d .(S10) [/formula]
The size of the fluctuation δN = N −N * can be estimated using the saddle-point method and the expression
[formula] d 2 log Z N dN 2 ≈ − w 2 N (w − N (d − 1))(w − N d) → − (d + ) 2 ( + 1) 1 N , [/formula]
giving Var[N ] ∝ N * . The fractional fluctuation in particle number thus behaves like δN/N * ∝ 1/ √ N * and is negligible for N * 1. A GCE with the chemical potential µ in Eq. S10 is thus equivalent to a CE of N * nucleosomes with mean gap length .
## S.1.4 statistical positioning with dynamic barriers
The previous section described statistical positioning in its original form, corresponding to non-overlapping nucleosomes distributed between two fixed barriers. The assumption of fixed barriers, however, may not be realized in nature, and we are led to consider a more general model with dynamic barriers: we here consider the statistical positioning of N free nucleosomes (indexed sequentially as 1, ..., N ) flanked by two partially restricted nucleosomes (indexed as 0 and N + 1) that function as moving barriers. For example, the restricted nucleosomes could be subjected to chromatin remodeling or a free energy barrier caused by a Poly(dA:dT) stretch. Let X k denote the center location of the k-th nucleosome. The effect of restrictions on the flanking nucleosomes is encoded in the marginal distribution P (X 0 , X N +1 ). The in-between nucleosomes are then positioned according to statistical positioning conditioned on X 0 and X N +1 . The goal of this section is to express Var[X k ] in terms of P (X 0 , X N +1 ) and compare the resulting expression to the fixed barrier case in Eq. S8.
The evaluation of Var[X k ] is facilitated by first conditioning on X 0 and X N +1 . For example, consider any quantity Q(X 0 , X N +1 , Y 1 , . . . , Y N +1 ), where as defined in the previous section, Y k denotes the partial sum of gap lengths to the left of the k-th nucleosomes. Then, we can compute the expectation of Q as
[formula] E[Q] = E X 0 ,X N +1 E Y |X 0 ,X N +1 [Q]. [/formula]
Because the flanking nucleosomes can be thought of as barriers delimiting the N in-between nucleosomes, the first step of computing the conditional expectation uses the fixed barrier results from the previous section with
[formula] L = X N +1 − X 0 − (N + 1)d. [/formula]
To average over the free nucleosomes, we use
[formula] X k = X 0 + kd + Y k to expand Var[X k ] = Var[X 0 ] + Var[Y k ] + 2 Cov[X 0 , Y k ].(S11) [/formula]
Here only the last two terms depend on the in-between nucleosomes through Y k . To evaluate Var
[formula] [Y k ], we rewrite E[(Y k − E[Y k ]) 2 ] = E X 0 ,X N +1 E Y |X 0 ,X N +1 [Y 2 k ] − E X 0 ,X N +1 E Y |X 0 ,X N +1 [Y k ] 2 .(S12) [/formula]
Using the fixed barrier results in Eq. S7 and Eq. S8, we get
[formula] E X 0 ,X N +1 E Y |X 0 ,X N +1 [Y k ] = E X 0 ,X N +1 kL N + 1 = kE[L] N + 1 E X 0 ,X N +1 E Y |X 0 ,X N +1 [Y 2 k ] = E X 0 ,X N +1 k(N + 1 − k)L(L + N + 1) (N + 1) 2 (N + 2) + kL N + 1 2 = k(N + 1 − k)E[L(L + N + 1)] (N + 1) 2 (N + 2) + k 2 E[L 2 ] (N + 1) 2 . [/formula]
Substituting these expressions into Eq. S12, we get
[formula] Var[Y k ] = k(N + 1 − k)E[L](E[L] + N + 1) (N + 1) 2 (N + 2) + k(k + 1)Var[L] (N + 1)(N + 2) . [/formula]
The last term in Eq. S11 is similarly simplified using the results from the previous section:
[formula] 2 Cov[X 0 , Y k ] = 2E X 0 ,X N +1 [(X 0 − E[X 0 ])(E Y k |X 0 ,X N +1 [Y k ] − E[Y k ])] = 2k N + 1 Cov[X 0 , X N +1 − X 0 ] = k(Var[X N +1 ] − Var[X 0 ] − Var[L]) N + 1 .(S13) [/formula]
Putting everything together finally gives
[formula] Var[X k ] = k(N + 1 − k) N + 2 E[ ](E[ ] + 1) − (N + 1)Var[ ] + (N + 1 − k)Var[X 0 ] + kVar[X N +1 ] N + 1 [/formula]
where we defined = L/(N + 1), so that E[ ] is the mean gap length. In the limit of large N with E[ ] and Var[L] kept fixed, this expression reduces to
[formula] Var[X k ] → Var[X 0 ] + kE[ ](E[ ] + 1) ,(S14) [/formula]
which is just Eq. S5, with replaced by its expectation. In this limit, the only effect of replacing the fixed barriers with partially constrained flanking nucleosomes is a constant offset to the fixed barrier formula in Eq. S9. As a result, the variance gradient is again constant in the semi-infinite limit or near the boundaries of a long confining interval.
## S.1.5 directionally packed nucleosomes
To model packing, we here reformulate statistical positioning in terms of an isothermal-isobaric ensemble and then generalize the case by adding packing forces. Statistical positioning models nucleosomes as dense bidirectional fluid, where movement is restricted only by steric constraints and barriers. Because the nucleosomes are assumed to be free, they are on average evenly spaced. Alternatively, ATP-dependent directional packing could increase nucleosome density near one of the barriers. We will thus model such packing as position-independent, but nucleosome-specific, forces acting on the nucleosomes and derive the variance profile in Eq. (4-6).
The probability distribution of fixed-barrier statistical positioning is, in the language of statistical physics, the canonical ensemble (CE) with zero and infinite energy for non-overlapping and overlapping nucleosomes, respectively. The single barrier case can be equivalently expressed in terms of the isothermalisobaric ensemble (IIE). In this distribution, the region length w is allowed to fluctuate and the weight of an allowed configuration is e −wP , where P is the pressure (we here set the temperature T = 1 for convenience). The single barrier case is retrieved in the thermodynamic limit w, N, L 1 with = L/N kept fixed. The IIE and the CE are equivalent near the barrier in this limit, and their respective parameters are related through P = ln +1 .
We now model packing as constant (i.e. position-independent) and nucleosome-specific forces f k pulling the nucleosomes towards a single barrier. In terms of the IIE, each allowed configuration has weight e −P w− k f k X k , where the second term in the exponent is the work done by the forces. To calculate P (L k |X k−1 ), note that incrementing L k by a suppresses the weight of a state by P (L 1 , ...L k−1 , L k + a, L k+1 , .., w + a) P (L 1 , ...L k−1 , L k , L k+1 , .., w) = e −p k a ,
[formula] p k = P + ∞ k =k f k , [/formula]
where p k is a nucleosome-specific pressure. Normalizing gives the geometric distribution
[formula] P (L k |X k−1 ) = (1 − e −p k )e −p k L k . [/formula]
Using this distribution, and noting that X k−1 and L k are independent, gives
[formula] Var[L k ] = E[L k ](E[L k ] + 1), e −p k = E[L k ] E[L k ] + 1 . (S15) [/formula]
The independence of X k−1 and L k also implies that Eq. S3 is true for this type of packing forces, finally giving
[formula] Var[X k ] = Var[X k−1 ] + E[L k ](E[L k ] + 1).(S16) [/formula]
Thus, the effect of constant packing forces is to modulate the nucleosome pressure p k . While this modulation in turn changes the nucleosome spacing and variance gradient, the relationship between the two is the same as in statistical positioning.
## S.2 supplementary methods: enrichment analysis
The enrichment analysis considers an ordered set of N genes and tests whether a fixed subset preferentially contains either high-ranking or low-ranking genes. To assess statistical significance, the N genes are first ranked between 0 and 1 as i/(N − 1), i = 0, . . . , N − 1, according to their associated data values. Then, the median rank x of the n genes in the fixed subset is calculated. More precisely, consider a data set Ω = {y 0 , y 1 , . . . , y N −1 } of N distinct values, sorted such that y 0 < y 1 < · · · < y N −1 . The rank of y i is thus i. To get a measure of rank that is independent of the size of the data set, we define the rescaled rank of y i to be r i = i/(N − 1), so that 0 ≤ r i ≤ 1. Let S = {y i 1 , . . . , y in } ⊂ Ω be a subset of n elements, where y i 1 < y i 2 < . . . < y in , and define the median rank of S to be median{r i 1 , . . . , r in }. We will develop an enrichment analysis that tests whether S preferentially contains either high-ranking or low-ranking values by comparing its observed median rank with the null distribution of the median rank of all possible distinct subsets of size n, where each distinct subset has uniform probability N n −1 of being sampled.
If n = |S| is odd, then the median rank of S is equal to x = r i (n+1)/2 . Note that x(N − 1) = i (n+1)/2 is an integer. The probability of x is thus
[formula] P (x) = (N − 1)x (n − 1)/2 (N − 1)(1 − x) (n − 1)/2 N n , [/formula]
where the numerator gives the total number of ways of picking (n − 1)/2 values less than y i (n+1)/2 and (n − 1)/2 values greater than y i (n+1)/2 . If n = |S| is even, then the median rank x of S is
[formula] x = r i n/2 + r i 1+n/2 2 = i n/2 + i 1+n/2 2(N − 1) . [/formula]
Thus, x is an integer multiple of 1 2(N −1) , and the possible values of x(N − 1) are n−1 2 , n 2 , n+1 2 , . . . , N − n+1 2 . Defining k = x(N − 1), it can be shown that the probability of x is given by
[formula] P (x) = N n −1 M m=0 k − 1 − m n/2 − 1 N − 1 − ( k + m + 1) n/2 − 1 , [/formula]
where M = min( k − 1 + 1 − n/2, N − n/2 − k − 1), z is the greatest integer less than or equal to z, and z is the smallest integer greater than or equal to z. Two-sided p-values were calculated from these probability functions. w 2 /24, the observed variance agrees well with the true value of σ 2 (dashed line), but it deviates from the true value for larger σ 2 . This analysis justifies our choice of w 2 /24 as a maximal cutoff σ 2 max for calling positioned nucleosomes whose variance we can confidently estimate. : Relation between chemical potential µ and nucleosome spacing for single barrier statistical positioning. The chemical potential was calculated from the formula e µ = ( +1) d−1 / d , using that + d = 167bp is the mean nucleosomes spacing. This relation is derived in Supplementary Methods by taking the thermodynamic limit of the grand canonical ensemble. Figure S7: Nucleosome fuzziness and spacing for short-and medium-long genes. (A) Thin lines show median nucleosome fuzziness σ 2 k of the k-th nucleosome for genes with 2 ≤N ≤ 15 nucleosomes, |ŵ −ŵN | ≤ 40bp and σ 2 +1 , σ 2 −1 < 1200bp 2 . The extrapolationŵN = −164 + 169N was used forN > 9. Dashed lines correspond to the double-barrier model in . Solid red line corresponds single barrier prediction with = 9bp . (B) Variance calculated as in (A), but without the constraint σ +1 , σ −1 < σ max . (C) Median variance σ 2 k for genes in (A). (D) Spacing beween nucleosomes in genes in . (E) Doublebarrier model from .
# S.3 supplementary figures
[fig] Figure S1, Figure S2: Properties of statistical positioning and the grand canonical ensemble (GCE). (A) Probability of most likely nucleosome count (i.e. max N P (N ) where P (N ) ∝ e µN w−N d+N N ) as a function of region length w and chemical potential µ, using d = 147. Red, blue and green dots correspond to values of µ and w used in (B), (D) and (E) respectively. (B) Example of a nucleosome distribution (solid) that is a mixture of two values of N (N = 4 dashed and N = 5 dotted). The length w was chosen such that P (N = 4) = P (N = 5) = 0.5. (C) Fraction of w-values (in the range [200, 1000]) for which one value of N dominates the GCE (max N P (N ) > 0.9) as a function of µ. (D) Values of Var[X k ] calculated from theoretical nucleosome distribution using peak calling (solid blue) or Eq. (7) (dashed), using µ = 2 and varying w. The nine w-values where chosen such that P (N = 5) > 0.9. Var[X k ] is truncated at 1200bp 2 , as described in Methods. (E) Same as in (D) but with µ = Modeled effect of nucleosome overlap on variance estimate. (A)The sampling density of reads associated with nucleosomes (dashed lines) are represented by an array of Gaussians, offset by w and with variance σ 2 (here set to w 2 /24, which is half the variance of a uniform random variable defined on [0, w]). In practice, only the marginal read distribution (solid line) is observed. Reads are assigned to the closest nucleosome peak (e.g. blue filled region for the middle nucleosome). (B) Solid line shows the estimated variance σ 2 observed of the reads in the shaded region in (A), for varying values of the actual variance σ 2 of the middle peak in (A). For σ 2 [/fig]
[fig] Figure S3: Principal component analysis of the fuzziness profiles of 8 nucleosomes in the 5'and 3'-ends of long genes. Left: Cumulative variance explained as a function of principal components. Right: Distribution of fuzziness in the first two principal components. Genes are grouped based on the quartiles of the first principal component. The groups A and A' correspond to the first quartile, B and B' to the two central quartiles, and C and C' to the last quartile. [/fig]
[fig] Figure S5: Plot of variance gradient vs. goodness-of-fit R 2 of statistical positioning for long genes. Scatter plots show the distribution of variance gradient (x-axis) and R 2 -fit (y-axis) calculated from 8 nucleosomes in the 5'-end and 3'-end of long genes (16 or more nucleosomes). Solid lines indicate the marginal distributions. Dashed lines show the marginal distributions obtained by randomly permuting nucleosomes 10 times and refitting. Vertical lines in the top marginal plots indicate the theoretical variance gradient in a single-barrier statistical positioning model for two different values of d, using an internucleosome distance of 168bp. [/fig]
[fig] Figure S6: Distribution of the distance between the last (−1) nucleosome in a gene and the TSS of the nearest downstream gene. The colored lines correspond to genes with fuzziness of the last nucleosome Var[X −1 ] in the intervals [0, 400) (blue), [400, 800) (purple), [800, 1200) (yellow), and [1200, ∞) (green). This figure shows that well-positioned last nucleosomes tend to be near the TSS of an adjacent gene. [/fig]
|
SEPPA 3.0—enhanced spatial epitope prediction enabling glycoprotein antigens
B-cell epitope information is critical to immune therapy and vaccine design. Protein epitopes can be significantly affected by glycosylation, while no methods have considered this till now. Based on previous versions of Spatial Epitope Prediction of Protein Antigens (SEPPA), we here present an enhanced tool SEPPA 3.0, enabling glycoprotein antigens. Parameters were updated based on the latest and largest dataset. Then, additional micro-environmental features of glycosylation triangles and glycosylationrelated amino acid indexes were added as important classifiers, coupled with final calibration based on neighboring antigenicity. Logistic regression model was retained as SEPPA 2.0. The AUC value of 0.794 was obtained through 10-fold cross-validation on internal validation. Independent testing on general protein antigens resulted in AUC of 0.740 with BA (balanced accuracy) of 0.657 as baseline of SEPPA 3.0. Most importantly, when tested on independent glycoprotein antigens only, SEPPA 3.0 gave an AUC of 0.749 and BA of 0.665, leading the top performance among peers. As the first server enabling accurate epitope prediction for glycoproteins, SEPPA 3.0 shows significant advantages over popular peers on both general protein and glycoprotein antigens. It can be accessed at http://bidd2.nus.edu. sg/SEPPA3/ or at http://www.badd-cao.net/seppa3/ index.html. Batch query is supported.
# Introduction
Antigens can be specifically bound by corresponding antibodies through interacting with epitope residues. The identification of B-cell epitopes is of primary importance to vaccine design, immune-diagnostics and antibody produc-tion. Pioneering work was started by CEP in 2005, then continuous efforts have been paid to epitope prediction in recent years. Currently, several popular tools are available online, such as PEPITO (3), DiscoTope 2.0 (4), Epitopiaand SEPPA, which were reviewed previously. In 2014, SEPPA 2.0 (7) introduced new features of 'relative ASA preference of unit patch' and 'consolidated amino acid index' to establish a logistic regression model by considering immune host and subcellular localization, demonstrating both increasing accuracy and low false positive rate. Recently, BepiPred 2.0 (8) employed random forest algorithm to predict sequence-based epitopes, achieving an AUC value of 0.596 on external validation set. Besides antigen structures, several methods considered additional information for epitope prediction, such as antibody informationor experimental data, which is only beneficial to those antigens with prior knowledge. Thus, despite substantial efforts, conformational epitope prediction for protein antigens is still challenging and worthy of further devotion.
In subsequent analysis of 897 immune complexes from PDB database, we found that almost 70% of protein antigens contain N-linked glycosylation sites, indicating the potential to be post-translational modification (PTM) of Nlinked glycosylation. It is known that antigenicity of protein antigens, particularly among virus antigens, can be dramatically affected by N-linked glycosylation. For example, it was reported that key glycan in HIV-1 gp120 surface could block antibody binding and inhibit antibody recognition. Also, a group of broadly neutralizing antibodies could directly target the glycan-dependent epitopes located in the V1/V2 region on HIV-1 gp120. In addition, effective tumor vaccines are under development targeting tumorassociated carbohydrate antigens (TACA), which resulted from aberrant glycosylation.
It is noted that N-linked glycosylation is the most common forms of protein glycosylation, which may influence the local environment physio-chemically, particularly to those neighboring amino acids on protein surface. Yet how to describe the micro-environment featured by glycosylation sites is highly challenging, which has largely hindered the study of epitope prediction involving glycoprotein antigens. So far, no algorithm has considered the effect of glycosylation sites for epitope prediction. Here, we present a latest and enhanced version of SEPPA 3.0, filling the blanks for N-linked glycoproteins. In this version, parameters were refreshed based on the largest dataset till now. Then, the ratio of glycosylation triangles and glycosylation related amino acid indexes were developed and further integrated as new classifiers for glycoprotein antigens. Finally, a calibration parameter was introduced to reduce false positive rate. The performance of SEPPA 3.0 was rigorously evaluated on both glycoprotein and general protein antigens and compared with popular peers.
## Dataset
All datasets were extracted from antigen-antibody complexes in Protein Data Bank (PDB). For quality control, those protein antigens with over 50 residues were remained. Solvent accessible surface areas (SASA) were calculated and epitope residues were defined as the same as SEPPA 2.0 (7). Finally, 832 PDB ID with 897 unique epitope patches were remained. For model construction, 767 protein antigens (520 with N-glycosylation sites) deposited before year 2016 were selected as internal training dataset. Totally, 767 unique epitope structures including 16 544 epitope residues as positive training dataset and 172 975 nonepitope residues as negative training dataset. Two sets were adopted as independent external testing datasets. One refers to general protein antigens without discrimination of Nlinked glycosylation or not. This set contains 130 conformational epitopes structures after year 2016, including 2598 positive residues and 31 144 negative ones. Those with Nglycosylation sites were referred as second testing dataset of glycoprotein antigens covering 106 antigens. (Supplementary .
# Materials and methods
Parameters in SEPPA 2.0 (7) were firstly updated based on the latest training dataset. Two critical parameters, the ratio of glycosylation triangles and glycosylation related amino acid indexes, were added as new classifiers to improve the prediction performance. Then, different machine learning approaches were screened, and logistic regression algorithm was selected to establish our model. Finally, calibration was introduced to reduce false positive rate. The algorithm of SEPPA 3.0 and definition of each parameter are given as below:
Algorithm of SEPPA 3.0
The functions of sub-model recommendation were remained the same as in SEPPA 2.0 (7). In addition, the prediction algorithm of SEPPA 3.0 was updated as below:
Step 1: determine all the surface residues in the protein antigen;
For each surface residue r :
Step 2: search all possible unit triangles within 15Å atom distance and calculate three triangle-related parameters: propensity index avg r (see SEPPA 1.0), relative ASA Apr e f r (see SEPPA 2.0) and ratio of glycosylation triangles Glytri r using Equation (2); Step 3: calculate clustering coefficient CC r (see SEPPA 1.0), consolidated AAindex value Index r (see SEPPA 2.0) and glycosylation related AAindex (Glyindex r ) using Equation (3);
Step 4: integrate above six parameters via logistic regression model to present raw antigenicity score for each residue;
Step 5: calibrate raw prediction score using Equation (4) to give the final antigenic score for residue r ;
Step 6: output the final antigenicity score, and highlight those residues with scores higher than defined threshold. Visualize the subset of predicted epitopes graphically.
## Ratio of glycosylation triangles
The N-glycosylation sites of Asn were defined by sequons of Asn-X-Ser/Thr, where X can be any amino acid apart from proline. Surface residue triangle involving Asn of glycosylation sequons was defined as glycosylation triangle (gl ytri ). Further, epitope glycosylation triangle (epi gl ytri ) was defined when it contains at least two epitope residues. By consolidating 20 amino acids into 13 functional subgroups, 71 different glycosylation triangles patterns of subgroups were observed based on training dataset. Ratio of glycosylation triangles (ratio i ) was intended as its general enrichment in epitope areas as Equation (1) described:
[formula] ratio i = N epi glytri i N glytri i (i = 1, 2, . . . , 71)(1) [/formula]
where N epi glytri i represents the number of epitope glycosylation triangles pattern i in training dataset, while N glytri i is the number of glycosylation triangles pattern i in training dataset.
For each surface residue r , search all possible gl ytri within 15Å atom distance and define the occurrences times of specific glycosylation triangle pattern i as N occur (i ) . The weighted score for glycosylation triangle pattern i (gl ytri i ) was defined as W ri = ratio i * N occur (i ) and scores of 71 existing gl ytri pattern were integrated by artificial neural networks (ANNs) into a consolidated ratio of glycosylation triangles as Equation (2) described:
[formula] Glytri r = NN{W r 1 , W r 2 , . . . , W ri , . . . , W r 71 } (i = 1, 2, . . . , 71)(2) [/formula]
where W ri means the weighted score of gl ytri pattern i , while Glytri r indicates the consolidated ratio of glycosylation triangles via two-layer and ten-node ANN.
## Glycosylation related amino acid indexes
Statistical analysis was done to derive four glycosylationrelated amino acid indexes from AAindex databaseas initial features classifying epitope residues from non-epitope surface ones, including KIMC930101, LAWE840101, RICJ880102 and ROBB760113 (Supplementary. Then, shell structure model was generated to summarize different amino acid indexes into each layer. Based on the radius of 10Å and step size of 2Å, five layers were generated for each residue r . For each index i of residue r , the amino acid index of all residues within layer j were averaged to generate a glycosylationrelated index as gl yi ndex i j . Finally, 20 indexes (4 indexes * 5 layers) were optimized through iterative filtering, and the optimized combination of glycosylation-related amino acid indexes were consolidated by ANNs as Equation (3) illustrated: where i represents four types of AAindexes and j represents five layers of shell model, gl yi ndex i j means the averaged index i within layer j of shell model, and Glyindex r indicates glycosylation-related AAindexes of residue r .
[formula] Glyindex [/formula]
## Calibration parameter
The idea of calibration is to adjust the raw score of individual residues by the overall tendency of neighboring residues, such as a low-score/high-score residue sitting in high-epitope/low-epitope environment. The adjusted score of residue r was defined by the average score of all neighboring surface residues as Equation (4) illustrated:
ad j ust score r = raw score i N
where raw score i represents the sum of raw predicted scores for each neighboring surface residue within 5Å atom distance of target residue r , while N means the total number of above residues.
# Results
## Internal validation of seppa 3.0
The performance of SEPPA 3.0 was rigorously validated through both internal and external validation. As being illustrated in SEPPA 2.0 (7), area under curve (AUC) value and balanced accuracy (BA) were adopted as evaluation parameters. For internal validation, five commonly used machine learning approaches were screened, including logistic regression, naïve Bayes, random forest, support vector machine and decision tree. The results of 10-fold crossvalidation showed that logistic regression gave the best prediction results with AUC value over 0.79 for general protein antigens (Supplementary, and was chosen for model construction of SEPPA 3.0.
## Comparing with peer methods
Further, SEPPA 3.0 was evaluated through independent testing dataset and further compared with six available peers on-line with two sets of independent testing data. Peers include Epitopia (5), Discotope-2.0 (4), Pepito (3), CBTOPE (24), SEPPA 2.0 (7) and BepiPred-2.0 (8). Two independent testing sets include: (i) group 1 contains 130 general protein antigens (G1) and (ii) group 2 contains 106 glycoprotein antigens (G2). Note that there is no recommended threshold from Epitopia (5), the optimized threshold with maximum BA on different testing set was purposely selected for Epitopia. The results of evaluation on G1 general protein antigens were illustrated in SupplementaryFurther performance was evaluated on G2 glycoprotein antigens. The AUC was pushed up to 0.749 by SEPPA 3.0, leading the most accurate prediction for glycoprotein antigens, while the second winner is Pepito (3) with AUC of 0.676. Thus, SEPPA 3.0 shows significant advantages over popular peers on both general protein and glycoprotein antigens. Detailed performance of SEPPA 3.0 and available peers can be found in .
## Case study
Here, the well-known gp120 glycoprotein antigen (PDB ID: 5IF0, Chain: G) was selected as input and the results were illustrated in. Gp120 contains 30 epitope residues according to the complex structures from PDB, out of which, 29 were predicted as true positive by SEPPA 3.0. Under the default threshold of 0.089, though 60 residues were calculated as potential epitope residues, most of true epitope residues were ranked in the top list. For example, by raising the threshold from 0.089 to 0.29, top 16 ranking marked in red contains 15 true epitope residues, indicating the excellent performance of SEPPA 3.0. The prediction results of 5IF0 G could reach to an AUC of 0.936 with BA of 0.872. Also, the predicted epitopes of SEPPA 2.0 and other peers including Bepipred-2.0, CBTOPE, Discotope-2.0, Epitopia and Pepito were illustrated into I, under default cut-offs respectively. Results illustrated that, compared with other peers includ-ing SEPPA 2.0, SEPPA 3.0 gave the best results which are most similar to reference epitopes. Meanwhile, those false positive of discrete candidates predicted by SEPPA 2.0 can be successfully adjusted through the algorithm of SEPPA 3.0. Besides the case of HIV-1 gp120, another example was also illustrated (Supplementary based on human glycoprotein antigens of CD27 (PDB ID: 5TLK Chain: X), which is an important antibody drug targets for autoimmune diseases and cancers. Model performance of 130 individual structures from testing dataset can be found in .
Nucleic Acids Research, 2019, Vol. 47, Web Server issue W393 USAGE Input SEPPA 3.0 (http://bidd2.nus.edu.sg/SEPPA3/ or http:// www.badd-cao.net/seppa3/index.html) accepts two types of input files: (i) Existing PDB IDs with chain name, and (ii) Local files in PDB format. Like SEPPA 2.0 (7), users are recommended to select subcellular localization of protein antigen and species of immune host if available. Also, batch query submission is encouraged. Users can submit multiple PDB IDs in batch query with specified PDB IDs, subcellular localization, species of immune host and chain name as input. After successful submission, SEPPA 3.0 will automatically recommend the best model based on user's specifications.
## Output
The output results of SEPPA 3.0 will be either presented in .html format or be sent back to users via email. The .html format will provide result summary of input files in sequence level, in which capital letter and lowercase represent the surface and core residues respectively. Possible epitope residues predicted were marked in red capital lettersand detailed antigenicity scores calculated were recorded in score file for each residue in query antigen. Users can visualize the prediction results through online plug-in Jsmol components or local PDB files downloadable. Here, HIV-1 gp120 (PDB ID: 5IF0, Chain: G) is shown as an example to illustrate the predicted results of SEPPA 3.0. Residues were colored according to predicted scores. More information can be found in the HELP page of SEPPA 3.0.
# Discussion
In recent decade, the progress in conformational epitope prediction goes smoothly but slowly. The space to improve may lie in several aspects, including accumulation of more epitope structures, appropriate classifying features, more refined models for specific datasets, and so on. In order to maintain the top performance, SEPPA 3.0 updated the largest datasets, the latest features, and designed two new classifiers for N-linked glycoprotein antigens. Our statistical analysis showed that N-glycosylation sites were significantly enriched in epitope areas rather than in non-epitope surface regions (paired t test, P = 3.577e-14, 95% CI = [0.032, 0.051]), suggesting that antibodies seem to prefer certain surface region containing N-glycosylation sites. In other words, comparing to those in non-epitope surface areas, N-glycosylation sites in epitope regions might have different local context residues layout, as well as physio-chemical fields resulted from neighboring residues. Thus two types of parameters were designed to test the potential difference of the micro-environment nearing glycosylation sites between epitope areas and non-epitope surface regions. After statistical analysis, the unit patch of residue triangles (6) involving N-glycosylation sites and four amino acid indexes were selected out to indicate residues layout and physio-chemical fields respectively. Meanwhile, by introducing shell structure to construct the glycosylation-related AAindexes, the layers of micro-environmental variations were fully considered and summarized for each residue. Despite of the diverse features invented till now, predicting results can only be lifted by combining multiple of them, suggesting the inherent complexity of epitope nature.
Another task for epitope prediction is to reduce the high False Positive Rate (FPR). Most available algorithms calculated epitope scores based on each individual residue. As structural epitope areas are surface patches being recognized and bound by CDR regions of antibodies, collective effects might play much more important roles than we had expected. Differed from other popular peers, SEPPA 3.0 designed the parameters of the local structural layout and micro-environment around each target residue to describe the collective effects in epitope regions. More importantly, a calibration procedure was elaborated to further reduce FPR based on neighboring influence. In fact, after calibration, the FPR of SEPPA 3.0 was significantly decreased from 0.35 to 0.26 for general protein antigens and from 0.34 to 0.25 for glycoprotein antigens.
For binary classification problems, AUC value was often adopted to evaluate the model performance under different thresholds. Another evaluation parameter we took is balanced accuracy, which fully considered the balance between sensitivity and specificity. Using the above two parameters, the performance of different tools can be fairly compared on independent datasets. It is noted that the structure information of conformational epitope is rapidly accumulating. Coupled with new classifiers and FPR reducing methods, it is expected that epitope prediction models could better serve the biological needs and assist the process of potential vaccine and immune therapeutic development.
## Data availability
PDB ID: 5IF0, 5TLK.
## Supplementary data
Supplementary Data are available at NAR Online. |
An unusual unifocal presentation of Castleman’s disease in a young woman with a detailed description of sonographic findings to reduce diagnostic uncertainty: a case report
Background: Castleman's disease is a rare lymphoproliferative disorder. It typically presents as mediastinal masses and causes a wide range of clinical symptoms. Histologically, Castleman's disease is classified as either a hyalinic vascular or plasma cell variant. The prognosis mainly depends on the histological type and broadly varies. We herein report our sonographic findings in a patient with Castleman's disease, including gray-scale ultrasonography, color Doppler ultrasonography, and sonoelastography ultrasonography, which have not been previously reported in the literature. These findings allowed for a preoperative diagnosis and avoidance of overly aggressive therapy.Case presentation: A 28-year-old European female patient with unicentric Castleman's disease of hyalinic vascular type (HV) restricted to the axilla was referred to us because of a 4-month history of a painless, solitary mass located in the left axilla. The patient's medical history was unremarkable.Conclusion: Castleman's disease is a pathologic entity of unknown etiology and pathogenesis. In this case report of unicentric HV-type CD, we demonstrate that typical sonographic findings can lead to a preoperative diagnosis of Castleman's disease. Core needle biopsy usually allows for a final diagnosis and helps to avoid unnecessary operations and overtreatment.
# Background
Castleman's disease (CD), or giant lymph node hyperplasia, is a rare lymphoproliferative disorder that typically presents as mediastinal masses. It was first described by Castleman and colleagues as a localized mass of mediastinal lymphoid follicles in 1954. Two years later, it was defined as a pathologic entity of unknown etiology and pathogenesis [bib_ref] Records of the Massachusetts General Hospital-weekly clinicopathological exercises (case 40011), Castleman [/bib_ref] [bib_ref] Localized mediastinal lymphnode hyperplasia resembling thymoma, Castleman [/bib_ref].
Clinically, CD may be localized with no major symptoms and present as a solitary mass or swelling, or it may be a generalized, symptomatic disease with fever, weight loss, anemia, hepatosplenomegaly, and generalized lymphadenopathy.
Histologically, the disease is also classified into two separate subtypes: the hyalinic vascular (HV) variant (80%-90% of cases) and the plasma cell (PC) variant (10%-20% of cases). Intermediate and mixed types have also been reported.
The prognosis mainly depends on the histological type and shows a broad variety. Treatment can range from curative surgery for the solitary form to the use of steroids, monoclonal antibodies, chemotherapy, and radiotherapy for the multicentric type [bib_ref] Castleman disease, Dham [/bib_ref].
We herein report on a 28-year-old female patient with unicentric CD restricted to the axilla. We describe the findings and imaging features of gray-scale ultrasonography (US), color Doppler US, sonoelastography US, and contrast-enhanced dynamic computed tomography (CT). A pathway to a preoperative diagnosis, management of the disease, and the clinical course are presented. A review of the literature and differential diagnoses are also presented.
## Case presentation
A 28-year-old European female patient was referred to us because of a 4-month history of a painless, solitary mass located in the left axilla. She had no accompanying complaints, history of fatigue, night sweats, or weight loss. Her medical history was unremarkable.
On routine physical examination, a solitary enlarged lymph node was detected in the left axilla in the absence of any breast pathology. No generalized lymphadenopathy or other organomegaly was noted. Peripheral blood counts and the erythrocyte sedimentation rate were within normal limits. Interestingly, the levels of lymphoproliferative markers such as serum soluble IL-2R, beta 2microglobulin, and immunoglobulins were also normal; however, the C-reactive protein level was slightly increased.
The lymph node in the left axilla measured 4 cm and was mobile, nontender, and soft in consistency. US examination of the breast and axilla was performed with an iU22 (Philips Healthcare, Bothell, WA) and ProSound 7 (Aloka, Hitachi, Zug, Switzerland) using a 12-MHz linear array transducer.
High-frequency, high-resolution gray-scale US revealed a well-defined, uniformly hypoechoic, ovoid axillary mass, 38 × 17 × 28 mm in size. The longitudinal diameter was greater than the transverse diameter with a longitudinal to transverse axis ratio of more than 2. A hyperechoic fatty hilum could not be detected and was totally replaced by cortical thickening. Although soft in consistency, the lesion could only be slightly deformed by compression with the transducer. Color Doppler flow was performed with optimized color Doppler parameters set at a low wall filter (80-100 Hz) and low velocity scale (pulse repetition frequency, 1000 Hz). Color gain was adjusted dynamically to maximize depiction of blood vessels while avoiding artifactual color noise. Bizarre and multifocal peripheral flow was detected, whereas central or central perihilar flow was not revealed [fig_ref] Figure 1: Color Doppler sonogram shows peripheral vascular flow within an ovoid hypoechoic axillary... [/fig_ref]. A three-dimensional and multislice imaging scan with the capability of reproducing high-resolution images confirmed these B-mode findings, but could not provide additional important information.
Spectral Doppler analysis along the periphery of the node showed both arterial and venous pulse wave patterns. The blood flow profile of the arteries indicated a broad range in the resistance index, pulsatility index, and peak systolic velocities varying from low to high pulsatility. Thus, no further information could be drawn on these indices.
Sonoelastography US confirmed the clinical examination findings: the lesion was characterized by soft tissue with some less elastic regions of higher stiffness.
An US-guided fine needle biopsy with multiple passages of the needle tip through the nodal cortex was made to sample as much of the nodal cortex as possible. Fine needle aspiration cytology (FNAC) only revealed a mixed population of small and large lymphoid cells. In particular, prominent vascularity with hyalinized capillaries was not detected. The FNAC results were subsequently reported as "negative for malignant cells," and histopathologic examination of the lymph node was advised. Therefore, US-guided core needle biopsy using a 14-G automated gun was performed, and a diagnosis of HVtype CD was confirmed: microscopic examination revealed many variably sized hyperplastic follicles, progressive vascular proliferation, and hyalinization [fig_ref] Figure 2: Histopathologic specimen shows a continuum of abnormal germinal centers [/fig_ref]. A multislice CT scan of the head, thorax, and abdomen was subsequently performed and allowed for the exclusion of multicentric type CD. The lymph node in the left axilla on CT was described as a well-circumscribed, homogeneous mass lesion with moderate to intense enhancement and rapid washout [fig_ref] Figure 3: CT scan shows enlarged lymph node [/fig_ref]. The patient underwent open biopsy by a surgical gynecologist, and the enlarged axillary lymph node was completely excised [fig_ref] Figure 4: Macroscopic aspect of affected lymph node measuring 4 [/fig_ref]. The postoperative course was uneventful, clinical follow-ups were unremarkable, and there has been no evidence of recurrence.
# Discussion
CD is a rare, benign lymphoproliferative disorder of unknown etiology. The two main hypotheses for its development are an abnormal immune response and viral infection. Human herpes virus 8 and interleukin 6 are regarded to be linked to the pathogenesis [bib_ref] Castleman disease in the 21st century: an update on diagnosis, assessment, and..., Van Rhee [/bib_ref]. Microscopically, as stated above, two histological subtypes are known: the HV type and PC type. Depending on the clinical presentation, CD can also be divided into a localized and multicentric type. About 90% of the localized type belongs to the HV subgroup, as seen in our patient, and almost all of the multicentric type is histologically the PC subtype.
CD can develop anywhere that lymphoid tissue is found, most commonly in the mediastinum (60%), but also in the abdomen, neck, lung, and retroperitoneum. Less than 4% of cases present as a lymph nodal mass in the axilla [bib_ref] Castleman's disease with isolated extensive cervical involvement, Yildirim [/bib_ref]. Patients with the HV type are usually asymptomatic, as our patient was, whereas patients with the PC type typically present with a broad variety of symptoms such as fever, weight loss, generalized lymphadenopathy, night sweats, and hepatosplenomegaly. As also demonstrated in our case, localized CD is normally cured after excision of the tumor with an excellent prognosis and 5-year survival of approximately 100%. On the other hand, successful treatment of the multicentric type often requires multimodal management including radiotherapy, chemotherapy, and surgery. The prognosis is generally less favorable [bib_ref] Castleman disease in the 21st century: an update on diagnosis, assessment, and..., Van Rhee [/bib_ref].
In most cases of CD, as in our patient, US shows a hypoechoic, well-circumscribed homogeneous mass lesion [bib_ref] A rare case of peripancreatic Castleman's disease diagnosed preoperatively by endoscopic ultrasound-guided..., Khashab [/bib_ref]. In all reported cases, the longitudinal to transverse axis ratio of involved lymph nodes was more than 2 and was significantly higher in benign than in malignant lymph nodes [bib_ref] Axillary node staging by ultrasonography and fine-needle aspiration cytology in patients with..., Baruah [/bib_ref]. Color Doppler findings are characteristic for the diagnosis of CD: prominent peripheral vascular proliferation in the node is not seen in healthy or reactive lymph nodes and is absent from lymph nodes affected by malignancy. Reactive lymph nodes are more likely to preserve a normal vascularity pattern with central hilar vessels, whereas lymph nodes in patients with CD show bizarre new blood vessels in the periphery due to neovascularization, as in our patient [bib_ref] Doppler findings in castleman disease-a rare case, Raniga [/bib_ref]. Histologically, polymorphous lymphoreticular infiltrates containing numerous capillaries are seen at the periphery of the lymph node. Malignant lymph nodes typically present a mixed vascular distribution including both central and peripheral flow.
These US and Doppler findings, although nonspecific, seem to be characteristic for the diagnosis of this uncommon disease entity and may help to differentiate this benign process from reactive lymph nodes and nodal metastases. These US findings must be proven to be efficacious, and larger studies of patients with CD are required to determine the role of US and sonoelastography in this group. Whether the distribution of nodal vascularity and Doppler flow characteristics can help to achieve a better understanding of CD must be assessed.
As in our case, CD is difficult to diagnose based on aspirate material. FNAC as the initial investigation method may be misleading because no specific cytomorphological criteria for a definitive diagnosis have been described, nor are there any cytomorphological features pathognomonic for the disorder [bib_ref] Castleman's disease -hyaline vascular type -clinical, cytological and histological features with review..., Ghosh [/bib_ref].
Another technique for preoperative axillary node diagnosis is US-guided core biopsy. Although more expensive and invasive, resulting in a higher complication rate, core biopsy has the advantage of sampling the nodal tissue more extensively than using FNAC. Using core needle biopsy and excision biopsy, all cases reported in the literature were diagnosed as CD [bib_ref] Castleman's disease -hyaline vascular type -clinical, cytological and histological features with review..., Ghosh [/bib_ref]. As in our patient, core needle biopsy was superior to FNAC and gave the correct definitive diagnosis.
The differential diagnoses of an axillary mass include metastases, lymphoid neoplasms such as Hodgkin's lymphoma and non-Hodgkin lypmphoma, and a number of reactive, inflammatory, and nonmalignant conditions such as rheumatoid arthritis, Wiskott-Aldrich syndrome, tuberculosis, sarcoidosis, syphilis, and other disorders of immune regulation in patients with acquired immune deficiency syndrome and Kaposi's sarcoma. Because of its variable clinical presentation, CD should be considered as a differential diagnosis of any enlarged lymph node.
# Conclusion
In conclusion, although it is probably not possible to render a definitive diagnosis of CD based on US findings, the presence of a hypoechoic lymph node with many prominent peripheral vessels on Doppler sonogram should at least raise the diagnostic possibility. This case report highlights the difficulty in diagnosing unicentric CD in FNA samples. Core needle biopsy, which usually achieves the final diagnosis, should be given preference. As shown in our case report, unicentric CD should be a differential diagnosis of an enlarged lymph node, especially in asymptomatic and young patients. Surgical removal of the affected lymph node is curative in localized HV-type CD. Confirmation of CD should be based upon the combination of clinical, sonographic, CT, and histopathological findings.
## Consent
Written informed consent was obtained from the patient for publication of this manuscript and accompanying images. A copy of the written consent is available for review by the Editor-in-Chief of this journal.
In this original case report, we first describe the findings and imaging features of Castleman's disease based on gray-scale ultrasonography (US), color Doppler US, sonoelastography US, and contrast-enhanced dynamic computed tomography. The description of the sonographic findings in this unique case of Castleman's disease of the axilla will certainly advance our understanding of this illness.
[fig] Figure 1: Color Doppler sonogram shows peripheral vascular flow within an ovoid hypoechoic axillary lymph node. [/fig]
[fig] Figure 2: Histopathologic specimen shows a continuum of abnormal germinal centers (GC) with prominent CD 23-positive follicular dendritic cells (dense black brown staining) and subtle vascular proliferation (original magnification, ×200). MZ, mantle zone. [/fig]
[fig] Figure 3: CT scan shows enlarged lymph node (arrow) in the left axilla. [/fig]
[fig] Figure 4: Macroscopic aspect of affected lymph node measuring 4.5 cm. [/fig]
|
Efficient and secure three-party mutual authentication key agreement protocol for WSNs in IoT environments
In the Internet of Things (IoT), numerous devices can interact with each other over the Internet. A wide range of IoT applications have already been deployed, such as transportation systems, healthcare systems, smart buildings, smart factories, and smart cities. Wireless sensor networks (WSNs) play crucial roles in these IoT applications. Researchers have published effective (but not entirely secure) approaches for merging WSNs into IoT environments. In IoT environments, the security effectiveness of remote user authentication is crucial for information transmission. Computational efficiency and energy consumption are crucial because the energy available to any WSN is limited. This paper proposes a notably efficient and secure authentication scheme based on temporal credential and dynamic ID for WSNs in IoT environments. The Burrows-Abadi-Needham (BAN) logic method was used to validate our scheme. Cryptanalysis revealed that our scheme can overcome the security weaknesses of previously published schemes. The security functionalities and performance efficiency of our scheme are compared with those of previous related schemes. The result demonstrates that our scheme's security functionalities are quantitatively and qualitatively superior to those of comparable schemes. Our scheme can improve the effectiveness of authentication in IoT environments. Notably, our scheme has superior performance efficiency, low computational cost, frugal energy consumption, and low communication cost.OPEN ACCESSCitation: Chen C-T, Lee C-C, Lin I-C (2020) Efficient and secure three-party mutual authentication key agreement protocol for WSNs in IoT environments. PLoS ONE 15(4): e0232277.Wireless sensor networks (WSNs) are crucial in these IoT applications [2, 3]. WSNs have become increasingly used in providing services for monitoring environments and activities because of their low cost, flexibility, ease of deployment, and wide range of applications(Fig 1)[4]. As illustrated inFig 1,WSNs comprise numerous sensor nodes scattered arbitrarily over a certain region. Sensor nodes can sense, process, and transmit information (e.g., temperature and traffic information). Remote users are required to reach a specific sensor node via the gateway node (GWN)[5,6]. Each scattered sensor node can collect data and route data back to the GWN. Remote users may communicate with a GWN through the Internet. When data from WSNs are made available to users, the legitimacy of each user must be verified before the system can grant access to the data, and the sensor nodes reserved for access must be confirmed to be legitimate. Hence, remote user authentication is necessary and critical for secure information transmission in WSNs[2,5,7,8]. The following basic design criteria must be considered when designing a remote user authentication scheme for WSNs [2, 5, 8]:1. Mutual authentication. Users and sensor nodes must mutually authenticate each other.After they have authenticated each other, they must arrange a session key for information transmission.2. Masquerade attack resistance. An adversary cannot impersonate a legal user to log in to WSNs. In addition, the adversary cannot masquerade as a sensor node to spoof the user.3. Replay attack resistance. The adversary cannot attempt to replay previously intercepted messages to spoof the GWN.4.Guessing attack resistance. The adversary cannot obtain useful information to devise an offline check of the correctness of guessed passwords.
# Introduction
Internet of Things (IoT) is an emerging technology, which is the extension of Internet connectivity into various devices such as sensors, vehicles, and mobile phones. These devices can interact with each other over the Internet. A wide range of applications connecting objects that can communicate with each other have been deployed; applications include transportation systems, healthcare systems, smart buildings, smart factories, and smart cities. time of a user's temporal credential is regulated by the GWN. A user's temporal credential is related to the identity of user and can be securely stored in a smart card. The temporal credential of a sensor node is also related to its identity and confidentially written in its storage. Based on the issuing and signing of temporal credential, the mutual authentication between the user and the GWN is achieved through the verification of temporal credential for the user. The mutual authentication between the sensor node and the GWN is achieved by the verification of the temporal credential for the sensor node. Each dynamic ID is temporarily assigned by the system and mapped to a specific user. A dynamic ID is a combination of its user's information and a random nonce. The random nonce is an arbitrary number; it is used only once during the communication. In the authentication process, the login message of the user i contains a dynamic ID, called DID i . The login message is dynamic for each login. For all i, the parameter DID i is associated with nonce N i and changed dynamically for each login. The use of a dynamic ID in each login message can avoid the risk of ID-theft. Our scheme introduces dynamic ID to anonymize users.
# Motivation and contribution
Typical IoT installations allow remote users to access data from sensor nodes in WSNs through the Internet. Researchers have been developing effective approaches for merging WSNs into IoT environments. Because of the resource constraints of sensor nodes, to design an efficient and secure authentication scheme for WSNs in IoT environments constitutes a nontrivial challenge. In IoT environments, the security effectiveness of remote user authentication is crucial for trustworthy information transmission. Computational efficiency and energy consumption are crucial because of the limited energy resources of WSNs. Moreover, time synchronization is a critical and challenging problem for WSNs; the system must provide a synchronized logical time clock for all devices and objects in IoT environments. Any adversary and any malicious node in IoT environments can attack clock synchronization. The communication errors, frequent topological changes, lowcost clocks, and limited energy levels of IoT nodes are other factors that can affect time synchronization. A timestamp-based authentication scheme requires trustworthy timestamps and synchronized time clocks to verify any device's legitimacy. When a system has a serious time synchronization problem, no device can be synchronized with any another device, and thus the system cannot verify any device's legitimacy. Therefore, any serious time synchronization failure causes mutual authentication failure. The time synchronization problem should be contemplated as designing a remote user authentication scheme for WSNs in IoT environments. Moreover, when a given user's ID is revealed, an adversary can determine any information concerning the user's identity and monitor the user's activities. An exposed user ID is also useful to the adversary because it provides login information. Therefore, anonymous access for each login should be required. Although several previously published studies have proposed diverse remote user authentication schemes, they have been neither highly secure nor efficient sufficiently to satisfy the requirements of WSNs in IoT environments (Related work in Section 2). This paper proposes a more efficient and secure authentication scheme for WSNs in IoT environments to ameliorate these security weaknesses.
The major contributions of our work are as follows:
1. We propose a new three-party scheme on the basis of temporal credentialand dynamic IDfor WSNs in IoT environments to achieve security, mutual authentication, and session key agreement. Cryptanalysis revealed that the security functionalities of the proposed scheme qualitatively and quantitatively superior to those of previous schemes; the proposed scheme can advance the field of authentication schemes. The Burrows-Abadi-Needham (BAN) logic methodwas used to validate our scheme.
2. The proposed scheme performs efficiently in IoT environments, with low computational cost, frugal energy consumption, and little communication cost.
3. Our scheme uses temporal credentials and random nonce instead of the timestamps to verify mutual authentication among U i , the GWN, and S j . Therefore, our scheme can avoid the time synchronization problem for WSNs in IoT environments. Moreover, dynamic ID technologyis applied in our scheme. User identities are consequently anonymous and can be confirmed only by the service provider.
## Organization of the paper
The remainder of this paper is organized as follows: Section 2 introduces a brief review of the related work in WSNs and explains the security weaknesses of the Ostad-Sharif et al. schemefor WSNs in IoT environments; Section 3 details the proposed efficient secure authentication scheme for WSNs in IoT environments; Section 4 presents the security analysis of the proposed scheme; Section 5 discusses the effectiveness and efficiency of the proposed scheme; and finally, Section 6 presents the study's conclusion.
## Related work in wsns
To satisfy the security requirements of WSNs, many remote user authentication schemes have been proposed. In 2004, Benenson et al.described the security issues of user authentication in WSNs and proposed a protocol for them, in which the user can achieve successful authentication with any subset of sensors from a set of n sensors (n being the average number of sensors within a broadcast distance of the user). Watro et al.proposed a TinyPK authentication protocol with the Rivest-Shamir-Adleman (RSA) public key cryptosystemand Diffie-Hellman key agreement algorithm. However, this authentication protocol has the disadvantage of the masquerade attack, in which an adversary can masquerade as a sensor node to spoof the user. Wong et al.proposed a less complex lightweight user authentication protocol for WSNs by using hash function operations. However, the scheme cannot protect against stolen-verifier, replay, and forgery attacks. Moreover, the passwords in the scheme can be revealed easily by any of the sensor nodes, and users cannot change their passwords freely. In 2009, to eliminate the weaknesses of the Wong et al. scheme, Dasproposed a two-factor user authentication scheme for WSNs. The scheme implements passwordbased authentication with the assistance of a GWN to access resource-constrained sensor nodes. However, this scheme is vulnerable to insider, masquerade, offline password-guessing, stolen smart card, and GWN bypassing attacks. The scheme does not provide mutual authentication, a key agreement, and a password change phase for users to change or update their password. Khan et al., Chen et al., and Yeh et al.have subsequently proposed new schemes for improving the inherent security weaknesses of the Das scheme. Khan et al.proposed a user authentication scheme for rectifying the susceptibilities of the Das scheme and achieving a more secure user authentication in WSNs. Afterward, Chen et al.provided a secrecy-improved mutual user authentication scheme for WSNs by applying hash functions. Yeh et al.proposed a new mutual user authentication protocol by using elliptic curves cryptography (ECC) and smart cards for WSNs. Xue et al.showed that the Khan et al. scheme is vulnerable to stolen smart card and GWN bypassing attacks. In addition, the Chen et al. scheme is vulnerable to insider, masquerade, stolen smart card, and GWN bypassing attacks. By contrast, the Yeh et al. scheme is vulnerable to stolen smart card and replay attacks. Xue et al.proposed a temporal-credential-based mutual authentication scheme for users, GWNs, and sensor nodes. With the assistance of password-based authentication, the GWN in the Xue et al. scheme can issue a temporal credential to each user and sensor node. However, the Xue et al. scheme is vulnerable to insider attacks and stolen smart card attacks; the scheme does not offer password protection. In 2016, Chang et al.proposed a flexible authentication scheme for WSNs which operates in two modes. The first mode provides a lightweight authentication scheme, and the second mode is an advanced protocol based on ECC. In 2018, Amin et al.demonstrated that the Chang et al. scheme is insecure against stolen smart card attack and cannot provide password protection. Amin et al.then proposed a robust authentication scheme using smartcards for WSNs. However, the Amin et al. scheme has higher energy consumption, computational costs, and communication costs than those published previously (Section 5). In healthcare applications, Challa et al.proposed a secure user authentication scheme for wireless healthcare sensor networks. The three factor authentication scheme is designed with ECC. The proposed scheme has several functionality features including dynamic sensor node addition, password updates, biometrics updates, and smart card revocation for WSNs. On the basis of ECC, Li et al.also proposed an anonymous authentication scheme for WSNs in IoT environments. In the scheme, they used fuzzy commitment schemeto handle user biometric information. In 2019, Harbi et al.proposed an ECC-based mutual authentication scheme to secure communication in IoT-enabled WSNs. The sensor network in the system is arranged into clusters to diminish the energy consumption of sensors. Each cluster has a cluster head, which is a leader sensor node. However, Challa et al. scheme, Li et al. scheme, and Harbi et al. scheme are all based on an ECC for WSNs. The ECC approach is a public key cryptography approach based on elliptic curves. According to a related study, the time cost of an ECC point multiplication is much larger than that of hash function operations, and the energy consumption for executing an asymmetric ECC cryptosystem is much higher than that for executing a hash function. Currently, researchers are designing effective remote user authentication schemes for WSNs in IoT environments. In 2019, Ostad-Sharif et al.proposed an efficient user authentication scheme and claimed that their scheme is appropriate for WSNs in IoT environments. However, in this section, we argue that the login and authentication phase of the Ostad-Sharif et al. scheme has design faults. Moreover, their scheme cannot provide password change and update a password in its password change phase. Their scheme also has the time synchronization problem. The details are presented as follows.
## Authentication design faults of the ostad-sharif et al. scheme in iot environments
Design faults exist in the login and authentication phase of the Ostad-Sharif et al. scheme. We illustrate this security weakness in the subsequent passages. When a registered user U i wants to access the information of sensor node S j , the login and authentication phase of the Ostad-Sharif et al. scheme must be executed in advance. At first, a registered user U i inserts a smart card into the smart card reader and imprints his/her fingerprint B i on the sensor device. The smart card contains the secret parameters {D i , C i , E i , SCN i , BK()}, in which SCN i denotes unique smart card number and BK() denotes biometric key generation/extraction function. The smart card reader first extracts masked biometric C i from the smart card and computes RN 0 i = BK(h(B i ))�C 0 i . After finding C 0 i , the smart card reader must validate whether C 0 i and C i are equal. If C 0 i 6 ¼ C i , then the smart card reader terminates the request. However, in the equation above, the smart card reader does not know random number RN 0 i and masked biometric C 0 i . Therefore, it cannot obtain RN 0 i and C 0 i from the equation. Finally, a legitimately registered user U i cannot pass the verification to access the system. This problem will happen to all legitimately registered users. The Ostad-Sharif et al. scheme has design faults in the login and authentication phase.
## Failure to provide password change capability in the ostad-sharif et al. scheme
The Ostad-Sharif et al. schemecannot provide password change capability. We demonstrate this weakness in the following passages. When a registered user U i wants to update the password PW i , the password change phase in the scheme must be executed. U i first inserts a smart card into the smart card reader. He or she then inputs identity ID i and password PW i . The smart card contains the secret parameters {D i , C i , E i , SCN i , BK()}. After the legitimacy of U i is verified, U i enters a new password PW new i . The smart card computes the following equations: (4) and (5), we obtain the following results:
[formula] 1. RPW new i = h(ID i k PW new i kRN i ), in which RN i denotes random number. 2. A 0 i = D i �RPW i 3. D new i = A new i � RPW i 4. L 0 i = E i � RPW i 5. E new i = L 0 i � RPW iE new i ¼ L 0 i � RPW i ¼ E i � RPW i � RPW i ¼ E i Finally, [/formula]
## Time synchronization and authentication problem of the ostad-sharif et al. scheme in iot environments
The Ostad-Sharif et al. scheme uses a timestamp T i to verify mutual authentication among U i , the GWN, and S j for WSNs in IoT environments. Therefore, the Ostad-Sharif et al. scheme must provide synchronized time clocks to all devices in IoT environments for timestamp comparison. However, as mentioned, both adversaries and malicious nodes can attack time synchronization. Frequent topological changes, low-cost clocks, and limited energy of the sensor nodes in IoT environments can also affect time synchronization. The time synchronization of all WSN devices in IoT environments is a nontrivial challenge in itself. When a serious time synchronization problem arises in Ostad-Sharif et al. scheme, the GWN, U i , and S j cannot be synchronized with each other and then the legitimacy values of the GWN, U i , and S j cannot be verified. Hence, the Ostad-Sharif et al. scheme may enter a state such that mutual authentication among the GWN, U i , and S j cannot be achieved].
## Proposed scheme
In this section, we propose an efficient and secure authentication scheme for WSNs in IoT environments. The WSN environment contains three participants: the user (U i ), sensor node (S j ), and gateway node (GWN). The scheme applies dynamic ID to achieve security and user anonymity (identity protection). The scheme applies temporal credential to achieve mutual authentication and session key agreement. Temporal credentials are securely protected and stored in smart cards. The scheme can withstand stolen smart card attacks (Section 4.2). The system protects passwords against off-line password guessing attacks (Section 4.2). The system need not maintain any password or verification table; therefore it can resist the stolen verifier attacks and insider attacks. The scheme can withstand masquerade attacks, replay attacks, GWN bypassing attacks, and GWN spoofing attacks (Section 4.4 and 4.8). Before the registration, users are not obliged to share their IDs and passwords with the GWN; hence, the scheme provides a convenient functionality of adding new users (Section 4.6). To solve the password-changing problem in previous schemes, we also introduce a new password change phase to update the password. In the new password change phase, U i can freely select and update the password without requiring the communication with any other participants (the GWN and S j ), such that it can avoid additional communication message overhead. Hash function is operated in our scheme for providing security and computational efficiency.lists the definition of the notations in our scheme. The GWN chooses the private keys K GWN-U and K GWN-S , and only the GWN knows them. The proposed scheme consists of four phases: (1) registration phase, (2) login phase, (3) authentication and key agreement phase, and (4) password change phase. They are described as follows:
## Registration phase
The registration phase comprises two parts, one for users and the other for sensor nodes. We first describe the registration phase for users. In this phase, when a new user U i undertakes to register, he or she selects the identification ID i and password PW i . Subsequently, U i generates a random number r i and sends ID i and h(r i �PW i ) to the GWN for registration through a secure channel. After receiving the messages from U i , the GWN selects the expiration time TE i of the temporal credential of U i . The GWN computes the temporal credential TC i and verification information R i for U i . The GWN then issues a smart card with the temporal credential TC i , expiration time TE i , and verification information R i to U i through a secure channel. The steps are detailed as follows:
Step U1. U i freely chooses identification ID i and password PW i .
Step U2. U i generates a random number r i and calculates h(r i �PW i ).
Step U3.
[formula] U i ) GWN: {h(r i �PW i ), ID i }. U i transmits h(r i �PW i ) and ID i to the GWN through a secure channel. Step U4. GWN ) U i : {ID GWN , PTC i , TE i , B i , R i , h(.)}. [/formula]
After receiving the message from U i , the GWN selects the expiration time TE i of the temporal credential of U i and computes the following equations to issue the temporal credential TC i for U i .
[formula] P i = h(ID i kID GWN kTE i ), TC i = h(P i kK GWN-U kTE i ), PTC i = TC i �h(r i �PW i ), Q i = h(ID i kK GWN-U ), B i = Q i �h(ID i kh(r i �PW i )), and R i = h(Q i ). [/formula]
The GWN then issues a smart card with the secret parameters
[formula] {ID GWN , PTC i , TE i , B i , R i , h (.)} to U i through a secure channel. [/formula]
Step U5. U i stores r i in the smart card, after which the smart card holds the parameters
[formula] {ID GWN , PTC i , TE i , B i , R i , r i , h(.)}. [/formula]
We now describe the registration phase for sensor nodes. In this phase, each sensor node S j is pre-configured with SID j . After deployment, the sensor node S j generates a random number r j and then sends SID j and h(r j �SID j ) to the GWN for registration through a secure channel. After receiving the messages from S j , the GWN issues a temporal credential TC j to S j through a secure channel. The steps are detailed as follows:
Step S1. S j is pre-configured with SID j .
Step S2. S j generates a random number r j and computes h(r j �SID j ).
Step S3. S j ) GWN:
[formula] {SID j , h(r j �SID j )}. [/formula]
S j sends SID j and h(r j �SID j ) to the GWN through a secure channel.
Step S4. GWN ) S j : {RTC j }. After receiving the message from S j , the GWN computes TC j = h (K GWN-S kSID j ) to issue the temporal credential TC j for S j and then calculates RTC j = TC j �h (h(r j �SID j )kSID j ). The GWN sends RTC j to S j through a secure channel.
Step S5. After receiving the message from the GWN, S j computes TC j = RTC j �h(h(r j �SID j )k SID j ) to find its temporal credential TC j and then stores it.
[formula] ) Secure channel b h(- ) [/formula]
One-way hash function c a A common channel is a channel allocated in common to participants. b A secure channel is a channel of delivering messages that can withstand tampering and overhearing. c A hash function has a one-way property that it is computationally infeasible to find a data object to map to a hash result.
https://doi.org/10.1371/journal.pone.0232277.t001
## Login phase
U i first inserts a smart card into the smart card reader to log in to the system. U i then gives (ID i , PW i ) that correspond to the smart card. The smart card of U i computes verification information R � i and then verifies it with the stored R i in the smart card. After passing verification, the legitimacy of U i is ensured. Afterward, U i can read the information stored in the smart card and find its temporal credential TC i . The steps are detailed as follows:
Step L1. User U i inserts a smart card into the smart card reader and provides keys (ID i , PW i ).
The smart card of user U i then computes Q i = B i �h(ID i kh(r i �PW i )) and R � i = h(Q i ). The smart card validates whether R � i and the stored R i in the smart card are equal. If the values are unequal, the smart card rejects the login request. Otherwise, the legitimacy of U i is ensured, and U i can read the information stored in the smart card.
Step L2. U i computes TC i = PTC i �h(r i �PW i ) to find its temporal credential TC i .
## Plos one
Three-party mutual authenticated key agreement protocol for WSNs
## Authentication and key agreement phase
After ensuring the legitimacy of U i and finding the temporal credential TC i , the system must complete mutual authentication among U i , the GWN, and S j . The first step of the mutual authentication phase involves identity verification for U i , which is conducted by the GWN. Afterward, the second step entails identity verification of the GWN, which is conducted by S j . The third step involves identity verification for S j , which is conducted by U i as well as the GWN. Finally, a session key KEY ij is negotiated between U i and S j to conduct encryption during data transmission later on. The steps are detailed as follows:
[formula] Step V1. U i ! GWN: {DID i , q 1 , PKS i , TE i , P i , N i }. U i generates a nonce N i and computes P i = h (ID i kID GWN kTE i ), DID i = ID i �h(TC i kID GWN kN i ), and q 1 = h(ID i kTC i kN i ). [/formula]
Afterward, U i randomly chooses a secret sharing key K i and computes PKS i = K i �h(TC i kN i ). After computation, U i sends the login request message
[formula] m 1 = {DID i , q 1 ,PKS i ,TE i , P i , N i } to the GWN. [/formula]
Step V2. GWN! S j : h(ID i kTC i kN i ). The GWN then verifies whether q � 1 and q 1 are equal. If q � 1 6 ¼ q 1 , then the GWN terminates the request and sends a reject message to U i . Otherwise, the legitimacy of U i is ensured, and the GWN accepts the login request. The GWN then records the login status of U i to indicate that Ui is logging in to the system. The GWN computes K i = PKS i �h (TC i kN i ). At this point, the GWN selects a proper sensor node S j with identification SID j and calculates its temporal credential TC j = h(K GWN-S kSID j ). The GWN then generates a nonce N GWN and computes DID
[formula] {DID i , DID GWN , q 2 , PKS GWN , ID GWN , N i , N GWN }. After obtaining message m 1 , the GWN computes TC i = h(P i kK GWN-U kTE i ), ID i = DID i �h(TC i kID GWN kN i ), and q � 1 == ID i �h(TC j kDID i kN GWN ), q 2 = h(ID i kTC j kN GWN ), and PKS GWN = K i �h(TC j kN GWN ). After computation, the GWN sends the message m 2 = {DID i , DID GWN , q 2 , PKS GWN , ID GWN , N i , N GWN }to S j . [/formula]
Step V3.
[formula] S j !U i , GWN: {SID j , q 3 , PKS j , N i , N GWN }. After receiving message m 2 , S j assesses ID GWN to verify whether the GWN is a participant. If verification is true, S j computes ID i = DID GWN �h(TC j kDID i kN GWN ) and q � 2 = h(ID i kTC j kN GWN ) [/formula]
. S j then verifies whether q � 2 and q 2 are equal. If q � 2 6 ¼ q 2 , then S j terminates the request and returns a reject message. Otherwise, the legitimacy of the GWN is ensured, and S j accepts the request. S j computes K i = PKS GWN �h(TC j kN GWN ). Afterward, S j randomly selects a secret sharing key K j . S j computes q
[formula] = h(ID i kSID j kK i kN i kN GWN ) and PKS j = K j �h(K i kN i kN GWN ). After computation, S j sends the message m 3 = {SID j , q 3 , PKS j , N i , N GWN }to U i and the GWN. [/formula]
Step V4. After receiving the message m 3 , U i and the GWN separately compute q � 3 = h (ID i kSID j kK i kN i kN GWN ). After computation, the GWN verifies whether q � 3 and q 3 are equal. If q � 3 = q 3 , then the GWN can verify the legitimacy of S j . User U i also verifies whether q � 3 and q 3 are equal. If q � 3 = q 3 , then U i can verify the legitimacy of S j and the GWN. Afterward, U i and the GWN separately compute K j = PKS j �h(K i kN i kN GWN ). Finally, after ending the mutual authentication phase, U i , the GWN, and S j separately generate the shared session key KEY ij by computing
[formula] KEY ij = h(K i kK j kN i kN GWN kSID j ). [/formula]
## Password change phase
To update or change the password, a user U i must insert his/her smart card into the smart card reader. Afterward, U i gives ID i and PW i , which correspond to the smart card. In the first step of the password change phase, the smart card of U i computes verification information R � i and then verifies it with the stored R i in the smart card. After passing verification, the legitimacy of U i is ensured. U i can then read the information stored in the smart card.
[formula] ID i , PW i ). The smart card of U i calculates Q i = B i �h(ID i kh(r i �PW i )) and R � i = h(Q i ) [/formula]
and then verifies whether R � i and the stored R i in the smart card are equal. If the values are unequal, the smart card rejects the login request. Otherwise, the legitimacy of U i is ensured, and U i can read the information stored in the smart card.
Step P2. The user U i selects a new password PW new i , and then U i generates a random number r new i . Then, the smart card calculates B new
[formula] i = Q i �h(ID i kh (r new i �PW new i )), PTC new i = PTC i �h (r i �PW i )�h(r new i �PW new i [/formula]
# Security analysis
This section presents the security analysis of the proposed scheme and proves its security strength. Our scheme can overcome the weaknesses of previous schemes. Our proposed scheme has the following main security features.
## Mutual authentication and session key agreement
Mutual authentication is a critical feature for verifying mutual validity among the GWN, U i , and S j in WSNs. Because encryption and a message authentication code (MAC) are required to protect data transmission between U i and S j , a session key must be negotiated in advance between these two participants. In this section, we first illustrate the mutual authentication analysis of the proposed scheme, then we present the formal proofs. In the authentication and key agreement phase of the proposed scheme, mutual authentication between the GWN and S j is accomplished by calculating verification information q 2 and q 3 . In Step V3, S j can verify the legitimacy of the GWN after determining whether q 2 and q � 2 are equal, where q 2 = h (ID i kTC j kN GWN ). Temporal credential TC j is included in verification information q 2 . This shows that the sensor node S j can authenticate the validity of the GWN. In Step V4, the GWN can verify the legitimacy of S j after confirming whether q 3 and q � 3 are equal, where
[formula] q 3 = h (ID i kSID j kK i kN i kN GWN ). A secret sharing key K i is included in verification information q 3 . [/formula]
This shows that the GWN can authenticate S j . By contrast, mutual authentication between U i and the GWN is accomplished by calculating verification information q 1 and q 3 . In Step V2, the GWN can verify the legitimacy of U i after determining whether q � 1 and q 1 are equal, where q 1 = h(ID i kTC i kN i ). Temporal credential TC i is included in verification information q 1 . This shows that the GWN can authenticate the user U i . In Step V4, U i can verify the legitimacy of S j after confirming whether q 3 and q � 3 are equal, where q 3 = h(ID i kSID j kK i kN i kN GWN ). A secret sharing key K i is included in verification information q 3 . This shows that the user U i can authenticate the sensor node S j . In addition, because S j has authenticated the validity of the GWN, the user U i further authenticates the validity of the GWN as well. Therefore, on the basis of temporal credential signing and the secret sharing key, U i , S j , and the GWN can mutually authenticate each other in the proposed protocol. In Step V4, after completing the mutual authentication phase, U i , the GWN, and S j can separately generate the shared session key KEY ij by computing KEY ij = h(K i kK j kN i kN GWN kSID j ), where secret sharing key K i and K j are selected randomly. This shows that U i , S j , and the GWN can share a common session key after finishing the mutual authentication phase. The common session key is validated by U i , the GWN, and S j . This illustration indicates that our scheme provides session key agreement and mutual authentication. The formal proofs are given in the following lemmas and Proposition 1. We use the BAN logic methodto formally validate the mutual authentication and session key agreement of our scheme. The BAN logic method is widely used to validate authentication and key establishment protocols. The BAN logic method accomplishes to introduce the logic of authentication and explain the protocols step-by-step. The notations of BAN logic are presented in. In, the symbols X and Y range over statements; Q and P are principals.
The essential logical postulates for the BAN logic are listed as follows]:
1. Freshness-propagation rule: Pj�ðXÞ Pj�ðX;YÞ . That is, if P is entitled to believe that one part of a formula (X,Y) is fresh, then he also is entitled to believe that the entire formula (X,Y) must also be fresh.
## 2.
Receiving rule: P⊲ðX;YÞ P⊲X and P⊲hXi Y P⊲X . That is, if a principal P can receive and read a formula (X,Y) or formula hXi Y , then he also can receive and read its components X.
## Nonce-verification rule: pj�ðxþ; pj�qj�x
Pj�Qj�X . That is, if P is entitled to believes that X is a fresh statement and that Q once said X, then P believes that Q believes X.
## Jurisdiction rule: pj�qj)x;pj�qj�x
Pj�X . That is, if P believes that Q has jurisdiction over X and P believes that Q believes X, then P believes X.
[formula] 5. Message-meaning rule: Pj � Q !Y P;P⊲hXi Y Pj�Qj�X [/formula]
. That is, if P is entitled to believe that the key Y is shared with Q, and P sees X encrypted under Y, then P is entitled to believe that Q once said X.
## Session-key rule: pj�ðkþ; pj�qj�x
Pj�P $ K Q , where statement X is an element of the combination session key K. That is, if P is entitled to believe that K is a fresh statement and that Q believes X, then P believes that P and Q share a common key K.
To validate the proposed protocol, we first summarize our scheme in the generic form:
[formula] Message m 1 . U i ! GWN: {DID i , q 1 , PKS i , TE i , P i , N i } = {ID i �h(TC i kID GWN kN i ), h(ID i kTC i kN i ), K i �h(TC i kN i ), TE i , [/formula]
## Notation definition
P⊲X P sees X : P can receive and read X (possibly after doing some decryption).
P|~X P said X : P once said X. P once sent a message including the statement X.
P|)X P controls X : P has jurisdiction over X.
P|�X P believes X : P is entitled to believe X.
#(X) fresh(X) : X is regarded as a fresh statement.
hXi Y X is combined with Y; Y is a secret.
(X,Y) X and Y are said simultaneously.
P$ K Q P and Q share a common key K.
[formula] P Y ! Q Statement Y ish(ID i kID GWN kTE i ), N i }. Message m 2 . GWN! S j : {DID i , DID GWN , q 2 , PKS GWN , ID GWN , N i , N GWN } = {ID i �h(TC i kID GWN kN i ), ID i �h(TC j kDID i kN GWN ), h(ID i kTC j kN GWN ), K i �h(TC j kN GWN ), ID GWN , N i , N GWN }. Message m 3 . S j !GWN: {SID j , q 3 , PKS j , N i , N GWN } = {SID j , h(ID i kSID j kK i kN i kN GWN ), K j �h(K i kN i kN GWN ),N i , N GWN }. Message m 3 . S j !U i : {SID j , q 3 , PKS j , N i , N GWN } = {SID j , h(ID i kSID j kK i kN i kN GWN ), K j �h(K i kN i kN GWN ), N i , N GWN }. [/formula]
Subsequently, we transform the generic form into the idealized form:
[formula] I 1 . U i ! GWN: hN i i TC i , hN i i TC i , hhN i i TC i i K i I 2 . GWN! S j : hN i i TC i , hN GWN i TC j , hN GWN i TC j , hhN GWN i TC j i K i I 3 . S j ! GWN: hN i ; N GWN i K i , hhN i ; N GWN i K i i K j I 4 . S j !U i :hN i ; N GWN i K i , hhN i ; N GWN i K i i K j [/formula]
To analyze our scheme, we use the following assumptions: , the S j returns (N i , N GWN ) to the GWN and U i .
[formula] A 1 . GWNj � U i ! TC i GWN A 2 . GWNj � S j ! K i GWN A 3 . S j j � GWN ! TC j S j A 4 . U i j � S j ! K i U i A 5 . GWN|�#(N i ) A 6 . S j |�#(N GWN ) A 7 . U i |�#(N i ,N GWN ) A 8 . GWN|�#(N i ,N GWN ) A 9 . GWN|�U i |)N i A 10 . S j |�GWN|)N GWN A 11 .GWN|�S j |)(N i ,N GWN ) A 12 . U i |�S j |)(N i ,N GWN )Lemma [/formula]
To prove that the GWN can authenticate S j , the following belief must be demonstrated:
[formula] B 3 . GWN|�(N i ,N GWN ) [/formula]
To prove that the U i can authenticate S j , the following belief must be demonstrated:
[formula] B 4 . U i |�(N i ,N GWN ) [/formula]
The steps of the proof for B 3 :
[formula] S 9 . GWN sees hN i ; N GWN i K i ( [/formula]
## Lemma 3. in our scheme, the gwn, u i , and s j can coordinate the common session key key ij .
Proof: To prove that U i , the GWN, and S j in our scheme can share a session key KEY ij = h (K i kK j kN i kN GWN kSID j ), the following beliefs must be demonstrated: Consequently, the GWN believes that GWN shares the session key KEY ij with S j . Proposition 1. U i , the GWN, and S j in our scheme can mutually authenticate each other; they can share a common session key.
[formula] B 5 . U i j � U i $ KEY ij GWN B 6 . GWNj � GWN$ KEY ij U i B 7 . S j j � S j $ KEY ij GWN B 8 . GWNj � GWN$ KEY [/formula]
Proof: From Lemma 2, U i in our scheme can authenticate S j . In addition, S j can authenticate the GWN (Lemma 1). Thus, U i can further authenticate the GWN as well. Conversely, the GWN can authenticate U i (Lemma 1). Consequently, the GWN and U i in our scheme can mutually authenticate each other. The GWN can authenticate S j (Lemma 2). Conversely, S j can authenticate the GWN (Lemma 1). Consequently, the GWN and S j in our scheme can mutually authenticate each other. Mutual authentication can be provided in our scheme. After finishing the mutual authentication, U i , the GWN, and S j can share a session key KEY ij = h (K i kK j kN i kN GWN kSID j ) (Lemma 3). Session key agreement can also be provided in our scheme.
## Password protection, guessing attack resistance, and stolen smart card attack resistance
When a user's smart card is stolen or lost in a stolen smart card attack, an adversary can acquire information from the smart card. Then, the adversary masquerades as an authorized user to access to the GWN. However, password protection functionality can prevent the leakage of password information, such that the adversary cannot obtain useful information to perform an off-line password guessing attack.
## Proposition 2. the proposed scheme can provide password protection, guessing attack resistance, and stolen smart card attack resistance.
Proof: In our scheme, the password presents with the h(r i �PW i ) form, in which PW i and r i are hidden. h(r i �PW i ) is not stored in the smart card, the GWN, or any other device. Thus, the adversary cannot directly obtain PW i by performing an off-line password guessing attack on h(r i �PW i ). Therefore, the proposed scheme can provide password protection and guessing attack resistance. Moreover, smart card secrets can be breached by monitoring power consumption or by analyzing leaked information. When the adversary has a smart card that has been lost by its legitimate owner, the adversary can acquire the secret parameters from that smart card by applying the previously discussed method. We can prove that the proposed scheme can also provide stolen smart card attack resistance. That is, in the proposed scheme, the adversary cannot masquerade as a legitimate user to log in to the GWN when the adversary has obtained a legitimate user's smart card. Suppose that when the smart card of user U i is stolen or lost, the adversary obtains that the smart card. The adversary can obtain the secret parameters {ID GWN , PTC i , TE i , B i , R i , r i , h(.)} from the smart card. To impersonate a legitimate user, the adversary must produce a new N } i , randomly choose an imitative secret sharing key K } i , and create an imitative login request message
[formula] {DID } i , q } 1 , PKS } i , TE i , P i , N } i } for the GWN. The imitative parameters {DID } i , q } 1 , PKS } i , P i } [/formula]
are obtained using the following equations:
[formula] DID } i = ID i �h(TC i kID GWN kN } i ), q } 1 = h(ID i kTC i kN } i ), PKS } i = K } i �h(TC i kN } i ), P i = h(ID i kID GWN kTE i ). [/formula]
Therefore, to obtain the imitative parameters {DID } i , q } 1 , PKS } i , P i }, the adversary must first obtain TC i and ID i by using the following equations:
[formula] TC i = h(P i kK GWN-U kTE i ), TC i = PTC i �h(r i �PW i ), ID i = DID i �h(TC i kID GWN kN i ). [/formula]
Nevertheless, the adversary cannot acquire TC i and ID i because he/she does not possess K GWN-U and PW i . Only the GWN knows the private key K GWN-U in our scheme. As previously discussed, the proposed scheme can provide password protection, and that the adversary cannot acquire PW i by executing an off-line password guessing attack. Therefore, the imitative parameter set {DID } i , q } 1 , PKS } i , P i }of a login request message is not acquired. The adversary cannot masquerade as an authorized user by only using a smart card.
## Two-factor security
By involving a smart card and a password in the login phase, two-factor security in our scheme can be achieved. Proposition 3. Two-factor security can be provided in our scheme. Proof: First, assume that the adversary only has the smart card of U i . Let us even assume that the adversary can intercept login request message m 1 = {DID i , q 1 , PKS i , TE i , P i , N i }. As mentioned in Proposition 2, the adversary can obtain the secret parameters {ID GWN , PTC i , TE i , B i , R i , r i , h(.)} from the smart card. To impersonate a legitimate user, the adversary must produce a new N } i , randomly choose a new sharing key K } i , and create an imitative
[formula] login request message {DID } i , q } 1 , PKS } i , TE i , P i , N } i } for the GWN, where DID } i = ID i �h (TC i kID GWN kN } i ), q } 1 = h(ID i kTC i kN } i ), and PKS } i = K } i �h(TC i kN } i ) [/formula]
. Consequently, to gain the parameter set {DID } i , q } 1 , PKS } i }, the adversary must acquire TC i and ID i by applying the following equations: TC i = h(P i kK GWN-U kTE i ), TC i = PTC i �h(r i �PW i ), and ID i = DID i �h (TC i kID GWN kN i ). Nevertheless, the adversary cannot acquire TC i and ID i because he/she does not possess K GWN-U and PW i . Only the GWN knows the private key K GWN-U in our scheme, and we have proven that the proposed scheme can provide password protection to prevent the leakage of PW i information (Section 4.2). Therefore, the parameter set {DID } i , q } 1 , PKS } i } of the login request message is not acquired, and the adversary cannot disguise as an authorized user by only using the smart card. Secondly, assume that the adversary only has the password PW i and identification ID i of U i . Under this condition, the adversary also cannot acquire TC i to calculate the parameters {DID } i , q } 1 , PKS } i } because he/she does not know K GWN-U and PTC i (which are not stored in the smart card). Therefore, the adversary cannot impersonate an authorized user when he/she either acquires information from the smart card or knows {ID i , PW i }. Our scheme can withstand this type of masquerade attack and provide two-factor security.
## Masquerade attack resistance and replay attack resistance
Protection against masquerade attacks is a principal security feature for any remote user authentication scheme. Replay attack resistance means that the adversary cannot attempt to replay any previously intercepted message to spoof the GWN. Proposition 4. Our scheme can provide masquerade attack resistance and replay attack resistance.
Proof: Proposition 3 has demonstrated that our scheme can protect against masquerade attacks caused by either the loss of a smart card or the revelation of sensitive identification and password details {ID i , PW i }. The reliability of our scheme against other masquerade attacks must be demonstrated. We can even assume that the adversary is a legitimate user L and undertakes to impersonate a user U i . Adversary L may intercept the login request message m 1
[formula] = {DID i , q 1 , PKS i , TE i , P i , N i }.} i , q } 1 , PKS } i , TE i , P i , N } i } for the GWN, where DID } i = ID i �h (TC i kID GWN kN } i ), q } 1 = h(ID i kTC i kN } i ), and PKS } i = K } i �h(TC i kN } i ). [/formula]
Nevertheless, adversary L still cannot acquire TC i and ID i to calculate the parameters {DID } i , q } 1 , PKS } i } because he/she does not possess K GWN-U and PW i (Proposition 2). In addition, adversary L cannot compute the shared session key KEY ij = h(K i kK j kN i kN GWN kSID j ) because he or she does not know K i and K j in KEY ij . Thus, adversary L cannot impersonate any other legitimate user. Consequently, our scheme can protect against masquerade attacks when an adversary impersonates any other legitimate user. Adversary L can undertake to replay the intercepted message {DID i , q 1 , PKS i , TE i , P i , N i } to the GWN. However, after receiving message m 3 = {SID j , q 3 , PKS j , N i , N GWN }, adversary L cannot compute the shared session key KEY ij = h(K i kK j kN i kN GWN kSID j ) because he or she cannot obtain K i and K j in KEY ij . Consequently, resistance to replay attacks is guaranteed as well. Next, we prove that an adversary cannot masquerade as a sensor node to spoof the user. Suppose adversary L has intercepted message m 2 when the GWN attempts to send it to S j ; that is, the message {DID i , DID GWN , q 2 , PKS GWN , ID GWN , N i , N GWN }. To masquerade as a sensor node to spoof the user, the adversary must randomly choose an imitative secret sharing key K } j and send an imitative response message {SID j , q 3 , . To obtain the parameters {q 3 , PKS } j }, the adversary must first know K i . Moreover, K i can be obtained by using the equation K i = PKS GWN �h(TC j kN GWN ). Nevertheless, the adversary cannot acquire K i because he/she does not possess the temporal credential TC j . Therefore, the parameters {q 3 , PKS } j } cannot be acquired, and the adversary cannot send an imitative response message {SID j , q 3 , PKS } j , N i , N GWN } to the GWN. Consequently, our scheme can protect against masquerade attacks when an adversary masquerades as a sensor node to spoof the user.
[formula] PKS } j , N i , N GWN } to the GWN, where q 3 = h(ID i kSID j kK i kN i kN GWN ) and PKS } j = K } j �h(K i [/formula]
## Stolen verifier attack resistance and insider attack resistance
The stolen verifier attack means that the adversary steals the verification table from the GWN or S j . By contrast, an insider attack involves any privileged insider of the GWN purposely obtaining a user password, which leads to security defects in the remote user authentication scheme.
Proposition 5. Our scheme can protect against stolen verifier attacks and insider attacks. Proof: The GWN and S j in our scheme do not retain any verification table for verifying the legitimacy of registered users or sensor nodes. Therefore, the adversary cannot find any verifiable information in the GWN or S j to impersonate a legitimate user. Consequently, our scheme can protect against stolen verifier attacks. Moreover, because U i presents h(r i �PW i ) to register with the GWN. r i and PW i are hidden from the GWN. In addition, the GWN does not store any verifier h(r i �PW i ). The privileged insider of the GWN cannot acquire PW i by executing any off-line password guessing attack. Consequently, our scheme can resist insider attacks.
## Password updating, adding new user functionality, and time synchronization avoidance
In our scheme, users are not obliged to share their IDs and passwords with the GWN before the registration. During the registration process, a new user U i can freely choose some identification string ID i and password PW i as favorite strings without requiring assistance from the GWN. Any new legitimate user can be freely added to the system after the registration. Therefore, the proposed scheme provides a convenient functionality for adding new users. Moreover, as mentioned, it is strongly recommended that for security policy, users update or change their passwords frequently to protect against compromise. In the password change phase of our scheme, a legitimate user U i can freely choose his/her new password to update or change the password without requiring extra communication message overhead to exchange messages with the GWN. Consequently, our scheme provides the functionalities of freely chosen passwords and efficient password updating. Finally, our scheme does not require any timestamp to verify mutual authentication among U i , the GWN, and S j because our scheme is a nonce-based scheme. Consequently, our scheme is not obliged to provide synchronized time clocks for all devices, and it can avoid the time-synchronization problem for WSNs in IoT environments.
## User anonymity (identity protection)
The user anonymity (identity protection) means that the identity of any user is disclosed only to service providers. Proposition 6. Our scheme can provide user anonymity to protect user identity. Proof: The adversary can intercept message m 1 = {DID i , q 1 , PKS i , TE i , P i , N i } to acquire the identification string of U i ,. The parameters DID i , q 1 , PKS i , and P i are obtained using the following equations:
[formula] DID i = ID i �h(TC i kID GWN kN i ), q 1 = h(ID i kTC i kN i ), PKS i = K i �h(TC i kN i ), P i = h(ID i kID GWN kTE i ). [/formula]
However, in Proposition 2, we show that the adversary cannot obtain ID i and TC i because he or she does not know K GWN-U and PW i . The identification string ID i also cannot be derived from the equations above. Therefore, an adversary cannot acquire ID i to identify the user U i , and our scheme can provide user anonymity to protect user identity.
## Gwn bypassing attack resistance and gwn spoofing attack resistance
A GWN bypassing attack occurs when an adversary can bypass the GWN to forge a verification message straight to the sensor node S j without passing the GWN login. By contrast, a GWN spoofing attack occurs when an adversary may impersonate the GWN to obtain private login information of U i . Proposition 7. Our scheme can protect against GWN bypassing attacks and GWN spoofing attacks.
Proof: To bypass the GWN, an adversary must send an imitative verification message m 2 = {DID i , DID GWN , q 2 , PKS GWN , ID GWN , N i , N GWN } straight to S j , where q 2 = h(ID i kTC j kN GWN ). However, the adversary cannot obtain q 2 to create an imitative message m 2 because he or she does not know the temporal credential TC j ; thus, the adversary cannot bypass the GWN to forge m 2 to S j . Without m 2 , S j cannot respond with any other messages. Consequently, our scheme can prevent GWN bypassing attacks. By contrast, the adversary may attempt to impersonate the GWN to acquire the secret login information of U i . To pose as the GWN, the adversary can intercept some login request message m 1 = {DID i , q 1 , PKS i , TE i , P i , N i } and respond with an imitative message m . Verification information q 3 includes a secret sharing key K i . However, as mentioned in Proposition 4, the adversary cannot acquire K i because he/she does not know temporal credential TC j . Therefore, the adversary cannot obtain q 3 ; thus, the adversary cannot send an imitative message m 3 = {SID j , q 3 , PKS j , N i , N GWN } to respond to U i . The adversary cannot convince U i that he/she is a legitimate GWN. Consequently, our scheme can protect against GWN spoofing attacks.
[formula] = {SID j , q 3 , PKS j , N i , N GWN } to U i , where q 3 = h (ID i [/formula]
## Performance evaluation and functionality comparison
Performance and functionality evaluations are critical to establish validity for practical deployment. In this section, the performance and functionality of our scheme are evaluated. The performance efficiency and functional effectiveness of our authentication scheme are demonstrated.presents a functionality comparison of our scheme versus previous related schemes. In, Yes denotes the scheme has a security feature; No denotes the contrary. The weaknesses of the previous related schemes for WSNs are mentioned in Section 2 and summarized in. We present a practical scenario to show that the proposed scheme can provide secure functionality and effectiveness for WSNs in IoT environments. Suppose that an adversary, Eve, undertakes to damage our scheme by executing the following attacks: guessing attack, stolen smart card attack, masquerade attack, replay attack, stolen verifier attack, insider attack, user anonymity attack, or GWN bypassing attack. Section 4 has shown that our scheme has the following abilities. Eve cannot directly obtain a user's password by executing a password guessing attack. When Eve steals a user's smart card, she cannot impersonate an authorized user to access the system. When Eve is even a legitimate user who pretends to be a different legitimate user, our scheme can protect against this masquerade attack. Moreover, Eve may undertake to replay some intercepted message to the GWN. Our scheme can provide resistance to replay attacks. Eve cannot breach the system by stealing the verification table.
## Functionality comparison
Even as an insider, Eve cannot acquire a password by executing any password guessing attack.
Eve may intercept a login request message from the user to acquire the identification information, but the identification information of a user cannot be derived. Finally, Eve cannot forge an imitative message and send it straight to the sensor node to bypass the GWN. Moreover, our scheme has other security functionalities, which include updating passwords, choosing passwords freely, adding new users, and time synchronization avoidance. Our scheme provides a secure common session key and mutual authentication. Our scheme can thus protect against all listed attacks from Eve.
## Performance evaluation
The proposed scheme comprises four phases: registration phase, login phase, authentication and key agreement phase, and password change phase. In a WSN environment, the performance of the authentication scheme is affected mainly by the authentication and key agreement phase. This phase is the main part of the authentication scheme and is what chiefly distinguishes it from the various authentication schemes in WSNs. Therefore, we focus our discussion on the performance comparison of the authentication and key agreement phase in the authentication schemes. The performance comparison is usually separated into communication costs and computational costs. The computational costs are defined as the time spent by the user and service provider in the process. By contrast, the communication costs are defined as the number of messages dispatched by the user and service provider in the process. The performance comparison of our scheme and previous related schemes is shown in . presents the computational 【Computational cost】 authentication phase costs and communication costs of the authentication and key agreement phase in each authentication scheme run without the consideration of interference and packet loss. The notation T h is defined as the time complexity of the hash function; T ecc is the time complexity of the encryption/decryption operation in elliptic curve cryptography (ECC) algorithm. The computational costs of the exclusive-or operation are usually neglected because it necessitates minimal computations. We first analyze the computational costs of the authentication and key agreement phase for each scheme as follows:
[formula] User 4T h 10T h 13T h 3T h 5T h 2T ecc +1T h 3T h 4T h 3T h GWN 8T h 14T h 14T h 5T h 11T h 4T ecc +3T h 5T h 5T h 4T h Sensor node 3T h 3T h 2T h 1T h 3T h 2T ecc +2T h 2T h 2T h 1T h key agreement phase User 3T h 2T h 1T h 3T h 3T h 1T h − � − � − � GWN 3T h 3T h 3T h 3T h 3T h 1T h − � − � − � Sensor node 3T h 2T h 2T h 4T h 3T h 1T h − � − � − � Total 24T h 34T h 35T h 19T h 28T h 8T ecc +9T h [/formula]
1. In the authentication phase of the Ostad-Sharif et al. scheme, the user requires 10T h to compute the parameters of the login request message and the response message. The GWN must spend 14T h to compute the parameters in a response message for the user and a request message for the sensor node. The sensor node must expend 3T h to confirm whether the verification equations hold. In addition, the user, GWN, and sensor node must expend 2T h , 3T h , and 2T h separately to negotiate the shared session key in the key agreement phase. Accordingly, the total computational costs for the user, GWN, and sensor node are 12T h , 17T h , and 5T h , respectively.
2. In the authentication phase of the Amin et al. scheme, the user requires 13T h to compute the parameters of the login request message and the response message. The GWN must spend 14T h to compute the parameters in a request message for the sensor node and a response message for the user. The sensor node must expend 2T h to confirm whether the verification equations hold.
In addition, the user, GWN, and sensor node must expend 1T h , 3T h , and 2T h separately to negotiate the shared session key in the key agreement phase. Accordingly, the total computational costs for the user, GWN, and sensor node are 14T h , 17T h , and 4T h , respectively.
3. In the authentication phase of the Chang et al. scheme, the user requires 3T h to compute the parameters of the login request message. The sensor node must expend 1T h to compute the parameters in a message for the GWN. The GWN must spend 5T h to verify the login request. In addition, the user, GWN, and sensor node must expend 3T h , 3T h , and 4T h separately to negotiate the shared session key in the key agreement phase. Accordingly, the total computational costs for the user, GWN, and sensor node are 6T h , 8T h , and 5T h , respectively.
4. In the authentication phase of the Xue et al. scheme, the user requires 5T h to compute the parameters of the login request message. The GWN must spend 11T h to verify the login request message and compute the parameters of the request message for the sensor node. The sensor node must expend 3T h to confirm whether the verification equations hold. Moreover, the user, GWN, and sensor node must expend 3T h , 3T h , and 3T h separately to negotiate the shared session key in the key agreement phase. Accordingly, the total computational costs for the user, GWN, and sensor node are 8T h , 14T h , and 6T h , respectively.
5. In the Khan et al. scheme, the user must expend 3T h to generate a login request message. The GWN must expend 5T h to confirm whether the verification equations hold and to calculate the parameters of the request message for the sensor node. The sensor node requires 2T h to confirm whether the verification equations hold and to generate a response message for the GWN. However, the Khan et al. scheme does not provide the key agreement phase for the session key agreement.
6. In the Chen et al. scheme, the user must expend 4T h to produce a login request message and to validate a response message. The GWN requires 5T h to validate a login request message and to respond to a user's request. The sensor node must expend 2T h to verify the request message from the GWN and to generate a response message for the user. However, the Chen et al. scheme also does not provide any key agreement phase.
7. In the Das scheme, the user requires 3T h to generate the login request message. The GWN must expend 4T h to confirm whether the verification equations hold and to calculate the parameters of the request message for the sensor node. The sensor node requires 1T h to confirm whether the verification equations hold and to generate a response message for the user. The Das schemedoes not provide the key agreement phase as well.
8. The Yeh et al. schemeuses elliptic curve cryptography (ECC) to provide both the authentication phase and session key agreement phase. That scheme requires that the user, GWN, and sensor node expend 2T ecc + 1T h , 4T ecc + 3T h , and 2T ecc + 2T h separately to complete the authentication phase. Moreover, the user, GWN, and sensor node must expend 1T h , 1T h , and 1T h separately to compute a shared session key in the key agreement phase. Accordingly, the total computational costs of the user, GWN, and sensor node are 2T ecc + 2T h , 4T ecc + 4T h , and 2T ecc + 3T h , respectively.
9. Our proposed scheme provides both the authentication phase and key agreement phase. In the authentication phase of our scheme, the user requires only 4T h to calculate the parameters of a login request message. The GWN expends only 8T h to verify the login request and to calculate the parameters of the request message for the sensor node. The sensor node requires only 3T h to confirm whether the verification equations hold. In the key agreement phase, the user, GWN, and sensor node expend only 3T h , 3T h , and 3T h , respectively, to negotiate the shared session key. Accordingly, the total computational costs for the user, GWN, and sensor node are 7T h , 11T h , and 6T h , respectively.
Our proposed scheme uses only the hash function and XOR operations to design a simple authentication and key agreement scheme. However, the Yeh et al. schemeprovides a authentication and key agreement scheme which is established by an asymmetric encryption algorithm (specifically, an ECC). According to an experimental finding obtained in a related study, the one-way hash function is computationally efficient. The time complexity of the hash function is less than that of an asymmetric ECC encryption operation. The following is a practical example for the computational costs: In an environment with a CPU of 3.2 GHz and with 3.0 GB of RAM, completing a one-way hash operation requires 0.02 ms on average when using SHA-1, and completing an asymmetric ECC encryption operation requires 0.45 ms on average when using ECC-160.
For the user in each scheme run, the . By contrast, our scheme can perform the run in 0.12 ms for 6T h . Therefore, the computational load of the sensor node in the proposed scheme is reduced to 12.5% compared with the Yeh et al. scheme.
In .
The energy consumption of the Yeh et al. schemeis ascribed chiefly to the asymmetric ECC cryptosystem and hash functions. By contrast, the energy consumption of our scheme is principally attributed to the hash functions. As mentioned, the energy consumption for executing the hash function is much lower than that for executing an asymmetric ECC cryptosystem. A practical example follows: While using SHA-1 to compute the hash value, a 1-byte data packet requires 0.76 μJ of energy. Nevertheless, a 163-bit ECC asymmetric cryptosystem requires 134.2 mJ of energy. As previously discussed, the total computational costs of the schemes of . Because the total energy consumption of the Yeh et al. scheme is excessive relative to other schemes, it cannot be shown in Although the total energy consumption of our scheme (18.2 μJ) is slightly greater than that of the Chang et al. scheme , our scheme provides superior security functionality to overcome the weaknesses of previous schemes.
As mentioned, the communication cost accounts for the number of messages transmitted. A low number of transmitted messages results in less consumption for the message overhead In this subsection, we demonstrate that our scheme is highly efficient because of the superior performance: low computational cost (0.14 ms for the user, 0.12 ms for the sensor node, and 0.22 ms for the GWN), low energy consumption (18.2 μJ for the authentication and key agreement phase), and low communication cost (4 transmitted messages for the authentication and key agreement phase, 0 transmitted messages for the password change phase).
# Conclusions
This paper analyzes the security weaknesses of related authentication schemes and proposes a more efficient and secure authentication scheme for WSNs in IoT environments. The BAN logic method is used to prove our scheme. Finally, we compare the functional effectiveness and performance efficiency of our scheme with those of previously published schemes. Cryptanalysis revealed that our scheme overcomes the security weaknesses of the previously published schemes. Our scheme satisfies the requirement of basic design criteria for the authentication scheme as well. Consequently, our scheme can enhance security effectiveness in real-world IoT environments and provide additional security functionalities compared with the other discussed schemes. Moreover, performance analysis revealed that our scheme demonstrates high efficiency and superior performance.
Our future work and challenges include attempting to find security risks in heterogeneous IoT environments. Various heterogeneous IoT applications can cause serious challenges in securing networks. Future studies will further evaluate the reliability and scalability of the proposed scheme in heterogeneous IoT environments. Moreover, we also study highly secure machine learningbased authentication schemes for WSNs in intelligent IoT environments. The integration of Big Data with intelligent IoT networks will be challenging due to the limited resources of WSNs. |
Oculomotor inhibition precedes temporally expected auditory targets
## Reviewer #1:
Consistent with previous work, the authors demonstrate that saccade execution is modulated by the predictability of a target. In a novel expansion of prior studies, this manuscript presents a case in which these relationships exist in response to auditory (rather than visual) stimuli. Thus, saccades are inhibited before auditory targets are expected. This is high-quality research that provides further evidence that the oculomotor system is recruited in circumstances where visual information is irrelevant. The work has been conducted to a high methodological standard, the findings appear sound, and the conclusions are appropriate. The article is well-written, clear, professional, and detailed.
-Thanks for your positive evaluation and constructive comments; addressing them has strengthened our manuscript.
Minor comments: 1. One small point about the language used: Describing the inhibition as a "supramodal inhibitory mechanism" that "goes beyond vision" seems problematic, because the evidence points to a mechanism inhibiting the oculomotor system and not the auditory system. The inhibition itself is not supramodal here, though it is in response to a supramodal anticipatory process. In many places (for example, lines 406 and after), the authors describe the findings as oculomotor inhibition and a supramodal temporal expectation, which fits my understanding of their discoveries. I think that this is an interesting study addressing a potentially important issue about temporal expectation. The paper very well written, the experiment is conducted with care and the data analysis is appropriate. The theoretical impact of the paper is less clear-cut as there is no obvious theoretical reason to expect oculomotor inhibition prior to auditory targets.
-Thanks for your positive evaluation. Clarifying the issues you pointed out below has strengthened our manuscript. Moreover, we have clarified that our finding has theoretical relevance, notwithstanding the fact that there were no obvious reasons either to expect it or to not expect it.
Although there is much to like about this paper, there is one fundamental issue that needs to be resolved. As the main message the paper claims that "These findings document a novel link between oculomotor inhibition and temporal processing and thus reveal a global supramodal mechanism of temporal expectations" (abstract). It is argued that "temporal expectations are formed by the cognitive system" (p.2). The question is whether this claim really holds given the experimental paradigm. My main concern is the use of the notion of "temporal expectation". The paradigm generates a particular rhythm of tones that are precisely presented at fixed intervals of one another. What happens is that the human body reacts to this rhythmic sequence generating all kind of physiological reactions that may coincide with these events. This may be heart rate variability, blood volume pulse, GSR and maybe, as in the current case, eye movements. This is interesting but do we believe that these automatic physiological reactions of the body represents anything about "expectation". I do not think so and I do not think that such physiological reactions are that interesting. Now, there are ways to really show that it is not some kind of low level autonomic reaction of the nervous system but instead represents true "expectation". The experiment that needs to be done is the following: Instead of having one constant foreperiod during a whole block there should two foreperiods of different length; each foreperiod is announced by a 100% valid cue (for example a high tone for foreperiod 1 sec and a low tone for foreperiod 2.5 sec). If it is really temporal expectation then we expect to see basically the same results as what was shown here. The cue generates an expectation for a particular interval and we should see this back in the eye movement data. If it is just a low level physiological reaction we expect that the results will not hold. This experiment seems to be crucial and without this experiment the effect reported here may be a trivial physiological reaction of the human body to a rhythmic tone.
-Please note that the procedure we used in this study does not produce a rhythm of tones. Our design is a non-rhythmic design based on the presentation of a cue followed by a predictable or non-predictable target with a variable ISI between trials. We further clarify this issue in the Introduction [lines 71-72]. -We have clarified the logic of this study and stressed the novelty of the findings to highlight its theoretical relevance [lines 70-72; 409-412].
## Reviewer #3:
Frequency of saccades and blinks reduce prior to expected onset of visual targets, a process assumed to be used by the brain to improve the detection/processing of newly presented visual information. Abeles and colleagues use a psychophysical paradigm to show that this oculomotor inhibition does also occur prior to the onset of auditory stimuli. This suggests that rather than a modality-specific system, the oculomotor inhibition relies on a multi-modal module.
-Thanks for your comments. Clarifying the following points in response to your comments has strengthened our manuscript.
Major comments: 1) It is not clear to me how this new finding is changing our perspective regarding the purpose of oculomotor inhibition. As pointed out by the authors the exact mechanism by which the "expectation" alters performance is not clear but I am not aware of any study showing that expectations stay within their specific modality and are not transferred across modalities. In other words, at least based on the literature cited in the paper and my knowledge, the results were not against any dogma and were expected.
-This is the first study to investigate this issue. Thus, the results cannot be in agreement or against any dogma; however, they were not expected (as we indicated and Reviewer 2 points out).
We have clarified the logic of this study and stressed the novelty of the findings to highlight its theoretical relevance [lines 54-55; 61-62; 66-72; 409-412].
2) The authors mention that "Finding no oculomotor inhibition prior to predictable auditory targets would support the hypothesis that the functional role of the oculomotor inhibition is merely to facilitate vision. Alternatively, finding an oculomotor inhibition effect for auditory targets would imply that its functional role goes beyond vision." If "it" in the last sentence refers to "oculomotor inhibition", then the rationale is not true. The alternative result does not essentially mean that functional role of oculomotor inhibition goes beyond vision. It could simply mean that "expectation" causes oculomotor inhibition to improve visual performance and at the same time alters processing in other domains to improve performance in other modalities. Therefore, my main concern is that the results do not provide any insight regarding the functional significance of oculomotor inhibition, except showing that the its controlling "expectation" module controlling is a supramodal one, which was expected (or not counter-argument was provided).
-We have clarified the logic of this study and stressed the novelty of the findings [lines 409-412].
Minor: 1) Why would the authors choose 400-500 ms after cue onset as the post cue window? They can slide a window (with different widths) through time to see which period shows most significant effect.
## --we have clarified this choice [lines 179-180].
Please note that sliding a window would not necessarily be a better alternative; to estimate the slope it is necessary to calculate two separated time intervals.
2) There is no reference to in the text. The authors choose to not address my concern claiming that I did not understand the experiment. I did understand the experiment (The figure with the method was clear enough) and there is a rhythmic sequence of cue and target (irrespective of the ISI, the target is always followed by the cue at a particular fixed time interval). I believe it is an autonomic reaction of the nervous system that has nothing to do with "expectation". The experiment I suggested would have solved this but the authors choose not to carry out this experiment. That is too bad because only that experiment would have told us whether "temporal expectations by the cognitive system" are really involved.
Reviewer #3: Remarks to the Author: Unfortunately, I am not able to find the answer to my concern in the rebuttal or the modified article. My concern is that the results were expected. An anticipated auditory stimulus prepares the attentional system through which it can pause other modalities and actions in order to better process the upcoming stimulus. In response to the same objection from Reviewer 2, the authors argue that since their design is not rhythmic it is not drawing an autonomic reaction of the nervous system. I would like to draw their attention to the findings of Zhao, Al-Aidroos, Turk-Browne that "Attention is spontaneously biased toward regularities." The regularity within the predictable condition is all what it takes to prepare the subjects to draw attention and I couldn't expect any result except what is shown.
I can completely understand that maybe there are crucial aspects that I am missing and other reviewers see that. I just couldn't imagine seeing any other result: an anticipated stimulus is expected to draw attention and takes the sensory resources toward its processing and put a dent in the processing of other signals and execution of motor commands. I would leave it for other reviewers and editors to decide if the paper is beyond this result or if this result was not known already.
## Reviewer #1
The authors have addressed all my previous comments. Congratulations on an excellent study! -Thank you for your support.
## Reviewer #2
The authors choose to not address my concern claiming that I did not understand the experiment. I did understand the experiment (The figure with the method was clear enough) and there is a rhythmic sequence of cue and target (irrespective of the ISI, the target is always followed by the cue at a particular fixed time interval). I believe it is an autonomic reaction of the nervous system that has nothing to do with "expectation". The experiment I suggested would have solved this but the authors choose not to carry out this experiment. That is too bad because only that experiment would have told us whether "temporal expectations by the cognitive system" are really involved.
-Thank you for your comments.
We respectfully disagree with the claim that the procedure is rhythmic. A rhythm should contain at least three consecutive repetitions of an interval pattern. The fact that in our procedure there was a fixed interval between the cue and the target makes this a predictable or regular condition but the procedure is not rhythmic by any definition. Each trial in the predictable condition included a constant interval as well as a jittered inter-trial interval (ITI), a varying fixation stabilization period, and varying response interval; thus, the outcome was not rhythmic.
In the following analysis we demonstrate computationally that the fixed condition was not more rhythmic than the random condition.
For each block, we calculated the time interval between consecutive target onsets. This difference captures all timing related factors: ITI +gaze contingent compliance time + foreperiod + RT. For each block we computed the standard deviation of these time differences to produce a measure of rhythmicity. The standard deviation from each block was averaged according to condition (fixed/random). A paired samples t-test did not yield a significant difference (fixed: mean = 1.32, SD = 2.33, random: mean = 1.66, SD = 1.57; t(19)=-0.53, p=0.7). Moreover, a Bayesian equivalent paired samples t-test (using JASP with default parameters) resulted in a BF01 of 3.79, supporting the null hypothesis that rhythmicity is similar between conditions.
We agree that the reported effect is likely to involve an autonomic reaction, but we respectfully disagree with the belief that it has "nothing to do with expectation". The system driving the observed effects should be able to learn the temporal meaning of the cue, i.e. to "understand" that a brief tone indicates a specific timing of the target. As the time intervals among trials vary, the cue has to be cognitively interpreted to be used for target prediction. We cannot rule out that the autonomic system is capable of this kind of complicated top-down inferences. In fact, this would have been a ground-breaking finding. However, considering that we have not used a rhythmic procedure, it is highly unlikely that the effect is driven solely by the autonomic reaction of the nervous system, without the involvement of top-down expectations mechanisms.
Considering what is known on the autonomic system, it could likely be involved in reducing eye movements, but in this study, it could not do so independently, without getting temporal information from higher expectation mechanisms.
Despite our disagreement with this critique, we have now included a new experiment, in which we reduced the number of interval repetitions in the predictable condition: 80% fixed intervals and 20% of a different interval (chosen to not be a harmonic of the first). We still found the same oculomotor inhibition effect with this reduced regularity. This new experiment provides converging evidence and further clarifies that our effects are not due to rhythmicity.
We note that the experiment that you have suggested is very interesting, but unfortunately, it is not relevant to our present study because temporal attention rather than expectation would have been manipulated. For a distinction between temporal attention and temporal expectation see . Moreover, we know from previous experiments in our lab, that using similar procedures to the task that you suggested would require an explicit association between cue identity and interval, and this would be inconsistent with our goal in the present study to investigate implicit expectations.
## Reviewer #3
Unfortunately, I am not able to find the answer to my concern in the rebuttal or the modified article. My concern is that the results were expected. An anticipated auditory stimulus prepares the attentional system through which it can pause other modalities and actions in order to better process the upcoming stimulus. In response to the same objection from Reviewer 2, the authors argue that since their design is not rhythmic it is not drawing an autonomic reaction of the nervous system. I would like to draw their attention to the findings of Zhao, Al-Aidroos, Turk-Browne that "Attention is spontaneously biased toward regularities." The regularity within the predictable condition is all what it takes to prepare the subjects to draw attention and I couldn't expect any result except what is shown.
I can completely understand that maybe there are crucial aspects that I am missing and other reviewers see that. I just couldn't imagine seeing any other result: an anticipated stimulus is expected to draw attention and takes the sensory resources toward its processing and put a dent in the processing of other signals and execution of motor commands. I would leave it for other reviewers and editors to decide if the paper is beyond this result or if this result was not known already.
-Thank you for your comments.
We agree that regularities modulate expectation is known from numerous studies (including the one you mentioned) and was the basic assumption of our study and the motivation or our experimental design. In fact, we open our manuscript stating, "Temporal expectations are formed by the cognitive system based on temporal regularities, and are used to distribute processing resources effectively across time." However, there is a very large gap between those studies and our current finding that very small involuntary fixational eye movements are reduced as part of this preparation, and that this reduction happens not only when expecting a visual stimulus but also when expecting an auditory one. This link between the auditory modality and the oculomotor system regarding temporal expectations is the novelty of our study and its contribution to this field.
We respectfully note that your view that this is an expected finding is not held by everyoneindeed, Reviewer 2 wrote, "there is no obvious theoretical reason to expect oculomotor inhibition prior to auditory targets". More importantly, there is no doubt that the finding is novel, as it had never been examined before. Our study contributes to the field both methodologically, as it can be used as an index for temporal expectation, and mechanistically, as a first step in understanding the interaction between modalities in expectation processes.
We agree with your comment that the effect could be driven by a crossmodal interaction between the modalities, and does not necessarily reflect a supramodal expectation mechanism, as we had suggested in the previous version. This interpretation is consistent with our view (described in the conclusion, page 21) that the oculomotor inhibition effect demonstrates how visual exploration pauses momentarily to allow for the processing of a predictable stimulus, regardless of its modality. Accordingly, we have modified the manuscript considerably to account for this hypothesis (e.g., page 2 line 62; page 3 line 89). Furthermore, we have added a section in the discussion describing this possible interpretation (page 21, lines 544 -561 & 575-577). We hope that you will find this new text to be a good discussion of the potentially involved mechanisms. |
Vaginal Bleeding in an Infant with Extreme Prematurity
Background. Minipuberty of infancy refers to the transient activation of the hypothalamic-pituitary-gonadal (HPG) axis during the first few months of life. Studies have documented a more exaggerated and prolonged gonadotropin surge in preterm infants compared with term infants. We present a case of minipuberty presenting with vaginal bleeding at the corrected age of 3 months of life. Case Presentation. A former 23 + 6-week infant presented with intermittent vaginal bleeding in the diaper at the corrected age of 3 months. Physical exam showed bilateral breast buds of 0.5 cm-1 cm with no signs of pubarche. Investigations showed pubertal levels of luteinizing hormone (LH), follicle-stimulating hormone (FSH), and estradiol. As she was impressed to have exaggerated minipuberty due to extreme prematurity, no intervention was given. Repeated hormonal workup at the corrected age of 8 months showed decreasing trend of gonadotropin and estradiol levels. Vaginal bleeding resolved, and breast buds also regressed clinically. Conclusion. Our case illustrated that the robust surge of gonadotropin in an ex-premature infant can in fact result in endometrial maturation and present as vaginal bleeding. ough the mechanism of this alteration in the HPG axis in prematurity is not clearly understood, pediatricians should be aware of the benign and self-limiting nature of this phenomenon and avoid unnecessary intervention.
# Introduction
Minipuberty of infancy refers to the transient activation of the hypothalamic-pituitary-gonadal (HPG) axis during the first 3-6 months of life. e rise in gonadotropin and sex steroid levels in both sexes allows maturation of sexual organs. Studies have documented a stronger and more prolonged gonadotropin surge in preterm compared with term females, suggestive of an alteration in pituitary-ovarian function in this group of infants. Here, we report a case where the elevated estradiol (E2) level results in endometrial maturation and shedding.
## Case presentation
A former 23 + 6-week infant, born with a birth weight of 540 grams, presented with intermittent fresh vaginal spotting on the diaper for 2-3 days at the corrected age of 3 months. Her neonatal history was significant for bronchopulmonary dysplasia (BPD), patent ductus arteriosus, retinopathy of prematurity, and grade 1 intraventricular haemorrhage which resolved on repeated ultrasound scan. As treatment for BPD, she was given inhaled fluticasone for 42 days until the corrected age of 36 weeks. She was on formula feeding with no medications except diuretics for BPD and vitamin supplement. Physical exam showed bilateral breast buds of 0.5 cm-1 cm with no other signs of puberty. Vaginal mucosa was not estrogenized, and there was no evidence of external trauma or foreign body. ere was no growth spurt, and she was growing along the <3rd centile. Systemic exam was otherwise unremarkable, and there were no cafe-au-lait spots. Investigations showed luteinizing hormone (LH) 3.7 IU/L, follicle-stimulating hormone (FSH) 18 IU/L, and estradiol (E2) 167 pmol/L. Cortisol and thyroid functions were normal. Prolactin, alpha-fetoprotein, and beta-human chorionic gonadotropin levels were unremarkable. Vulval swab yielded commensals only. Ultrasound pelvis showed a pear-shaped uterus with a uterus : cervix ratio of 1.7 and a smooth endometrial echo measuring 0.14 cm in thickness, suggestive of hormonally stimulated uterus. ere was no intrauterine mass or abnormal adnexal mass. Vaginal bleeding spontaneously subsided in 4 days. e gonadotropin-releasing hormone (LHRH) test performed at the corrected age of 4 months showed a predominant FSH response with a rise of LH/FSH (IU/L) from 0.49/4.3 to 16/ 27 at 20 mins and 13/28 at 60 mins. Magnetic resonance imaging of the brain and pituitary was unremarkable except the finding of a Rathke cleft cyst. As she was impressed to have exaggerated minipuberty due to extreme prematurity, no intervention was given. Hormonal workup at the corrected age of 8 months showed a decreasing trend of gonadotropin and estradiol levels. Breast buds regressed clinically, and there was no recurrence of vaginal bleeding. Repeated laboratory evaluation at the corrected age of 12 months showed a prepubertal value of LH < 0.1 IU/L and E2 <18.4 pmol/L. Ultrasound pelvis also revealed more tubular uterus with a fundus to cervical ratio of 1. She was last seen at the corrected age of 13 months, and there was no recurrence of pubertal development.
# Discussion
Vaginal bleeding during minipuberty is reported only in a few case reports in the literature. Similar to the reported cases, our patient belongs to the extreme prematurity group and presented during the first 6 months of life. e postnatal gonadotropin surge in these infants resembles that of central precocious puberty and results in gonadal steroidogenesis and clinical symptoms.
Onset of puberty is preceded by 2 periods of temporary HPG axis activation in early life. e first happens during midgestation and the second happens during the first 3-6 months of life, known as minipuberty. With a fall in circulating placental hormones, gonadotropin levels begin rising between days 6 and 10 after birth. ough females have lower peak LH values compared to boys, the FSH rise is more marked and sustained, and the level remain elevated until 3-4 years of age.
is brief period of postnatal pituitary activation is an important phase in reproductive development in both sexes. In boys, it is associated with testicular testosterone secretion, penile and testicular growth, and an increase in the numbers of Sertoli and germ cells. In girls, positive association have been demonstrated between elevated E2 levels and an increase in mammary gland size and uterine growth. Postnatal HPG axis activation also occurs in preterm infants and is even stronger and more prolonged. In preterm female infants, studies have shown that peak FSH and LH levels are more marked and prolonged compared to term infants, and the measurement of E2 levels in 3-month-old girls was also higher. Moreover, there is a stronger association between the postnatal E2 surge and the growth of the mammary gland diameter and uterine length, as well as an increase in folliculogenesis after the FSH surge. In our case, the initial LH and FSH levels represented central activation of the HPG axis, and she also demonstrated a prolonged and significant rise in the E2 level. As in other reported cases, the more intensive E2 surge results in breast growth and vaginal bleeding as a result of estrogenic effect on the endometrium. e condition is self-limiting, and clinical resolution is expected with a fall in gonadotropin and E2 levels.
Other causes of vaginal bleeding due to foreign body, infection, trauma, and even nonaccidental injury have to be considered in our case, and it is important to exclude these through physical examination, laboratory evaluation, and imaging. Our patient had no features of external trauma, and vulval swab was unremarkable. McCune-Albright syndrome (MAS), characterized by café-au-lait spots, gonadotropinindependent precocious puberty, various endocrinopathies, and fibrous dysplasia, should also be considered in any prepubertal female presenting with recurrent menstrual bleed. With hormonal profile suggesting central activation of the HPG axis, as well as absence of ovarian cysts on
# Conclusion
Pseudomenstruation shortly after birth is well known to parents and clinicians. On the other hand, vaginal bleeding beyond the neonatal period is an alarming symptom and should be evaluated carefully. Our case illustrated that the robust surge of gonadotropin in an ex-premature infant can result in endometrial maturation and present as vaginal bleeding. Treatment with luteinizing hormone-releasing hormone (LHRH) analogs is not necessary, and clinical resolution is expected with a gradual return of gonadotropin levels to the prepubertal range. As we encounter more and more survivors of extreme prematurity nowadays, we should be aware of this phenomenon and the proper management with a conservative approach.
## Abbreviations
HPG: Hypothalamic-pituitary-gonadal LHRH: Gonadotrophin-releasing hormone LH:
Luteinizing hormone FSH:
Follicle-stimulating hormone E2:
Estradiol BPD: Bronchopulmonary dysplasia MAS: McCune-Albright syndrome.
## Data availability
No data were used to support this study.
## Consent
Written consent has been obtained from the patient's mother regarding this publication. |
The Role of Alternative Crops in an Upcoming Global Food Crisis: A Concise Review
[bib_ref] Preventing food crises using a food policy approach, Timmer [/bib_ref] [bib_ref] The food crisis isn't over, Braun [/bib_ref] [bib_ref] Effect of COVID-19 on agricultural production and food security: A scientometric analysis, Okolie [/bib_ref] [bib_ref] Act now before Ukraine war plunges millions into malnutrition, Osendarp [/bib_ref] [fig_ref] Figure 1: The share of Ukraine, the Russian Federation, and the rest of the... [/fig_ref] [bib_ref] Act now before Ukraine war plunges millions into malnutrition, Osendarp [/bib_ref] [bib_ref] Effect of COVID-19 on agricultural production and food security: A scientometric analysis, Okolie [/bib_ref] [bib_ref] Impact of COVID-19 in food supply chain: Disruptions and recovery strategy, Barman [/bib_ref] [bib_ref] Effect of COVID-19 on agricultural production and food security: A scientometric analysis, Okolie [/bib_ref] [bib_ref] COVID-19 risks to global food security, Laborde [/bib_ref] [bib_ref] Effect of COVID-19 on agricultural production and food security: A scientometric analysis, Okolie [/bib_ref] [bib_ref] Impact of COVID-19 in food supply chain: Disruptions and recovery strategy, Barman [/bib_ref] [bib_ref] Effect of COVID-19 on agricultural production and food security: A scientometric analysis, Okolie [/bib_ref] [bib_ref] COVID-19 risks to global food security, Laborde [/bib_ref] [bib_ref] COVID-19 risks to global food security, Laborde [/bib_ref] [bib_ref] Effect of COVID-19 on agricultural production and food security: A scientometric analysis, Okolie [/bib_ref] [bib_ref] COVID-19 risks to global food security, Laborde [/bib_ref]
## Alternative crops and food security
Initially, we need to define what an alternative (or novel, innovative, retrovative, etc.) crop is. In most cases, the alternative crops (ACs) are described as crops that can be introduced into a new agroecosystem in lieu of "traditional" crops that are usually more susceptible to biotic (e.g., pests) and abiotic stress(e.g., salinity). For instance, heritage cereals that are usually more resilient to extreme weather events compared to the modern cereal varieties [bib_ref] A return to the genetic heritage of durum wheat to cope with..., Slama [/bib_ref] , could be characterized as ACs. According to Isleib, ACs are crops (re)introduced in a particular geographic area due to their potential high value or other benefits to the farming systems of that area. ACs are frequently mixed up with the underutilized species (NUS) (also known as neglected, orphan, or niche crops). NUS were primarily being cultivated in their center of origin; however, at some point in time they lost favor and now have regained interest (locally or in a wider scale). The concepts of ACs and NUS have apparent differences, yet some NUS could be perceived as ACs, provided they have been proposed as promising crops to be (re)introduced to an area of adaptation. Subsequently, there is no strict classification of a group of crops as ACs [fig_ref] Table 1: List of crops that have been proposed as ACs [/fig_ref]. As a case in point, teff is regarded as an AC in the Mediterranean Basin, yet as a traditional crop in Ethiopia [bib_ref] Tef: A tiny grain with enormous potential, Girija [/bib_ref].
In order to evaluate the potential role of ACs in an upcoming GFC, initially we have to examine their beneficial effects on food security (FS). According to the definition by FAO, FS is the state when "all people have physical and economic access to sufficient, safe, and nutritious food that meets their dietary needs and food preferences for an active and healthy life". Of course, "access" is not the only factor that defines FS. FS is founded on four pillars: availability, access, utilization, and stability.
The availability of food is both a quantitative, and a qualitative indicator, as it refers to the existence of sufficient amounts of domestically produced and/or imported nutritious food [bib_ref] Climate change and food security in Sri Lanka: Towards food sovereignty, Gunaratne [/bib_ref]. To enhance food availability, the introduction of an AC to a region should aim to increase the quantities of produced food commodities in that region, and offer high-quality, nutritious food alternatives. Several crops rich in micro-and macronutrients have been proposed as ACs [fig_ref] Table 1: List of crops that have been proposed as ACs [/fig_ref]. The literature also highlights the acclimatization and adaptation potential of some of them under high salinity, drought conditions, water logging, and in soils with low fertility. The introduction of ACs to such low-productivity areas could increase food availability., their common and scientific names, family, area of origin, and nutritional value. These crops have been proposed as ACs due to their acclimatization potential to marginal environments and/or their tolerance to biotic (pests and diseases) or abiotic stress (high salinity and sodicity, droughts, and high temperatures). Excelent source of essential unsaturated fatty acids, particularly OMEGA-3 fatty acids [bib_ref] Oil-seed crop: Camelina sativa, Zubr [/bib_ref] As marginal areas constitute a significant portion of available land, their exploitation via the introduction of stress tolerant ACs would increase their agricultural productivity, and thus enhance FS. For instance, in Iran, high salinity and droughts are major obstacles to agriculture. The experimental incorporation of quinoa, a salinity tolerant crop, in Iran reported promising results as irrigation with 14 dS m −1 of saline water resulted in grain yields of 2-3 t/ha.
Access to food is mainly determined by economic factors [bib_ref] Food security and the dynamics of wheat and maize value chains in..., Grote [/bib_ref]. Typically, the balance between food prices and household income/assets influences food access. The prices of agricultural commodities usually depend on their supply and demand [bib_ref] Retail prices of nutritious food rose more in countries with higher COVID-19..., Bai [/bib_ref]. Shifts in their supply/demand equilibrium tend to alter their price. This affects FS, as extreme downward or upward price oscillations have been proven to be detrimental to food access. From a financial point of view, ACs can boost food access to rural areas by improving the agriculture household income, especially in the case of small-and medium-scale farmers. Moreover, ACs promote food/crop diversity. Crop diversification can stabilize the farmers' flow of income, especially in a small-farm scale [bib_ref] Can crop diversity strengthen small-scale farmers' resilience?, Kozicka [/bib_ref]. Additionally, they provide resilience to income shocks and opportunities for improved incomes, due to the potentially favorable prices of ACs [bib_ref] Diversification for enhanced food systems resilience, Hertel [/bib_ref].
The FAO defines food utilization as "the proper biological use of food", under the context of a healthy diet that provides sufficient energy and essential nutrients [bib_ref] Food security, Rivera [/bib_ref]. Modern agriculture relays heavily on a few staple food crops to meet the global demand. It is estimated that wheat, maize, and rice provide more than half of the world's plant-derived calories. That being the case, a lack in dietary variety can lead to malnutrition despite adequate caloric intake [bib_ref] Food security and the dynamics of wheat and maize value chains in..., Grote [/bib_ref]. Enriching crop diversity would benefit both the FS and the natural agroecosystems. The introduction of ACs could increase the versatility and improve the nutritional content of meal plans, as the grains of several ACs are rich in proteins, fats, crude fibers, micro-and macronutrients, etc.. Moreover, crop diversity is known to benefit nutritional stability [bib_ref] Global relationships between crop diversity and nutritional stability, Nicholson [/bib_ref]. Admittedly, due to their poor market presence, the advancements in processing and storage methods of AC final products are often lackluster. Proper storage is essential for tackling food insecurity [bib_ref] Improved on-farm storage reduces seasonal food insecurity of smallholder farmer households-Evidence from..., Brander [/bib_ref].
Food stability can be achieved when food availability, access, and utilization are consistent through time. This dimension of FS comprises an expression of the need for sustainable food production and sustainable agricultural systems. Sustainable agriculture itself is compromised of the changing climate, the loss of agricultural biodiversity, soil degradation, and water and air pollution. The intensification of agriculture and the mainstay inputs of conventional agricultural systems only aggravate the situation. On the contrary, the adoption of ACs could tackle these constraints on sustainable agriculture. Besides the enhancement of crop diversity, many ACs often require low (compared to traditional crops) chemical inputs in the form of fertilizers and pesticides. Similarly, the drought tolerant ones are characterized by reduced irrigation needs. Due to their low input needs, they can perform adequately under organic systems. As a result, the introduction and cultivation of ACs could further reduce environmental degradation. In a recent study by Mazac et al. [bib_ref] Incorporation of novel foods in European diets can reduce global warming potential,..., Mazac [/bib_ref] , the authors estimated that the incorporation of novel foods in European food systems would contribute to global warming mitigation, as well as improve water and land use by over 80%.
## Food for thought
The introduction of ACs should be dealt with caution, otherwise, not only will they not contribute to FS, but they could also be unprofitable for the farmers and damaging to the agricultural systems they are introduced into. To fully understand this dynamic, one can refer to the example of quinoa. Quinoa originates from the South American Andes, where it has been cultivated for more than 8000 years. Following the 1950s, quinoa gained international attention, due to its high nutritional value, that peaked around the mid-2010s [bib_ref] When neglected species gain global interest: Lessons learned from quinoa's boom and..., Andreotti [/bib_ref]. As the demand for quinoa grew, the crop was introduced to many countries in Europe and North America, though they continued to import significant amounts of quinoa from the three major producers (Peru, Bolivia, and Ecuador) [bib_ref] When neglected species gain global interest: Lessons learned from quinoa's boom and..., Andreotti [/bib_ref]. Quinoa markets boomed, the demand increasing rapidly, and the international prices were elevated. In the Andes, the attractive prices of quinoa shifted its cultivation from small-farm "traditional" models to large-scale, market-oriented farming. Cultivation of quinoa was intensified to the point that land use changes, land degradation, extensive monocropping, and the loss of genetic diversity threatened the sustainability of both the production of quinoa and the local agroecosystems [bib_ref] When neglected species gain global interest: Lessons learned from quinoa's boom and..., Andreotti [/bib_ref]. Soon, the production of quinoa in the Andes doubled, the supply of quinoa exceeded the global demand, and the prices fell [bib_ref] When neglected species gain global interest: Lessons learned from quinoa's boom and..., Andreotti [/bib_ref]. This phenomenon is known as "boom and bust", a chain reaction catalyzed by market trends, that results in acute price oscillations and a shift towards less sustainable agricultural practices [bib_ref] When neglected species gain global interest: Lessons learned from quinoa's boom and..., Andreotti [/bib_ref].
Andreotti et al. [bib_ref] When neglected species gain global interest: Lessons learned from quinoa's boom and..., Andreotti [/bib_ref] acknowledged that the "boom and bust" of such crops can be divided in stages: promotion, boom, bust, and transition to a new system. These stages could function as the pillars upon which policymakers can design frameworks for the introduction of ACs to new areas. Here, we will highlight key features of the ACs integration process that, based on the literature, as well as empirical knowledge, should always be regarded. Based on the work of Andreotti et al. [bib_ref] When neglected species gain global interest: Lessons learned from quinoa's boom and..., Andreotti [/bib_ref] , this process could be divided into three phases: promotion of ACs, incorporation to food systems, and sustainable production, and they should aim to avoid the "boom and bust" phenomenon [fig_ref] Figure 2: Highlights of the three stages of introducing ACs [/fig_ref]. Promotion of ACs: Promoting an AC is based on the simple, yet admittedly challenging task of raising awareness and educating the public. This can be done via mainstream and social media, workshops and living labs, educational initiatives, or even peer-to-peer interactions. National governments possess the means (e.g., taxations and subsidies) to motivate farmers to adopt ACs. Farmers on their behalf should comprehend every beneficial aspect of the ACs on FS and sustainable agricultural practices, instead of focusing solely on their potential short-term economic returns. This tendency has also been observed on the adoption rates of integrated pest management strategies. However, wellinformed farmers have been reportedly more likely to change their attitudes towards these practices [bib_ref] Towards Pesticide-Free Farming? Sharing Needs and Knowledge Promotes Integrated Pest Management, Lucchi [/bib_ref]. Well-informed farmers might also be more likely to adopt ACs. Consumers, on the other hand, might already be more willing to embrace ACs. A recent study by Wendin et al.found that, in the case of heritage cereals, women and elders amongst the different age/sex groups were the most concerned regarding the origin and health benefits of the AC, and the elderly were more willing to pay higher prices for the AC products. They also reported that in the majority of the participants in their study were aware of the heritage cereals (to least at some extent). The authors attributed this finding partially to the recent health trends that have been related with such ACs. However, consumers (mainly in developed countries) should adopt proper attitudes and not simply follow food trends.
of quinoa from the three major producers (Peru, Bolivia, and Ecuador) [bib_ref] When neglected species gain global interest: Lessons learned from quinoa's boom and..., Andreotti [/bib_ref]. Quinoa markets boomed, the demand increasing rapidly, and the international prices were elevated. In the Andes, the attractive prices of quinoa shifted its cultivation from small-farm "traditional" models to large-scale, market-oriented farming. Cultivation of quinoa was intensified to the point that land use changes, land degradation, extensive monocropping, and the loss of genetic diversity threatened the sustainability of both the production of quinoa and the local agroecosystems [bib_ref] When neglected species gain global interest: Lessons learned from quinoa's boom and..., Andreotti [/bib_ref]. Soon, the production of quinoa in the Andes doubled, the supply of quinoa exceeded the global demand, and the prices fell [bib_ref] When neglected species gain global interest: Lessons learned from quinoa's boom and..., Andreotti [/bib_ref]. This phenomenon is known as "boom and bust", a chain reaction catalyzed by market trends, that results in acute price oscillations and a shift towards less sustainable agricultural practices [bib_ref] When neglected species gain global interest: Lessons learned from quinoa's boom and..., Andreotti [/bib_ref].
Andreotti et al. [bib_ref] When neglected species gain global interest: Lessons learned from quinoa's boom and..., Andreotti [/bib_ref] acknowledged that the "boom and bust" of such crops can be divided in stages: promotion, boom, bust, and transition to a new system. These stages could function as the pillars upon which policymakers can design frameworks for the introduction of ACs to new areas. Here, we will highlight key features of the ACs integration process that, based on the literature, as well as empirical knowledge, should always be regarded. Based on the work of Andreotti et al. [bib_ref] When neglected species gain global interest: Lessons learned from quinoa's boom and..., Andreotti [/bib_ref] , this process could be divided into three phases: promotion of ACs, incorporation to food systems, and sustainable production, and they should aim to avoid the "boom and bust" phenomenon [fig_ref] Figure 2: Highlights of the three stages of introducing ACs [/fig_ref]. Promotion of ACs: Promoting an AC is based on the simple, yet admittedly challenging task of raising awareness and educating the public. This can be done via mainstream and social media, workshops and living labs, educational initiatives, or even peer-to-peer interactions. National governments possess the means (e.g., taxations and subsidies) to motivate farmers to adopt ACs. Farmers on their behalf should comprehend every beneficial aspect of the ACs on FS and sustainable agricultural practices, instead of focusing solely on their potential short-term economic returns. This tendency has also been observed on the adoption rates of integrated pest management strategies. However, well-informed farmers have been reportedly more likely to change their attitudes towards these practices [bib_ref] Towards Pesticide-Free Farming? Sharing Needs and Knowledge Promotes Integrated Pest Management, Lucchi [/bib_ref]. Well-informed farmers might also be more likely to adopt ACs. Consumers, on the other hand, might already be more willing to embrace ACs. A recent study At this point, a paradox needs to be addressed. As mentioned above, nearly 200 million people are currently in food crisis all over the world. Yet more than 700 million are on the opposite side of the nutritional spectrum, being overweight or obese. The rates of obesity are expected to increase, and by 2030, more than 1 billion people are estimated to be living with obesity. Obesity is rapidly turning into an epidemic, especially in the developed countries. Recent studies report that more than 40% of North Americans are obese and approximately 60% of Europeans are either overweight or obese . The literature strongly suggests that, besides the plethora of health problems that have been attributed to obesity, there is a link between it and COVID-19 high mortality rates [bib_ref] Association between obesity and hospital mortality in critical COVID-19: A retrospective cohort..., Plourde [/bib_ref]. As health experts call for a solution, the timing is perfect for proposing more diverse and healthy diets. ACs could offer viable solutions to meal plans and their promotion could be part of a healthier "new food agenda", that could also target younger audiences (e.g., inclusion of zero food waste and healthy diet-related lectures in school curriculums) and help them develop healthy eating habits from their early years.
Incorporation into food systems: Initially, the selection of the introduced AC is premises on meticulous planning based on region-specific studies. Several factors must be considered, including environmental and pedoclimatic niches, regional food preferences and needs, and the societal benefits of the AC to local communities. The involvement of National Agricultural Research Systems (NARS) is vital. Research should also include breeding programs to improve the crop if needed (e.g., reduce seed heterogeneity). However, the importance of the genetic diversity of the crop should not be neglected, as it relates to the crop's adaptability. The lesson learned from quinoa's boom and bust is that when ACs transition from smallholding to an industrial agriculture model, the crops' genetic variety gets disregarded, due to market pressure. The industrialization of the ACs might also ignore the empirical farming knowledge passing down from generation to generation (especially in the case of NUCs). Ex situ gene banks and the utilization of cultivation-practices related to traditional knowledge will be essential for the improvement of the ACs' performance. Finally, it is crucial to ensure the ACs market presence. This requires the assessment of the dynamics of regional agricultural development, the existing markets and supply chains, and the logistics costs.
Sustainable production: Research on the optimization of cultivation practices (e.g., fertilization, irrigation, etc.) and food processing should be constant and not limited during the introduction of ACs in a new food system. The effects of ACs on the environment and on the everyday lives of rural populations should be regularly evaluated. Governments, civil society organizations, and private partners should monitor the value chains of ACs and interfere when needed to avoid any boom-and-bust scenarios. Overall, the cooperation of both the public and the private sectors (public-private partnerships) would be beneficial for the sustainability of ACs, and the agricultural systems as a whole. A recent report by the UN Food Systems Summit, the World Bank, the International Food Policy Research Institute (IFPRI), and the Food and Land Use Coalition presented the Food Finance Architecture, a five food finance imperatives-based policy for sustainable food systems. Under this context, governments and private sector partners could mutually finance investments with social and environmental impact, such as the incorporation of ACs in food systems.
ACs are very promising for the future of agriculture. After all, the introduction of ACs also complies with the Sustainable Development Goals (SDGs) set by the UN [fig_ref] Figure 3: ACs and the SDGs set by the UN [/fig_ref]. The sustainability of agriculture has become a major challenge for policy makers all around the globe. Both the EU and the USA, two of the most significant economic regions of the world with vastly contrasting approaches in agriculture, have designed their strategies to achieve that [bib_ref] Sustainable Agriculture in the US vs. the EU: A Comparative Look at..., Hutchins [/bib_ref]. ACs seem to be fitting for the Special Objectives of the EU's Common Agricultural Policy 2023-2027(climate change mitigation, biodiversity enhancement, sustainable food production), the aims of the European Green Deal(reduction of chemical inputs, creation of sustainable food labeling, reduction of greenhouse gases emissions), as well as the aspirational goals of the US Agriculture Innovation Strategy(market expansion and diversity). The adoption of ACs could also facilitate the implementation of the Agenda 2063 that aims to enhance Africa's collective FS by 2063, and the 2030 Strategy of the Asian Development Bank that intends to tackle climate change while strengthening FS in Asia by 2030. ACs can be grown sustainably, but at the same time, they have much to offer to the concept of agricultural sustainability itself. duction of chemical inputs, creation of sustainable food labeling, reduction of greenhouse gases emissions), as well as the aspirational goals of the US Agriculture Innovation Strategy(market expansion and diversity). The adoption of ACs could also facilitate the implementation of the Agenda 2063 that aims to enhance Africa's collective FS by 2063, and the 2030 Strategy of the Asian Development Bank that intends to tackle climate change while strengthening FS in Asia by 2030. ACs can be grown sustainably, but at the same time, they have much to offer to the concept of agricultural sustainability itself.
[fig] Figure 1: The share of Ukraine, the Russian Federation, and the rest of the world in the global ex ports of sunflower-seed or oil (crude) on the top, and wheat on the bottom, during 2021. Data ob tained from the official website of the Observatory of Economic Complexity (OEC (https://oec.world/en) (accessed on 14 October 2022). [/fig]
[fig] Figure 2: Highlights of the three stages of introducing ACs. All the actors (yellow color) of the ACs value chain and their key interactions are briefly depicted in the figure. The actions and interactions of each stage are depicted with different colors (white, blue, and green). As the three stages are not necessarily successive, actions and interactions of the actors from different stages could be simultaneous. [/fig]
[fig] Figure 3: ACs and the SDGs set by the UN. Each SDG is depicted with a different color and is attributed its corresponding number. The contribution of ACs to the SDGs is depicted in the center of the figure next to the corresponding number of each SDG. Author Contributions: Conceptualization, A.M. and I.K.; methodology, I.K.; software, I.R.; validation, A.M., I.R. and I.K.; formal analysis, I.R.; investigation, A.M.; resources, I.K.; data curation, I.R.; writing-original draft preparation, A.M.; writing-review and editing, A.M., I.R. and I.K.; visualization, A.M. and I.K.; supervision, I.K.; project administration, I.K. All authors have read and agreed to the published version of the manuscript. [/fig]
[fig] Funding: This research received no external funding. [/fig]
[table] Table 1: List of crops that have been proposed as ACs [/table]
|
Obesity may increase survival, regardless of nutritional status: a Swedish cohort study in nursing homes
Background: To investigate the associations between the body mass index (BMI), Mini Nutritional Assessment-Short Form (MNA-SF) scores, and 2-year mortality.Methods: A nationwide cohort study using data from a national quality register of older (age ≥ 65 years) nursing home residents (N = 47,686). Individuals were categorized according to BMI as underweight (< 18.5 kg/m 2 ), normalweight (18.5-24.9 kg/m 2 ), overweight (25.0-29.9 kg/m 2 ), and obese (class I, 30.0-34.9 kg/m 2 ; class II, 35.0-39.9 kg/m 2 ; class III, ≥ 40.0 kg/m 2 ). Participants' nutritional status were categorized as good (MNA-SF score 12-14), at risk of malnutrition (MNA-SF score 8-11), or malnutrition (MNA-SF score 0-7). Associations with mortality were analysed using Cox proportional-hazards models.Results: At baseline, 16.0% had obesity, and 14.6% were malnourished. During 2 years of follow-up, 23,335 (48.9%) individuals died. Compared with normal-weight individuals, mortality was greater among underweight individuals [hazard ratio (HR) 1.62, 95% confidence interval (CI) 1.55-1.69] and lesser among individuals with class I (HR 0.63, 95% CI 0.60-0.66), class II (HR 0.62, 95% CI 0.56-0.68), and class III (HR 0.80, 95% CI 0.69-0.94) obesity. Compared with individuals with good nutritional status, mortality was increased for those with malnutrition (HR 2.98,95% CI 2.87-3.10). Lower mortality among obese individuals was also seen in subgroups defined according to MNA-SF scores.Conclusions: Among older nursing home residents, obesity, including severe obesity, was associated with lower 2-year mortality. Higher BMIs were associated with better survival, regardless of nutritional status according to MNA-SF.
# Introduction
The prevalence of obesity is increasing among older adults [bib_ref] Prevalence and trends of overweight and obesity in older adults from 10..., Peralta [/bib_ref] [bib_ref] Prevalence of obesity among older adults in the United States, Fakhouri [/bib_ref] , and was present in 14% of older European nursing home residents [bib_ref] Health determinants and survival in nursing home residents in Europe: results from..., Vetrano [/bib_ref]. Obesity has several adverse health-related outcomes in older adults [bib_ref] Obesity in nursing homes: a critical review, Bradway [/bib_ref]. Furthermore, obesity is a well-established risk factor for many diseases, including cardiovascular disease (CVD) [bib_ref] Association of Body Mass Index with Lifetime Risk of cardiovascular disease and..., Khan [/bib_ref] , cancer [bib_ref] Associations of body mass index with cancer incidence among populations, genders, and..., Wang [/bib_ref] , and type 2 diabetes mellitus [bib_ref] Prevalence of obesity, diabetes, and obesity-related health risk factors, Mokdad [/bib_ref] , and for death [bib_ref] Body mass index and mortality: a meta-analysis based on person-level data from..., Mcgee [/bib_ref] [bib_ref] Association of all-cause mortality with overweight and obesity using standard body mass..., Flegal [/bib_ref] [bib_ref] Prospective studies collaboration. Body-mass index and cause-specific mortality in 900 000 adults:..., Whitlock [/bib_ref]. However, the latter association is not straightforward; obesity has been associated with reduced mortality in patients with established CVD (e.g., heart failure and coronary heart disease) [bib_ref] Obesity and prevalence of cardiovascular diseases and prognosis -the obesity paradox updated, Lavie [/bib_ref] , chronic obstructive pulmonary disease (COPD) [bib_ref] Body mass index and mortality in chronic obstructive pulmonary disease: a meta-analysis, Cao [/bib_ref] , and cancer [bib_ref] The obesity paradox in Cancer: a review, Lennon [/bib_ref]. This phenomenon has been called the obesity paradox. In older nursing home residents, overweight and obesity have been associated with lower mortality compared to normal-weight [bib_ref] Health determinants and survival in nursing home residents in Europe: results from..., Vetrano [/bib_ref] [bib_ref] Inverse relationship between body mass index and mortality in older nursing home..., Veronese [/bib_ref] , and lower mortality has been observed in both shorter and long-term follow-up for obese individuals [bib_ref] Obesity can benefit survival-a 9-year prospective study in 1614 Chinese nursing home..., Lee [/bib_ref]. Obesity is often defined as a Body Mass Index (BMI [kg/m 2 ]) ≥ 30.0 kg/m 2 and less is known about the association between obesity class I-III and mortality in the population of older nursing home residents. In a study of the association between obesity class I-III and mortality showed inconsistent results among older adults in nursing homes. They reported no significant associations between obesity and mortality in the whole sample, however, those newly admitted with obesity class II-III or obesity class III had a higher risk of mortality, whereas lower risk of mortality was found for a BMI > 28 kg/m 2 in those already living in the nursing homes [bib_ref] Obesity and mortality in elderly nursing home residents, Grabowski [/bib_ref].
On the other side of the spectrum, underweight has been associated with higher mortality in many previous studies [bib_ref] Health determinants and survival in nursing home residents in Europe: results from..., Vetrano [/bib_ref] [bib_ref] Inverse relationship between body mass index and mortality in older nursing home..., Veronese [/bib_ref]. Underweight is often caused by malnutrition, which is common [bib_ref] The nutritional situation in Swedish nursing homes -a longitudinal study, Bolmsjo [/bib_ref] [bib_ref] Body mass index, Mini nutritional assessment, and their association with five-year mortality..., Burman [/bib_ref] [bib_ref] Prevalence of proteinenergy malnutrition risk in European older adults in community, residential..., Leij-Halfwerk [/bib_ref] , and associated with increased mortality [bib_ref] The nutritional situation in Swedish nursing homes -a longitudinal study, Bolmsjo [/bib_ref] [bib_ref] Screening for malnutrition among nursing home residents -a comparative analysis of the..., Diekmann [/bib_ref] [bib_ref] Does undernutrition still prevail among nursing home residents?, Torma [/bib_ref] [bib_ref] The Mini Nutritional Assessment-Short Form and mortality in nursing home residents--results from..., Lilamand [/bib_ref] in older nursing home residents. However, the relationship between body weight and nutritional status is complicated. For example obesity is a risk factor for several diseases and conditions that negatively affect nutritional status, thus obesity can co-exist with malnutrition and/or sarcopenia [bib_ref] Double burden of malnutrition in persons with obesity, Barazzoni [/bib_ref]. Furthermore, weight loss despite a normal/high BMI can indicate malnutrition. Meanwhile it is possible to be underweight and not malnourished [bib_ref] GLIM criteria for the diagnosis of malnutrition -a consensus report from the..., Cederholm [/bib_ref]. Considering the complicated association between BMI and nutritional status, it may be important to study both their individual and aggregate associations with mortality to better understand nursing home mortality.
Consequently, it is arguably useful to investigate the association between BMI and mortality in a large sample that makes it possible to compare the different obesity classes, and to also investigate the combined effects of BMI and nutritional status, in a population of older nursing home residents. The demographic shift, leading to a larger population of older adults in the future [bib_ref] The future population of Sweden, Sweden [/bib_ref] , increase the need for further research of these important health issues, especially in this population of frail older nursing home residents. Therefore, the aim of this large nationwide cohort study was to investigate the association between obesity and mortality, including the potential heterogeneity for obesity class I-III. Furthermore, the combined effect of BMI and nutritional status according to the Mini Nutritional Assessment-Short Form (MNA-SF) among older nursing home residents in Sweden was investigated.
# Methods
## Study design and setting
This was a nationwide cohort study based on data from the National quality register Senior Alert (SA) that contains data on older adults who have contact with the healthcare system, most commonly in nursing homes, hospitals and via home care services. Senior Alert has been described in greater detail previously, in summary, several risk assessments, including a risk assessment for malnutrition, are performed and findings are recorded in the register as part of preventive care in which those at malnutrition risk should be further assessed and preventive action plans should be planned and later evaluated [bib_ref] Senior alert: a quality registry to support a standardized, structured, and systematic..., Edvinsson [/bib_ref].
## Inclusion and exclusion criteria
Eligible participants were all nursing home residents aged ≥ 65 years who were registered in SA between 1 January 2012 and 31 December 2013 and who did not fulfil exclusion criteria with missing BMI or MNA-SF data and those with weights < 20 kg or > 210 kg, heights < 100 cm or > 210 cm, BMIs < 10 kg/m 2 or > 70 kg/m 2 , or documented death before SA registration. The date of inclusion was the date of SA registration, and participants were followed until death or for a maximum of 2 years. Caregivers were obliged to inform all persons assessed about their SA registration and all persons registered were able to withdraw from the register (opt-out) at any point. This study was conducted in accordance with the principles of the Declaration of Helsinki. The study was approved by the Regional Ethical Review Board in Umeå (2013-86-31 M and 2013-456-32 M).
## Participants' characteristics
In SA, the risk of malnutrition was assessed with the six-item MNA-SF [bib_ref] Screening for undernutrition in geriatric practice: developing the short-form mini-nutritional assessment (MNA-SF), Rubenstein [/bib_ref] [bib_ref] Validation of the Mini nutritional assessment short-form (MNA-SF): a practical tool for..., Kaiser [/bib_ref] , which covers respondents' decline in food intake, unintentional weight loss in the past 3 months, mobility, psychological stress and/or acute disease in the past 3 months, neuropsychological problems (dementia or depression), and BMI (BMI; < 19 kg/ m 2 , 0 points; 19 to < 21 kg/m 2 , 1 point; 21 to < 23 kg/m 2 , 2 points; ≥ 23 kg/m 2 , 3 points). The sum of these items' score gives the total screening score. The maximum possible MNA-SF score is 14; scores of 0-7 indicate malnutrition, scores of 8-11 indicate the risk of malnutrition, and scores of 12-14 indicate good nutritional status [bib_ref] Screening for undernutrition in geriatric practice: developing the short-form mini-nutritional assessment (MNA-SF), Rubenstein [/bib_ref] [bib_ref] Validation of the Mini nutritional assessment short-form (MNA-SF): a practical tool for..., Kaiser [/bib_ref]. Body Mass Index is also registered in SA and according to the World Health Organization's (WHO) categories, participants' BMIs were used to classify them as individuals with underweight (< 18.5 kg/ m 2 ), normal-weight (18.5-24.9 kg/m 2 ), overweight (25.0-29.9 kg/m 2 ), and obesity (class I, 30.0-34.9 kg/m 2 ; class II, 35.0-39.9 kg/m 2 ; class III, ≥ 40.0 kg/m 2 ). Information about participants' diagnoses (dementia, hip fracture, COPD, renal failure, rheumatoid arthritis, myocardial infarction, stroke, and diabetes) was collected from the National Patient Register (NPR), which has full national coverage of patients discharged from hospitals since 1987. In 2001, the registry of diagnoses made during outpatient specialized care was introduced [bib_ref] External review and validation of the Swedish national inpatient register, Ludvigsson [/bib_ref]. Diagnoses are coded in the NPR according to the Swedish version of the International Classification of Diseases and Related Health Problems, Tenth Revision. Prevalent diagnoses were determined based on diagnosis occurrence in the registry up to the date of inclusion. Information about participants' level of education and disposable income was collected from Statistics Sweden, a government agency providing official statistics in Sweden. Mortality data were collected from the Cause of Death Register, which has full national coverage since 1961.
## Statistical analyses
Analyses of potential differences in sex and age between participants and non-participants were performed using the chi-squared test and the independent-samples t test, respectively. For comparisons among participant groups, defined by MNA-SF categories, the chi-squared test was used for categorical variables and the Welch one-way analysis of variance test was used for continuous variables. Associations between BMI, MNA-SF scores, and mortality were explored using Cox proportional-hazards models, including the combined effect of BMI and MNA on mortality. The regression of Schoenfeld residuals was used to test the proportional hazards assumption for all included variables against time. As the BMI and MNA-SF data violated this assumption, the observation period was divided into intervals (0 to < 6 months, 6 to < 12 months, 12 to < 18 months, 18-24 months, and 0-24 months) and separate regression analyses were performed. All models were adjusted for age, sex, level of education, disposable income, dementia, hip fracture, COPD, renal failure, rheumatoid arthritis, myocardial infarction, stroke, and diabetes, and variables that violated the proportional hazards assumption were included by stratification (except BMI and MNA-SF). Kaplan-Meier survival curves were used to visualize the associations between BMI, MNA-SF scores, and mortality. Separate analyses were performed for obese participants overall (BMI ≥30.0 kg/m 2 ) and by obesity classes. Subgroup analyses were performed to explore differences between women and men, and these were adjusted for age, level of education, disposable income, dementia, hip fracture, COPD, renal failure, rheumatoid arthritis, myocardial infarction, stroke, and diabetes. To investigate the interaction with sex, likelihood ratio tests were performed of models with the interaction term (BMI(continuous)/BMI(categorized)/ MNA(continuous)/MNA(categorized) × sex) and models without interaction terms. P values < 0.05 were considered to be significant. The statistical analyses were using R (version 3.5.0, The R Foundation for Statistical Computing, Vienna, Austria).
# Results
Of 49,604 individuals identified in the SA registry, the cumulative dropouts were; death before SA registration (n = 273), missing MNA-SF (n = 1612), weight < 20 kg (n = 1), missing weight (n = 17), height < 100 cm (n = 11), height > 210 cm (n = 1), BMI < 10 kg/m 2 (n = 2) and BMI > 70 kg/m 2 (n = 1). In total, 1918 individuals were excluded resulting in a final sample of 47,686 individuals (70% women) with a mean age of 86.3 ± 7.4 years. Excluded individuals were older than participants (86.8 ± 7.3 years, p = 0.006), with no difference in sex distribution (69.6% women).
About one-third of participants were diagnosed with dementia, 21.8% had had hip fractures, and 22.6% had had strokes at baseline. Overall, 8.3% of participants were underweight, 30.7% were overweight, and 16.0% were obese (12.1, 3.0, 0.9% had obesity class I, II, III, respectively). According to MNA-SF scores, 14.6% of participants were malnourished and 45.0% were at risk of malnutrition; 15.0% of the individuals with malnutrition were overweight or obese. During the 2-year followup period, 23,335 (48.9%) individuals died and 31.7% of these deaths occurred during the first 6 months, resulting in a mean follow up time of 1.4 years. The last documented death in the cohort occurred on 20 March 2015. Most participant characteristics differed among MNA-SF categories [fig_ref] Table 1: Baseline characteristics of participantsDifferences between groups [/fig_ref].
Kaplan-Meier survival curves show the associations of 2-year mortality with BMI and MNA-SF scores . Adjusted associations are presented in [fig_ref] Table 2: Adjusted Cox proportional hazards for associations of BMI and MNA-SF scores with... [/fig_ref] (unadjusted results, which were similar, are presented in the Appendix, . The associations of mortality with BMI and MNA-SF scores persisted throughout all follow-up periods, except for obesity class III where lower mortality was found in analyses of the entire study period of 24 months and in the first 6 months of followup [fig_ref] Table 2: Adjusted Cox proportional hazards for associations of BMI and MNA-SF scores with... [/fig_ref].
Adjusted associations between BMI and 2-year mortality according to MNA-SF categories are presented in [fig_ref] Table 3: Adjusted Cox proportional hazards for the association of BMI according to MNA-SF... [/fig_ref] (unadjusted results, which were similar, are presented in the Appendix, [fig_ref] Table 2: Adjusted Cox proportional hazards for associations of BMI and MNA-SF scores with... [/fig_ref]. Compared with individuals with obesity and good nutritional status according to MNA-SF scores, the mortality rate was greater among overweight, and normal-weight individuals with MNA-SF scores 12-14. Also compared to obese individuals with good nutritional status, greater mortality was found in those at risk of malnutrition (MNA-SF scores of 8-11) irrespective of BMI categories, with an increasing trend seen with lower BMIs. Greater mortality was found with all BMI categories among individuals with malnutrition according to MNA-SF scores, and especially in those with underweight according to BMI.
The same pattern was seen in all follow-up periods except among individuals with a good nutritional status according to MNA-SF scores and overweight where some of the follow-up periods did not reach significance [fig_ref] Table 3: Adjusted Cox proportional hazards for the association of BMI according to MNA-SF... [/fig_ref]. Adjusted and unadjusted subgroup analyses revealed slightly greater mortality among men with low MNA-SF scores and BMIs compared with their female counterparts (in the Appendices, [fig_ref] Table 3: Adjusted Cox proportional hazards for the association of BMI according to MNA-SF... [/fig_ref] and A4, respectively).
# Discussion
In this nationwide cohort study of older nursing home residents in Sweden, overweight and obesity according to the BMI, and malnutrition and the risk thereof according to MNA-SF were common. During the 2-year follow-up period, about half of the population died. Underweight according to the BMI and malnutrition according to MNA-SF were associated independently with greater mortality, compared to normal-weight and MNA-SF score indicating good nutritional status, respectively. Overweight and obesity were associated with lower mortality than was normal-weight. Obesity was associated with reduced mortality, irrespective of MNA-SF scores. The lower mortality rates observed for obese individuals and those with good nutritional status are in agreement with previous findings [bib_ref] Health determinants and survival in nursing home residents in Europe: results from..., Vetrano [/bib_ref] [bib_ref] Inverse relationship between body mass index and mortality in older nursing home..., Veronese [/bib_ref] [bib_ref] Screening for malnutrition among nursing home residents -a comparative analysis of the..., Diekmann [/bib_ref] [bib_ref] Does undernutrition still prevail among nursing home residents?, Torma [/bib_ref]. The associations appeared to be stronger during the first 6 months of follow-up, which might be explained by the protection that a larger energy reserve offers during acute illness, in fact, deteriorating health is one of the situations where risk assessments and SA registration is recommended. Researchers have suggested that the main concern for older adults and people with chronic diseases is wasting disease, rather than traditional CVD risk factors [bib_ref] Risk factor paradox in wasting diseases, Kalantar-Zadeh [/bib_ref] , and that longer follow-up periods are needed to observe the negative effects of obesity [bib_ref] From the BMI paradox to the obesity paradox: the obesity-mortality association in..., Antonopoulos [/bib_ref]. However, the lower mortality among obese individuals and those with good nutritional status according to MNA-SF persisted throughout the follow-up period in this study, reducing the potential for reverse causality. Hence, in this study cohort, obesity and good nutritional status according to MNA-SF were favourable in the short and long terms.
In the literature, obesity is often defined as a BMI ≥ 30.0 kg/m 2 , which has been associated with lower mortality compared to normal-weight among nursing home residents [bib_ref] Health determinants and survival in nursing home residents in Europe: results from..., Vetrano [/bib_ref] [bib_ref] Inverse relationship between body mass index and mortality in older nursing home..., Veronese [/bib_ref]. Using the subdivision for obesity, lower mortality has been found with class I obesity while results for obesity class II and III has been inconclusive in studies of older adults in the general population [bib_ref] Association of body mass index with all-cause and cardiovascular disease mortality in..., Wu [/bib_ref] [bib_ref] Body mass index and all-cause mortality among older adults, Cheng [/bib_ref]. Grabowski et al. [bib_ref] Obesity and mortality in elderly nursing home residents, Grabowski [/bib_ref] reported, in a study conducted in 1996, higher mortality for obesity class II and III in newly admitted nursing home residents while lower mortality was seen for BMI > 28 kg/m 2 in those who were already residing in the nursing homes. The present study did not have information on when the SA registration was made, or if any preventive actions were planned and their potential effect. It is possible that the care preventive process might have had an effect on the associations with mortality. None the less, the present findings suggest a survival benefit among older nursing home residents with obesity, including class I, II, and III.
To the best of our knowledge, it has not been shown before that obesity reduces mortality risk independently of nutritional status according to MNA-SF. We can only speculate about the factors underlying the observed associations, as the obesity paradox is not fully understood and many contributing factors have been proposed [bib_ref] From the BMI paradox to the obesity paradox: the obesity-mortality association in..., Antonopoulos [/bib_ref] [bib_ref] Obesity paradox does exist, Hainer [/bib_ref]. Higher BMI reflect a protective energy reserve and indicate a preserved muscle mass in a state of malnutrition. Also, older, obese nursing home residents are a selected group of survivors [bib_ref] From the BMI paradox to the obesity paradox: the obesity-mortality association in..., Antonopoulos [/bib_ref].
Furthermore, obesity can contribute to a decline in nutritional state [bib_ref] Double burden of malnutrition in persons with obesity, Barazzoni [/bib_ref]. The present study results do contribute to the complex field of obesity and malnutrition and indicate that higher BMI regardless of the assessed nutritional status according to MNA-SF is protective.
## Study limitations and strengths
This study has some limitations that should be addressed. As it was observational, causality could not be assessed. In addition, the BMI is a limited measure [bib_ref] BMI-related errors in the measurement of obesity, Rothman [/bib_ref] ; it does not provide information about muscle mass, body fat distribution, or cardiorespiratory fitness [bib_ref] From the BMI paradox to the obesity paradox: the obesity-mortality association in..., Antonopoulos [/bib_ref]. It is however, a well-established and accessible measure. Moreover, a small proportion (0.9%) of participants in this study had class III obesity, which limited the ability to draw conclusions for this group. Furthermore, the BMI is included in the MNA-SF score; thus these variables are not independent of each other, which must be considered when interpreting the results. Dementia, hip fracture, COPD, and stroke are expected to be more common among those classified as malnourished according to MNA-SF scores as they affect items of the MNA-SF questionnaire directly as a diagnosis or indirectly by affecting food intake, mobility, weight and/or BMI. These diagnoses are also associated with increased mortality and could thus affect the associations of mortality with the BMI and MNA-SF score. Also, there are other diagnoses and conditions than those available for this study, that might affect nutritional status and BMI. Finally, the BMI and MNA-SF data violated the proportional hazards assumption in this study, which was addressed by performing separate analyses for follow-up intervals.
To our knowledge, this study is the largest to examine associations of the BMI, and MNA-SF score with mortality among older nursing home residents and due to the inclusion criteria has a good representativity and generalizability of the findings. Almost half of the population died during the 2-year follow-up period, in agreement with findings from previous studies conducted in Sweden and Europe [bib_ref] Health determinants and survival in nursing home residents in Europe: results from..., Vetrano [/bib_ref] [bib_ref] The nutritional situation in Swedish nursing homes -a longitudinal study, Bolmsjo [/bib_ref]. In addition, the prevalence of underweight, overweight, and obesity according to the BMI, and malnutrition, and the risk of malnutrition according to MNA-SF were in agreement with previously reported values [bib_ref] Health determinants and survival in nursing home residents in Europe: results from..., Vetrano [/bib_ref] [bib_ref] The nutritional situation in Swedish nursing homes -a longitudinal study, Bolmsjo [/bib_ref]. Furthermore, the prevalence of dementia, and higher prevalence of dementia among malnourished individuals, concurs with findings from previous studies of older adults in nursing homes in Sweden [bib_ref] The nutritional situation in Swedish nursing homes -a longitudinal study, Bolmsjo [/bib_ref]. Thus, the study cohort seemed to be representative of older adults in nursing homes, thus increasing the generalizability of the present findings.
[fig] Figure 1 2, Figure 2: Kaplan-Meier survival curves showing the association between body mass index (BMI) and 2-year mortality. (1) = BMI < 18.5 kg/m 2 ; (2) = BMI 18.5-24.9 kg/m 2 ; (3) = BMI 25.0-29.9 kg/m 2 ; (4) = BMI 30.0-34.9 kg/m 2 ; (5) = BMI 35.0-39.9 kg/m 2 ; (6) = BMI ≥ 40.0 kg/m Kaplan-Meier survival curves showing the association between Mini Nutritional Assessment-Short Form (MNA-SF) scores and 2-year mortality. (1) = MNA-SF 12-14; (2) = MNA-SF 8-11; (3) = MNA-SF 0-7 [/fig]
[table] Table 1: Baseline characteristics of participantsDifferences between groups (defined by MNA-SF categories) were analysed using chi-squared test for categorical variables and using Welch one-way analysis of variance tests for continuous variables. BMI Body mass index (kg/m 2 ), COPD Chronic obstructive pulmonary disease, MNA-SF Mini Nutritional Assessment-Short Form, SD Standard deviation. a Disposable income, in 1000 SEK per year. b Analyses of all BMI categories with obesity divided in class I, II, and III [/table]
[table] Table 2: Adjusted Cox proportional hazards for associations of BMI and MNA-SF scores with 2-year all-cause mortality during follow-up intervals Analyses were of MNA-SF scores (ref.[12][13][14], BMIs with obesity defined as BMI ≥ 30.0 kg/m 2 (ref.) and BMIs with obesity divided into class I, II, III (ref.18.5-24.9 kg/m 2 ). They were adjusted for age, sex, education level, disposable income, dementia, hip fracture, chronic obstructive pulmonary disease, renal failure, rheumatoid arthritis, myocardial infarction, stroke and diabetes. BMI Body mass index (kg/m 2 ), MNA-SF Mini Nutritional Assessment-Short Form. a P value for number of deaths in follow-up intervals. b P values for numbers of deaths in BMI categories and MNA-SF categories, respectively [/table]
[table] Table 3: Adjusted Cox proportional hazards for the association of BMI according to MNA-SF score with 2-year all-cause mortality [/table]
|
Phylogenetic analysis and development of molecular markers for five medicinal Alpinia species based on complete plastome sequences
Background: Alpinia species are widely used as medicinal herbs. To understand the taxonomic classification and plastome evolution of the medicinal Alpinia species and correctly identify medicinal products derived from Alpinia species, we systematically analyzed the plastome sequences from five Alpinia species. Four of the Alpinia species: Alpinia galanga (L.) Willd., Alpinia hainanensis K.Schum., Alpinia officinarum Hance, and Alpinia oxyphylla Miq., are listed in the Chinese pharmacopeia. The other one, Alpinia nigra (Gaertn.) Burtt, is well known for its medicinal values.Results: The four Alpinia species: A. galanga, A. nigra, A. officinarum, and A. oxyphylla, were sequenced using the Nextgeneration sequencing technology. The plastomes were assembled using Novoplasty and annotated using CPGAVAS2. The sizes of the four plastomes range from 160,590 bp for A. galanga to 164,294 bp for A. nigra, and display a conserved quadripartite structure. Each of the plastomes encodes a total of 111 unique genes, including 79 protein-coding, 28 tRNA, and four rRNA genes. In addition, 293-296 SSRs were detected in the four plastomes, of which the majority are mononucleotides Adenine/Thymine and are found in the noncoding regions. The long repeat analysis shows all types of repeats are contained in the plastomes, of which palindromic repeats occur most frequently. The comparative genomic analyses revealed that the pair of the inverted repeats were less divergent than the single-copy region. Analysis of sequence divergence on protein-coding genes showed that two genes (accD and ycf1) had undergone positive selection. Phylogenetic analysis based on coding sequence of 77 shared plastome genes resolves the molecular phylogeny of 20 species from Zingiberaceae. In particular, molecular phylogeny of four sequenced Alpinia species (A. galanga, A. nigra, A. officinarum, and A. oxyphylla) based on the plastome and nuclear sequences showed congruency. Furthermore, a comparison of the four newly sequenced Alpinia plastomes and one previously reported Alpinia plastomes (accession number: NC_048461) reveals 59 highly divergent intergenic spacer regions. We developed and validated two molecular markers Alpp and Alpr, based on two regions: petN-psbM and psaJ-rpl33, respectively. The discrimination success rate was 100 % in validation experiments. Conclusions: The results from this study will be invaluable for ensuring the effective and safe uses of Alpinia medicinal products and for the exploration of novel Alpinia species to improve human health.
plants and provide new probes for species identification. Their comparatively conserved and well-defined genome structures allow the investigation of a wide range of crucial issues. Initially, genetic studies focused on understanding each plastid genome, particularly of the overview features, such as genome size, gene content, and sequence repetition. Lately, the crucial role of the plastomes in the evolution and impact for speciation has become obvious demonstrated by the sequence divergence, large inversion, differences in coding and intergenic regions, and evolutionary analysis.
To date, complete plastomes are available from more than 100 Zingiberaceae species, including four Alpinia species. Recently, a complete plastome of A. oxyphylla (NC_035895) was analyzed, and the plastome shared the highest sequence similarity of > 90 % to that of A. zerumbet. Based on the single nucleotide polymorphism (SNP) matrix among 28 whole plastomes, including a plastome (NC_048461) of A. hainanensis and two plastomes (NC_035895, MK262729) of A. oxyphylla, a phylogenetic analysis showed that Alpinia and Amomum are closely related in the family Zingiberaceae. Such results provided useful information to understand the Alpinia evolution. However, they have not focused on the species that are widely used for their medicinal values and there is no phylogenetic analysis using nuclear markers in Alpinia species.
In previous reports, phylogeny, biodiversity assessment within populations, and the authentication of Alpinia species have been studied using several molecular markers. Nuclear ribosomal DNA internal transcribed spacers (ITS) sequences have been used as markers to distinguish A. galanga from its adulterants. Efficacy of DNA barcode internal transcribed spacer 2 (ITS2) was tested on species identification of Alpinia species from Peninsular Malaysia. Also, the information of genetic relatedness was developed using seven plastid barcoding loci among wild Alpinia nigra (Gaertn.) B.L. Burtt populations.
Lately, chloroplast-derived DNA markers were developed to authenticate medicinal plants. One example is SNPs and insertion-deletion mutations (Indels) of the intergenic regions in the plastome of Panax ginseng species. However, there are no systematic studies to develop molecular markers for medicinal Alpinia species. Our short-term goal is to understand the taxonomic relationship of medicinal Alpinia species and develop molecular markers for their discrimination. And our long-term goal is to develop a method for ensuring the efficacy and safety of Alpinia medicinal products and identify new Alpinia species for medicinal uses. In this study, we reported and compared the four complete plastome sequences of A. galanga, A. nigra, A. officinarum, and A. oxyphylla sampled from Guangxi, China. The phylogenetic relationships of medicinal Alpinia species were studied based on plastome sequences and single-copy nuclear genes. Molecular markers based on plastomes were furtherly developed for the discrimination of the five Alpinia species and were validated successfully.
# Results
## Features of the alpinia species plastomes
The plastomes are circular structures of 160,590 bp (A. galanga), 164,294 bp (A. nigra), 162,140 bp (A. officinarum), and 161,394 bp (A. oxyphylla) long. The schematic representation of the plastomes is shown inand Figures S1, S2 and S3, respectively. The four plastomes display the typical quadripartite characters and show a high degree of conservation in organization and structure. They consist of a Large Single-Copy (LSC) region (87,267 − 88,970 bp) and a Small Single-Copy (SSC) region , and are somewhat higher than those of the whole plastomes.
All of the four Alpinia plastomes encode a set of 111 unique genes with identical gene order and gene clusters. Seventy-nine of these are protein-coding genes, 28 are tRNA genes, and four are rRNA genes. Fourteen genes (atpF, ndhA, ndhB, petB, petD, rpl2, rpoC1, rps16, trnA-UGC, trnC-ACA, trnE-UUC, trnK-UUU, trnL-UAA, trnS-CGA) contain one intron, while three genes, clpP, ycf3 and rps12, possess two introns . In particular, the rps12 is generated by trans-splicing and has three exons, S6 and S7, lower panels). We also detected 2-262 heteroplasmic sites with minor allele frequency (MAF) of 0.6-1 % in four sequenced Alpinia species (Figures S8, S9, S10 and S11).
## Sequence repetition in the alpinia plastomes
Comparative analysis of sequence repetition between all four plastomes found that the overall distribution, types, and numbers of repeats are highly similar among the plastomes. Simple sequence repeats (SSRs) are sequences composed of repeats with motifs from 1 to 6 bp in length. They are widespread in plastomes and widely utilized for species identification, genetic linkage construction, and molecular breeding. A total of 293-296 SSRs were found in the Alpinia plastomes. The most abundant mononucleotide SSRs are polyadenine or polythymine repeat types. Interestingly, hexanucleotide SSRs were not found in the plastomes of A. galanga and A. oxyphylla but were detected in the other two Alpinia plastomes. Further analysis of the size and location of the different SSR units and comparison revealed that the composite SSR was variable among the four species, while the dinucleotide repeat of AT was conserved (Tables S10, S11, S12, S13 and S14).
Long repeat analyses of four sequenced plastomes showed that 45-49 dispersed repeats were detected, which belong to forward, reverse, complementary and palindromic repeats. Forward (direct) and palindrome (inverted) repeats were considerably higher in number than reverse and complement repeats. The majority of these repeats with the repeat length range from 30 to 49 bp were located in intergenic spacer (IGS) regions (Tables S15, S16, S17 and S18). We found the dispersed repeats within those genes were mostly located in the exons but not in the introns. They can potentially facilitate structural rearrangements and develop variability among plastomes in a population.
On average, the numbers of detected tandem repeats range from 28 in A. officinarum up to 33 in A. oxyphylla. The copy numbers of these repeats range from 1.9 to 5.3 copies per tandem repeat, and the repeat sizes range from 30 to 158 bp per copy (Tables S19, S20, S21 and S22). The tandem repeats were found extensively in the IGS regions.
## Expansion of the ir regions in alpinia plastomes
The variations in the single-copy and IR regions' sizes and boundaries commonly cause evolutionary events such as contraction and expansion in the plastome architecture. We compared the IR and single-copy region boundaries among six species, including one Zingiber species and the five Alpinia plastomes, the four Alpinia sequenced in our research, and A. hainanensis. Two A. oxyphylla genomes sequences previously were included in the analysis. Some divergences were identified among the plastomes of four Alpinia species and Z. spectabile. Particularly, IR expansions were found in the LSC/IRa boundary of the four Alpinia species, which included the complete rps19 gene in the IRs of these species.
In contrast, the rps19 gene is located in the LSC region of Z. spectabile. The distances between the border of rps19 and the IR/LSC junction were 13, 160, 119, 129, and 129 bps in the plastomes of Z. spectabile, A. galanga, A. nigra, A. officinarum, and A. oxyphylla, respectively. Another interesting observation is that the ycf1 gene is localized in the IRb region. The ycf1 gene sequence is significantly longer in the Alpinia species, 3944 bp for A. nigra, 1428 bp for A. galanga, and 3944 bp for A. nigra, compared with that of Z. spectabile (924 bp).
## Hypervariable regions
We compared the plastome sequences of five Alpinia species, among them, A. oxyphylla with three accessions, and Zingiber species to determine the overall variations among the Alpinia and Zingiber species. As shown in , the plastomes are highly conserved among these species. The IR regions were less divergent than the LSC and SSC regions. The coding regions were more conserved than the noncoding regions. However, ndhA, petB, ycf1, and ycf2 genes showed a relatively high degree of sequence divergence. In contrast, the IGS regions were highly diverse, particularly in the following regions: rps16-trnQ, petN-psbM, psaC-ndhE, accD-psaI, psaJ-rpl33, matK-rps16, psbH-petB . Hypervariable regions can be used to resolve phylogenies and to discriminate closely related plant species. The pairwise comparison of intergenic spacer regions was conducted to identify divergence hotspot regions among the five Alpinia species using the Kimura 2-parameter (K2p) model. The average K2p distance ranged from 0.00 to 6.793 among 59 IGSs extracted from these species. Among them, the IGS regions psbE-petL, petN-psbM, accD-psaI, petD-rpoA showed the largest distances of 6.79, 6.32, 5.51, and 5.27, respectively,.
## Phylogenomic analyses based on plastome data
The availability of more complete plastome sequences of Alpinia species allows us to conduct phylogenomic analyses with higher resolution in Zingiberaceae. We performed a phylogenetic analysis using the Maximum likelihood (ML) method based on DNA sequences of 77 genes shared among 20 species from Zingiberaceae, including the four Alpinia species sequenced in the study. The sister genus of Alpinia is Amomum with a Bootstrap score (BS) of 100. The species of Alpinia are distributed in two main clades. The first clade (BS: 100) is formed by A. galanga and A. nigra, both medicinal species. The second clade (BS: 100) contains most of the species sampled to date. These species are from the tropical and subtropical geographic regions and many species are of medicinal value. Two accessions of A. oxyphylla are clustered together (BS: 99), which are subsequently clustered together with A. officinarum. Also, the phylogenetic positions of A. galanga and A. nigra were reported for the first time based on the plastomes. The bootstrap scores are high for all branches indicating the high degree of reliability of the phylogenetic tree.
## Phylogenetic analysis based on nuclear markers
The low-coverage sequence data generated from this study allowed us to perform phylogenetic analysis using additional nuclear markers. We extracted nuclear genes from sequence data among the Angiosperms-mega 353 gene set. Among these genes, 352, 353, 353, 352 genes had mapped reads, and the reads mapped to 173, 28, 93, 59 genes were assembled into contigs for A. galanga, A. nigra, A. officinarum, and A. oxyphylla, respectively. Among these assembled contigs, only four genes (AT4G04780, AT3G53760, AT5G53800, AT1G06240) were shared among the four species. These four genes were used to construct a phylogenetic tree using the same method as that for the complete plastome sequences. The reconstructed ML tree with these four genes was well resolved overall. And two of the nodes were supported with bootstrap values of 75 and 69 % . Among the four Alpinia species, A. galanga was sister to A. nigra, and A. officinarum was sister to A. oxyphylla. To compare if the relationships in both the nuclear and plastome trees are consistent, the phylogenetic analysis of plastomes with the same taxon sampling as the nuclear tree was conducted. The relevant result was consistent with the results of phylogenetic inferences obtained with nuclear markers . This approach enabled us to define further the phylogenetic relationship between the four Alpinia species using nuclear genes.
## Variation and evolutionary selection of protein-coding genes
Purifying/positive selection analyses of 77 protein-coding genes in the Alpinia plastomes showed that most genes exhibited ω values less than 0.5. Five genes (psbI, petN, psbM, petL, and psbT) had the lowest ω ratios close to 0. In contrast, the ω values of ycf2, accD, rpl23, rps7, and ycf1 were more than 1.00, respectively. The results showed that the genes accD and ycf1 were under positive selection. The likelihood ratio test identified three and five amino acid sites in accD and ycf1 that were positively selected (under posterior probability > . These sites are also highly polymorphic in the two genes.
## Molecular marker development based on alpinia plastomes
To discriminate the five medicinal Alpinia species, we selected two hypervariable IGS regions, petN-psbM, and psaJ-rpl33, to develop two DNA markers named Alpp and Alpr, respectively. The PCR primers used to amplify these two markers are shown in. PCR amplification of total DNAs from all five medicinal species samples resulted in products having expected size ,,. The DNA fragments were extracted from each band and then subjected to Sanger sequencing. The sequencing results were identical to the expected sequences (Figures S13 and S14). Marker Alpp, derived from the petN-psbM IGS region, has two specific SNP loci and one Indel loci. These three variable loci can be used to differentiate three of the five Alpinia species, except A. officinarum and A. oxyphylla. The marker Alpr, derived from the psaJ-rpl33 IGS region. It has two SNP loci and one Indel loci. When using the SNP and Indel loci from both Alpp and Alpr, all the five species can be differentiated successfully. We also have tested the new primers on all ten available Alpinia plastomes obtained from NCBI and this study in silico. These markers can discriminate all eight species based on the SNP and Indel loci from both Alpp and Alpr.
# Discussion
Here, we studied five medicinal Alpinia species, Alpinia galanga, A. hainanensis, A. officinarum, A. oxyphylla, and A. nigra. We sequenced the four plastomes of these five species. Three of them belonging to Alpinia galanga, A. officinarum, and A. nigra were reported for the first time. Two plastomes of A. oxyphylla were released during the study period. We carried out a detailed analysis of the genome features, performed the phylogenetic analysis with plastid proteomes and nuclear makers. Lastly, we developed a set of two primers that can distinguish these five medicinal species. Compared to the plastomes of previously published Alpinia species, all the plastomes presented in this study exhibited consistent genomic structure, gene order, and content. And there are no significant structural rearrangements, such as inversions or gene relocations. The size of the A. oxyphylla plastomes (MK940824) in this study is almost identical to the other two reported plastomes, which were 161,394 bp (MK940824), 161,410 bp (MK262729), and 161,351 bp (NC_035895). We found that the most abundant mononucleotide SSRs are of polyadenine or polythymine repeat types in the four Alpinia species, consistent with those reported previously. Plastomes are well-arranged, except for the expansion of the IR regions in the Alpinia species. Judged from comparative analysis with the plastome of Z. spectabile as a reference, the IR lengths of all the four Alpinia species plastomes were all increased to ≥ 160 kbp. Also, one evidence supporting this expansion is that the rps19 gene has moved to the IR regions. In other species of Alpinia, plastomes reported so far, the entire rps19 gene is also localized in the IR region, which is consistent with our findings. The analysis revealed that the four Alpinia species sequenced in our study have heteroplasmy sites in their plastomes. However, the positions of these detected heteroplasmic sites and two developed molecular markers did not overlap.
Classifications and phylogenetic analysis among Zingiberaceae were previously reported based on morphological features and DNA sequences of the nuclear internal transcribed spacer (ITS) and plastid matK regions. We use four new plastome sequences to define the position of four Alpinia species in Zingiberaceae. The new accession of A. oxyphylla sequenced (MK 940,824) in this study was most closely related to the other one A. oxyphylla plastomes reported previously (NC_035895). To date, the phylogenetic inference of Alpinia species has mainly relied on plastid markersand few multi-copy nuclear ribosomal regions such as ITS. Our phylogenetic analysis results create reliable phylogenies of the four Alpinia species sequenced by us using the nuclear markers for the first time. In addition, phylogenetic analysis using plastome and nuclear sequences revealed the identical phylogenetic relationships for the four Alpinia species.
Because of the lack of mobility, plants must deal with the challenge of abiotic stresses, such as soil salinity, drought, and extreme temperature. Many genes from plastomes, such as clpP, rbcL, and matK, ycf1 and ycf2, have been positively selected. The positive selection of the plastome genes may serve as an adaptive evolution for adjusting to environmental changes. In the selective pressure analysis, five genes were positively selected, and their selection might reflect the adaptive evolution of these Alpinia species. The results are consistent with the reports that accD and ycf1 evolved under positive selection in the Zingiber plastomes. Particular amino acids were identified to have been positively selected in two genes, accD, and ycf1. For example, the plastid accD is an essential gene required for leaf development, and the ycf1 is crucial for plant viability. In the current research, all four Alpinia species studied distributed in tropical and Sequence identity plot of the seven Alpinia plastomes with Zingiber spectabile as a reference by mVISTA. The species names are shown to the left. The grey arrows above the alignment indicate the transcription direction of genes. In the alignment box, the blue color box indicates protein-coding, the pink color box shows the conserved noncoding sequence, and the light green box indicates tRNAs and rRNAs. The x-axis represents the positions in the cp. genome, and the Y-scale represents the percent identity ranging from 50-100 % subtropical areas. Their living environment's high temperature and humidity may be the reason for the positive selection of the accD and ycf1 genes.
One of our goals is to develop markers that can distinguish the five medicinal Alpinia species. DNA markers derived from the plastomes have been widely used and are considered highly discriminatory for species identification such as Panax and Cruciata, including SNPs and InDels. So far, these plastome-derived DNA markers are usually used to analyze intraspecies level diversity and phylogenetic analysis in Alpinia. The most variable regions of the complete plastome can be used for DNA barcoding of closely related plant species. Therefore, we developed the specific markers for discriminating Alpinia species based on the plastomes' hypervariable regions. The hypervariable regions identified in our study, such as petN-psbM, psaC-ndhE, accD-psaI, were similar to those reported previously. We found two markers derived from the petN-psbM and psaJ-rpl33 IGS regions that successfully distinguished the five Alpinia species. The marker Alpp1 can't discriminate between A. officinarum and A. oxyphylla, because they are more closely related than with the other studied species. It has to be used combined with the marker Alpr1 for successful discrimination of the five Alpinia species.
Only a handful of Alpinia plastomes are sequenced and available in databases. Because the genus includes more than 200 spp., the information on the phylogeny of the genus is still rather limited. The complete Alpinia plastome sequences provided in this study expanded the taxonomic sampling and subsequently formulated new hypotheses about new potential relationships among Alpinia taxa. From this point forward, additional plastomes of Alpinia species should be sequenced, which allow us to take a broad view of the evolutionary relationship and evolutionary processes of Alpinia species, lay the foundation for the further usage of these plants for the benefit of human lives. In this study, we developed molecular markers for the five Alpinia species that are of economic importance. With the identification of additional economically important Alpinia species, the same methodology can be used to identify their corresponding differentiating markers.
# Conclusions
The complete plastomes of A. galanga, A. nigra, and A. officinarum are reported for the first time in this study. In addition, two molecular markers were developed from the hypervariable regions that can distinguish these five medicinal Alpinia species. The results obtained from these studies will contribute to our understanding of Alpinia classification, plastome evolution, and the discrimination of medicinal products derived from Alpinia species. Tribes to which each species belongs were shown to the right side of the tree. Bootstrap values were calculated with 1000 replicates Phylogenetic trees based on common genes identified using HybPiper pipeline and the shared DNA sequences of 77 protein-coding genes in the plastomes for the same five species. The phylogenetic tree on the left panel was constructed with the sequences of 4 shared contigs for nuclear genes present in 4 Alpinia species found by the HybPiper pipeline using the maximum likelihood method implemented in Phylosuite. The Oryza sativa L. was used as the outgroup. Bootstrap support scores were calculated from 1000 replicates. And the phylogenetic tree on the right panel was constructed with the shared DNA sequences of 77 protein-coding genes in the plastomes of the same five species in the nuclear tree using the same methods in phylogenomic analysis
The samples were silica-dried and stored at the Herbarium of the Institute of Medicinal Plant Development (voucher numbers: Implad201910413, Implad201910414, Implad20180327, and Implad20180362). To develop molecular markers of Alpinia species, we collected fresh leaves of another group from Guangxi Medicinal Plant Garden in Nanning, Guangxi, China, and the ginger garden of South China Botanical Garden, China (113°36' E, 23°18' N, 510,650). All samples were collected with permission from the Garden authorities. Detailed information is shown in. A plant genomic DNA kit (Tiangen Biotech, Beijing, Co., Ltd.) was used to extract total DNAs. The purity of total DNA was evaluated using electrophoresis on 1.0 % agarose gels. And the concentration was measured using a Nanodrop spectrophotometer 2000 (Thermo Fisher Scientific Inc., Waltham, MA, USA). This study complies with relevant institutional, national, and international guidelines and legislation.
## Plastome sequencing, assembly, and annotation
The sequencing libraries of total DNA from each species were prepared using the TruSeq DNA Sample Prep Kit (Illumina, Inc., San Diego, CA, USA) following the manufacturer's instructions. The total DNA was sheared into fragments at approximately 500 bp long for pairedend library construction. The libraries were sequenced on an Illumina HiSeq 3000 platform (Illumina Inc., San Diego, CA, USA). After obtaining the paired-end reads (2 × 250 bp), we downloaded the plastid genomes from the GenBank database (https://www.ncbi.nlm.nih.gov/ genome/organelle/). These plastome sequences were used to search against Illumina paired-end reads using BLASTn with an E-value cutoff of 1e-5. The filtered reads were considered plastome-related and used for the downstream genome assembly. SPAdes (v. 3.10.1)and CLC Genomics Workbench (v. 7, QIAGEN, Aarhus, Denmark) were used for de novo assembly. The dot plot of the contigs and reference genome were constructed and visualized for evaluating the assembly quality. The contigs were subjected to reassembly using the Seqman module of Lasergene (v. 11.0, Madison, Wisconsin). Only one contig was obtained for each of the Alpinia species.
We used the CpGAVAS2 web serverwas used to annotate the four genomes. Cutoffs for the E-values of BLASTn and BLASTx were set to 1e-10. After the prefiltering step, the number of top hits for annotation included in the reference gene sets was set to 10. Manual Likelihood ratio tests to identify positively selected sites within the accD and ycf1 genes across 21 Alpinia plastomes The gel electrophoresis results of the amplification of DNA barcodes using designed primers. Lane M was the marker of DL1000. The lanes from left to right corresponded to products amplificated from the first individual of A. galanga, A. hainanensis, A. nigra, A. officinarum, and A. oxyphylla by primer Alpp and Alpr, respectively. The original uncropped image is shown incorrections were performed to determine the positions of the start and stop codons and the intron/exon boundaries. Codon usage frequency and GC content (i.e., the relative content of guanines and cytosines) were calculated using custom scripts. Their circular gene maps were drawn by the cpgview web server (http://www. herbalgenomics.org/cpgview/). The raw data and the annotated plastomes have been submitted to GenBank. The accession numbers of raw data were SRR9072115 (A. galanga), SRR9072120 (A. nigra), SRR9080445 (A.
# Repeat sequence analysis
SSRs were detected using the MISA Perl Script (http:// pgrc.ipk-gatersleben.de/misa/). The minimal numbers of repeat units are eight for mononucleotide repeats, four for di-and trinucleotide repeats, and three for tetra-, penta-, and hexanucleotide repeats. Long repeat sequences with a minimal length of 30 bp and hamming distance = 3 were predicted using REPuter. Tandem repeat structures were scanned with Tandem Repeats Finder. We set the parameters to 2 for matches and 7 for mismatches and Indels. In contrast, we set the minimum alignment score and maximum period size to 50 and 500, respectively. The minimum repeat size was 30 bp, and the cutoff for similarities among the repeat units was 90 %. All of the identified repeat structures were verified manually. Nested or redundant repeats were removed.
# Phylogenomic analysis
For phylogenetic analyses, the DNA sequences of 77 protein-coding genes from 21 species of the Zingiberaceae family were extracted from the whole plastome sequences and aligned using MAFFT v.7. The 77 genes included accD, atpA, atpB, atpE, atpF, atpH, atpI, ccsA, cemA, clpP, infA, matK, ndhA, ndhB, ndhC, ndhD, ndhE, ndhF, ndhG, ndhH, ndhI, ndhJ, ndhK, petA, petB, petD, petG, petL, petN, psaA, psaB, psaC, psaI, psaJ, psbA, psbC, psbD, psbE, psbF, psbH, psbI, psbJ, psbK, psbL, psbM, psbN, psbT, rbcL, rpl14, rpl16, rpl2, rpl20, rpl22, rpl23, rpl32, rpl33, rpl36, rpoA, rpoB, rpoC1, rpoC2, rps11, rps12, rps14, rps15, rps16, rps18, rps19, rps2, rps3, rps4, rps7, rps8, ycf1, ycf2, ycf3, and ycf4. All aligned gene sequences were concatenated, and the bestfit evolutionary model (JTT + F + I + G4) was selected following the Bayesian information criterion (BIC) scores computed by ModelFinder. The maximum likelihood (ML) tree was constructed by IQTREE v1.6.10with 1000 non-parametric bootstrap replications, and Costus pulverulentus, Costus viridis, and Canna indica as the outgroup taxa. Finally, the consensus tree was visualized using the MEGA X software.
## Identification of nuclear markers for phylogenetic analysis
To explore the phylogenetic relationship implied by single-copy nuclear markers, we used the HybPiper v1.2to identify nuclear markers among the Angiosperms-mega 353 gene setfrom our sequencing reads for the four Alpinia species and then used them for phylogenetic analysis. The command line is "./reads_first.py -b mega353.fasta -r sample_001.fastq sample_002.fastq --prefix sample_result -bwa". The HybPiper package contains an internal reference set of 353 genes. It can identify genes from high-throughput sequencing results that are homologous to these 353 genes and extract them for phylogenetic analysis. In particular, the expanded Angiosperms353 target file, which is a drop-in replacement for the original Angiosperms353 filein the HybPiper analyses, was used to capture loci in our sequence reads. We identified the potential genes for phylogenetic analysis as follows. Firstly, we used the retrieval script in HybPiper to identify contigs matching each probe (https://github.com/ mossmatters/HybPiper). This was done using the reads_ first.py script. Secondly, the common genes among the four species were selected. Finally, the contigs of these genes were used to create a phylogeny. Briefly, a phylogenetic tree was constructed with the contigs of these nuclear genes as above by IQTREE v1.6.10with 1000 non-parametric bootstrap replications, except the best-fit model is HKY + F and the outgroup is Oryza sativa L. The phylogenetic analyses based on these nuclear sequences were conducted for the sample taxa as those based on the shared coding sequences of 77 protein-coding genes in the plastomes.
# Selective pressure analysis
The levels of selective pressure for a protein-coding gene are measured by the ratio of nonsynonymous to synonymous substitutions (ω). To detect the Alpinia plastid genes that were under positive selection, we extracted 77 protein-coding genes common to the 21 Zingiberaceae plastomes, performed multiple sequence alignment using MAFFT, and constructed a maximum likelihood (ML) tree using IQTREE v1.6.10. Then we calculated the ratio of nonsynonymous (dN), synonymous (dS) and ω (dN/dS) values were using CodeML in PAML Version 4.9with a One-ratio model (model = 0, seqtype = 1, NSsites = 0). If the ω value is > 1, the Bayes empirical Bayes (BEB) method implemented in the program EasyCodeML which is called site models (seqtype = 1, model = 0, NSsites = 0, 1, 2, 3, 7, 8)were used to identify positively selected sites.
## Identification of the hypervariable regions
We conducted a comparative genome analysis for the complete Alpinia plastomes using the software mVISTA (http://genome.lbl.gov/vista/mvista/submit.shtml) in the Shuffle-LAGAN mode. The annotated Z. spectabile plastome (NC_020363) was used as the reference in the analysis.
To identify the most divergent regions, we wrote a custom script to extract the start and end of the IGS regions from the GenBank files for the five plastomes, together with the plastome of A. hainanensis. A total of 59 IGSs shared by the five Alpinia plastomes were identified. The sequences were extracted and aligned using the ClustalW2 (v. 2.0.12) program with options "-type = DNA -gapopen = 10 -gapext = 2". Pairwise distances were calculated using the K2p evolution model implemented in the distmat program from the EMBOSS package (v. 6.3.1).
## Supplementary information
The online version contains supplementary material available at https://doi. org/10.1186/s12870-021-03204-1.
Additional file 1:. Schematic representation of the A. nigra plastome features.. Schematic representation of the A. officinarum plastome features. . Schematic representation of the A. oxyphylla plastome features.. The schematic diagram of position and length of introns and exons for the splitting genes in the plastome of A. galanga. The gene rps12 was a trans-splicing gene.. The schematic diagram of position and length of introns and exons for the splitting genes in the plastome of A. nigra. The gene rps12 was a trans-splicing gene. . The schematic diagram of position and length of introns and exons for the splitting genes in the plastome of A. officinarum. The gene rps12 was a trans-splicing gene. . The schematic diagram of position and length of introns and exons for the splitting genes in the plastome of A. oxyphylla. The gene rps12 was a trans-splicing gene.. The VCF output for the A. officinarum. . The VCF output for the A. oxyphylla.. The VCF output for the A. oxyphylla.. The VCF output for the A. nigra.. The original and full-length gel electrophoresis results of the amplification of DNA barcodes using designed primers.. The alignment of amplicons produced by designed Alpp primers.. The alignment of amplicons produced by designed Alpr primers.. The alignment of amplicons in 10 Alpinia plastomes produced by designed Alpp primers in silico.. The alignment of amplicons in 10 Alpinia plastomes produced by designed Alpp primers in silico.
Additional file 2:. Base composition in the plastomes of four Alpinia species. List of genes annotated in the plastome of A. galanga. Numbers in parentheses represented the repetition of genes. Superscript T: trans-splicing gene.. List of genes annotated in the plastome of A. nigra. Numbers in parentheses represented the repetition of genes. Superscript T: trans-splicing gene. . List of genes annotated in the plastome of A. officinarum. Numbers in parentheses represented the repetition of genes. Superscript T: trans-splicing gene. . List of genes annotated in the plastome of A. oxyphylla. Numbers in parentheses represented the repetition of genes. Superscript T: trans-splicing gene. . The length of introns and exons for the splitting genes in the plastome of A. galanga. The gene rps12 was a trans-splicing gene. . The length of introns and exons for the splitting genes in the plastome of A. nigra. The gene rps12 was a transsplicing gene. . The length of introns and exons for the splitting genes in the plastome of A. officinarum. The gene rps12 was a transsplicing gene. . The length of introns and exons for the splitting genes in the plastome of A. oxyphylla. The gene rps12 was a transsplicing gene.. SSR identified in the plastome of A.galanga. P1 = Mononucleotide; P2 = Di nucleotide; P3 = Tri nucleotide; P4 = Tetra nucleotide; P5 = Penta nucleotide; 6 = Hexa nucleotide repeats and c = Compound repeat microsatellites.. SSR identified in the plastome of A.nigra. P1 = Mononucleotide; P2 = Di nucleotide; P3 = Tri nucleotide; P4 = Tetra nucleotide; P5 = Penta nucleotide; 6 = Hexa nucleotide repeats and c = Compound repeat microsatellites.. SSR identified in the plastome of A. officinarum. P1 = Mononucleotide; P2 = Di nucleotide; P3 = Tri nucleotide; P4 = Tetra nucleotide; P5 = Penta nucleotide; 6 = Hexa nucleotide repeats and c = Compound repeat microsatellites.. SSR identified in the plastome of A. oxyphylla. P1 = Mononucleotide; P2 = Di nucleotide; P3 = Tri nucleotide; P4 = Tetra nucleotide; P5 = Penta nucleotide; 6 = Hexa nucleotide repeats and c = Compound repeat microsatellites.. Comparison of SSR markers found among four Alpinia species and one outgroup species of Zingiber spectabile. Zisp: Zingiber spectabile; Alga: Alpinia galanga; Alni: Alpinia nigra; Alof: Alpinia officinarum; Alox: Alpinia oxyphylla.. Dispersed repeat sequences in the plastome of A. galanga.. Dispersed repeat sequences in the plastome of A.nigra.. Dispersed repeat sequences in the plastome of A. officinarum.. Dispersed repeat sequences in the plastome of A. oxyphylla.. Tandem repeat sequences identified in the plastome of A. galanga.. Tandem repeat sequences identified in the plastome of A. nigra.. Tandem repeat sequences identified in the plastome of A.officinarum.. Tandem repeat sequences identified in the plastome of A.oxyphylla. a: coding sequences; b: intergenic spacers.. The distances among the shared intergenic spacer (IGS) regions from the five Alpinia plastomes. Alga: Alpinia galanga; Alha: Alpinia hainanensis; Alni: Alpinia nigra; Alof: Alpinia officinarum; Alox: Alpinia oxyphylla.. The list of accession numbers of the plastome sequences used in the phylogenetic analyses of the Zingiberaceae. The dN, dS and dN/dS (ω) value of 77 commom protein-coding genes from plastomes of 21 Alpinia species.. The two pairs of primers for the ampilification of DNA barcodes.. The list of sample numbers of the samples used in the species discrimination analyses of the Alpinia. |
Sediment Facilitates Microbial Degradation of the Herbicides Endothall Monoamine Salt and Endothall Dipotassium Salt in an Aquatic Environment
Endothall dipotassium salt and monoamine salt are herbicide formulations used for controlling submerged aquatic macrophytes and algae in aquatic ecosystems. Microbial activity is the primary degradation pathway for endothall. To better understand what influences endothall degradation, we conducted a mesocosm experiment to (1) evaluate the effects of different water and sediment sources on degradation, and (2) determine if degradation was faster in the presence of a microbial community previously exposed to endothall. Endothall residues were determined with LC-MS at intervals to 21 days after endothall application. Two endothall isomers were detected. Isomer-1 was abundant in both endothall formulations, while isomer-2 was only abundant in the monoamine endothall formulation and was more persistent. Degradation did not occur in the absence of sediment. In the presence of sediment, degradation of isomer-1 began after a lag phase of 5-11 days and was almost complete by 14 days. Onset of degradation occurred 2-4 days sooner when the microbial population was previously exposed to endothall. We provide direct evidence that the presence and characteristics of sediment are of key importance in the degradation of endothall in an aquatic environment, and that monoamine endothall has two separate isomers that have different degradation characteristics.submersed aquatic weeds, it is applied to the water column in which the weeds are growing to achieve a target concentration throughout the water column in the target area. Absorption of endothall into the foliage then occurs from the water. A consequence of this application method is that all other non-target organisms in the water are exposed to endothall.An emerging use-pattern is the application to flowing irrigation canals to remove submersed weeds that obstruct flow. In this use, the target concentration is achieved by metering a volume of endothall into the canal at a rate proportional to the discharge for a designated period to provide the target exposure time. This pulse of endothall-treated water then moves down the irrigation canal. As irrigation canals have many connections to natural rivers and wetlands, this endothall-treated water can discharge into other aquatic ecosystems, or it can be irrigated directly onto crops.The concentration of endothall in the water and the length of time plants are exposed to endothall (exposure time) determine the effectiveness of endothall against the target weeds[15,21]and the toxic effect on any non-target organism. Once applied, the aquatic ecosystem is exposed to endothall until it is completely removed by natural degradation processes or dissipation. Therefore, the length of time required for complete degradation of endothall is critical for the safety of non-target organisms and any other uses of the water or aquatic ecosystem (e.g., agriculture, recreation, drinking). Endothall degradation rate and therefore, persistence time, can vary greatly among different systems, depending on the prevailing physical, chemical and hydrodynamic properties[16,[22][23][24]. Persistence time has been reported to differ greatly between the formulations of endothall [7], therefore, it is essential to improve our understanding of the patterns and processes of degradation of both formulations of endothall to better predict the persistence time in aquatic ecosystems.Most studies on endothall dissipation and persistence were conducted during the 1960s and 1970s, following the first aquatic use registration of endothall in 1960 in the USA, and the results have been reviewed and summarized elsewhere[16,[22][23][24]. Most of those studies focused on a single formulation of endothall, mainly the dipotassium salt. A major limitation in most previous studies is that the endothall concentrations in water were analyzed using a bioassay technique originally developed in 1962[25], which is indirect and less precise than modern quantitative techniques such as LC-MS[26]. In addition, none of these early studies systematically considered the effects of environmental variables on endothall persistence.The transformation and degradation of endothall in aquatic environments is performed by microbes, particularly bacteria[26,27], and the loss of endothall by other means such as volatilization, sorption, photolysis, hydrolysis and oxidation are negligible[16,23,28]. Earlier studies suggest that provision of an environment that supports microbial growth will enhance the rate of endothall degradation[25][26][27]. Arthrobacter bacteria isolated from a lake hydrosoil have been shown to use endothall as the sole source of carbon and energy[26].In this paper, we present the results of an experiment conducted in mesocosms under controlled conditions to determine the effects of different sources of water and sediment on the degradation of monoamine endothall and dipotassium endothall. The experiment was designed to (1) evaluate the effects of three different water and sediment sources on endothall degradation, and (2) determine if degradation was faster in the presence of a microbial community that had previously been exposed to endothall.Together, these experiments represent an important advancement because it is the first study that uses a fully replicated experimental design coupled with direct quantification (LC-MS) to systematically determine how (1) presence or absence of sediment, (2) a range of native sediment and non-sterile water sources, (3) formulation of endothall, and (4) pre-exposure of microbial communities to endothall, jointly impact the degradation of endothall. This randomized design allows us to undertake a cause and effect analysis of the effects of these environmental variables on the decay of both formulations of endothall. The methodology provides waterbody managers with reliable information on the effect of environmental variables on decay of contemporary, commercially available, formulations of endothall, and thus enables them to manage endothall-treated water in an informed manner. Int. J. Environ. Res. Public Health 2018, 15, 2255 3 of 16Materials and MethodsExperimental DesignAn experiment was established to examine the effects of water and sediment variables on endothall persistence. The experiment was conducted in mesocosm systems created by filling plastic mesocosms with different combinations of water, sediment and aquatic plants, collected from different sources. The factors and levels are shown inTable 1. The sources of water, sediment and plants were: (i) Melbourne city potable water with and without garden soil (Greensborough, Melbourne, Victoria); (ii) Central Goulburn Irrigation District (Victoria) irrigation channel water with and without sediment; and (iii) Coleambally (New South Wales) irrigation channel water with and without sediment. Ribbon weed plants (Vallisneria australis S.W.L. Jacobs and Les) were collected from each irrigation district and included in the treatments with sediment from irrigation channels.
# Introduction
Endothall is a herbicide used for controlling submerged aquatic weeds and algae. Endothall is available in two formulations: Endothall dipotassium salt and endothall dimethylalkylamine salt (hereafter monoamine endothall). The active ingredient for both formulations is endothall acid (7-oxabicyclo[2.2.1]heptane-2,3-dicarboxylic acid), which is free when endothall salts are applied to water [bib_ref] Chemical, toxicological, and biological properties of endothall, Keckemet [/bib_ref]. Endothall is a contact herbicide that severely affects plant physiological processes, disrupting plant cell membranes [bib_ref] Probing mode of action in plant cell cycle by the herbicide endothall,..., Tresch [/bib_ref] , damaging plant tissues within 2-5 days after application [bib_ref] Effects of endothall and other aquatic herbicides on chlorophyll fluorescence, respiration and..., Macdonald [/bib_ref]. Endothall is highly effective against many submerged aquatic weed species [bib_ref] Chemical, toxicological, and biological properties of endothall, Keckemet [/bib_ref] [bib_ref] Evaluation of selected herbicides for the control of exotic submerged weeds in..., Hofstra [/bib_ref] [bib_ref] Determining the efficacy of the herbicides endothal and diquat on the aquatic..., Clements [/bib_ref] [bib_ref] Control of delta arrowhead (Sagittaria platyphylla) in Australian irrigation channels with long..., Clements [/bib_ref] [bib_ref] Potential new herbicides for submerged aquatic weeds in Victoria, Dugdale [/bib_ref] [bib_ref] Concentration-exposure time relationships for controlling fanwort (Cabomba caroliniana) with endothall amine salt..., Hunt [/bib_ref]. Endothall has been widely used in the USA for decades [bib_ref] Endothall concentration and exposure time relationships for the control of Eurasian watermilfoil..., Netherland [/bib_ref] , has been registered in New Zealand [bib_ref] Endothall for aquatic weed control in New Zealand, Wells [/bib_ref] , and registration is currently being pursued in Canada and Australia. Both formulations of endothall are used to control nuisance aquatic vegetation in a variety of aquatic systems including ponds, lakes, pools and irrigation channels [bib_ref] Endothall species selectivity evaluations: Southern latitude aquatic plant community, Skogerboe [/bib_ref] [bib_ref] Critical review of actually available chemical compounds for prevention and management of..., Jančula [/bib_ref] [bib_ref] Efficacy of endothall dimethylalkylamine salt applied to static irrigation channels during winter..., Clements [/bib_ref].
In terrestrial situations, herbicides are typically applied directly to the foliage of weeds, where absorption occurs and subsequently the weed dies. In contrast, when endothall is used to control . Description of treatments (source of water; presence/absence and source of sediment; endothall formulation; microbe augmentation). Endothall was applied at 2.4 mg ae L −1 ; Monoamine endothall was added as Teton™, Dipotassium endothall was added as Cascade™. The sources of water and sediment, and microbe augmentation are described in the text.
## Treatment no.
Water Sediment Endothall Form Microbe [fig_ref] Table 2: Water quality data taken over the duration of the experiment [/fig_ref] Central [fig_ref] Table 4: Analysis of variance for treatment effects on endothall isomer-1 concentration at day-0 [/fig_ref] Central Goulburn Central Goulburn --7
Potable -Monoamine -8
Coleambally -Monoamine -9
Central Goulburn -Monoamine - [bib_ref] Evaluation of selected herbicides for the control of exotic submerged weeds in..., Hofstra [/bib_ref] Potable Garden Monoamine - [bib_ref] Determining the efficacy of the herbicides endothal and diquat on the aquatic..., Clements [/bib_ref] Coleambally Coleambally Monoamine -12
Central Goulburn Central Goulburn Monoamine - [bib_ref] Potential new herbicides for submerged aquatic weeds in Victoria, Dugdale [/bib_ref] Potable -Dipotassium - [bib_ref] Concentration-exposure time relationships for controlling fanwort (Cabomba caroliniana) with endothall amine salt..., Hunt [/bib_ref] Coleambally -Dipotassium - [bib_ref] Endothall concentration and exposure time relationships for the control of Eurasian watermilfoil..., Netherland [/bib_ref] Central Goulburn -Dipotassium -Potable [fig_ref] Table 2: Water quality data taken over the duration of the experiment [/fig_ref] Central Goulburn Central Goulburn Monoamine + A total of 72 mesocosms (24 treatments × 3 replicates) with 18 L capacity [(240 × 240 × 300 mm), black, polypropylene, square bucket, Garden City Plastics FV300] were numbered serially from 1 to 72 and randomly assigned to each of the treatments and replicates, i.e., a fully randomized design. Designated mesocosms were first supplied with 2 L of sediment from the respective sediment source and then 12 L of water was added. Mesocosms not containing sediment were filled with 14 L of water only. Three treatments each had two 50 mL centrifuge tubes half filled with the same sediment and pushed into the sediment substrate at the base of the mesocosms. The tubes with sediments remained in situ until collected as sediment samples later during the experiment. Mesocosms containing irrigation channel water and sediment were then supplied with a single ribbon weed plant suspended in the water column.
To generate a microbial population pre-exposed to endothall, a mesocosm was created three months prior to the experiment. To create this mesocosm, sediment was added to a 100 L polyethylene tank to a depth of~100 mm, covered with~50 L of water and two ribbon weed plants added (all sourced from the Central Goulburn Irrigation District). The sediment and water were not autoclaved. Monoamine endothall was applied to a target concentration of 2.4 mg ae L −1 and then the tank was left in a glasshouse for three months. The endothall concentration and degradation in the mesocosm was not measured, nor was any additional carbon source added. Treatments that included augmentation with microbes pre-exposed to endothall were dosed with 100 mL of water and 50 mL of sediment from this tank (as a mixed slurry), after addition of water and sediment to the experimental mesocosms. Hereafter, these treatments are referred to as "augmented" versus all the other treatments that are "non-augmented".
The mesocosms were placed in a row along the length of a temperature controlled (~18 - C) glasshouse at floor level. Each bucket was individually aerated, such that it created a gentle movement within the water column only, and without agitating the sediment surface. Each air line was fitted with a reverse flow valve to prevent water flowing back into the air delivery system. The mesocosms were left to stabilize for one week before endothall was added. Herbicide, dipotassium endothall (Cascade™) or monoamine endothall (Teton™), was applied to achieve a target concentration of 2.4 mg ae L −1 , by pouring the required volume of stock solution into each of the designated mesocosms.
## Endothall sampling
## Water column
To determine the concentration of endothall in each mesocosm, 25 mL water samples were collected 0, 1, [bib_ref] Effects of endothall and other aquatic herbicides on chlorophyll fluorescence, respiration and..., Macdonald [/bib_ref] [bib_ref] Determining the efficacy of the herbicides endothal and diquat on the aquatic..., Clements [/bib_ref] , and 21 days after endothall application. Day-0 samples were collected 3 to 4 h after endothall application, to provide an indication of starting concentration. The sampling times were chosen to end at 21 days because the current permit for endothall use in Australia (issued by the Australian Pesticide and Veterinary Medicines Authority) specifies a 21-day withholding period before the endothall-treated water can be used for watering livestock, preparing agricultural sprays and irrigation. Therefore, it is important to know what the likely endothall concentration is at this time. The transitional sampling times were compressed in the first 12 days because prior studies indicate that most endothall usually degrades by 7 to 14 days [bib_ref] Effects and persistence of endothall in the aquatic environment, Holmberg [/bib_ref] [bib_ref] Persistence of endothall in aquatic environment as determined by gas-liquid chromatography, Sikka [/bib_ref].
Water samples were collected from the middle of the water column using a 50 mL syringe fitted with a plastic tube at the tip of the syringe. Each mesocosm had a dedicated sampling syringe and tube which remained in the mesocosm for the duration of the experiment. Samples were acidulated with one drop of 40% HCl solution and stored in a refrigerator until analysis.
For the treatments with endothall added, endothall acid concentration was determined by LC-MS. Chromatographic separation of endothall was achieved using an Eclipse XDB-C8 column (150 × 2.1 mm, 3.5 µm, Agilent Technologies, Santa Clara, CA, USA) on a Vanquish UPLC system (Thermo Scientific™, Waltham, MA, USA). The mobile phase was composed of 0.5% formic acid (A) and acetonitrile containing 0.1% of formic acid (B). The flow rate was 0.25 mL/min with a gradient elution of 5 to 60% B over 10 min. The injection volume was 5 µL.
The detection of endothall was by LTQ-Orbitrap Elite mass spectrometer (Thermo Scientific™) operated in electrospray ionization (ESI) negative Fourier transform mode. The heated capillary was maintained at 350 - C with a source heater temperature of 300 - C, and the sheath, auxiliary and sweep gases were at 40, 15 and 5 units, respectively. The source voltage was set to 3.2 kV and the resolution was set to 60,000. Endothall (deprotonated ion) was extracted from a full scan spectrum and quantified using an external calibration curve.
## Sediment
One of the two tubes containing sediment, that were placed at the bottom of selected treatments was removed 7 days after endothall application, and the other tube was removed 16 days after application. Upon removal, excess water was tipped out of the tubes, the tube was then capped and stored at −18 - C. The residual water in the sediment was collected after centrifugation and the level of endothall in the water determined by LC-MS. To verify the possibility of endothall being adsorbed to the sediment, the sediment was washed with 60% acetonitrile solution and the concentration of endothall in the washing liquid was measured by LC-MS.
At the end of the experiment, sediment samples from all sources and from the control treatments (not exposed to endothall) were collected in 250 mL plastic containers and stored in a refrigerator until they were analyzed at a commercial laboratory.
## Water quality monitoring
Electrical conductivity (EC, µS/cm 2 ), dissolved oxygen (DO, mg ae L −1 ), turbidity (NTU) and pH were measured three times per week by collecting a small volume (25-40 mL) of water from each treatment. EC, DO and pH, were measured using a water quality meter (Hach HQ40D Portable Multi Meter; HACH COMPANY, 389 Loveland, CO, USA), and turbidity was measured using a turbidity meter (Hach 2100Q Portable Turbidimeters; HACH COMPANY, 389 Loveland, CO, USA). A temperature logger was installed in one mesocosm.
## Data analyses
Two isomers of endothall were detected in the samples (see Results for description), which had different degradation dynamics. Therefore, analyses were split into isomer-1 and isomer-2.
## Isomer one
Statistical analyses were carried out using the 15 treatments with endothall applied . A general pattern for the responses over time (days after endothall application) of each mesocosm was for the response to be reasonably stable at the start of the experimental period (i.e., only a small difference between the concentration at day 0 and day 1) and near the end of the experimental period (i.e., only a small difference between the concentration at day 16 and day 21), but sometimes large and quick (e.g.,~90% reductions in~2 days) changes between these times. During times of large and quick changes, for mesocosms in a treatment, the between mesocosm variability of the isomer-1 concentration increased substantially, which makes statistical analysis less straightforward. Near the end of the experiment, one treatment, namely treatment 14, was not stable (i.e., the isomer-1 concentrations were declining quickly from days 16 to 21), and hence the isomer-1 concentration at day 21 had much greater variability between mesocosms than the between mesocosm variability for other treatments. Therefore, the treatment 14 mesocosms were not included in some statistical analyses, and in these cases the individual concentrations for the three mesocosms of treatment 14 were presented instead.
The isomer-1 response curves were statistically examined by calculating and analyzing the following five summary statistics, that have been carefully chosen to summarize the majority of treatment differences in the isomer-1 response curves with monoamine or dipotassium endothall:
## 1.
Isomer-1 concentration at day-0 2.
Percent decline in the isomer-1 concentration over the 21-day period (i.e., 100 × (isomer-1 concentration at day-0 − isomer-1 concentration at day-21)/(isomer-1 concentration at day-0) 3.
Days to 25% of observed reduction (i.e., number of days since application on the first day when the isomer-1 concentration was first observed to achieve at least a 25% reduction of the full isomer-1 concentration reduction between day-0 and day-21) 4.
Days to 50% of observed reduction (i.e., number of days since application on the first day when the isomer-1 concentration was first observed to achieve at least a 50% reduction of the full isomer-1 concentration reduction between day-0 and day-21) 5.
Days to 75% of observed reduction (i.e., number of days since application on the first day when the isomer-1 concentration was first observed to achieve at least a 75% reduction of the full isomer-1 concentration reduction between day-0 and day-21)
The days to 25, 50, or 75% of observed reduction can only take on the values 1, 3, 5, 7, 9, 11, 16 or 21 days.
The isomer-1 concentration at day-0 and the percent decline measure key components of the isomer-1 concentration at day-21. The days to 25%, 50% and 75% reduction collectively measure the relative speed of breakdown of isomer-1.
Each measurement was analyzed using an analysis of variance with mesocosm as the unit of analysis, and that included factorial and nested treatment terms that elucidated the effects present. The isomer-1 concentration at day-21 was analyzed after the mesocosm data had been log(y + 0.3) transformed, and the percent decline in the isomer-1 concentration over the 21-day period was log(1.1 − (y/100)) transformed. These transformations prevented the residual variation being substantially different between treatments with large differences in isomer-1 concentrations near the end of the study period. Additionally, the treatment 14 mesocosms were excluded from these two analyses, for the reason explained earlier. There was no reason to transform the data for the concentration at day-0, or to exclude the mesocosms from treatment 14, and thus neither of these were carried out for this measurement.
The analyses of days to 25%, 50% and 75% reduction were restricted to mesocosms from endothall treatments that had sediment added because these were the only endothall treatments with large reductions in isomer-1 concentrations by day-21. Since these measurements only allowed a restricted number of concentrations (i.e., data was discrete), non-parametric analyses of variance were carried out. Since the times between measurement intervals were not equally spaced, which tends to strongly negate equal variation between treatments, the data was firstly ranked with ties. These ranked data were analyzed using standard analysis of variance methodology. However, p values were obtained from the usual F values using a permutation test with 100,000 random permutations.
When calculating F values, the residual mean square was calculated from the saturated model (i.e., a model including an effect for all treatments included in the analysis). Additionally, the standard errors of difference were calculated using the residual mean square from the saturated model.
Statistical analyses were calculated using GenStat 18, including the ANOVA (generally balanced analysis of variance) and FIT (general linear models, not necessarily balanced) directives and the RANK (ranking data) and APERMTEST (permutation tests in analysis of variance) procedures.
## Isomer two
Isomer-2 concentrations were very low in the mesocosms that received dipotassium endothall over the entire study period, and consequently the statistical analyses are restricted to the nine treatments that received monoamine endothall (i.e., treatments 7 to 12 and 22 to 24). Similar to isomer-1, the isomer-2 response curves were statistically examined by calculating and analyzing the following two summary statistics, that have been carefully chosen to summarize the majority of treatment differences in the isomer-2 response curves with monoamine endothall:
## 1.
Isomer-2 concentration at day-0 2.
Percent decline in the isomer-2 concentration over the 21-day period (i.e., 100 × (isomer-1 concentration at day-0 − isomer-1 concentration at day-21)/(isomer-1 concentration at day-0)) Similar to isomer-1, each measurement was analyzed using an analysis of variance with mesocosm as the unit of analysis, and that included factorial and nested treatment terms that elucidated the effects present. At day-21 the mesocosms from the three treatments with sediment added, but no augmentation, appeared not to be stable. Hence the isomer-2 concentration at day-21 had much greater variability between mesocosms in these three treatments, than the other six treatments. This greater variability was also reflected in the percent decline in the isomer-2 concentration over the 21-day period. To address this issue, the isomer-2 concentration at day-21 and the percent decline in the isomer-2 concentration over the 21-day period were analyzed using non-parametric analyses of variance, using a similar technique to the 'speed of decline' summary measurements for isomer-1. Similar to the isomer-1 concentration at day-0, the isomer-2 concentration at day-0 was analyzed using a parametric analysis of variance, without any transformation.
Since, by day-21, the isomer-2 concentrations had stabilized to a low concentration for only three treatments, the analyses of days to 25%, 50% and 75% reduction provided limited information on the 'speed of decline' for isomer-2, thus, analyses are not reported.
As with isomer-1, all statistical inference used the residual mean square from the saturated model. Statistical analyses were calculated using GenStat 18, including the ANOVA directive and the RANK and APERMTEST procedures.
# Results
It is known that endothall exists as a mixture of three stereoisomers of which the (1R,2S,3R,4S)-isomer is the most herbicidally active. The number and identity of isomers in endothall herbicide formulations have never been reported to our knowledge. Two isomers of endothall were detected in this study. Isomer-1 was dominant in both endothall formulations, while isomer-2 was present in substantial amounts in monoamine endothall but at trace levels in dipotassium endothall [fig_ref] Figure 1: Extracted ion chromatogram of endothall, in two formulations [/fig_ref]. The differences in isomer composition between these two formulations have not been described before. treatments. This greater variability was also reflected in the percent decline in the isomer-2 concentration over the 21-day period. To address this issue, the isomer-2 concentration at day-21 and the percent decline in the isomer-2 concentration over the 21-day period were analyzed using nonparametric analyses of variance, using a similar technique to the 'speed of decline' summary measurements for isomer-1. Similar to the isomer-1 concentration at day-0, the isomer-2 concentration at day-0 was analyzed using a parametric analysis of variance, without any transformation.
Since, by day-21, the isomer-2 concentrations had stabilized to a low concentration for only three treatments, the analyses of days to 25%, 50% and 75% reduction provided limited information on the 'speed of decline' for isomer-2, thus, analyses are not reported.
As with isomer-1, all statistical inference used the residual mean square from the saturated model. Statistical analyses were calculated using GenStat 18, including the ANOVA directive and the RANK and APERMTEST procedures.
# Results
It is known that endothall exists as a mixture of three stereoisomers of which the (1R,2S,3R,4S)isomer is the most herbicidally active. The number and identity of isomers in endothall herbicide formulations have never been reported to our knowledge. Two isomers of endothall were detected in this study. Isomer-1 was dominant in both endothall formulations, while isomer-2 was present in substantial amounts in monoamine endothall but at trace levels in dipotassium endothall [fig_ref] Figure 1: Extracted ion chromatogram of endothall, in two formulations [/fig_ref]. The differences in isomer composition between these two formulations have not been described before.
## Water quality and sediment properties
Water temperature, pH and dissolved oxygen in the mesocosms were similar between the treatments [fig_ref] Table 2: Water quality data taken over the duration of the experiment [/fig_ref] and represent an environment suitable for aerobic microbial growth. Turbidity and electrical conductivity varied by water source and addition of sediment [fig_ref] Table 2: Water quality data taken over the duration of the experiment [/fig_ref].
The sediments from the irrigation channels were both clays, with low organic matter and pH, in contrast to the garden soil, which was a loam with high organic matter and neutral pH [fig_ref] Table 3: Properties of the sediment used in the experiment [/fig_ref]. Another important difference was that the garden soil was much more biologically active (as measured by total colony forming units per g soil; [fig_ref] Table 3: Properties of the sediment used in the experiment [/fig_ref].
## Endothall degradation
Endothall (isomer-1) persisted for the 21-day duration of the experiment in the treatments that did not contain sediment. In contrast, where sediment was present, there was an initial period of endothall persistence, followed by rapid degradation commencing by days five to seven. After this period of rapid degradation, endothall concentration remained above zero [fig_ref] Figure 2: Endothall isomer-1 [/fig_ref]. Isomer-2 followed the same trend in relation to sediment presence for treatments that contained monoamine endothall, although it did not always degrade to low levels, even in the presence of sediment.
At day-0, the starting concentrations of monoamine endothall (isomer-1) only differed with endothall formulation and presence of sediment [fig_ref] Table 4: Analysis of variance for treatment effects on endothall isomer-1 concentration at day-0 [/fig_ref]. The starting concentrations were 3.05 (sediment present) and 3.08 (sediment absent) mg ae L −1 , and of dipotassium endothall were and 2.27 (sediment present) and 2.50 (sediment absent) mg ae L −1 , with the standard error of difference ranging between 0.038 and 0.044.
During the 21-day period of the experiment endothall did not degrade in water without sediment; its concentration increased by 19% and 24% for monoamine and dipotassium endothall, respectively [fig_ref] Table 5: Analysis of variance for treatment effects on endothall isomer-1 as % decline... [/fig_ref]. These increases in concentration were likely due to evaporation of water from the mesocosms. In contrast, 96% of monoamine endothall (with or without augmentation) had decayed by day-21, and 98% of dipotassium endothall had decayed by day-21 in the treatments with sediment added (Tables 5 and 6; [fig_ref] Figure 2: Endothall isomer-1 [/fig_ref]. The endothall remaining was lower when sediment was sourced from irrigation channels (97% reduction) than from a garden (96% reduction, [fig_ref] Table 6: Effect [/fig_ref]. Endothall decline was also slightly greater with sediment that was augmented with sediment slurry pre-exposed to endothall than when it was not augmented (Table 6; [fig_ref] Figure 2: Endothall isomer-1 [/fig_ref]. from the mesocosms. In contrast, 96% of monoamine endothall (with or without augmentation) had decayed by day-21, and 98% of dipotassium endothall had decayed by day-21 in the treatments with sediment added (Tables 5 and 6; [fig_ref] Figure 2: Endothall isomer-1 [/fig_ref]. The endothall remaining was lower when sediment was sourced from irrigation channels (97% reduction) than from a garden (96% reduction, [fig_ref] Table 6: Effect [/fig_ref]. Endothall decline was also slightly greater with sediment that was augmented with sediment slurry pre-exposed to endothall than when it was not augmented [fig_ref] Table 6: Effect [/fig_ref] ; [fig_ref] Figure 2: Endothall isomer-1 [/fig_ref]. Degradation of isomer-2 in monoamine endothall treatments was similar to isomer-1, i.e., it decayed only when sediment was added and to a greater degree when sediment was augmented with sediment slurry pre-exposed to endothall [fig_ref] Figure 2: Endothall isomer-1 [/fig_ref]. Notable differences were that (1) there was no effect of sediment on the concentration at day-0 (F = 1.71, p = 0.21), (2) the concentration of endothall remaining at day-21 was not affected by source of water or sediment , and (3) that isomer-2 decayed to zero when the sediment was augmented [fig_ref] Table 8: Effect of treatment on the percent decline of isomer-2, from day-0 to... [/fig_ref] , which did not occur for isomer-1 in any treatments [fig_ref] Table 6: Effect [/fig_ref]. . Analysis of variance for isomer-2 at day-21, as % decline at day-0. The data is analyzed after a rank with ties transformation on treatments with monoamine endothall. All p values are calculated using permutation tests on the F statistic on rank transformed data. DF = Degrees of freedom. Mean endothall isomer-1 and isomer-2 concentrations in the sediment were 0.07 (range 0.01-0.17), and 0.23 (range 0.04-0.45) mg ae L −1 , respectively, indicating that endothall did not accumulate in the sediment.
## Rate of endothall degradation
For the treatments that included sediment, degradation of endothall was characterized by a period of slow degradation followed by rapid degradation to near zero endothall [fig_ref] Figure 2: Endothall isomer-1 [/fig_ref]. In this situation, the time to first observing 25%, 50% and 75% of the observed degradation can be used to characterize both the onset of degradation and the rapidity of degradation during the rapid degradation period.
Degradation was earlier when sediment was augmented with microbes, by up to 4 days [fig_ref] Table 10: Analyses of variance for treatment effects on the first day in which... [/fig_ref]. Monoamine endothall with microbe augmentation decayed by 25% by day-5 for all water sources, while the treatments without augmentation (amine and dipotassium endothall) took 5-9 days to reach this threshold. Likewise, most treatments with pre-exposure to endothall reached 50 and 75% degradation 2-4 days before unaugmented treatments . Whilst potable water + garden sediment did not necessarily lead to earlier degradation, once the rapid degradation phase had commenced the degradation in potable water + garden sediment was substantially more rapid than in irrigation water. In fact, the predicted day for first observing 25% reduction was the same as the predicted day for first observing 75% reduction in all garden soil treatments . In general, dipotassium endothall reached these degradation thresholds at the same time as monoamine endothall, except in potable water + garden sediment, where it decayed to these thresholds two days later . Onset of degradation for isomer-2 occurred several days later than isomer-1 and decayed to trace concentrations only when augmented with sediment slurry pre-exposed to endothall [fig_ref] Figure 2: Endothall isomer-1 [/fig_ref]. . Effect of treatments on the first day in which endothall isomer-1 concentration was observed to decay by 25, 50 or 75% of its total decay. Analysis only includes mesocosms treated with sediment. Days are transformed to ranks with ties. BT = Back transformed. SED = Standard error of the difference.
# Discussion
The presence of the two isomers of endothall have not been reported previously. We have found these in similar relative proportions in all batches of dipotassium and monoamine endothall that we have characterized with LC-MS, from laboratory and field studies of endothall. The concentration of isomer-1 correlates with the concentration of endothall measured by ELISA (RaPID Assay ® Endothall Test Kit, Strategic Diagnostics Inc.; authors unpublished data) and correlates closely with the target concentration in the water column after dosing. The relative herbicidal potency of these two isomers remains unknown and it is puzzling that isomer-2 is present in a substantial proportion only in the monoamine formulation. To properly understand the degradation of endothall it is critical that we know what its component parts are and their degradation dynamics. Our data show that the two isomers showed clear differences in environmental fate, i.e., isomer-2 was much more persistent than isomer-1 [fig_ref] Figure 2: Endothall isomer-1 [/fig_ref]. However, our experiment was not designed to determine the degradation products of endothall and we saw no evidence that isomer-1 was converted to isomer-2. In common with isomer-1, the degradation of isomer-2 was influenced by presence of sediment and augmentation with microbes.
The environmental conditions in the mesocosms were aerobic and suitable for microbial growth, demonstrated by the biological activity that was recoded from the sediment substrate [fig_ref] Table 2: Water quality data taken over the duration of the experiment [/fig_ref]. Therefore, degradation of endothall should occur in all mesocosms, but persistence of endothall varied greatly according to the environmental factors that were manipulated in the mesocosms. The presence or absence of a sediment substrate had the largest influence on endothall persistence. When no sediment substrate was present there was no degradation detected over the 21 days, and this was consistent with all three water sources tested. We did not detect any accumulation of endothall in the sediment, therefore we surmise that the removal of endothall from the water column was due to a process of degradation, rather than adsorption to soil. Hiltibranfound that endothall (disodium salt) applied at 5 mg ae L −1 took 40-61 days to reach 0.1 mg L −1 in lake water in the absence of sediment, in contrast to 13 days when mud was added to the same lake water. The same study also compared tap water which alone took 21 days to reduce endothall from 1.0 to 0.1 mg L −1 but the same water took eight days with the addition of mud. Our results, which are based on a fully replicated experimental design that used LC-MS to determine endothall residues, corroborate the findings of Hiltibran, who determined disodium salt endothall residues from unreplicated mesocosms using a flaxseed bioassay method, and extend them to the two formulations of endothall that are currently used to manage aquatic weeds.
Since we concluded our experiment after 21 days, we are not able to suggest what would happen to the endothall in water-only treatments beyond this 21-day time frame. It is, however, obvious that the non-sterile water itself has very little or no effect on endothall degradation, and that endothall persists longer in water without the presence of a sediment substrate. Degradation will occur whenever sediment or soil is present, even when the soil is taken from a residential garden, suggesting either or both of the following: (1) The microbe(s) responsible for degradation are commonly present in soil and are cosmopolitan; [bib_ref] Probing mode of action in plant cell cycle by the herbicide endothall,..., Tresch [/bib_ref] there are a wide range of microbial taxa which can degrade endothall. Despite a number of studies that have investigated endothall degradation in the presence of a range of microbial inoculum from a range of sources [bib_ref] Persistence of endothall in the aquatic environment, Simsiman [/bib_ref] [bib_ref] Persistence of endothall in aquatic environment as determined by gas-liquid chromatography, Sikka [/bib_ref] , none have attempted to determine if the taxa responsible for degradation are the same when endothall is degraded in different habitats (or in mesocosms with constituents collected from different habitats). The only two studies that report on attempts to isolate the microbes responsible for degradation both found Arthrobacter are able to utilize endothall as their sole source of carbon [bib_ref] Metabolism of endothall by aquatic microorganisms, Sikka [/bib_ref] [bib_ref] Studies on soil bacteria (Arthrobacter globiformis) capable of decomposing the herbicide endothal, Jensen [/bib_ref]. The validity of postulates 1 and 2, above, can be verified by undertaking experiments to identify and isolate the taxa which are responsible for degradation (e.g., time series analysis of microbial community response, enrichment culture techniques and isotopic labelling).
The water collected from irrigation channels and used in these mesocosms was turbid (mean turbidity 22.0 NTU in both water sources) and contained suspended particles, so it was likely to contain a microbial population. Despite this, endothall degradation did not occur in the water column, in our study or that of Hiltibran. We do not know what aspect of the sediment allows degradation to occur, possible explanations include provision of microbial inoculum, a nutrient source, or a habitat substrate for microbes. Assuming degradation occurs on or in the sediment, this finding suggests the possibility that degradation will be proportional to the depth of water overlaying the sediment (i.e., degradation will be greater where the water volume to sediment surface area ratio is low, such as occurs in shallow water bodies).
Where degradation occurred, it began after a lag phase of 5-9 days, depending on treatment combination. Degradation was slow during the lag phase, followed by a relatively rapid decline in endothall concentration until a stable concentration was achieved. However, this low stable concentration was not always achieved in the 21-day duration of the experiment. This lag phase probably reflects a lag phase in the growth of the microbes responsible for degradation, which is typical of microbial growth. Holmberg and Lee [bib_ref] Effects and persistence of endothall in the aquatic environment, Holmberg [/bib_ref] reported a similar pattern of degradation of dipotassium endothall applied in a pond (average depth 1 m) at a rate of 5 mg L −1 . Endothall disappeared slowly for the first 12-13 days, followed by a sharp decline to less than 1 mg L −1 during the next few days, and reaching the level of detection in 18 days. A slightly different degradation pattern was reported by Sikka and Rice [bib_ref] Persistence of endothall in aquatic environment as determined by gas-liquid chromatography, Sikka [/bib_ref] , who found that endothall (dipotassium salt) concentrations decreased in the water column following treatment at 2 mg L −1 , with a corresponding increase in the top inch of the hydrosoil up to 22 days after treatment. Accumulation in the hydrosoil was rapid during the first 3 days, becoming more gradual through day 22; it was suggested that the initial rapid decline in the water column was because of sorption to the hydrosoil. Endothall disappeared completely from the hydrosoil in 44 days after treatment. In our study, dipotassium endothall concentration was lower by a small amount (0.25 mg ae L −1 ) a few hours after application for the mesocosms which contained sediment. It is possible that this was due to adsorption, like that reported by Sikka and Rice [bib_ref] Persistence of endothall in aquatic environment as determined by gas-liquid chromatography, Sikka [/bib_ref] , but this cannot be confirmed as we did not determine endothall residues in the sediment until day-7 and day-16 (at which times mean pore water concentrations were 0.09 and 0.05 mg L −1 , respectively, which were 6 and 33% of the concentration in the mesocosms).
Degradation of isomer-1 occurred at a similar rate between the two formulations, at least for irrigation water, although the dipotassium endothall persisted at a concentration lower than monoamine endothall (0.02-0.06 versus 0.1-0.2 mg ae L −1 , respectively, [fig_ref] Figure 2: Endothall isomer-1 [/fig_ref]. For monoamine endothall, this level is only marginally below the minimum dose specified on the product label (0.3 mg ae L −1 ;. Nevertheless, as found in this study, the general characteristics of degradation of dipotassium and monoamine endothall are similar. Only trace concentrations of isomer-2 were present in dipotassium endothall so we did not compare degradation of isomer-2 between the formulations. We do not know of any study that compares the degradation of the two formulations.
It is concerning that endothall did not degrade fully during this study because, if this also occurs in the field, it indicates there is potential for chronic, low concentration exposure to plants after field applications. This persistence is unlikely to be detected in field applications because dissipation and dilution will occur in natural water bodies. Nevertheless, further mesocosm studies of longer duration are required to determine how long it takes for endothall to degrade completely.
Augmentation with a sediment and water slurry from a tank previously exposed to endothall resulted in a significantly shorter lag phase and quicker onset of rapid degradation of isomer-1 of monoamine endothall, typically 2-4 days sooner than treatments without augmentation. With isomer-2, it was observed that the 21-day decline increased from about 50 to 100% with augmentation. It is likely that these effects occurred because the microbial taxa had evolved to utilize endothall during their single previous exposure to it. A similar finding has been reported, where bacteria capable of degrading endothall were isolated from soil as late as one year after treatment, but not from untreated soil [bib_ref] Persistence of endothall in aquatic environment as determined by gas-liquid chromatography, Sikka [/bib_ref]. These data indicate that a single exposure to endothall is enough for the microbial community to adapt to using endothall as a carbon source and that repeated use of endothall at a single location may result in faster decay. An undesirable aspect of this is that it could reduce endothall efficacy where long exposure times are important. This is supported by reports that endothall efficacy has reduced after successive application of endothall in some lakes in the USA, where adaption of the microbial communities in the lakes to use endothall as a carbon source has been hypothesized as a potential explanation (Michael Netherland, US Army Core of Engineers, pers. comm.). Where it is important to minimize off-site movement of endothall-treated water, faster degradation rates are desirable. If the microbes responsible for degradation of endothall can be isolated and cultured, then we may be able to pre-emptively inoculate areas with pre-adapted microbes to rapidly degrade endothall from these areas.
Temperature (together with moisture content) is the most important environmental factor affecting microbial growth and activity in soils [bib_ref] Comparison of temperature effects on soil respiration and bacterial and fungal growth..., Pietikäinen [/bib_ref]. As degradation of endothall is a biologically driven process, we expect that temperature will therefore also have a large influence on degradation rates. We are not aware of any other studies relating temperature and endothall degradation in the aquatic environment. However, several studies investigating the degradation of other microbially degraded herbicides when applied to soil report a positive relationship with temperature [bib_ref] Influence of temperature, soil moisture and soil characteristics on the persistence of..., Walker [/bib_ref] [bib_ref] Soil depth and temperature effects on microbial degradation of 2, 4-D, Veeh [/bib_ref] [bib_ref] Effect of temperature and moisture on the degradation and sorption of florasulam..., Krieger [/bib_ref]. Characterization of this for endothall will be the subject of further investigations.
From a management perspective, the precise degradation rate of endothall in a waterbody will be difficult to predict because degradation rate varies according to sediment source, pre-exposure to endothall, endothall formulation, and likely temperature. Despite this, the general trend of degradation and persistence were similar and predictable.
In our study, when sediment is present most isomer-1 endothall persisted in the water column for 7-14 days. While this persistence time is comparable with some previous studies [bib_ref] Effects and persistence of endothall in the aquatic environment, Holmberg [/bib_ref] [bib_ref] Persistence of endothall in aquatic environment as determined by gas-liquid chromatography, Sikka [/bib_ref] , reported persistence times vary widely (see reviews [bib_ref] Diquat and endothall: Their fates in the environment, Simsiman [/bib_ref] [bib_ref] Fate and persistence of aquatic herbicides, Reinert [/bib_ref]. Persistence of isomer-2 was much longer but there are no prior studies to compare it to. Field studies are necessary to confirm if results of this mesocosm study are applicable in the natural waterbodies that are normally deeper and more dynamic than laboratory mesocosms.
This study provides useful insight into safer and better aquatic plant management using endothall. The findings of this study can be used by waterbody managers and regulatory authorities to better understand endothall decay in a range of aquatic environments. This understanding will underpin strategies to manage endothall-treated water, so that it does not move into sensitive, downstream environments. Further, this understanding can be used by these organizations to initiate studies to resolve the critical knowledge gaps we have identified, e.g., the relative herbicidal properties of the isomers, the time required for complete degradation of endothall and the mechanisms involved in endothall decay.
# Conclusions
In conclusion, we provide direct evidence that the presence and characteristics of sediment are of key importance in the degradation of endothall in an aquatic environment, and we provide prima facie evidence that this importance is associated with the microbial community in the sediment. We have also identified that monoamine endothall has two sperate isomers that have different degradation characteristics. The herbicidal properties of these need to be investigated.
[fig] Figure 1: Extracted ion chromatogram of endothall, in two formulations. Endothall concentrations of isomer-1 are 10 mg L −1 in both cases. [/fig]
[fig] Figure 2: Endothall isomer-1 (a) and isomer-2 (b) concentrations in mesocosms for each treatment. Values represent average over three replicates. Plots represent various combinations of endothall formulation, source of water and sediment, presence or absence of sediment, and microbe augmentation. Note different y-axis scales between left and right panels. (1) Treatments containing dipotassium endothall without microbe augmentation. (2) Treatments containing monoamine endothall (without microbe augmentation). (3) Treatments containing monoamine endothall with sediment, +/− microbe augmentation. Note: a subset of treatments is shown in both (2) and (3). Legend provided in top right panel: Water sources represented by line style; endothall form (DE = Dipotassium endothall, ME = Monoamine endothall), sediment presence or absence (+/−sed) and preexposure to endothall (+/−aug), represented by shape and full of symbols. [/fig]
[table] Table 2: Water quality data taken over the duration of the experiment. Values are mean ± one standard deviation. EC = electrical conductivity; DO = dissolved oxygen. All treatments with the same combination of water and sediment are grouped. [/table]
[table] Table 3: Properties of the sediment used in the experiment. CEC = Cation exchange capacity. [/table]
[table] Table 4: Analysis of variance for treatment effects on endothall isomer-1 concentration at day-0. DF = Degrees of freedom. [/table]
[table] Table 5: Analysis of variance for treatment effects on endothall isomer-1 as % decline between day-0 and day-21. The data is analyzed after a log10(1.1-(%decline/100)) transformation. Treatment 14 is excluded from analysis (see methods). DF = Degrees of freedom. [/table]
[table] Table 6: Effect [/table]
[table] Table 8: Effect of treatment on the percent decline of isomer-2, from day-0 to day-21. SED = Standard error of the difference. [/table]
[table] Table 10: Analyses of variance for treatment effects on the first day in which endothall isomer-1 had decayed by 25, 50 or 75% of its total decay. Analysis only includes mesocosms with sediment. Days are transformed to ranks with ties. All p values are calculated using permutation tests on the F statistic on rank transformed data. DF = Degrees of freedom. [/table]
|
Effects of Glenoid and Humeral Bone Defects on Recurrent Anterior Instability of the Shoulder
## Effect of glenoid bone defects
Glenoid bone defects are known as one of the important risk factors for recurrence after surgical treatment of instability.The size of a glenoid bone defect should be preoperatively estimated to determine the optimal surgical procedure for recurrent shoulder instability. On the basis of the defect size, surgeons choose either an arthroscopic soft-tissue stabilization procedure or a bone graft procedure.However, the cutoff value for determining the surgical procedure remains controversial. The general consensus is that greater than 20%-25% glenoid bone loss requires a bony procedure such as the Latarjet procedure.Bigliani et al.first emphasized the importance of bone defects that affect more than 25% of the glenoid width in the choice of a treatment method for recurrent anterior shoulder instability; they recommended bony reconstruction procedures as primary surgical treatment. In a recent clinical study including 223 patients who underwent arthroscopic Bankart repair, the postoperative recurrence rate increased up to 21% when the glenoid defect was more than 20% of the glenoid width.However, some recent studies have reported that the critical cutoff value should be lower than 20%. 14-16) Shaha et al.evaluated clinical outcomes and recurrence rates among 72 patients with anterior shoulder instability divided according to the extent of bone defects. They suggested that a glenoid bone defect above 13.5% led to a clinically significant decrease in the Western Ontario Shoulder Instability score even in patients without postoperative recurrence. Another clinical study also proposed that the cutoff value of glenoid bone defects for surgical failure should be 17.3% based on the assessment of 169 patients with anterior glenoid erosion.In the study, compared to patients with a bone defect of less than 17.3%, patients with a glenoid bone defect of more than 17.3% showed a significantly higher rate of surgical failure. These results were also supported biomechanically by cadaveric studies. In a biomechanical study, Shin et al.simulated 10%, 15%, 20%, and 25% glenoid bone defects with osteotomies. After Bankart repair of each shoulder, they found that shoulders with a glenoid bone defect of 15% or more had significantly higher anterior glenohumeral translation than shoulders with less glenoid loss. Therefore, even a glenoid bone defect less than 20% should be considered as a critical amount of bone defect that requires surgical procedures.
## Effect of humeral bone defects
Humeral bone defects are frequently accompanied by glenoid bone defects especially in patients who undergo recurrent instability events. 1) Because of their interaction during shoulder abduction and external rotational movement, it is important to address these two lesions simultaneously. The glenoid track concept is an excellent tool to evaluate bipolar bone defects and predict postoperative recurrence by using three-dimensional computed tomography (3D CT) or magnetic resonance imaging.Metzger et al.showed that the glenoid track concept is a good preoperative predictor of humeral head engagement and may help to guide surgical decision making. They classified patients on the basis of the glenoid track concept and then compared with clinical evidence of humeral head engagement found during arthroscopic examination. Among patients with an off-track lesion, 84.5% had clinical evidence of engagement, whereas only 12.4% of patients with an on-track lesion showed clinical engagement. Many studies regarding the glenoid track concept demonstrated that the off-track lesion is a strong predictor of postoperative recurrence. One clinical study showed a 75% recurrence rate in eight patients with off-track lesions and only an 8% recurrence rate in 49 patients with ontrack lesions.The positive predictive value of off-track measurement for postoperative recurrence was 75%-the value was more than 20% higher than that of glenoid bone defects (44%), although the sample size of that study was relatively small. In another study evaluating 100 patients who underwent an arthroscopic stabilization procedure, the postoperative recurrence rate was significantly higher in patients with off-track lesions (33%) than in patients with on-track lesions (6%). 22) Therefore, Di Giacomo et al.recommended arthroscopic Bankart repair with additional remplissage procedure in patients with off-track lesions if glenoid bone loss was less than 20%. Park et al.evaluated clinical outcomes of 23 patients with off-track lesions after arthroscopic Bankart repair combined with additional remplissage procedure. 23) They performed the additional remplissage procedure only in patients with an engaging humeral head in arthroscopic examination after Bankart repair with capsular plication. As a result, compared with patients with on-track lesions, patients with off-track lesions showed satisfactory clinical outcomes and recurrence rates after arthroscopic Bankart repair with the remplissage procedure. They emphasized the greater importance of the glenoid bone defect than the off-track lesion as a predictor of postoperative surgical failure.
There would be some weaknesses of the glenoid track concept, because on-or off-track lesions are determined only by the bone defect size of the humeral head and glenoid although soft-tissue conditions such as cap-Clinics in Orthopedic Surgery - Vol. 12, No. 2, 2020 - www.ecios.org sulolabrum and glenohumeral ligaments are also important factors for postoperative stability.In patients with high-quality anterior capsule, the anterior capsule usually becomes tight and the humeral head rotates internally after Bankart repair with capsular plication. The location of the Hill-Sachs lesion is then altered far from the glenoid anterior margin, leaving little chance to reach the glenoid anterior margin. On the other hand, in patients with poorquality anterior capsule, this phenomenon would not occur even after proper capsular plication. The shape of Hill-Sachs lesions would also affect the glenoid track concept. In patients with a wide and shallow Hill-Sachs lesion, there would be little possibility of humeral head engagement, even if it is an off-track lesion by the glenoid track concept.
## Treatment strategy for recurrent anterior instability with bone defects
The size of a glenoid bone defect could be calculated on the en face view of preoperative 3D CT by using the Sugaya method.Surgeons should determine whether a bony procedure is needed on the basis of the size of a glenoid bone defect. The critical value of a glenoid defect size has been widely considered as 20%-25%; however, that value should be lowered to 13%-17% according to recent clinical and biomechanical studies.In the treatment of patients with borderline glenoid bone defects, surgeons should discuss surgical options with patients because compared with arthroscopic Bankart repair, bony procedures have inherent advantages and disadvantages. In a recent study comparing clinical outcomes after the Latarjet procedure and arthroscopic Bankart repair in patients with a borderline glenoid bone loss of 15%-20%, both procedures provided satisfactory clinical outcomes and pain relief; however, the recurrence rate was significantly lower after the Latarjet procedure than arthroscopic Bankart repair.For patients with more than 20%-25% glenoid bone loss, many clinical studies recommend bony procedures such as the Latarjet procedure because of the low recurrence rate: 5%-17% after the Latarjet procedure vs 11%-16% after the arthroscopic stabilization procedure.According to a meta-analysis study comparing clinical outcomes and recurrence rates after the Latarjet procedure and Bankart repair, the Latarjet procedure offered greater postoperative stability although differences in the glenoid bone defect size were not considered in the study. However, compared with arthroscopic stabilization surgery, the bony procedure has been associated with a higher incidence of surgical complications. One meta-analysis reported a 13.4% overall complication rate after glenoid bone grafting, such as iatrogenic nerve palsy, graft nonunion, fracture, hematoma formation, and screw loosening.Compared to the Latarjet procedure, arthroscopic Bankart repair showed a lower incidence of postoperative complications (0%-3.1%) even when combined with a complex bone incorporation procedure.Therefore, patient's demand and physical activity level should be considered first when the surgeon determines surgical options in patients with a glenoid defect around 20%. A recent clinical analysis provided satisfactory clinical outcomes after an arthroscopic stabilization procedure combined with the remplissage procedure, which is considered as a primary treatment option even in patients with a glenoid bone defect more than 20%.The authors of the study proposed that compared to nonanatomic reconstruction such as bony procedures, anatomical reconstruction using soft-tissue procedures including labral repair with capsular plication would result in less complications.
For the Hill-Sachs lesion, the engagement into the anterior edge of the glenoid was the important factor for determining the necessity of additional procedures such as the remplissage procedure in the past. Nowadays, treatment of the Hill-Sachs lesion is based on the combined glenoid defect. It would be important to determine whether the patient has an on-or off-track lesion. According to the glenoid track concept, the width of the glenoid track and the width of the Hill-Sachs lesion are used to determine whether the patient has an on-or off-track lesion.
The width of the glenoid track is calculated as 83% of the normalized glenoid width minus the glenoid bone defect width.If the width of the Hill-Sachs lesion is greater than the width of the glenoid track, the patient is considered to have an off-track lesion, whereas an on-track lesion is considered present in the opposite case. Generally, in patients with recurrent instability who have an on-track lesion and less than 20% glenoid bone loss, isolated Bankart repair is recommended; the additional remplissage procedure is recommended in patients with an off-track lesion and less than 20% glenoid bone loss.Bone graft procedures are necessary in patients who have a glenoid bone defect of more than 20% regardless of the glenoid track lesion. This treatment algorithm was supported by a recent study comparing clinical outcomes after arthroscopic Bankart repair with selective remplissage procedure between patients with and without off-track lesions.In the study, patients with an off-track lesion showed clinical outcomes and recurrence rates comparable to those in patients with an ontrack lesion after arthroscopic Bankart repair with selective remplissage procedure. Selective remplissage procedure, performed only when engagement of the humeral head is observed in arthroscopic examination after arthroscopic Bankart repair with capsular plication, should be also considered as a treatment option for off-track lesions.
# Conclusion
In the treatment of recurrent anterior instability of the shoulder, the proper treatment strategy is determined on the basis of the evaluation of glenoid and humeral bone defects. However, there are various measurement methods, and the critical value of glenoid and humeral bone defects to determine surgical procedures has yet to be established. Therefore, the treatment strategy for recurrent anterior instability of the shoulder should be determined more flexibly by the patient's condition. For more individualized treatment, surgeons should discuss surgical options with patients, considering their demand and physical activity level. |
Noli Timere: The Role of Reassuring Adults in Dealing with COVID-19 Anxiety in Pediatric Age
# Introduction
In a recent study [bib_ref] Anxiety in Older Adolescents at the Time of COVID-19, Smirni [/bib_ref] unexpected high level of 'state-anxiety' was recorded in a group of Italian healthy older adolescents during the most restrictive period of the Corona Virus Disease-19 (COVID- [bib_ref] Communication with children and adolescents about the diagnosis of a life-threatening condition..., Dalton [/bib_ref] pandemic. Surprisingly, the highest anxiety symptoms were breathing difficulties. Therefore, since the COVID-19 mainly affects respiratory functions and the sample consisted of healthy non-clinical subjects, and the assessment tools measured state and non-trait anxiety, the observed high level of anxiety was associated with a temporary feeling of apprehension that favors an increase in anxiety responses, as during the period of spread of COVID-19.
Over the past twenty years, a large body of literature had widely documented emotional reactions in children who experienced pandemic and subsequent disease-containment measures (e.g., the Severe Acute Respiratory Syndrome (SARS) outbreak; Hemagglutinin Type 1 and Neuraminidase Type 1 (H1N1-Avian Influenza A) strain; Ebola virus) [bib_ref] Risk Perception and Compliance with Quarantine during the SARS Outbreak, Cava [/bib_ref] [bib_ref] The psychological impact of quarantine and how to reduce it: Rapid review..., Brooks [/bib_ref] [bib_ref] The immediate psychological and occupational impact of the 2003 SARS outbreak in..., Maunder [/bib_ref] [bib_ref] The Role of Fear-Related Behaviors in the 2013-2016 West Africa Ebola Virus..., Shultz [/bib_ref].
Taken together, several recent studies from China reported that the COVID-19 pandemic can worsen existing mental health problems, and destabilize emotionally fragile children and adolescents [bib_ref] Behavioral and Emotional Disorders in Children during the COVID-19 Epidemic, Jiao [/bib_ref] [bib_ref] The psychological impact of the COVID-19 epidemic on college students in China, Cao [/bib_ref] [bib_ref] Analysis of influencing factors of anxiety and emotional disorders in children and..., Li [/bib_ref] especially in the emotional dimension [bib_ref] Behavioral and Emotional Disorders in Children during the COVID-19 Epidemic, Jiao [/bib_ref] [bib_ref] Online mental health services in China during the COVID-19 outbreak, Liu [/bib_ref] [bib_ref] Incremental validity of coronaphobia: Coronavirus anxiety explains depression, generalized anxiety, and death..., Lee [/bib_ref] [bib_ref] Mental Health Status Among Children in Home Confinement During the Coronavirus Disease, Xie [/bib_ref] [bib_ref] Clinical course and risk factors for mortality of adult inpatients with COVID-19..., Zhou [/bib_ref] [bib_ref] Understanding the school community's response to school closures during the H1N1 2009..., Braunack-Mayer [/bib_ref] [bib_ref] A Support Group for Home-Quarantined College Students Exposed to SARS: Learning from..., Pan [/bib_ref] [bib_ref] The mediator role of self-disclosure and moderator roles of gender and social..., Wang [/bib_ref] Moreover, it is very likely that it should have long-term psychological effects, even when everyone's life is back to normal [bib_ref] Online mental health services in China during the COVID-19 outbreak, Liu [/bib_ref] [bib_ref] Incremental validity of coronaphobia: Coronavirus anxiety explains depression, generalized anxiety, and death..., Lee [/bib_ref].
Children and adolescents suddenly have to face an unknown and incomprehensible faceless enemy, which force them to distance themselves from their peers [bib_ref] Incremental validity of coronaphobia: Coronavirus anxiety explains depression, generalized anxiety, and death..., Lee [/bib_ref] to drastically change their daily reassuring habits, and to lose all certainty for their future and loved ones. Moreover, they experience the emotions, insecurity, fear and worries of adults, although the younger they are, the less they are able to cognitively understand what is happening around them. However, they feel emotionally that something very important is happening and that many behaviors around them are inexplicably changing [bib_ref] Disaster Preparation and Recovery: Lessons from Research on Resilience in Human Development, Masten [/bib_ref] [bib_ref] Communication with children and adolescents about the diagnosis of a life-threatening condition..., Dalton [/bib_ref] [bib_ref] Parenting as primary prevention, Perrin [/bib_ref]. Probably, widespread anxiety, uncertainties of parents and familyfear of being infected, long-term home isolation, and forced removal from schoolmates and relatives have supported an increase in anxious responses [bib_ref] The interior life of the family: Looking from the inside out and..., Fiese [/bib_ref].
However, from the earliest stages of the spread of COVID-19, the elderly has been identified as the most vulnerable people at risk of contagion and fatal complications. Consequently, health authorities around the world have rightly focused on such frailest population. Currently, the debate on the effects of the pandemic and restrictive measures are still focused on the elderly population, the general health organization and the economic productivity.
Minor attention was paid to pediatric populations and their emotional reactions, and despite children and adolescents faced severe anxiety, fear and stress, become a neglected minority.
On these assumptions, pandemic planning must take into account specific strategies to address the behavioral and emotional reactions of the pediatric populations, ensuring that they do not experience long-term trauma from either the pandemic disease or public health response strategies.
# Aims and methods
The current narrative review aimed to report the scientific literature evidence focused on emotional reactions to pandemic in the pediatric population and on the role of adults in containing COVID-19 pandemic anxiety. Such a review, according to recommendations by World Health Organizationmay be of the highest importance to inform health authorities in order to plan measures, including rapid guidelines, to contain and mitigate the impact of the pandemic on children's mental health and adolescents.
To this end, around a hundred articles regarding recent epidemics have been reviewed following the search criteria by keywords such as: management of anxiety during epidemics; psychological and behavioral problems during epidemic stress; anxiety disorders in COVID-19 time; pandemic anxiety; pediatric populations state-anxiety; child-parent communication; parental stress management. Among these, about forty recent articles on the COVID-19 pandemic have been further investigated, because they were specific to the pediatric age and related to the current pandemic. References were identified through electronic database searching in Ovid MEDLINE, Embase, PsycINFO, Scopus and Web of Science.
Studies inclusion criteria were as follows: (1) empirical studies and reviews; (2) written in English; (3) data on psychological factors collected during epidemics and in COVID-19 pandemic; (4) sample < 18 years of age; (5) sample of parents of children/adolescent < 18 years of age; (6) data on the prevalence of anxiety symptoms and/or factors associated with anxiety and fear management; (7) studies focused on management and possible interventions to contain children's anxieties, to help them to express themselves and communicate their experiences from the epidemic; and (8) studies that showed dealing with the effects of COVID-19 stress. 21 articles were excluded because they not refer to key topics directly, nor to health emergencies, or if full-text copies could not be obtained. General risk communication materials, such as pamphlets, posters, and infographics, were excluded as they do not provide evidences about their effectiveness. Lack of transparency due to missing methodology information was also grounds for exclusion. [fig_ref] Figure 1: Selection of studies flowchart [/fig_ref] shows the selection of studies flowchart.
The final database search was run on September 2020.
# Results
After an extensive literature search, twelve studies have been thorough. Tables 1 and 2 schematically summarize the observations shown by these studies: [fig_ref] Table 1: Studies on the psychological impact on children and adolescents in pre-COVID-19 time [/fig_ref] shows the studies that investigated the psychological impact on children and adolescents during the pandemics predating COVID-19; [fig_ref] Table 2: Studies on the psychological impact on children and adolescents during COVID-19 [/fig_ref] shows the studies that investigated the psychological impact on children and adolescents during COVID-19 pandemic.
# Results
After an extensive literature search, twelve studies have been thorough. Tables 1 and 2 schematically summarize the observations shown by these studies: [fig_ref] Table 1: Studies on the psychological impact on children and adolescents in pre-COVID-19 time [/fig_ref] shows the studies that investigated the psychological impact on children and adolescents during the pandemics predating COVID-19; [fig_ref] Table 2: Studies on the psychological impact on children and adolescents during COVID-19 [/fig_ref] shows the studies that investigated the psychological impact on children and adolescents during COVID-19 pandemic.
## Psychological impact on children and adolescents in pre-covid-19 time
The literature of epidemics, in pre-COVID-19 time, has already documented the psychological dimension in epidemics and, namely, the emotional reactions of children and adolescents and the protective factors that can mitigate the psychological impact of the epidemic and containment measures such as social isolation.
A cross-sectional studyinvestigated post-traumatic stress reactions to pandemic and disease-containments in children and parents in areas (USA and Canada) severely affected by H1N1 or SARS. Using a mixed method approach (survey, focus groups, and interviews), data from a group of 398 parents were collected. Isolated children (about 30%) and their parents (about 25%) met criteria for post-traumatic stress disorder (PTSD). Posttraumatic stress scores were four times higher in isolated children than in not quarantined, while 28% of parents quarantined reported sufficient symptoms of a trauma-related mental health disorder, compared with 6% of parents who were not quarantined.
Very interesting, a strong relationship was found between PTSD symptoms in parents and children. Almost 86% of parents who met the clinical symptoms for PTSD had children who also met the clinical cut-off score. Moreover, the PTSD was confirmed by public health services, in more than two-thirds of subjects. Such evidence suggested that pandemic disasters and disease-containments are affecting the lives of children and their families also socially and psychologically. Therefore, public health programs should consider the PTSD in parents and children to minimize the risk of adverse biopsychosocial consequences.
A review, conducted by a multidisciplinary professional team [bib_ref] The Role of Fear-Related Behaviors in the 2013-2016 West Africa Ebola Virus..., Shultz [/bib_ref] provides a conceptual overview of the role of fear-related behaviors and their potential impacts on epidemiologic outcomes during the 2013-2016 West Africa Ebola virus disease outbreak. The authors showed that fear and related behaviors can amplify the psychological impact of the epidemic and reduce the ability to cope with it. The invisible viral agent can create anxious uncertainty regarding risk, exposure, and infection that can lead to intense fear and dread. Consequently, fear and fear-related behaviors may play an important role in the spread of the epidemic, in the use of life-saving therapeutic measures and in increasing the risk for psychopathological behaviors. Therefore, the authors pointed out the role of effective communication to inform without spreading panic and of psychological support, especially for the emotionally fragile subjects.
Three studies underline the emotional effects of the separation of children and parents which, although may be inevitable as a protective measure of the infection, may result in depressive reactions in children, parents and even health workers [bib_ref] When family-centered care is challenged by infectious disease: Pediatric health care delivery..., Koller [/bib_ref] [bib_ref] Paediatric pandemic planning: Children's perspectives and recommendations, Koller [/bib_ref] The studies suggest to treat child as a person in relation to the family as the primary source of strength and support [bib_ref] Pediatric epidemic crisis: Lessons for policy and practice development, Nicholas [/bib_ref].
In a wide programme of research, Koller et al. [bib_ref] When family-centered care is challenged by infectious disease: Pediatric health care delivery..., Koller [/bib_ref] [bib_ref] Paediatric pandemic planning: Children's perspectives and recommendations, Koller [/bib_ref] examined, according to an ethnographic qualitative approach, the experiences of a sample of children affected by SARS, their parents, and pediatric health care providers, within a Canadian pediatric hospital.
Using an in-depth interview method, all participants were asked to evaluate their experience, the way they deal with the pandemic and the infection control measures, and to provide suggestions for future outbreaks. Moreover, children and parents were asked specific questions about hospitalization and separation, while health care professionals were asked to describe work-related experiences while caring for SARS patients and the impact on their personal lives.
Both children, parents, and health care workers expressed feelings of emotional distress, sadness, loneliness, worry, fear and helplessness. The prevailing theme was the psychological impact of separation and isolation. Several health care providers expressed the emotional impact they felt by observing children who were separated from their families. Other recurring themes were communication difficulties due to isolation, the use of protective masks and clothing, and the limitation of family visits together with the loss of parental control and the changes in parental and professional roles. Parents expressed discomfort for the impossibility to care for their child, while the care workers covered parental role for children.
The children's narratives centered on four aspects, namely, increased attention to emotional reactions, collective responsibility for infection control, more effective communications and, finally, proper resources management. Additionally, the children stressed the importance of having consistent healthcare providers and allowing for therapeutic interventions that included supportive discussions and opportunities for normal play and activity. Based on these findings, the authors recommend improving patient participation by sharing information, recognizing children's emotional reactions, providing parents with regular information about their child's condition, and enabling them to contribute to decision making.
Nicholas et al. [bib_ref] Pediatric epidemic crisis: Lessons for policy and practice development, Nicholas [/bib_ref] address, in a pediatric perspective, health policy and practice implications resulting from SARS impacted on the health care facilities in Canada.
A series of semi-structured, descriptive qualitative interviews were conducted, two months after hospitalization, to examine the experiences, impacts, and implications of epidemic outbreak related health care policies for 23 participants: pediatric SARS patients, their parents, and frontline pediatric health care providers. Interview questions invited participants to identify their experiences, perceived policy and practice implications of SARS, and lessons for future outbreak.
Participants highlighted some key issues, including the development of communication strategies; releasing the vulnerability among all the subjects involved; and the development of practical guidelines. Therefore, planning strategies in pediatrics should include not only rapid containment of viruses, but also continuous and coordinated communication, and humane and accessible care.
Three very exciting studies document two different therapeutic approaches to mitigate the psychological effects of the epidemic: a support therapeutic group [bib_ref] A Support Group for Home-Quarantined College Students Exposed to SARS: Learning from..., Pan [/bib_ref] and an expressive arts program [bib_ref] A process description of playing to live! A community psychosocial arts program..., Decosimo [/bib_ref] [bib_ref] Playing to live: Outcome evaluation of a community-based psychosocial expressive arts program..., Decosimo [/bib_ref].
In line with these observations, there is also the evidence of an observational study on a little group of SARS home-quarantined college students in Taiwan [bib_ref] A Support Group for Home-Quarantined College Students Exposed to SARS: Learning from..., Pan [/bib_ref]. The authors described an experience with a support group lasted five sessions for a total of 500 min that included lectures, group tasks and activities, handouts, here-and-now interaction, and discussions. At an early stage, different structured relational activities were used as ice-breaking activities and to promote mutual knowledge. Subsequently, information was provided on the SARS pandemic and containment measures, and finally a structured activity was proposed in which everyone had to present 'the peaks and valleys' and name the main events of their life. In the second stage, the discussion centered on the concerns of the participants. They were invited to describe one's life events that expressed different emotions (happiness, sadness, anger, fear, shame, helplessness). Many students initially pointed out that they did not like to talk about their feelings. On the contrary, they admitted that they no longer had strong feelings and that they felt rather bored when talking about the isolation. Indeed, such reduced emotional resonance profoundly impacted on their joie de vivre, their relationship life and their projection into the future. However, gradually all the members of the group described their personal experiences showing growing awareness and management of emotions. In the final stage, the participants were asked to write positive thoughts, hopes and blessings for all other participants. The thoughts were read and discussed into the group.
The authors concluded that the support group proved a powerful way for students to connect when they felt most vulnerable. Each could process their own experience by sharing it with others and by accepting the emotions and similar experiences of others. In addition, the ability to receive direct information on updated SARS projects has reduced early anxiety and developed a new perspective to fully engage in campus social activities.
Decosimo et al. [bib_ref] A process description of playing to live! A community psychosocial arts program..., Decosimo [/bib_ref] [bib_ref] Playing to live: Outcome evaluation of a community-based psychosocial expressive arts program..., Decosimo [/bib_ref] analyse the efficacy of a therapeutic psychosocial expressive arts program ('playing to live') focused on the growing psychosocial and mental health needs of children from three countries in West Africa who experienced a large Ebola epidemic.
Two groups of a total of 870 children and adolescents (aged 3-18 years) who were Ebola-survivors were enrolled in an expressive arts program 'playing to live' for a 5-month or 3 months with the goal of supporting reintegration and decreasing stigma. The program's framework was that expressive arts provide for children a safe space to express themselves and communicate their experiences from the epidemic, learning trauma coping skills, exploring relationships and emotionsand giving meanings to their chaotic environment, confusion and fear through creativity, mentorship, and peer support.
The activities instructed children to use art, play, and storytelling to explore what they want, for example, for their future, and to help them to identify positivity within their current situation and build hope and future goals [bib_ref] A process description of playing to live! A community psychosocial arts program..., Decosimo [/bib_ref].
Results indicated that both treatment groups reached significant responses for the decrease of psychological stress symptoms, suggesting the urgent need for psychosocial support programming after a trauma. According to the authors, children who have access to psychosocial support and resources after trauma have a higher potential for recovery and resiliency [bib_ref] Psychological aspects of traumatic injury in children and adolescents, Caffo [/bib_ref] and reducing trauma stress symptoms.
## Psychological impact on children and adolescents during covid-19
Studies on the emotional dimension and protective factors of children and adolescents in the period of COVID-19 are still quite sparse, also due to the methodological limitations that do not allow adequate control groups, comparative baseline data and longitudinal research designs [bib_ref] Child and adolescent mental illness during COVID-19: A rapid review, Racine [/bib_ref].
A previously cited study [bib_ref] Anxiety in Older Adolescents at the Time of COVID-19, Smirni [/bib_ref] investigated state anxiety and emotion awareness in a healthy Italian sample of older adolescents during the pandemic lockdown, using the Self-rating Anxiety Scale (SAS) and the Italian Emotion Awareness Questionnaire. Over half of the SAS individual items reached a high anxiety score, and consequently the SAS total score reached an unusually high anxiety score. Analysing the single items, the item recording the highest score was item of breathing difficulties. Likewise, items referring to sleep disorder, anxiety, panic and a negative expectation of the future reached high average scores. Since the sample was a healthy, non-clinical one and the SAS measured state and non-trait anxiety, the unusually high anxiety scores observed would not appear to be attributed to the sample's stable emotional functioning, but it is likely to be due to a temporary condition or feeling of tension and apprehension that favors a leavening of anxious responses.
These findings supported the hypothesis that the COVID-19 pandemic may be a risk condition for an increased state-anxiety in older adolescents and suggested the need to provide 1. an effective, empathic communication system with the direct participation of older adolescents, 2. a psychological counselling service for the stress management of adolescents.
Four different studies from China examined large groups of children and adolescents and found significant levels of anxiety and depression or somatic disorders (bodily aches, pains or difficulties breathing) in response to the COVID-19 pandemic [bib_ref] Behavioral and Emotional Disorders in Children during the COVID-19 Epidemic, Jiao [/bib_ref] [bib_ref] Online mental health services in China during the COVID-19 outbreak, Liu [/bib_ref] [bib_ref] Mental Health Status Among Children in Home Confinement During the Coronavirus Disease, Xie [/bib_ref] [bib_ref] Clinical course and risk factors for mortality of adult inpatients with COVID-19..., Zhou [/bib_ref].
A cross-sectional survey [bib_ref] Somatic symptoms and concern regarding COVID-19 among Chinese college and primary school..., Liu [/bib_ref] explored the impact of the COVID-19 pandemic on somatic symptoms and concerns about pandemic in a sample of Chinese college students (n = 198) and of primary school students (n = 209), using the Somatic Self-rating Scale and a questionnaire aimed at examining three different apprehensions related to the Covid-19: daily needs, effectiveness of prevention and control measures, the threat to life and health for both the participants and their families.
Concerns about the outbreak were associated to somatic complaints in both groups, although the psychological impact was different. Primary school children were mainly concerned about the threat to life and health, and this concern was associated with anxiety and somatic symptoms. However, they had a lower incidence of somatic symptoms than college students (2.39% vs. 34.85%). College students were concerned for all the three domains, and all three of the concerns rated were linked to anxiety and depression and to somatic complaints. The authors argued that the differences between groups could be related to the fact that primary school students are still protected by parents who attend to their daily needs and control measures, while the majority of college students are independent and, as result, they tended to worry about all aspects of the pandemic and exhibit more somatic symptoms than primary school children. However, it is interesting to note that college students in the pandemic condition appear emotionally fragile, despite the independence they have achieved from their parents.
Therefore, data support the need to differentiate psychological measures in relation to different ages. Specifically, the authors pointed out that psychological health of children should begin with an appropriate health education for parents to protect their children from psychological distress [bib_ref] Testing, and Treatment of U.S. Health Care Personnel: Recommendations from the National..., Sosa [/bib_ref]. Similarly, for college students, the authors emphasize the importance of direct health education aimed at improving knowledge of COVID-19 to promote prevention and control measures [bib_ref] Prevalence of self-reported depression and anxiety among pediatric medical staff members during..., Chen [/bib_ref] [bib_ref] Mental health toll from the coronavirus: Social media usage reveals Wuhan residents'..., Zhong [/bib_ref].
Xie et al., [bib_ref] Mental Health Status Among Children in Home Confinement During the Coronavirus Disease, Xie [/bib_ref] in a cross-sectional study, investigated depressive and anxiety symptoms, by the Children's Depression Inventory-Short Form and the Screen for Child Anxiety Related Emotional Disorders, among a large group of primary school children (2330) in home confinement during the COVID-19 in a Province of China.
Overall, a percentage of students higher than other investigations in primary schools of China [bib_ref] Prevalence of depressive symptoms in primary school students in China: A systematic..., Xu [/bib_ref] reported depressive (22.6%) and anxiety symptoms (18.9%).
Interestingly, the levels of anxiety and depression were influenced by positive or negative attitudes towards the epidemic. Students who were slightly or less concerned about being affected by COVID-19 showed lower anxiety scores and lower depressive symptoms. Conversely, students who were not optimistic about the outbreak they showed higher anxiety scores and depressive symptoms. Similarly, anxiety and fear of contagion and feelings of life threat were higher in highly infected areas. No significant association was found between demographic characteristics and anxiety symptoms. Sex did not predict anxiety and depression. According to the authors, the increase in children's depressive symptoms must be associated with the reduction of outdoor activities and social relationships.
In a cross-sectional study, Zhou et al. [bib_ref] Clinical course and risk factors for mortality of adult inpatients with COVID-19..., Zhou [/bib_ref] using an online survey, assessed a very large number of students (8079), aged 12-18 years. Socio-demographic information and students' awareness of COVID-19 were collected. Participants were asked about their familiarity with COVID-19 prevention and control, whether they had taken all prevention and control measures to avoid contagion, and about their opinions towards the COVID-19 trend projections. Depressive and Anxiety symptoms were assessed by the Patient Health Questionnaire [bib_ref] Validation and utility of a self-report version of PRIME-MD: The PHQ primary..., Rl Skroenke [/bib_ref] and the Chinese version of the Generalized Anxiety Disorder scale [bib_ref] A brief measure for assessing generalized anxiety disorder: The GAD-7, Spitzer [/bib_ref].
Overall, the prevalence of depressive and anxiety symptoms was 43.7% and 37.4%, respectively. Age was significant, the higher the grade, the greater the prevalence of depressive and anxiety symptoms. Senior high school was a risk factor, confirming Liu's data [bib_ref] Online mental health services in China during the COVID-19 outbreak, Liu [/bib_ref] documenting greater emotional fragility in college students compared to primary school students. Sex also was significant, as girls showed a higher level of anxiety and depression than boys. Similarly, anxiety and depressive symptoms in rural areas were significantly higher than in cities. This was consistent with previous studies, which found nearly twice as much emotional disorders among the poor as the rich [bib_ref] Culture and depression, Kleinman [/bib_ref] [bib_ref] Child and adolescent mental health problems in Tyva Republic, Russia, as possible..., Slobodskaya [/bib_ref]. Interestingly, the level of knowledge and awareness about COVID-19 and positive attitude towards the development of epidemic appeared as protective factors against depressive and anxiety symptoms. This was consistent with previous studies showing that wearing a mask and practicing hand hygiene reduces the level of anxiety and depression [bib_ref] Immediate Psychological Responses and Associated Factors during the Initial Stage of the..., Wang [/bib_ref]. According to the authors, provide accurate information, such as the progress of medicines and vaccines, and refute alarming news can reduce anxiety levels, and favor a better emotional stability [bib_ref] Immediate Psychological Responses and Associated Factors during the Initial Stage of the..., Wang [/bib_ref].
Age-related differences have been documented by a cross-sectional study conducted by Jiao et al., [bib_ref] Behavioral and Emotional Disorders in Children during the COVID-19 Epidemic, Jiao [/bib_ref] among 320 children and adolescents, aged 3-18 years, under protective isolation, using an online questionnaire, completed by the parents. Clinging, inattention, and irritability were the most severe symptoms showed by the sample in all age. The younger group (3-6 years) was more likely than older to manifest symptoms, such as clinginess and fear that family members could be infected, while children aged 6 to 18 years were more likely to show inattention and persistent inquiry. Moreover, the study confirmed Xie's data [bib_ref] Mental Health Status Among Children in Home Confinement During the Coronavirus Disease, Xie [/bib_ref] that fear and anxiety were higher in children residing in highly epidemic areas.
Based on the data analysis, the authors pointed out that resilience should be implemented in children and teens. Parents may support the child by increasing communication to address their fears and concerns, playing collaborative games to alleviate loneliness, encouraging physical activity, and using music, media entertainment, reading, and singing to reduce the worry, fear, and stress [bib_ref] Çullu Fostering Resilience in Children: The Essential Role of Healthcare Professionals and..., Pettoello-Mantovani [/bib_ref]. Moreover, the authors suggest to pay attention to sleep difficulties and nightmares, prevent increased daytime sleep and model a positive psychological attitude to reduce stress, and focus on positive activities.
# Discussion
The focus of the current overview has been to examine research studies, over the past two decades, reporting on the effects of COVID-19 and previous epidemic events on emotional responses in the pediatric population and on the role of adults in containing COVID-19 pandemic anxiety.
Overall, the reviewed studies, both in the pre-COVID-19 and in the COVID-19 time, all agree in supporting the devastating action of epidemic events and containments measures on psychological well-being and mental health of childhood and adolescence [bib_ref] Child and adolescent mental illness during COVID-19: A rapid review, Racine [/bib_ref] [bib_ref] Emotional-Behavioral Disorders in Healthy Siblings of Children with Neurodevelopmental Disorders, Caliendo [/bib_ref] [bib_ref] Noli Me Tangere: Social Touch, Tactile Defensiveness, and Communication in Neurodevelopmental Disorders, Smirni [/bib_ref]. Moreover, the studies broadly showed that parents' reassurance behaviours and social support may mitigate the negative effects of epidemics on stress response in developmental age [bib_ref] Stress hormones and human memory function across the lifespan, Lupienand [/bib_ref] [bib_ref] Maternal care during infancy regulates the development of neural systems mediating the..., Caldji [/bib_ref] [bib_ref] Environmental Enrichment Reverses the Effects of Maternal Separation on Stress Reactivity, Francis [/bib_ref].
The recurrent themes in pre-COVID studies were the isolation, social separation and the loss of reassuring communication and information. In all studies, prolonged separation from significant adults has been for children and adolescents a very serious experience of loss and rejection, resulting in mourning depressive reactions, anxiety and fear. The communication between children, relatives, peers, friends and parents became extremely complex. Isolated children and parents achieved significantly higher post-traumatic stress levels than children or parents not in isolation.
Probably, as evidenced by a multidisciplinary overview [bib_ref] The Role of Fear-Related Behaviors in the 2013-2016 West Africa Ebola Virus..., Shultz [/bib_ref] fear and fear-related behaviors could amplify the psychological impact of the epidemic and support a sort of circularity of anxiety, reducing the ability to cope with it. Therefore, fear and fear-related behaviors may play a role in increasing the risk for psychopathological behaviors. The authors pointed out the main role of effective communication to inform without spreading panic.
Interestingly, research reported that most parents reaching significant levels of posttraumatic stress have children with high levels of stress and vice versa. The limitation of visits or even their complete prohibition increased the experience of loneliness and helplessness [bib_ref] When family-centered care is challenged by infectious disease: Pediatric health care delivery..., Koller [/bib_ref] [bib_ref] Paediatric pandemic planning: Children's perspectives and recommendations, Koller [/bib_ref] [bib_ref] Pediatric epidemic crisis: Lessons for policy and practice development, Nicholas [/bib_ref]. It almost seems that the forced separation due to the epidemic undermines the child's attachment to the secure figure of the parent. The psychological literature fully agrees on the key role of attachment to the parental figure for the emotional and social growth of the child and on the confusion resulting from the loss of this bond. Even the relationship with an unknown health worker proposes an experience of relational stress that may refer to the paradigm of the strange situation [bib_ref] Developmental psychiatry comes of age, Bowlby [/bib_ref] [bib_ref] Separation anxiety: A critical review of the literature, Bowlby [/bib_ref] [bib_ref] The impact of isolation on hospitalized children during SARS, Koller [/bib_ref].
Healthcare providers also expressed feelings of helplessness and inadequacy in the face of the loneliness of their patients, unable to replace the parents, although they are almost engaged with a parental role. Moreover, they cannot ensure a consistent presence as they alternate according to mobile shifts, and they circulate protected by a mask and clothing that give them a not very familiar reassuring aspect. Parents, in turn, felt frustrated at not being able to exercise their parental role and not being able to actively participate in decision-making. All the protagonists, therefore, experience and complain of a persistent condition of loneliness and helplessness in the face of a mandatory and protective isolation to contain the contagion.
Studies documenting therapeutic interventions to alleviate the psychological impact of the pandemic describe activities just focused on recovering and improving communication.
Within psychological support groups [bib_ref] A Support Group for Home-Quarantined College Students Exposed to SARS: Learning from..., Pan [/bib_ref] or expressive arts groups [bib_ref] A process description of playing to live! A community psychosocial arts program..., Decosimo [/bib_ref] [bib_ref] Playing to live: Outcome evaluation of a community-based psychosocial expressive arts program..., Decosimo [/bib_ref] , students were invited to verbalize and reframe their experiences of fear, anxiety and depression together with their peers and qualified professionals. The therapeutic expressive or support group acts precisely on the recovery of a more effective communication and everyone can share with peers fears and stressful experiences and feel the support of sharing from those who have experienced similar emotions.
Strangely, in the psychological support group, many home-quarantined students, when talking about the SARS quarantine, in the early stages, manifested some sort of underlying emotional detachment and a widely emotional disengagement from events and people. However, under the therapeutic prompting of the leader, thereafter some began to tell their experience showing that, within the group, communication was becoming therapeutic by confronting others' similar emotions and experiences.
Likewise, the psychosocial expressive arts programs 'Playing to live' developed in Liberia to promote play and expression as a measure of healing after Ebola Virus Disease showed that children found enjoyment outside of art and play, increasing their hope and self-esteem, being with peers and receiving support from adults. A new, artistic unexplored communication modality makes children aware of their creative strength, and begins their individual's emotional and cognitive recovery [bib_ref] A process description of playing to live! A community psychosocial arts program..., Decosimo [/bib_ref] [bib_ref] Playing to live: Outcome evaluation of a community-based psychosocial expressive arts program..., Decosimo [/bib_ref]. The studies, therefore, show that improving communication is an effective and powerful way to reframe emotions of the pandemic and process them with a greater realistic view [bib_ref] A Support Group for Home-Quarantined College Students Exposed to SARS: Learning from..., Pan [/bib_ref].
A relevant topic in all studies was the matter of information to reduce anxiety and realistically review mistaken beliefs and fears. Both children and adolescents feel reassured to receive up-to-date information on the epidemic when the information is accompanied by a reassuring attitude about protective containment measures and progress by scientists who are studying the disease and possible treatments.
Studies on the psychological effects of COVID-19 in pediatric age are relatively fewer than those reported in previous epidemics. However, the studies, from different researchers and from different countries, confirmed the findings of the pre-COVID-19 searches on the impact of the current epidemic on the emotional life and behavior of children and adolescents.
An emerging finding concerns the different response in different ages. Older adolescents appeared as more sensitive to anxiety than younger children and more fragile and vulnerable in the face of the new and complex problems of protection against the spread of the virus [bib_ref] Online mental health services in China during the COVID-19 outbreak, Liu [/bib_ref]. Probably, according to researchers, the younger child can still rely on the protection and care of adults, while the older adolescent feels exposed in the first person, less protected from parents [bib_ref] Online mental health services in China during the COVID-19 outbreak, Liu [/bib_ref]. Moreover, according to some studies, younger children (3-6 years) manifest their discomfort with childlike regressive behaviors of excessive attachment and expressing the fear of losing their parents for the infection, while older children show more attention difficulties and persistently asking reassurance questions. All this would imply, according to the researchers, that the child's anxieties should be managed involving mainly the parents through parent training, while on older adolescents direct therapeutic treatments could be implemented [bib_ref] Behavioral and Emotional Disorders in Children during the COVID-19 Epidemic, Jiao [/bib_ref].
It is interesting to note that the studies showed a significant relationship between the attitude towards the pandemic, the level of knowledge and awareness of COVID-19 and the level of anxiety and depression. Students who were very worried about being infected had higher anxiety and depression scores, while, by contrast, students less concerned about the infection had lower scores [bib_ref] Clinical course and risk factors for mortality of adult inpatients with COVID-19..., Zhou [/bib_ref]. Similarly, those who routinely wore masks and regularly practiced hand hygiene had lower levels of anxiety and depression. In other words, a valid information, a realistic approach to the severity of the pandemic and an implementation of the available containment strategies appear as protective factors for anxiety control. In the same view, anxiety levels were higher in rural or low-wealth areas.
Probably, in such contexts, a lower ability to manage information in a critical way and to protect oneself from false or contradictory news favors a greater rise in anxiety [bib_ref] Culture and depression, Kleinman [/bib_ref] [bib_ref] Child and adolescent mental health problems in Tyva Republic, Russia, as possible..., Slobodskaya [/bib_ref].
Therefore, taken together, the studies reviewed, show that the pandemic is a catastrophic event for the general health of the population, for the economics and for the mental health of the youngest subjects and more developmentally fragile.
Any outbreak, according to the literature, appears as an uncontrollable, unpredictable and cumulative stress factor occurring outside the control of everyone, including the parents, and increasing fear, anxiety, uncertainty, and feelings of helplessness on the everyday life of children and parents. Therefore, pandemic and restrictive measures, in line with the learned helplessness stress model [bib_ref] Learned helplessness: Critique and reformulation, Abramson [/bib_ref] , may be viewed as an important cumulative stress source, in which a faceless and invisible enemy require full and persistent alert. In pediatric age, emergencies, as chronic and traumatic stress, may be severely dangerous because they stimulate an excessive release of stress hormone cortisol [bib_ref] Acute stressors and cortisol responses: A theoretical integration and synthesis of laboratory..., Dickerson [/bib_ref] [bib_ref] Stress physiology, health, and behavioral development, Gunnar [/bib_ref] , by the limbic-hypothalamic-anterior pituitary-adrenal cortex system [bib_ref] Pituitary-Adrenal and Autonomic Responses to Stress in Women after Sexual and Physical..., Heim [/bib_ref] [bib_ref] Early life experience alters response of adult neurogenesis to stress, Mirescu [/bib_ref] , and, over time, according to the cumulative stress hypothesis [bib_ref] Mismatch or cumulative stress: Toward an integrated hypothesis of programming effects, Nederhof [/bib_ref] , they may reduce long-term health and increase the vulnerability to stress-related pathologies [bib_ref] Physiology and Neurobiology of Stress and Adaptation: Central Role of the Brain, Mcewen [/bib_ref] [bib_ref] Self-esteem, locus of control, hippocampal volume, and cortisol regulation in young and..., Pruessner [/bib_ref]. Moreover, dysregulation of stress hormones could have lasting effects on cognitive processes [bib_ref] Stress hormones and human memory function across the lifespan, Lupienand [/bib_ref] [bib_ref] Self-esteem, locus of control, hippocampal volume, and cortisol regulation in young and..., Pruessner [/bib_ref] [bib_ref] Memory performances and personality traits in mothers of children with obstructive sleep..., Smirni [/bib_ref] [bib_ref] Pediatric neurofibromatosis 1 and parental stress: A multicenter study, Gallai [/bib_ref] [bib_ref] Executive functioning in autism spectrum disorders: A case-control study in preschool children, Carotenuto [/bib_ref] , given that the hippocampus, with its high density of glucocorticoid receptors, appears as a structure particularly involved in the negative feedback loop for the regulation of cortisol [bib_ref] Modulation of dendritic differentiation by corticotropin-releasing factor in the developing hippocampus, Chen [/bib_ref] [bib_ref] How do glucocorticoids influence stress responses? Integrating permissive, suppressive, stimulatory, and preparative..., Sapolsky [/bib_ref] and in cumulative exposure to elevated cortisol levels [bib_ref] Chronic social stress: Effects on limbic brain structures, Fuchs [/bib_ref]. Furthermore, high cortisol, corticotropin and epinephrine can affect the sensitivity of fear responses mediated by the amygdala and alter the hormonal response to social challenges, increasing the risk of immune dysfunction and other health problems.
In sum, according to the reviewed literature, children and adolescents are not indifferent to the dramatic impact of the COVID-19 epidemic, while an efficient management of the pandemic must take into account the pediatric population which cannot be neglected as a minor matter compared to the elderly, the economy and general health care [bib_ref] The psychological impact of quarantine and how to reduce it: Rapid review..., Brooks [/bib_ref]. Since the lockdown time is over, children and adolescents must recover sociality, return to living in the open air, rediscover playing, free time, aiming for the beauty of their everyday life, without forgetting basic rules for keeping COVID-19 away.
Therefore, in line with literature's data, the core values to prevent or mitigate the impact of COVID-19 and containment strategies must be focused on the verbal and nonverbal reassuring presence of the adult as an important protective factor. Specifically, planning to deal with the effects of pandemic stress should ensure a reassuring presence of the adult, an effective child-parent communication, a child-friendly day, and a long-lasting shared time with parents.
## Ensure reassuring presence of the parents
Researchers in child development widely described the negative impact of separation from parents on hospitalized children. Bowlby and Robertson [bib_ref] Developmental psychiatry comes of age, Bowlby [/bib_ref] [bib_ref] Separation anxiety: A critical review of the literature, Bowlby [/bib_ref] [bib_ref] The impact of isolation on hospitalized children during SARS, Koller [/bib_ref] associated intense separation anxiety with poor attachment between parent and child. According to Bowlby and Ainsworth's secure base theory, children must strongly fee the reassuring presence of the parents as 'secure base' [bib_ref] The impact of isolation on hospitalized children during SARS, Koller [/bib_ref]. Child or adolescent must feel they can always come back on a secure island where someone, strong and experienced, is ready to welcome him and to comfort and reassure him to face the world adequately.
The study by Liu et al. [bib_ref] Online mental health services in China during the COVID-19 outbreak, Liu [/bib_ref] documented that college students, in a pandemic condition, still appear quite emotionally fragile even if they have reached a level of autonomy and independence from adults. The study showed a greater presence of somatic symptoms in college students than in primary school children (34.85% vs. 2.39) and a wider range of concerns regarding the epidemic linked to anxiety and depression.
Moreover, research on dealing with stressful situations has broadly showed the effectiveness for children of social support, especially from significant adults, to markedly reduce the psychobiological responses to stress [bib_ref] Alloparental care and the ontogeny of glucocorticoid stress response among stepchildren. In..., Flinn [/bib_ref]. Some study found that parents with higher levels of anxiety and distress have reported a higher level of distress in their children.
Therefore, the first and most effective measure is to be empathically present and convey calm with the entire repertoire of verbal and non-verbal communication. Particular attention should be paid to children with neurodevelopmental disorders because they could have communication difficulties about their feelings and sensation of pain [bib_ref] Autism and Migraine: An Unexplored Association?, Vetri [/bib_ref] [bib_ref] Feeling physical pain while depressed: The effect of alexithymia, Smirni [/bib_ref] [bib_ref] I stay at home with headache. A survey to investigate how the..., Papetti [/bib_ref].
Anxiety and stress may be much more contagious than any virus, and parents are the most important example for the child [bib_ref] Disaster Preparation and Recovery: Lessons from Research on Resilience in Human Development, Masten [/bib_ref]. To convey calm, the key factor is feeling calm, to take care of yourself, to live all-around and to do any activity that makes feel good. In anxiety and distress conditions, children learn from adults how to manage stress [bib_ref] Parenting as primary prevention, Perrin [/bib_ref].
Moreover, children need to feel that their anxieties and fears are understood and shared by significant adults. According to the inhibition theory [bib_ref] Putting stress into words: Health, linguistic, and therapeutic implications, Pennebaker [/bib_ref] , when you can't talk about traumatic experiences for a long time, it becomes cumulative stress that makes you more vulnerable to stress-related diseases.
## Ensure effective child-parent communication
The studies examined in the current review document that the level of knowledge and awareness about COVID-19 may be a protective factor [bib_ref] Clinical course and risk factors for mortality of adult inpatients with COVID-19..., Zhou [/bib_ref]. However, it is important to note that any information can have a different value in relation to the credibility of the person who provides it. Sometimes, in pre-COVID pandemics, official recommendations have been met with skepticism [bib_ref] Understanding responses to government health recommendations: Public perceptions of government advice for..., Teasdale [/bib_ref]. Washing your hands, coughing under a mask or at the elbow, keeping social distances, cleaning surfaces, not touching your face can all be simple behaviors to adopt. However, they become acceptable if the proposer is as authoritative for the child as his parent can be. Therefore, an effective and honest communication childparents supports a proper understanding of what is going on and stimulates the child to safely express their feelings [bib_ref] Appealing to fear: A meta-analysis of fear appeal effectiveness and theories, Tannenbaum [/bib_ref]. Children of any age, even the youngest, feel the emotional states around them [bib_ref] Communication with children and adolescents about the diagnosis of a life-threatening condition..., Dalton [/bib_ref] and will get even more anxious if they don't perceive the adults as sincere and easy to understand. Unexplained or uncertain behaviours may be felt as a threat and as anxiety source.
Moreover, according to the magical thinking construct, children, because of confusion about mind and reality, attempt to make sense on their own [bib_ref] Current approaches to helping children cope with a parent's terminal illness, Christ [/bib_ref] and they may belief that their thoughts or behaviors have a causal effect on unrelated real events, such as disease. Therefore, providing children with an accurate meaningful explanation [bib_ref] Bluebol regression, Stein [/bib_ref] , will ensure they do not feel unnecessarily frightened or guilty, or inappropriately blame themselves or feeling that the illness is a punishment for previous bad behaviour and reinforce the child's positive and preventive behaviors [bib_ref] Effective health communication-A key factor in fighting the COVID-19 pandemic, Finset [/bib_ref].
## Ensure a child-friendly day
Containing children's anxieties means helping them not to experience boring and uninteresting days. Inactivity may easily feed parasitic thoughts, fears, and ritualistic and perseverative ideas and behaviors [bib_ref] Playing to live: Outcome evaluation of a community-based psychosocial expressive arts program..., Decosimo [/bib_ref]. On the contrary, several studies [bib_ref] Behavioral and Emotional Disorders in Children during the COVID-19 Epidemic, Jiao [/bib_ref] [bib_ref] A process description of playing to live! A community psychosocial arts program..., Decosimo [/bib_ref] [bib_ref] Playing to live: Outcome evaluation of a community-based psychosocial expressive arts program..., Decosimo [/bib_ref] pointed out that media entertainment, reading, physical exercise may be protective factors to mitigate distress resulting from pandemic condition. Being involved in stimulating activities promotes the safety, psychological well-being and mental health of children. Of course, it's not about forcing them, without any break, into 'demanding' activities, but live as children in a structured, stimulating and welcoming world in which the playfulrecreational dimension and commitment are integrated [bib_ref] Playing to live: Outcome evaluation of a community-based psychosocial expressive arts program..., Decosimo [/bib_ref].
Moreover, even careful use of television and, more generally, of informatic tools (videogames, play stations) can be an interesting and inspiring way into a child's day. It is important to limit exposure to media news by selecting only official information and make sure that correct messages are received from children. But it is the diligent and reassurance presence of the adult that makes the difference.
## Ensure a long-lasting shared time with the parents
The restrictions of the pandemic can be rethought as a challenge to improve relationships with children by staying close and sharing their life and their ways of expression. Not least, a climate of relaxation, mediated by significant and collaborative playful and recreational activities, becomes a favourable opportunity for the psychological well-being of the entire family, especially for more fragile people, though in adults can promote a redefinition of the forgotten everyday life relevant values.
A qualitatively significant time sharing represents a crucial response to the child's fears and anxieties and to child's need to feel loved and appreciated, while the lockdown experience may be a condition to disrupt this process on multiple levels [bib_ref] Case study: Play therapy and eye movement desensitization and reprocessing for pediatric..., Banbury [/bib_ref].
# Conclusions
The current narrative review focused on emotional reactions to pandemic in the pediatric population and the reassuring role of adults in dealing with pandemics.
Pandemic and restrictive measures, in line with the learned helplessness stress model [bib_ref] Learned helplessness: Critique and reformulation, Abramson [/bib_ref] , may be viewed as an important cumulative stress source. The literature, in pre-COVID-19 time, has already documented the psychological dimension in epidemic events and, namely, the emotional reactions of children and adolescents. Similarly, literature data, in the COVID-19 time, consistently confirmed an increase in depressive and anxious symptoms in children and adolescents [bib_ref] Behavioral and Emotional Disorders in Children during the COVID-19 Epidemic, Jiao [/bib_ref] [bib_ref] Online mental health services in China during the COVID-19 outbreak, Liu [/bib_ref] [bib_ref] Mental Health Status Among Children in Home Confinement During the Coronavirus Disease, Xie [/bib_ref] [bib_ref] Clinical course and risk factors for mortality of adult inpatients with COVID-19..., Zhou [/bib_ref].
Several studies pointed out that social isolation, separation, and communication difficulties were the recurring issues in complaints of both children and adults. Social isolation may be felt as a serious source of anxiety by the child, by the parents and also by health professionals. Research reported that most parents reaching significant levels of post-traumatic stress have children with high levels of stress, supporting a sort of circularity process of anxiety within the group. According to a conceptual overview, fear and fearrelated behaviors can amplify the psychological impact of the epidemic and reduce the ability to cope with it [bib_ref] The Role of Fear-Related Behaviors in the 2013-2016 West Africa Ebola Virus..., Shultz [/bib_ref].
Conversely, therapeutic approaches focused on peer communication, within psychological support groups or expressive arts groups, have proven to be powerful tools in pediatric age to recover emotional stability after stressful experiences due to pandemics.
Moreover, the level of knowledge and awareness about COVID-19 and positive attitude towards the development of epidemic appeared as protective factors against depressive and anxiety symptoms. Therefore, all searches emphasized the role of reassuring information and social support in preventing or mitigating psychological impact of pandemic and containment strategies [bib_ref] Stress hormones and human memory function across the lifespan, Lupienand [/bib_ref] [bib_ref] Maternal care during infancy regulates the development of neural systems mediating the..., Caldji [/bib_ref] [bib_ref] Environmental Enrichment Reverses the Effects of Maternal Separation on Stress Reactivity, Francis [/bib_ref].
Future research will have to focus, with rigorous and appropriate methodologies, on the long-term effects of the psychological impact of epidemics on the pediatric population.
[fig] Figure 1: Selection of studies flowchart. [/fig]
[table] Table 1: Studies on the psychological impact on children and adolescents in pre-COVID-19 time. [/table]
[table] Table 2: Studies on the psychological impact on children and adolescents during COVID-19. [/table]
[bib_ref] Mental Health Status Among Children in Home Confinement During the Coronavirus Disease, Xie [/bib_ref] |
Outbreak of Invasive Wound Mucormycosis in a Burn Unit Due to Multiple Strains of Mucor circinelloides f. circinelloides Resolved by Whole-Genome Sequencing
Mucorales are ubiquitous environmental molds responsible for mucormycosis in diabetic, immunocompromised, and severely burned patients. Small outbreaks of invasive wound mucormycosis (IWM) have already been reported in burn units without extensive microbiological investigations. We faced an outbreak of IWM in our center and investigated the clinical isolates with whole-genome sequencing (WGS) analysis. We analyzed M. circinelloides isolates from patients in our burn unit (BU1, Hôpital Saint-Louis, Paris, France) together with nonoutbreak isolates from Burn Unit 2 (BU2, Paris area) and from France over a 2-year period (2013 to 2015). A total of 21 isolates, including 14 isolates from six BU1 patients, were analyzed by whole-genome sequencing (WGS). Phylogenetic classification based on de novo assembly and assembly free approaches showed that the clinical isolates clustered in four highly divergent clades. Clade 1 contained at least one of the strains from the six epidemiologically linked BU1 patients. The clinical isolates were specific to each patient. Two patients were infected with more than two strains from different clades, suggesting that an environmental reservoir of clonally unrelated isolates was the source of contamination. Only two patients from BU1 shared one strain, which could correspond to direct transmission or contamination with the same environmental source. In conclusion, WGS of several isolates per patients coupled with precise epidemiological data revealed a complex situation combining potential cross-transmission
between patients and multiple contaminations with a heterogeneous pool of strains from a cryptic environmental reservoir. IMPORTANCE Invasive wound mucormycosis (IWM) is a severe infection due to environmental molds belonging to the order Mucorales. Severely burned patients are particularly at risk for IWM. Here, we used whole-genome sequencing (WGS) analysis to resolve an outbreak of IWM due to Mucor circinelloides that occurred in our hospital (BU1). We sequenced 21 clinical isolates, including 14 from BU1 and 7 unrelated isolates, and compared them to the reference genome (1006PhL). This analysis revealed that the outbreak was mainly due to multiple strains that seemed patient specific, suggesting that the patients were more likely infected from a pool of diverse strains from the environment rather than from direct transmission among them. This study revealed the complexity of a Mucorales outbreak in the settings of IWM in burn patients, which has been highlighted based on WGS combined with careful sampling. Mucor circinelloides, Mucorales, mucormycosis, strains, burn, clade, genome, genotype, mixed infection, outbreak, whole-genome sequencing, wound M ucormycosis is a rare and life-threatening infection caused by Mucorales belonging to the subphylum Mucoromycotina [bib_ref] A higher-level phylogenetic classification of the Fungi, Hibbett [/bib_ref]. These molds are ubiquitously distributed in the environment and mostly disseminated through airborne spores, which can be considered infective propagules responsible mainly for respiratory (lung and sinuses), wound, and skin infections [bib_ref] Epidemiology and outcome of zygomycosis: a review of 929 reported cases, Roden [/bib_ref].
Patients at risk for mucormycosis are immunocompromised (hematological malignancies, hematopoietic stem cell transplantation, solid organ transplant, and steroid therapy) or have diabetes mellitus, deferoxamine treatment, trauma, or severe burns [bib_ref] Epidemiology and clinical manifestations of mucormycosis, Petrikkos [/bib_ref] [bib_ref] A global analysis of mucormycosis in France: the RetroZygo Study, Lanternier [/bib_ref]. Among skin-related infections, contaminated materials (Elastoplast bandages, tape, tongue depressors, ostomy bags, and linens) have been implicated as the causes of local or disseminated infections in patients with various underlying diseases [bib_ref] Healthcare-associated mucormycosis, Rammaert [/bib_ref]. Specifically, in burn patients, invasive wound mucormycosis (IWM) has been reported in both small series [bib_ref] Report of two cases with extensive destruction of the face and nasal..., Rabin [/bib_ref] and epidemiological surveys [bib_ref] Epidemiology of fungal infection in burns: therapeutic implications, Schaal [/bib_ref] [bib_ref] Zygomycosis in Europe: analysis of 230 cases accrued by the registry of..., Skiada [/bib_ref] [bib_ref] Fungal burn wound infection. A 10-year experience, Becker [/bib_ref].
Over the past 10 years, outbreaks of mucormycosis have been increasingly reported in various environments. In humans, nosocomial outbreak cases occurring before 2008 were reviewed by . Antoniadou found 12 reported outbreaks and two pseudoepidemics of cases since 1977. Mucormycosis outbreaks have been reported in the United States, United Kingdom, and Europe [bib_ref] Outbreaks of zygomycosis in hospitals, Antoniadou [/bib_ref]. Since 2008, outbreaks or clustered cases have been reported after the tornadoes in Joplin, MO (13 patients) [bib_ref] Necrotizing cutaneous mucormycosis after a, Fanfair [/bib_ref] [bib_ref] Whole genome sequence typing to investigate the Apophysomyces outbreak following a tornado..., Etienne [/bib_ref] , in an intensive care unit (ICU) in France (3 patients) [bib_ref] Three cases of cutaneous mucormycosis with Lichtheimia spp. (ex Absidia/Mycocladus) in ICU...., Poirier [/bib_ref] , in adults (6 patients) [bib_ref] Hospital outbreak of pulmonary and cutaneous zygomycosis due to contaminated linen items..., Cheng [/bib_ref] or infants (5 patients) (16) exposed to contaminated linens in the United States, in infants in Egypt (5 patients) [bib_ref] Investigation and management of a Rhizomucor outbreak in a pediatric cancer hospital..., El-Mahallawy [/bib_ref] , and in patients undergoing arthroscopy in Argentina (40 patients) [bib_ref] Control of an outbreak of postoperative bone mucormycosis: an intervention study of..., Chaves [/bib_ref]. More specifically, in a Belgian burn unit, Christiaens et al. described an outbreak of Lichtheimia corymbifera associated with nonsterile Elastoplast bandage contamination in seven burn patients, including five with infection and two with colonization [bib_ref] An outbreak of Absidia corymbifera infection associated with bandage contamination in a..., Christiaens [/bib_ref]. In this study, the authors did not have evidence for the genotypic relatedness of the strains between patients and material strains. In addition, a large outbreak due to yogurt contamination in the United States responsible for digestive symptoms (nausea, cramps, vomiting, and diarrhea) in about 300 individuals has been described recently [bib_ref] Analysis of a food-borne fungal pathogen outbreak: virulence and genome of a..., Lee [/bib_ref]. In this study, phylogenetic analysis and whole-genome sequence (WGS) analysis yielded new information on the genetic structure of Mucor circinelloides. Mucor circinelloides is a single species consisting of four different formae (f. circinelloides, f. griseocyanus, f. janssenii, and f. lusitanicus), with forma circinelloides the most commonly involved in human mucormycosis [bib_ref] Analysis of a food-borne fungal pathogen outbreak: virulence and genome of a..., Lee [/bib_ref].
Between 2013 and 2015, we faced an outbreak of proven IWM due to M. circinelloides f. circinelloides in a burn unit (BU) in Hôpital Saint-Louis (SLS), Paris, France, involving six patients raising the hypothesis of a common source of contamination. The outbreak was not suspected until M. circinelloides was recovered from wounds of patient P04. Over the same period of time, 4 additional cases that occurred in a burn unit (Burn Unit 2 [BU2]) of another hospital in a Paris suburb (Hôpital d'Instruction des Armées, Clamart, France [PER]) were reported to the National Reference Center for Invasive Mycoses and Antifungals (NRCMA).
Our aim was to clarify the origin/source of infection in both BU1 and BU2. In the absence of genotyping markers for this organism, WGS analysis was performed on 21 isolates (14 from the outbreaks and 7 unrelated) to investigate the links between clinical isolates, understand the epidemiology of the outbreak, and identify and eliminate the potential source of the infections.
# Results
Clinical and microbiological investigations. Three patients (P03, P04, and P05) developed proven IWM due to M. circinelloides f. circinelloides within 18 days after admission in BU1 (between 18 August and 5 September 2014) and subsequently died from these infections [fig_ref] TABLE 1: Isolates sequenced in this study [/fig_ref]. The outbreak was suspected when a positive culture was observed in P04 (11 days after the first positive sample in P03). Sequential samples from the wounds were prospectively obtained starting with P03. A few months later (114 days), another patient (P06) developed proven IWM, and M. circinelloides f. circinelloides was also involved [fig_ref] FIG 1: Epidemiological map of 13 patients whose isolates were selected in our study [/fig_ref]. We retrospectively noticed isolates from the same species had already been identified in 2013 from two patients in BU1 (P01 with no infection and P02 with proven IWM). A total of seven patients were exposed to (P01 and P07) and/or infected with (P02 to P06) this species in BU1, with P01 as the putative index case. Infection control measures were implemented locally to avoid potential nosocomial transmission to other patients of the unit. More than 30 environmental samples were cultured. All were negative. DNA amplified with the Mucor/Rhizopus PCR test [bib_ref] Detection of circulating Mucorales DNA in critically ill burn patients: preliminary report..., Legrand [/bib_ref] was detected only in the Bair Hugger filters that were used during the hospitalization of P03, P04, and P05. To prevent transmission to other patients, we needed to investigate whether these strains were clonal and needed unrelated isolates from other geographic areas. The additional cases corresponded to an outbreak in BU2 (PER) involving four patients with IWM, as well as one case of proven invasive mucormycosis with kidney invasion in a transplant recipient in Strasbourg (STR) about whom the NRCMA was notified.
Overall, the 12 patients (21 clinical isolates) included 10 cases identified in two burn units [fig_ref] TABLE 1: Isolates sequenced in this study [/fig_ref] ; [fig_ref] FIG 1: Epidemiological map of 13 patients whose isolates were selected in our study [/fig_ref].
Phylogenetic analyses of three loci. Internal transcribed spacer (ITS), D1/D2, and RPB1 sequences were analyzed as separate (data not shown) and combined data sets. The topologies of the multilocus data set by 3 methods (neighbor-joining [NJ], maximum likelihood , and Bayesian inference) were comparable between the individual trees of the three genes analyzed. Four clades [fig_ref] FIG 2: PhyML tree constructed by the maximum likelihood method [/fig_ref] , respectively, denoted C1 (14 isolates, including 11 from BU1), C2 (4 isolates, including 2 from BU1), C3 (2 strains in addition to the reference strain, including 1 from BU1), and C4 (1 isolate), were identified from the analysis of the combined data set, which yielded a significant support (Ն95% bootstrap for NJ and ML; 1.0 for Bayesian inference). Isolates recovered from BU1 were distributed in three clades (C2, C3, and C4). All patients from BU1 had at least one isolate included in C1.
Whole-genome analysis. To better resolve the diversity of the strains within the four clades and because no further genotyping methods existed for this organism, whole-genome sequencing was performed. Because the biology and the genetics of this organism are poorly understood, we first checked the reproducibility of the sequencing process and the stability of the genome, to be able to define genetically identical strains.
Establishing the genetic threshold to determine genetically identical strains. For the three strains isolated from single-spore colonies (i.e., P05_600_BU1_SLS, P04_603_BU1_SLS, and P03_594_BU1_SLS), two complementary approaches were used to estimate the genetic distance between the parent and single-spore colony: estimates of both the evolutionary distance (defined as the proportion of nucleotide differences: i.e., 0.00035, 0.00044, and 0.00029, respectively) and the number of SNP differences (i.e., 4,731, 5,296, and 4,187, respectively). This information gave us the expected genetic proximity measures between pairs of genomes arising from identical strains and independently sequenced isolates. The largest of the three pairs of measures (i.e., 0.00044 mismatch per nucleotide and 5,296 SNPs) was therefore selected as a cutoff below which two compared isolates were defined as belonging to the same strain.
Experimental investigation of the potential genetic drift of M. circinelloides f. circinelloides. Experimental investigation of the 1006PhL genome upon iterative subculturing on agar (n ϭ 3) and three passages in mice (n ϭ 3) revealed no acquisition of SNPs during this process, suggesting that the genome of M. circinelloides f. circinelloides was stable upon iterative passages.
Whole-genome phylogenetic classification of the 21 clinical isolates. A phylogenetic classification of the whole genome of the 21 isolates was then performed [fig_ref] FIG 3: Minimum-evolution phylogenetic tree of the whole genome of 21 clinical isolates and... [/fig_ref]. This phylogenetic tree allows classification of the genomes in four main clades corresponding exactly to the same clades (C1 to C4) described for the analysis of three loci [fig_ref] FIG 2: PhyML tree constructed by the maximum likelihood method [/fig_ref]. Clades C2, C3, and C4 contained isolates that are clearly distinct from those inside clade C1: e.g., the average estimated evolutionary distances and SNP numbers between isolates from C1 and those inside C2, C3, and C4 are 0.0187, 0.0385, and 0.0390 and 291,467, 702,061, and 715,595, respectively, whereas among C1 isolates, the average pairwise distance and SNP number are 0.00146 and 19,588, respectively (see [fig_ref] TABLE 1: Isolates sequenced in this study [/fig_ref] in the supplemental material).
Furthermore, increased resolution of WGS allowed robust identification of strains (as defined above and in Materials and Methods) and understanding of which clinical isolates belong to which strain. As a result, the 21 investigated isolates and the reference 1006PhL could be partitioned into 14 distinct strains (S1 to S14 [ [fig_ref] FIG 3: Minimum-evolution phylogenetic tree of the whole genome of 21 clinical isolates and... [/fig_ref] ; [fig_ref] TABLE 1: Isolates sequenced in this study [/fig_ref]. Outbreak dynamics. In BU1, the isolate of the potential index case P01 (P01_617_ BU1_SLS, S12) was different from the isolates subsequently recovered in BU1. The isolate from P02, corresponding to the specific strain S5 clustered in C1 with isolates from P03, P04, P05, and P06. P03 and P06 were also infected with two isolates from S1 (594 and 592) and S6 (032 and 023), respectively. P04 was infected over 10 days with two strains, S9 and S7, belonging to C2 (isolates 601 and 559) and C1 (isolates 602 and 603), respectively. P05 was infected over 11 days with isolates belonging to three strains from C1, S8 (isolates 598 and 600), S3 (isolate 622), and S1 (isolate 599). Indeed, P04 and P05 had mixed infections during the course of their disease, suggesting initial contamination with a mixture of strains, the latter strain also recovered in P03. This suggests cross contamination or common infection in P3 and P5. This has also been observed in BU2, where two patients (P09 and P10) shared strain S2 (isolates 703 and 704), which clustered in C1. Two patients (P11 and P08) were also infected with strains that belonged to C2 and C3, respectively. The environmental (isolate 615) and the colonization (isolate 621) isolates from our hospital clustered in C2 and C4, respectively. The patient from eastern France clustered in C1 (S4), but with a specific strain different from the other C1 strains, as expected for a geographically unrelated infection.
# Discussion
Because isolation of Mucorales is rare in the hospital, the observation of the same species in two independent samples or patients has long been considered a sufficient criterion to suspect and assess transmission or common contamination. Here, we investigated further outbreak-related and -unrelated isolates based on WGS analysis. To our knowledge, this is the first time WGS analysis has been used to resolve an outbreak of invasive Mucorales infection in the context of nosocomial acquisition.
Because WGS was applied for the first time in this setting, we first sought to evaluate reproducibility of the sequencing process and intraculture variation/stability during in vivo passage. As the genome of three selected strains was sequenced and assembled twice, we were able to compare contig sets belonging a priori to the same strain. This method led to the definition of a pairwise distance cutoff. Therefore, if two genome sequences belonging to different isolates have a pairwise distance below this cutoff, it was inferred a posteriori that the two isolates corresponded to the same strain. This cutoff should vary as a function of the organism, the method of sequencing, the bioinformatics pipeline, and the pathophysiology of the disease, suggesting that such data should be obtained each time an investigation of an outbreak due to rare organisms is undertaken. Of note, using pairwise evolutionary distances could be considered as a fast but accurate alternative to the well-known average nucleotide identity (ANI) approach to compare genomes [bib_ref] DNA-DNA hybridization values and their relationship to whole-genome sequence similarities, Goris [/bib_ref] [bib_ref] Development of a Rhizoctonia solani AG1-IB specific gene model enables comparative genome..., Wibberg [/bib_ref] because both were shown to be linearly correlated [bib_ref] Mash: fast genome and metagenome distance estimation using MinHash, Ondov [/bib_ref].
Contrary to the initial hypothesis of a single-strain transmission in BU1, we observed that all of the patients from the BU1 outbreak (P01 to P06) were infected by different strains. Surprisingly, our data revealed that each strain was patient specific in BU1, except for S1, suggesting that the outbreak in BU1 was due to multiple strains present in and acquired from a local environmental "reservoir" containing clonally unrelated isolates. This hypothesis is reinforced by two patients (P04 and P05) with IWM coinfected by more than 2 genetically distinct strains. Another hypothesis is that the patients could have been exposed to specific strains or a mixture of strains before arriving in BU1. However, a delay between admission and the first positive culture was 16 days (median), making the hypothesis that exposure occurred in the environment of BU1 more likely. In the settings of severe burns, where invasive fungi do sporulate on the wounds, transmission by air and transmission by the hands of health care workers to other patients are both possible. A major point to emphasize is that only two patients from BU1 (P03 and P05) shared the same strain, S1. This feature has also been observed in BU2, suggesting that transmission between patients is possible. However, this does not rule out the hypothesis of a contamination by the same strain from the environment. Thus, while we cannot entirely exclude that mating and also recombination could occur on the surface of the wound, zygospores of Mucor circinelloides have never been observed in wound tissue and are not known to germinate under any conditions tested thus far in vitro. Moreover, the conditions of the host environment (elevated temperature and light) are not optimal for sexual reproduction, and thus it seems unlikely that mating and recombination would be occurring in the setting of infection.
Environmental investigation of the outbreak in BU1 failed to identify the source of infection using culture of multiple samples from the environment, as well as by PCR. DNA amplified with the Mucor/Rhizopus PCR [bib_ref] Detection of circulating Mucorales DNA in critically ill burn patients: preliminary report..., Legrand [/bib_ref] was detected only in the Bair Hugger filters that were used during the hospitalization of IWM patients P03, P04, and P05. Despite the negative result of this investigation, it is likely that the contamination came from a local source because this has already been described with linens or Elastoplast in burn units [bib_ref] Hospital outbreak of pulmonary and cutaneous zygomycosis due to contaminated linen items..., Cheng [/bib_ref] [bib_ref] Mucormycosis outbreak associated with hospital linens, Duffy [/bib_ref].
Our findings about the genetic structure of M. circinelloides f. circinelloides are reminiscent of the WGS investigations of the Apophysomyces sp. outbreak in Joplin [bib_ref] Whole genome sequence typing to investigate the Apophysomyces outbreak following a tornado..., Etienne [/bib_ref] or of the Saprochaete clavata outbreak in France [bib_ref] Multicenter outbreak of infections by Saprochaete clavata, an unrecognized opportunistic fungal pathogen, Vaux [/bib_ref] , which revealed that several genetic groups can be responsible for infections over the same period of time. In our case, in a given restricted area (BU1), we identified a large diversity of isolates responsible for IWM and were not able to find isolates belonging to unique strains recovered in different places. At the other end of this spectrum, for Apophysomyces trapeziformis several genetically identical isolates were recovered in different places at a distance of several miles [bib_ref] Whole genome sequence typing to investigate the Apophysomyces outbreak following a tornado..., Etienne [/bib_ref]. In the case of Exserohilum rostratum, all outbreak isolates have closely related genomes, suggesting that a unique strain was responsible for the outbreak [bib_ref] Whole-genome analysis of Exserohilum rostratum from an outbreak of fungal meningitis and..., Litvintseva [/bib_ref]. These differences could be explained by a completely different pathophysiology of the disease and mode of fungal transmission between these fungal organisms.
This outbreak is illustrative of the importance of the sampling strategy. Repeated sampling is paramount, as is avoiding the assumption that one patient should harbor only one strain. Indeed, in our case two patients were infected by mixtures of strains concomitantly, as already described for cryptococcosis. Mixed infections and the impact of genetic heterogeneity of isolates from single patients on the interpretation of transmission chains are an emerging theme in molecular epidemiology in the genomic era [bib_ref] Genomic infectious disease epidemiology in partially sampled and ongoing outbreaks, Didelot [/bib_ref]. Our investigation fully supports the view that multisampling is critical to decipher transmission patterns, especially in the context of outbreaks with such long time frames. Here, multiple sampling was only performed prospectively for four patients (P03, P04, P05, and P06), with one isolate stored and studied for all of the remaining patients. The proportion of mixed infection may thus have been higher than detected here if all of the isolates from all of the patients had been investigated. Based on our experience, to investigate a cluster of wound mucormycosis, we recommend sampling and analyzing several wound lesions and several isolates per patient and analyzing isolates recovered at different times. This will have the potential to find the maximal diversity of the strains if present and have a better view of the outbreak dynamics. In outbreaks involving rare organisms for which the biology and the genetics are mostly unknown, we recommend testing the stability of the genome under different conditions, including in vivo models to determine the threshold above which to strains can be delineated.
To prevent potential transmission between patients, all patients with a positive culture were isolated and had a dedicated nurse. To avoid persistence of spores in the environment, after a patient's death or discharge, rooms should be cleaned by using vaporization of 6% H 2 O 2 . In addition, our management of this outbreak led to the implementation of a twice-weekly screening of serum samples from patients hospitalized in BU1 using Mucorales PCR and by immediate prescription of antifungal treatment when the PCR test was positive [bib_ref] Detection of circulating Mucorales DNA in critically ill burn patients: preliminary report..., Legrand [/bib_ref] , in addition to isolation of all culture-positive and/or PCR-positive patients and dedicated nurses to prevent the risk of transmission to other patients. So far, with 2 years of hindsight, no additional cases of M. circinelloides f. circinelloides IWM have been observed. This illustrates how prevention through hygiene together with early diagnosis of mucormycosis could improve patient management and help control dramatic outbreaks in hospital settings among populations at risk.
# Materials and methods
Isolates and patients. In order to study the genetic relatedness between the BU1 clinical isolates, the following additional isolates identified as M. circinelloides from other sources were selected [fig_ref] TABLE 1: Isolates sequenced in this study [/fig_ref]. (i) The first group of isolates were sequential isolates prospectively collected from the skin lesions of 4 of the patients, as previously described [bib_ref] Detection of circulating Mucorales DNA in critically ill burn patients: preliminary report..., Legrand [/bib_ref]. (ii) The next isolates were collected in Hôpital Saint-Louis (SLS) but in another ward (one from the environment and another one colonizing a patient). (iii) The next isolates (n ϭ 4) were recovered in another outbreak that involved 4 patients hospitalized in another burn unit (BU2) in the Hopital d'Instruction des Armées, Clamart (PER), located in the suburb of Paris, over the same period of time. These isolates have been sent to the French National Reference Center for Invasive Mycoses and Antifungals (NRCMA). (iv) Finally, one isolate was obtained from invasive mucormycosis recovered from the kidney biopsy specimen of a kidney transplant recipient in Strasbourg (STR) in the eastern part of France.
All positive slants were subcultured once on Sabouraud dextrose agar with gentamicin and chloramphenicol (Bio-Rad, Marnes-la-Coquette, France) at 30°C using the bulk and never single colonies.
Overall, 21 isolates were selected for further analysis: 15 clinical isolates from BU1 (SLS) (12 bulk cultures, including 1 from colonization), 1 environmental isolate from SLS, 4 clinical isolates from BU2 (PER), and 1 clinical isolate from STR.
The sequence of the Mucor circinelloides 1006PhL strain was used as the reference genome [bib_ref] Topographic diversity of fungal and bacterial communities in human skin, Findley [/bib_ref]. Five of the six patients had proven IWM according to a modified version of the EORTC/MSG criteria [bib_ref] Detection of circulating Mucorales DNA in critically ill burn patients: preliminary report..., Legrand [/bib_ref] [bib_ref] Revised definitions of invasive fungal disease from the European Organization for Research..., De Pauw [/bib_ref] , and one was considered colonized and was not treated.
Environmental investigation in BU1. Extensive environmental sampling was performed, and mycological contamination was investigated by culture methods on Sabouraud agar (Bio-Rad, Marnesla-Coquette, France) and 2% malt extract incubated at 30 and 37°C. Overall, 30 specimens from nonsterile material, air, surfaces, and aeration machineries, the technical room, Bair Hugger machines, and dedicated nonsterile material were tested. The Bair Hugger machine allows active warming of the patient. It pulses warm air through a plastic pipe into a blanket applied on the patient.
Polyphasic identification of isolates. The 21 isolates were sent to the NRCMA, where the purity was verified and identification to the species-level performed using phenotypic and molecular identification on the bulk culture. In detail, microscopic examination was performed on 5-to 7-day-old cultures growth on 2% malt agar at 30°C. Amplification and sequencing of the ITS1-5.8S-ITS2 region and the D1/D2 region of the large subunit (LSU) ribosomal DNA (rDNA) were performed as described previously [bib_ref] Molecular and phenotypic evaluation of Lichtheimia corymbifera (formerly Absidia corymbifera) complex isolates..., Garcia-Hermoso [/bib_ref]. The amplification of the RPB1 gene (RNA polymerase II largest subunit) was made with primers RPB1Ac and RPB1Cr [bib_ref] The origin of red algae: implications for plastid evolution, Stiller [/bib_ref]. The PCR products were then sequenced, and the consensus sequences were obtained as previously described [bib_ref] Molecular and phenotypic evaluation of Lichtheimia corymbifera (formerly Absidia corymbifera) complex isolates..., Garcia-Hermoso [/bib_ref]. Sequences were subjected to pairwise alignments against curated fungal reference databases available at the online MycoBank database (http://www.mycobank.org/). Sequence alignment and phylogenetic analysis. Multilocus sequence alignments of partial sequences of 3 different loci (ITS, 28S, and RPB1) were performed [bib_ref] Molecular and phenotypic evaluation of Lichtheimia corymbifera (formerly Absidia corymbifera) complex isolates..., Garcia-Hermoso [/bib_ref] [bib_ref] Spectrum of zygomycete species identified in clinically significant specimens in the United..., Alvarez [/bib_ref].
The sequences were aligned using MAFFT v.7.308 with default settings. Data from each gene were analyzed separately and combined as a concatenated 3-locus data set. For the multilocus data set, neighbor-joining (NJ) phylogenetic trees were constructed using MEGA6 software (35) with the Tamura 3-parameter substitution model and 1,000 bootstrap replicates. The program PhyML v3.0.1 (36) was used to infer maximum likelihood (ML) phylogeny using the TN93 substitution model and 1,000 bootstrap repetitions. Bayesian analysis with default prior of MrBayes v.3.2 (37) was conducted to determine posterior probabilities. Two analyses were done by running 10 6 generations in four chains, sampling every 100 generations.
Whole-genome sequencing and assembly. Because genotyping methods for M. circinelloides f. circinelloides were lacking, WGS was performed to compare isolates. Libraries were constructed using a Nextera XT DNA sample preparation kit (Illumina) and sequenced with an Illumina NextSeq 500 sequencing system with a 2ϫ 150-nucleotide paired-end protocol. Two lanes from these tagged libraries resulted in~12.6 million read pairs per strain on average. Statistics of assembly for all the isolates are summarized in [fig_ref] TABLE 2: Statistics of assembly of the strains sequenced in this study [/fig_ref].
All sequenced reads were clipped and trimmed with AlienTrimmer v.0.4.0 (38), corrected with Musket v.1.1 [bib_ref] Musket: a multistage k-mer spectrumbased error corrector for Illumina sequence data, Liu [/bib_ref] , merged (if needed) with FLASH v.1.2.11 [bib_ref] FLASH: fast length adjustment of short reads to improve genome assemblies, Magoč [/bib_ref] , and subjected to a digital normalization procedure with Khmer v.2.0. For each sample, processed reads were finally assembled with SPAdes v.3.10 [bib_ref] SPAdes: a new genome assembly algorithm and its applications to single-cell sequencing, Bankevich [/bib_ref].
Whole-genome analysis. For each pair of assembled genomes, an evolutionary distance (proportion of aligned nucleotide differences) was estimated with Mash v.1.0.2 (sketch size ϭ 100,000) [bib_ref] Mash: fast genome and metagenome distance estimation using MinHash, Ondov [/bib_ref]. For each sample, sequenced reads were also aligned with minimap2 v.2.9 (https://github.com/lh3/minimap2) against the assembled contigs of the reference strain 1006PhL in order to estimate the number of SNPs between each pair of isolates. Both pairwise evolutionary distance and SNP number difference estimates are available in [fig_ref] TABLE 1: Isolates sequenced in this study [/fig_ref] in the supplemental material. Finally, the matrix of evolutionary distances was used to infer a minimum-evolution phylogenetic tree with FastME v.2.1.5 [bib_ref] Fast and accurate phylogeny reconstruction algorithms based on the minimum-evolution principle, Desper [/bib_ref] [bib_ref] FastME 2.0: a comprehensive, accurate, and fast distance-based phylogeny inference program, Lefort [/bib_ref].
Reproducibility of the sequencing process. In a specific experiment dedicated to determine the reproducibility of all parts of the sequencing process (from extraction to sequence analysis), single-spore isolation was performed for 3 isolates recovered from patients P03, P04, and P05 in BU1 (P05_600_BU1_SLS, P04_603_BU1_SLS, and P03_594_BU1_SLS [ [fig_ref] TABLE 1: Isolates sequenced in this study [/fig_ref]. Spore suspensions were prepared in sterile water with 0.05% Tween 80 and serially diluted, and spores were enumerated to approximately plate five spores per potato dextrose agar (PDA) plate. After 12 h of incubation at 30°C, plates were inspected under a microscope at ϫ10 magnification, and germinating spores were then transferred to new PDA plates using a stereomicroscope. One colony was thus selected from each isolate for additional sequencing.
Genome stability experiments. Genome stability during vegetative growth and host infections was analyzed with the 1006PhL strain. In brief, the strain was grown on PDA for 1 day. The colony was streaked to isolate a single colony, which was then transferred onto a new PDA plate. After 4 days of incubation at 30°C under the light, the spores were collected and subjected to the same procedures up to three passages (VEG1, VEG2, and VEG3 isolates).
For infection passages, spores were suspended in sterile phosphate-buffered saline (PBS). Male 8-week-old BALB/c mice were infected with 10 6 spores in 200 l of sterile PBS via tail vein injection. At day 3 postinoculation, the mice were sacrificed, and their brains were collected and placed onto PDA after brain homogenization. After 1 day of incubation, the fungal colony emerging from the brain was streaked to isolate single colonies. One colony was transferred onto PDA and incubated at 30°C under light for 4 days. Spores were then used as the inoculum for the next infection passage, up to three passages (INF1, INF2, and INF3 isolates). Genomic DNAs of these isolates were used to construct libraries with 180-base fragments and 2-to 3-kb jumps, which were sequenced by an Illumina HiSeq 2000 platform at the University of North Carolina High-Throughput Sequencing Facility (HTSF). Data were analyzed using the same pipelines as that implemented for the clinical isolates. Data availability. All FASTQ files were made available on GenBank under accession no. ERS2032288 to ERS2032317 in BioProject PRJEB23686.
# Supplemental material
Supplemental material for this article may be found at https://doi.org/10.1128/mBio .00573-18.
[fig] FIG 1: Epidemiological map of 13 patients whose isolates were selected in our study. Isolates from several geographic areas in France (from dark to light gray) have been studied: Burn Unit 1 (BU1) of hospital 1, Paris France; wards of SLS, Paris, France; BU2 of PER in the Paris area and eastern France (one isolate from Strasbourg [STR]). Isolates were prospectively collected for patients P03 to P06. Analyzed isolates are depicted in dark open circles. The index case from BU1 (solid arrow) was thought to be P01, and the outbreak was recognized after P04 got infected (dashed arrow). [/fig]
[fig] FIG 2: PhyML tree constructed by the maximum likelihood method. The tree was inferred from concatenated 3-locus data set (ITS, 28S, and RPB1). Bootstrap support values from PhyML greater than 70% (left) and a Bayesian posterior probability of Ͼ0.80 (right) are shown at the nodes. The 21 clinical isolates are grouped in clades C1 to C4. The scale bar indicates 0.001 nucleotide substitution per character. [/fig]
[fig] FIG 3: Minimum-evolution phylogenetic tree of the whole genome of 21 clinical isolates and the 1006PhL reference strain. The 21 isolates are grouped in different strains clustered in clades C1 to C4. Clinical isolates belonging to the same strain are highlighted in light gray. The scale bar indicates 0.004 nucleotide substitution per character. [/fig]
[table] TABLE 1: Isolates sequenced in this study [/table]
[table] TABLE S1 ,: XLSX file, 0.1 MB. [/table]
[table] TABLE 2: Statistics of assembly of the strains sequenced in this study [/table]
|
Oxidative Stress in Age-Related Neurodegenerative Diseases: An Overview of Recent Tools and Findings
Reactive oxygen species (ROS) have been described to induce a broad range of redoxdependent signaling reactions in physiological conditions. Nevertheless, an excessive accumulation of ROS leads to oxidative stress, which was traditionally considered as detrimental for cells and organisms, due to the oxidative damage they cause to biomolecules. During ageing, elevated ROS levels result in the accumulation of damaged proteins, which may exhibit altered enzymatic function or physical properties (e.g., aggregation propensity). Emerging evidence also highlights the relationship between oxidative stress and age-related pathologies, such as protein misfoldingbased neurodegenerative diseases (e.g., Parkinson's (PD), Alzheimer's (AD) and Huntington's (HD) diseases). In this review we aim to introduce the role of oxidative stress in physiology and pathology and then focus on the state-of-the-art techniques available to detect and quantify ROS and oxidized proteins in live cells and in vivo, providing a guide to those aiming to characterize the role of oxidative stress in ageing and neurodegenerative diseases. Lastly, we discuss recently published data on the role of oxidative stress in neurological disorders.
# Introduction
Oxidative stress can be defined as an imbalance between the generation of reactive oxygen species (ROS) and their quenching by antioxidants in a specific cell or tissue [fig_ref] Figure 1: Development of oxidative stress [/fig_ref]. One of the most studied ROS is hydrogen peroxide (H 2 O 2 ), but many other, more reactive species, such as superoxide anion radicals (O 2
- − ), hydroxyl radicals ( - OH), singlet oxygen ( 1 O 2 ), nitrogen dioxide (NO [bib_ref] Cellular Mechanisms and Physiological Consequences of Redox-Dependent Signalling, Holmström [/bib_ref] - − ), hypochlorous acid (HOCl) and peroxynitrite (ONOO − ) are generated by various enzymatic and non-enzymatic processes in biological systems [bib_ref] NOX Enzymes and the Biology of Reactive Oxygen, Lambeth [/bib_ref].
The excessive amount of ROS triggers the so-called redox-dependent signaling in physiological conditions. However, when it surpasses a threshold, ROS become detrimental. Apart from aberrant redox signaling, ROS can damage a myriad of cellular components, such as proteins, DNA and lipids. As several amino acid residues are amenable to oxidation (depending on the ROS involved), oxidative damage leads to a variety of oxidative modifications onto proteins. These modifications could alter cellular function by regulating stability, activity, subcellular localization or the protein-protein interaction of oxidized proteins [bib_ref] Cellular Mechanisms and Physiological Consequences of Redox-Dependent Signalling, Holmström [/bib_ref]. A wide range of cellular mechanisms are activated to eliminate ROS as well as the ROS-mediated oxidized biomolecules (such as the newly discovered protection of promoters from oxidative damage though the action of the nuclear mitotic apparatus protein, NuMA [bib_ref] A Mechanism for Oxidative Damage Repair at Gene Regulatory Elements, Ray [/bib_ref]. Nonetheless, upon persistent activation these mechanisms may either aggravate the cellular damage or become impaired, eventually leading to a vicious cycle that promotes the progression of several pathologies. In physiological conditions, ROS remain at low levels by the effect of antioxidants such as superoxide dismutase (SOD), catalase (CAT) and glutathione peroxidase (GSHPx). However, during oxidative stress, ROS levels rise, due to either higher production or impaired antioxidant defense (or both), and this results in the accumulation of damaged biomolecules. Such events are detrimental and either drive or exacerbate diseases such as neurodegenerative disorders.
Oxidative stress is implicated in a wide variety of age-related diseases ranging from diabetes and cardiovascular disease to cancer and neurodegenerative diseases [bib_ref] Guidelines for Measuring Reactive Oxygen Species and Oxidative Damage in Cells and..., Murphy [/bib_ref] [bib_ref] Chemistry and Biology of Reactive Oxygen Species in Signaling or Stress Responses, Dickinson [/bib_ref] [bib_ref] Targeting Oxidative Stress in Disease: Promise and Limitations of Antioxidant Therapy, Forman [/bib_ref]. A common characteristic of most neurodegenerative diseases is the formation of toxic protein aggregates that affect many functions, eventually leading to neuronal loss. Interestingly, several of these proteins can become oxidized. While this may not come as a surprise since protein oxidative damage is a non-discriminative process, studies on α-synuclein [bib_ref] Increased Oxidative Stress Exacerbates α-Synuclein Aggregation In Vivo, Scudamore [/bib_ref] , prion protein [bib_ref] Design of Anti-and pro-Aggregation Variants to Assess the Effects of Methionine Oxidation..., Wolschner [/bib_ref] and TDP-43 [bib_ref] TDP-43 Aggregation Induced by Oxidative Stress Causes Global Mitochondrial Imbalance in ALS, Zuo [/bib_ref] have shown that their oxidation is in fact required for their initial nucleation to take place. This may indicate that the oxidation of aggregation-prone proteins may constitute another common characteristic among neurological disorders. The role of oxidative stress in these diseases is further evident by the fact that oxidative damage has been consistently detected in the post-mortem brains of patients with Alzheimer's (AD), Parkinson's (PD) and Huntington's (HD) diseases [bib_ref] Cross-Talk between Redox Signalling and Protein Aggregation, Van Dam [/bib_ref]. While it is currently unknown whether oxidative stress is a driver or a consequence that further exacerbates them, these findings point to the dependency of disease progression on oxidative stress.
All the above highlight the need for further insight into the interplay between oxidative stress and neurodegenerative diseases. Several methods for assessing oxidative stress have been developed over the years. Beyond traditional (e.g., electron spin resonance (EPR)-based techniques and mass spectrometry [bib_ref] Taking up the Cudgels for the Traditional Reactive Oxygen and Nitrogen Species..., Daiber [/bib_ref] and commercial assays, these methods include novel tools for selectively detecting ROS and achieving global profiling of damaged proteins in live cells or organisms. While more reliable and informative, the exploitation of these tools is extremely limited and, when utilized, is often in fields other than neurodegenerative diseases. Hence, this review focuses on presenting both recent advanced tools for evaluating oxidative stress and findings that emphasize how oxidative stress may contribute to the progression of neurodegenerative disorders. In physiological conditions, ROS remain at low levels by the effect of antioxidants such as superoxide dismutase (SOD), catalase (CAT) and glutathione peroxidase (GSHPx). However, during oxidative stress, ROS levels rise, due to either higher production or impaired antioxidant defense (or both), and this results in the accumulation of damaged biomolecules. Such events are detrimental and either drive or exacerbate diseases such as neurodegenerative disorders.
Oxidative stress is implicated in a wide variety of age-related diseases ranging from diabetes and cardiovascular disease to cancer and neurodegenerative diseases [bib_ref] Guidelines for Measuring Reactive Oxygen Species and Oxidative Damage in Cells and..., Murphy [/bib_ref] [bib_ref] Chemistry and Biology of Reactive Oxygen Species in Signaling or Stress Responses, Dickinson [/bib_ref] [bib_ref] Targeting Oxidative Stress in Disease: Promise and Limitations of Antioxidant Therapy, Forman [/bib_ref]. A common characteristic of most neurodegenerative diseases is the formation of toxic protein aggregates that affect many functions, eventually leading to neuronal loss. Interestingly, several of these proteins can become oxidized. While this may not come as a surprise since protein oxidative damage is a non-discriminative process, studies on αsynuclein [bib_ref] Increased Oxidative Stress Exacerbates α-Synuclein Aggregation In Vivo, Scudamore [/bib_ref] , prion protein [bib_ref] Design of Anti-and pro-Aggregation Variants to Assess the Effects of Methionine Oxidation..., Wolschner [/bib_ref] and TDP-43 [bib_ref] TDP-43 Aggregation Induced by Oxidative Stress Causes Global Mitochondrial Imbalance in ALS, Zuo [/bib_ref] have shown that their oxidation is in fact required for their initial nucleation to take place. This may indicate that the oxidation of aggregation-prone proteins may constitute another common characteristic among neurological disorders. The role of oxidative stress in these diseases is further evident by the fact that oxidative damage has been consistently detected in the post-mortem brains of patients with Alzheimer's (AD), Parkinson's (PD) and Huntington's (HD) diseases [bib_ref] Cross-Talk between Redox Signalling and Protein Aggregation, Van Dam [/bib_ref]. While it is currently unknown whether oxidative stress is a driver or a consequence that further exacerbates them, these findings point to the dependency of disease progression on oxidative stress.
All the above highlight the need for further insight into the interplay between oxidative stress and neurodegenerative diseases. Several methods for assessing oxidative stress have been developed over the years. Beyond traditional (e.g., electron spin resonance (EPR)-based techniques and mass spectrometry [bib_ref] Taking up the Cudgels for the Traditional Reactive Oxygen and Nitrogen Species..., Daiber [/bib_ref] and commercial assays, these methods include novel tools for selectively detecting ROS and achieving global profiling of damaged proteins in live cells or organisms. While more reliable and informative, the exploitation of these tools is extremely limited and, when utilized, is often in fields other than neurodegenerative diseases. Hence, this review focuses on presenting both recent advanced tools for evaluating oxidative stress and findings that emphasize how oxidative stress may contribute to the progression of neurodegenerative disorders.
## Protein aggregation
Protein aggregation constitutes an example of this vicious cycle of negative effects, and it plays a key role in the pathology of neurodegenerative diseases. In the case of PD, AD and HD, the presence of aggregates of α-synuclein, Aβ peptide, and Huntingtin with polyglutamine expansion, respectively, alone or together with other known, or yetto-be-discovered peptides, has been linked to the symptomatology and progression of the diseases [bib_ref] Protein Aggregation and Neurodegenerative Disease, Ross [/bib_ref] [bib_ref] Mechanistic Models of Protein Aggregation Across Length-Scales and Time-Scales: From the Test..., Meisl [/bib_ref] [bib_ref] Role of Protein Aggregation in Neurodegenerative Diseases, Tutar [/bib_ref]. Compelling data suggest that the relationship between oxidative stress and protein aggregation is bidirectional, since aggregation-prone conformers exacerbate the production of ROS (e.g., Aβ has been shown to initiate lipid peroxidation [bib_ref] Redox Proteomics and Amyloid β-Peptide: Insights into Alzheimer Disease, Butterfield [/bib_ref] , whereas, in turn, oxidative stress promotes protein aggregation [fig_ref] Figure 2: Figure 2 [/fig_ref]. While molecular-level mechanistic details of the latter route are very sparse, in silico studies on prion protein have suggested that ROS could induce β-sheet formation (a form prone to aggregation) either directly or indirectly. In the first case, this is caused by destabilization of prion's physiological α-fold form due to the direct oxidation of methionine residues by ROS [fig_ref] Figure 2: Figure 2 [/fig_ref] , left) [bib_ref] Methionine Sulfoxides on Prion Protein Helix-3 Switch on the α-Fold Destabilization Required..., Colombo [/bib_ref]. In the second case, ROS (generated by copper reduction) initially induce cleavage of prion's N-terminus and, subsequently, the remaining copper promotes βsheet formation due to its chelation by neighboring histidine residues [fig_ref] Figure 2: Figure 2 [/fig_ref] , right) [bib_ref] Can Copper Binding to the Prion Protein Generate a Misfolded Form of..., Pushie [/bib_ref]. While these mechanisms need to be validated further and may not be representative for other disordered proteins, they certainly exemplify the complex relationship between oxidative stress and protein aggregation.
## Protein aggregation
Protein aggregation constitutes an example of this vicious cycle of negative effects, and it plays a key role in the pathology of neurodegenerative diseases. In the case of PD, AD and HD, the presence of aggregates of α-synuclein, Aβ peptide, and Huntingtin with polyglutamine expansion, respectively, alone or together with other known, or yet-to-bediscovered peptides, has been linked to the symptomatology and progression of the diseases [bib_ref] Protein Aggregation and Neurodegenerative Disease, Ross [/bib_ref] [bib_ref] Mechanistic Models of Protein Aggregation Across Length-Scales and Time-Scales: From the Test..., Meisl [/bib_ref] [bib_ref] Role of Protein Aggregation in Neurodegenerative Diseases, Tutar [/bib_ref]. Compelling data suggest that the relationship between oxidative stress and protein aggregation is bidirectional, since aggregation-prone conformers exacerbate the production of ROS (e.g., Aβ has been shown to initiate lipid peroxidation [bib_ref] Redox Proteomics and Amyloid β-Peptide: Insights into Alzheimer Disease, Butterfield [/bib_ref] , whereas, in turn, oxidative stress promotes protein aggregation [fig_ref] Figure 2: Figure 2 [/fig_ref]. While molecular-level mechanistic details of the latter route are very sparse, in silico studies on prion protein have suggested that ROS could induce β-sheet formation (a form prone to aggregation) either directly or indirectly. In the first case, this is caused by destabilization of prion's physiological α-fold form due to the direct oxidation of methionine residues by ROS [fig_ref] Figure 2: Figure 2 [/fig_ref] , left) [bib_ref] Methionine Sulfoxides on Prion Protein Helix-3 Switch on the α-Fold Destabilization Required..., Colombo [/bib_ref]. In the second case, ROS (generated by copper reduction) initially induce cleavage of prion's N-terminus and, subsequently, the remaining copper promotes βsheet formation due to its chelation by neighboring histidine residues [fig_ref] Figure 2: Figure 2 [/fig_ref] , right) [bib_ref] Can Copper Binding to the Prion Protein Generate a Misfolded Form of..., Pushie [/bib_ref]. While these mechanisms need to be validated further and may not be representative for other disordered proteins, they certainly exemplify the complex relationship between oxidative stress and protein aggregation. Interplay between protein aggregates and oxidative damage. (A) Increased protein aggregation in several neurodegenerative diseases is associated with increased ROS levels. This relationship is bidirectional; excessive ROS production is found to induce protein aggregation and vice versa. Two examples of this vicious cycle are shown: (1) a decrease in the mitochondrial levels of Superoxide dismutase 2 (SOD2) gives rise to elevated ROS, which in turn promotes Α-synuclein aggregation (right panel) and (2) the accumulation of Αβ oligomers induces lipid peroxidation, an oxidative damage that further results in protein carbonylation (left panel). (B) Proposed mechanisms of ROS-induced aggregation of prion protein. At physiological conditions, part of the prion protein exists in the α-fold form (depicted as a helix in the middle). However, in disease, its scrapie isoform forms β-sheets, which eventually drive fibrillation. The transition from α-fold to β-sheet is proposed to be mediated by ROS, with two different mechanisms: (1) by methionine oxidation (depicted as red circles), which destabilizes the α-fold form (left side) or (2) by ROS-mediated cleavage-induced by copper-of the protective N-terminus of prion, and subsequently, β-sheet However, in disease, its scrapie isoform forms β-sheets, which eventually drive fibrillation. The transition from α-fold to β-sheet is proposed to be mediated by ROS, with two different mechanisms: (1) by methionine oxidation (depicted as red circles), which destabilizes the α-fold form (left side) or (2) by ROS-mediated cleavage-induced by copper-of the protective N-terminus of prion, and subsequently, β-sheet formation due to copper chelation by neighboring histidine residues. The position of the residues shown is for illustration only and do not correspond to their true location.
In line with the above, it has been proposed that, during ageing, age-related damage in the brain activates stress response pathways that rely on redox signaling through the production of H 2 O 2 as a second messenger. It is widely accepted that this occurs as an attempt to restore homeostasis. Over time, the accumulation of damage leads to ROS concentrations that exceed the physiological levels required for activating signaling pathways. Hence, ROS shifts from its beneficial stress response-associated effects to its detrimental outcomes due to ROS-dependent damage through a vicious cycle that inevitably results in high ROS levels [bib_ref] Cross-Talk between Redox Signalling and Protein Aggregation, Van Dam [/bib_ref]. According to this hypothesis, damage due to the formation of aggregates, such as Aβ and tau oligomers in AD [bib_ref] Aβ Oligomers Induce Neuronal Oxidative Stress through an N-Methyl-D-Aspartate Receptor-Dependent Mechanism That..., De Felice [/bib_ref] [bib_ref] Du Mitochondrial Oxidative Stress Contributes to the Pathological Aggregation and Accumulation of..., Du [/bib_ref] , can induce ROS generation. While this has been shown for different types of aggregates, it is still debatable which aggregating species are responsible for inducing ROS generation. A study in induced pluripotent stem cells showed that both oligomeric and fibrillar α-synuclein were able to increase ROS levels [bib_ref] Alpha-Synuclein Oligomers Interact with Metal Ions to Induce Oxidative Stress and Neuronal..., Deas [/bib_ref]. However, another study in yeast using Aβ42 peptide variants with different aggregation propensity found that peptides that are unable to be recruited into inclusions induce ROS regeneration, while insoluble deposits do not [bib_ref] Protein Aggregation into Insoluble Deposits Protects from Oxidative Stress, Carija [/bib_ref]. Collectively, all the above underline the strong relationship between oxidative stress and protein aggregation in neurodegenerative diseases.
## Ros generation and ros-related processes
Oxidative stress is mediated by several exogenous (such as pathogens, dietary components, toxic substances, pharmaceutical drugs) and endogenous factors. With respect to endogenous pro-oxidant events, broadly speaking, this includes lowering the antioxidant defense (e.g., downregulation enzymatic or non-enzymatic antioxidants or loss-of-function mutations of enzymes) or inducing ROS generation and propagation (e.g., through the activation of mitochondrial NADPH oxidase). The dysfunction of mitochondria and peroxisomes (often interlinked), increase in free metal ions, as well as ROS themselves (through oxidative damage), could further promote oxidative stress. In the next sections we will focus on the relation between ROS and mitochondria, iron signaling and parthanatos.
## Major ros sources
Within cells, mitochondria produce ROS which are involved in both physiological signaling and the oxidative damage of different cell components, including the mitochondria itself. O 2
- − , the primal ROS produced in mitochondria, is generated by the one-electron reduction in molecular oxygen (O 2 ) during oxidative phosphorylation for the synthesis of ATP. However, mitochondria also generate H 2 O 2 , which is produced by the dismutation of O 2
- − catalyzed by mitochondrial SOD. Thus, diffusion-restricted O 2 - − derived from the electron transport chain (ETC) can be converted by SOD enzymes (especially SOD2) into highly diffusible secondary messenger H 2 O 2 [bib_ref] How Mitochondria Produce Reactive Oxygen Species, Murphy [/bib_ref]. Redox cofactors such as NADH are essential, both to combat oxidative stress and to facilitate reductive reactions in biosynthesis (for example, proline biosynthesis). The amounts and ratios of oxidized and reduced NAD(H) and NADP(H) affect biosynthetic capability and, subsequently, regulate cell proliferation [bib_ref] Escaping Death: Mitochondrial Redox Homeostasis in Cancer Cells, Ciccarese [/bib_ref] [bib_ref] Defective NADPH Production in Mitochondrial Disease Complex I Causes Inflammation and Cell..., Balsa [/bib_ref]. While mitochondria are vital to the cell survival, aberrant mitochondrial homeostasis (the so-called dysfunctional mitochondrial quality control, including dysregulated mitophagy) has been implicated in the pathogenesis of neurodegenerative diseases, such as AD [bib_ref] Dysfunctional Mitochondria and Mitophagy as Drivers of Alzheimer's Disease Pathogenesis. Front, Chakravorty [/bib_ref] , PDand HD [bib_ref] Mitochondrial Quality Control in Neurodegenerative Diseases: Focus on Parkinson's Disease and Huntington's..., Franco-Iborra [/bib_ref] , and it has, thus, been proposed as a promising therapeutic target [bib_ref] The Role of Mitophagy in Various Neurological Diseases as a Therapeutic Approach, Kaur [/bib_ref].
In addition to mitochondria, other sources of ROS exist inside the cell. Peroxisomes are organelles that participate in metabolic oxidative processes (β-oxidation of very-longchain fatty acids, α-oxidation of branched-chain fatty acids, oxidation of D-amino acids and polyamines), as well as the synthesis of ether phospholipids, bile acids and docosahexaenoic acid. These processes produce ROS as byproducts, which are subsequently neutralized by antioxidant enzymes, including catalases. Peroxisomes crosstalk with the nucleus, mitochondria, endoplasmic reticulum (ER) and lysosomes, and these inter-organelle communications are key to maintain peroxisome function. Of note, peroxisomal dysfunction has been described in disease and during ageing and is associated with many cellular alterations, highlighting its important roles in redox homeostasis.
## Ferroptosis
Iron is an important signaling molecule that is related to oxidative stress (for instance, it is required, along with H 2 O 2 , for the generation of - OH) and neurodegenerative disease. Consequently, the brains of patients with AD and PD contain high iron concentrations [bib_ref] Iron in Neuronal Function and Dysfunction, Salvador [/bib_ref] , and mechanistical studies highlight the importance of iron homeostasis for proper neuronal function, since both iron deficiency [bib_ref] Iron Deficiency on Neuronal Function, Muñoz [/bib_ref] and overload [bib_ref] Iron Exposure and the Cellular Mechanisms Linked to Neuron Degeneration in Adult..., Li [/bib_ref] negatively impact neuronal function.
Ferroptosis is an iron-dependent type of programmed cell death that involves oxidative stress, since it is accompanied by lipid peroxidation [bib_ref] Ferroptosis: Past, Present and Future, Li [/bib_ref]. It has been found that α-synuclein aggregation drives ferroptosis via aggregate-membrane interaction in human stem cell-derived models of synucleinopathy. Importantly, α-synuclein oligomers further induce lipid peroxidation, and the inhibition of lipid peroxidation blocks the aggregatemembrane interaction, indicating that the oxidation state of the lipid membrane is key for triggering ferroptosis in this experimental model of PD [bib_ref] Alpha Synuclein Aggregation Drives Ferroptosis: An Interplay of Iron, Calcium and Lipid..., Angelova [/bib_ref]. Therefore, stress-induced lipid peroxidation could promote ferroptosis and its accompanying neurodegeneration. Importantly, there is a complex interplay between iron regulation (and, subsequently, ferroptosis) and the alteration of mitochondrial homeostasis in a plethora of neurodegenerative diseases, including AD and PD [bib_ref] The Role of Ferroptosis in Neurodegenerative Diseases, Xu [/bib_ref] [bib_ref] Iron-Induced Energy Supply Deficiency and Mitochondrial Fragmentation in Neurons, Huang [/bib_ref]. The mechanisms underlying this crosstalk are under study. For instance, an in vivo study using peroxiredoxin 5 (Prx5)-knock out mice showed that Prx5 deficiency-induced iron overload causes neuronal toxicity and eventually activates neuronal death in the hippocampus through endoplasmic reticulum-mediated mitochondrial fission.
Numerous molecular mediators of ferroptosis have been suggested as pharmaceutical targets, including nuclear factor erythroid 2-related factor-antioxidant response element (Nrf2-ARE), n-acetylcysteine (NAC), Fe 2+ , NADPH, and its oxidases NOX, among others [bib_ref] Targeting Molecular Mediators of Ferroptosis and Oxidative Stress for Neurological Disorders, Li [/bib_ref] , underlying its significance in pathological conditions.
## Parthanatos
Parthanatos is a programmed cell death signaling pathway in which excessive oxidative damage to DNA by ROS (e.g., peroxynitrite (ONOO − ), among others) rapidly hyperactivates poly(ADP-ribose) polymerase (PARP). Hyperactivated PARP induces the downstream formation of large poly(ADP-ribose) polymers that subsequently promote the release of apoptosis-inducing factor (AIF) from the outer mitochondrial membrane into the cytosol, where it forms a complex with macrophage migration inhibitory factor (MIF). The complex translocates into the nucleus, where it triggers DNA degradation and cell death. Parthanatos underlies a wide spectrum of diseases, including those affecting the nervous system, such as AD, PD and stroke [bib_ref] Molecular Mechanisms of Parthanatos and Its Role in Diverse Diseases, Huang [/bib_ref].
## Recent developments for evaluating oxidative stress
During oxidative stress, high levels of ROS result in damage to proteins, nucleic acids and lipids [bib_ref] Targeting Oxidative Stress in Disease: Promise and Limitations of Antioxidant Therapy, Forman [/bib_ref]. Due to its implication in several diseases, the development of methods for evaluating oxidative stress has been an important aspect in the field. Oxidative stress can be assessed by determining ROS levels, the oxidative damage or the antioxidant status. Methods for measuring ROS [bib_ref] Chemistry and Biology of Reactive Oxygen Species in Signaling or Stress Responses, Dickinson [/bib_ref] [bib_ref] Taking up the Cudgels for the Traditional Reactive Oxygen and Nitrogen Species..., Daiber [/bib_ref] , oxidized biomolecules (proteins [bib_ref] Chemical Probes for Redox Signaling and Oxidative Stress, Abo [/bib_ref] , lipids [bib_ref] Detection and Structural Analysis of Lipid-Derived Radicals in Vitro and in Vivo, Matsuoka [/bib_ref] [bib_ref] Fluorescence Imaging for Visualizing the Bioactive Molecules of Lipid Peroxidation within Biological..., Su [/bib_ref] and nucleic acids [bib_ref] Potential Application of the Oxidative Nucleic Acid Damage Biomarkers in Detection of..., Guo [/bib_ref] and antioxidant status [bib_ref] Approaches and Methods to Measure Oxidative Stress in Clinical Samples: Research Applications..., Katerji [/bib_ref] [bib_ref] Conventional and Innovative Methods to Assess Oxidative Stress Biomarkers in the Clinical..., Gaggini [/bib_ref] have been previously reviewed and, importantly, good practice guidelines for such measurements were very recently set out by leading experts in the field [bib_ref] Guidelines for Measuring Reactive Oxygen Species and Oxidative Damage in Cells and..., Murphy [/bib_ref]. Herein, we focus on recent and notable developments and aim not only to inform the readers, but also to inspire further research into the field. We Antioxidants 2023, 12, 131 6 of 21 limit our discussion to methods for both the direct measurements of ROS and the oxidative damage (indirect effect) on proteins. All tools, along with their application method and their key features, reported in this section are summarized in [fig_ref] Table 1: Summary of tools reported in Section 4 [/fig_ref] to assist the reader in locating the most suitable tool for their purpose. HQ a Two-photon excitability [bib_ref] A Two-Photon Fluorescent Probe for Exogenous and Endogenous Superoxide Anion Imaging in..., Li [/bib_ref] MitoHCy-NH 2 a Detection of ROS generated by a certain protein [bib_ref] Ratiometric Near-Infrared Fluorescent Probe for Synergistic Detection of Monoamine Oxidase B and..., Wang [/bib_ref] Peroxynitrite (ONOO − )
## Mito-cc a
FRET-based, mitochondria targeting probe with two-photon excitability [bib_ref] Selective Visualization of the Endogenous Peroxynitrite in an Inflamed Mouse Model by..., Cheng [/bib_ref] sfGFP(66Thy) b
Turn-on reporters with non-natural amino acids [bib_ref] Genetically Encoding Thyronine for Fluorescent Detection of Peroxynitrite, Li [/bib_ref] pnGFP-Ultra b [bib_ref] A High-Performance Genetically Encoded Fluorescent Biosensor for Imaging Physiological Peroxynitrite, Chen [/bib_ref] Hydrogen Carbonylation m-APA a High reaction kinetics and selectivity [bib_ref] Quantitative Profiling of Protein Carbonylations in Ferroptosis by an Aniline-Derived Probe, Chen [/bib_ref] TFCH a Imaging High sensitivity and suitable for live cell imaging [bib_ref] A Novel Fluorogenic Assay for the Detection of Nephrotoxin-Induced Oxidative Stress in..., Mukherjee [/bib_ref] a Small molecule probe, b genetically encoded reporter.
## Direct probing of ros
To understand whether oxidative stress is implicated in a certain condition or disease, it is crucial to define ROS levels. A variety of ROS are produced either by natural cellular processes or by the reactions of ROS themselves with other molecules (e.g., hypohalous acids are generated by the enzymatic peroxidation of halogen ions with H 2 O 2 [bib_ref] Reactions and Reactivity of Myeloperoxidase-Derived Oxidants: Differential Biological Effects of Hypochlorous and..., Pattison [/bib_ref]. Deciphering which particular ROS is in excess in a certain disease background could offer an insight into the cellular mechanism involved. While the reaction rates of the different ROS vary (e.g., radical vs. non-radical ROS), ROS have overlapping chemical properties, rendering their specific identification challenging. As such, great effort has been made to develop small molecule probes and genetically encoded reporters for discriminating ROS within the cellular milieu. Since fluorescence microscopy is the preferred method in most biological experiments (for both live cell or in vivo samples), in this section we focus on fluorescence imaging tools, as they can be used together with other fluorescently labeled molecules or markers.
## Hypohalous acids (hox)
HOX are ROS which are produced by peroxidases in the presence of H 2 O 2 and any of chloride, bromide or thiocyanate (a pseudohalide). As opposed to hypothiocyanous acid (HOSCN), HOCl and hypobromous (HOBr) acids are strong oxidants and less selective [bib_ref] Guidelines for Measuring Reactive Oxygen Species and Oxidative Damage in Cells and..., Murphy [/bib_ref]. Among the three, HOCl has attracted a lot more attention as ROS since chloride plasma concentration is much higher compared to the others [bib_ref] Reaction of Myeloperoxidase Compound I with Chloride, Bromide, Iodide, and Thiocyanate, Furtmüller [/bib_ref].
Despite the variety of previously reported probes [bib_ref] Recent Development of Synthetic Probes for Detection of Hypochlorous Acid/Hypochlorite, Ma [/bib_ref] , HOCl probes with near-infrared (NIR) emission and high sensitivity (i.e., at low nanomolar range), both desired properties for in vivo application, were missing. To this end, Yi and co-workers reported the development of FDOCl-1, a turn-on probe with excellent selectivity and high sensitivity to HOCl. The probe was successfully used to monitor HOCl production in the joints of inflammation on a mouse model of arthritis [bib_ref] Deformylation Reaction-Based Probe for in Vivo Imaging of HOCl, Wei [/bib_ref]. The team subsequently coupled the probe onto a green-emitting naphthalene (generating probe FDOCl-18) to enable visualization of the probe distribution in the "off" mode [bib_ref] Release of Amino-or Carboxy-Containing Compounds Triggered by HOCl: Application for Imaging and..., Wei [/bib_ref]. Upon encountering HOCl, the naphthalene falls off and the NIR dye turns on [fig_ref] Figure 3: Tools for the visualization of hypohalous acids [/fig_ref]. Due to their high reactivity and low permeability, some ROS do not escape the site of their production [bib_ref] Chemistry and Biology of Reactive Oxygen Species in Signaling or Stress Responses, Dickinson [/bib_ref]. This is why it is highly desirable to have tools to track ROS at The methylene blue (top structure) is conjugated to naphthalene (indicated as a green star) through the HOCl-sensing group (highlighted by the gray oval, left structure). The methylene blue is non-fluorescent while the sensing group is attached. Upon sensing HOCl, a deformylation mechanism takes place that leads to the liberation of the methylene blue, which is now fluorescent. (B) Sensing mechanism of the organelle-specific turn-on probe for HOCl. The fluorescence properties of a two-photon dye (acedan) were quenched by HOCl-labile group (gray circle, left structure). The probe was further equipped with organelle specific-targeted moieties to drive it either to mitochondria (MITO-TP) or to lysosomes (LYSO-TP). (C) Structure of the HOBr ratiometric probe RhSN-mito. In the absence of HOBr, the probe emits at 530 nm. Reaction of the sensing moiety (gray oval, top structure) with HOBr leads to the formation of an additional ring that alters the fluorescence properties of the probe, which now emits at 663 nm. Due to their high reactivity and low permeability, some ROS do not escape the site of their production [bib_ref] Chemistry and Biology of Reactive Oxygen Species in Signaling or Stress Responses, Dickinson [/bib_ref]. This is why it is highly desirable to have tools to track ROS at the subcellular level. HOCl is one such ROS as it reacts rapidly with nearby thiolates [bib_ref] Reconciling the Chemistry and Biology of Reactive Oxygen Species, Winterbourn [/bib_ref]. Chang and co-workers therefore developed turn-on probes that enable monitoring HOCl distribution at specific organelles . The probes were constructed by combining a twophoton dye with chemical moieties that target either the mitochondria or the lysosomes (termed, MITO-TP and LYSO-TP, respectively, [fig_ref] Figure 3: Tools for the visualization of hypohalous acids [/fig_ref]. The probes were successfully applied on inflamed murine models and showed that both the mitochondria and lysosomes of macrophage cells generate higher levels of HOCl during inflammation.
While many HOCl-sensing probes have been shown to be selective against other ROS, selectivity is not usually assessed against HOBr or HOSCN. This may imply that such HOCl probes cannot discriminate between the different hypohalous acids. This is not the case for the HOBr-sensing probes reported by Tao and co-workers. This ratiometric probe, namely RhSN-mito, was initially developed based on a previous observation that HOBr, but not HOCl, crosslinks the sulfur of methionine with proximal amines on collagen IV [bib_ref] Bromine Is an Essential Trace Element for Assembly of Collagen IV Scaffolds..., Mccall [/bib_ref]. The probe was shown to be selective for HOBr against many ROS, including HOCl, and suitable for in vivo imaging in zebrafish. The authors later reported a derivative of the first probe, which now exhibited NIR emission upon encountering HOBr, an improvement better suited for in vivo application [fig_ref] Figure 3: Tools for the visualization of hypohalous acids [/fig_ref] [bib_ref] High-Quantum-Yield Mitochondria-Targeting Near-Infrared Fluorescent Probe for Imaging Native Hypobromous Acid in Living..., Liu [/bib_ref].
Genetically encoded fluorescent reporters for monitoring ROS are perhaps more often the method of choice, and such tools have offered valuable insight in the field. However, the development of such reporters for hypohalous acids was considered very challenging. This is because many ROS reporters rely on the oxidation of cysteines to modify their fluorescent properties (either switch on or off). While hypohalous acids react rapidly with cysteines, selectivity is hard to be achieved as cysteines are susceptible to oxidation by many other ROS. In order to develop a HOX-selective biosensor, Belousov and co-workers utilized the E. coli-derived transcription repressor NemR [bib_ref] Hypocrates Is a Genetically Encoded Fluorescent Biosensor for (Pseudo)Hypohalous Acids and Their..., Kostyuk [/bib_ref] , a protein that had previously been shown to contain a certain cysteine (Cys106) prone to oxidation by chlorine species [bib_ref] Does the Transcription Factor NemR Use a Regulatory Sulfenamide Bond to Sense..., Gray [/bib_ref]. The team therefore developed the so-called Hypocrates sensor by integrating a YFP protein between the DNA-binding and the sensory domain of NemR, the latter containing Cys106 [fig_ref] Figure 3: Tools for the visualization of hypohalous acids [/fig_ref]. The reporter was shown to be sensitive to all three hypohalous acids with unexpectedly very good selectivity over other ROS. The sensor was successfully applied to live cells and in vivo.
## Superoxide anion (o 2
- − )
While not very reactive, O 2 - − is one of the most important ROS; not only is it generated by two major ROS sources (ETC [bib_ref] Superoxide Flashes in Single Mitochondria, Wang [/bib_ref] and NADPH [bib_ref] NOX Enzymes and the Biology of Reactive Oxygen, Lambeth [/bib_ref] , it is also a precursor of other ROS such as H 2 O 2 and ONOO − [bib_ref] Targeting Oxidative Stress in Disease: Promise and Limitations of Antioxidant Therapy, Forman [/bib_ref]. Despite its importance, traditional methods for measuring O 2
- − have certain limitations and may not provide accurate measurements, as recently outlined [bib_ref] Guidelines for Measuring Reactive Oxygen Species and Oxidative Damage in Cells and..., Murphy [/bib_ref]. For instance, while the circularly permuted yellow fluorescent protein (cpYFP) was initially reported as an O 2
- − sensor [bib_ref] Superoxide Flashes in Single Mitochondria, Wang [/bib_ref] , it was later shown to be completely unresponsive to it [bib_ref] The 'Mitoflash' Probe CpYFP Does Not Respond to Superoxide, Schwarzländer [/bib_ref] , making it imperative that more reliable tools are developed. Hua and co-workers reported the development of DMPS-O, a turn-on small molecule probe that exhibited excellent selectivity and detection limit at low nanomolar range [bib_ref] A Turn-on Mitochondria-Targeted near-Infrared Fluorescent Probe with a Large Stokes Shift for..., Xu [/bib_ref]. DMPS-O has NIR emission, making it suitable for in vivo imaging, and, importantly, a mitochondriatargeting group that enables measuring O 2
- − at its production site, a crucial aspect given that it is short-lived. While it was not assessed in any living organism, such probes could be used in in vivo systems, as shown by Liu and co-workers, who developed a turn-on probe, namely HQ, using a dye with two-photon excitation that minimizes damage and has better penetration [bib_ref] A Two-Photon Fluorescent Probe for Exogenous and Endogenous Superoxide Anion Imaging in..., Li [/bib_ref]. The suitability of HQ was demonstrated in a lung inflammation mouse model to monitor endogenous O 2
- − fluxes. Lastly, an interesting concept for monitoring ROS production at the protein level was shown by Chen and co-workers [bib_ref] Ratiometric Near-Infrared Fluorescent Probe for Synergistic Detection of Monoamine Oxidase B and..., Wang [/bib_ref]. Monoamine oxidase B (MAO-B) is a mitochondrial-localized enzyme that generates ROS as a by-product through the oxidation of biogenic amines. MAO-B has attracted considerable interest, for it is found to be elevated during aging as well as in several neurodegenerative disorders. To facilitate the detection of ROS production by MAO-B, the authors developed a ratiometric cyanine-based probe (termed MitoHCy-NH 2 ) equipped with a MAO-B-and a mitochondria-targeting group. They successfully showed that their probe can monitor activity and ROS generation by MAO-B in cells and in an aging mouse model.
[formula] 4.1.3. Peroxynitrite (ONOO − ) [/formula]
As mentioned earlier, ONOO − is generated by the reaction of O 2 - − with nitric oxide (NO - ). Although the reaction is not catalyzed by any enzyme (in contrast to ROS already discussed), ONOO − production is likely restricted to sites where O 2
- − is present, since the latter is short-lived and impermeable [bib_ref] Peroxynitrite: Biochemistry, Pathophysiology and Development of Therapeutics, Szabó [/bib_ref]. ONOO − is highly reactive and apart from causing oxidative damage (e.g., tyrosine nitration and lipid peroxidation), it is also responsible for generating other reactive species such as nitrogen dioxide, carbonate and radicals [bib_ref] Genetically Encoding Thyronine for Fluorescent Detection of Peroxynitrite, Li [/bib_ref]. Tools that selectively detect ONOO − over other reactive ROS could contribute to deepening our understanding of its role in health and disease.
Various small molecule probes, equipped with different chemical groups for sensing ONOO − , have been developed over the years. However, selectivity was in many cases an issue, since those groups would also be targeted by other ROS, given the low in cellulo concentrations of ONOO − compared to other ROS. Recent efforts have yielded exceptionally selective turn-on [bib_ref] A Far-Red-Emitting Fluorescence Probe for Sensitive and Selective Detection of Peroxynitrite in..., Wu [/bib_ref] or -off [bib_ref] A Naphthalimide-Based Lysosome-Targeting Fluorescent Probe for the Selective Detection and Imaging of..., Qian [/bib_ref] probes, with additional desirable features for either in vivo application, such as NIR emission [bib_ref] A Far-Red-Emitting Fluorescence Probe for Sensitive and Selective Detection of Peroxynitrite in..., Wu [/bib_ref] , two-photon excitabilityor organelle-specific detection (e.g., lysosomal localization [bib_ref] A Naphthalimide-Based Lysosome-Targeting Fluorescent Probe for the Selective Detection and Imaging of..., Qian [/bib_ref]. A further improvement of the above examples constitutes the probe developed by Chang and co-workers. The team reported the development of MITO-CC, a FRET-based probe with two-photon excitability that can selectively detect ONOO − in mitochondria [bib_ref] Selective Visualization of the Endogenous Peroxynitrite in an Inflamed Mouse Model by..., Cheng [/bib_ref]. MITO-CC emits at the NIR spectrum in the absence of ONOO − but at the UV-vis range upon its detection, as the FRET acceptor falls off and decomposes [fig_ref] Figure 4: Peroxynitrite-selective tools and their mechanism of detection [/fig_ref]. The authors demonstrated that the probe could detect endogenous ONOO − in both live cells and a mouse inflammation model.
Enriching the tools available for the selective detection of ONOO − with genetically encoded reporters is equally challenging, as in the case of HOX-sensing reporters. A way to overcome the limitations related to protein manipulations permitted by traditional approaches (i.e., mutations) is by genetic code expansion [bib_ref] Designer Proteins: Applications of Genetic Code Expansion in Cell Biology, Davis [/bib_ref] , as exemplified by two studies. In an attempt to develop fluorescent protein reporters for ONOO − , two independent teams employed genetic code expansion to create GFPs with non-canonical amino acids, namely sfGFP(66Thy) and pnGFP-Ultra, respectively [bib_ref] Genetically Encoding Thyronine for Fluorescent Detection of Peroxynitrite, Li [/bib_ref] [bib_ref] A High-Performance Genetically Encoded Fluorescent Biosensor for Imaging Physiological Peroxynitrite, Chen [/bib_ref]. The non-canonical amino acids used were tyrosine mimics, and they were chosen to contain peroxynitrite-labile moieties. Genetic code expansion enabled them to replace the tyrosine that is part of GFP's chromophore with their non-canonical amino acids so that the protein is not fluorescent. Upon reaction with ONOO − , but not with other ROS, the moieties fall off and GFP turns on [fig_ref] Figure 4: Peroxynitrite-selective tools and their mechanism of detection [/fig_ref]. While none of the studies used their reporters in an in vivo system, genetic code expansion can be applied in animals [bib_ref] Using a Quadruplet Codon to Expand the Genetic Code of an Animal, Xi [/bib_ref].
nelle-specific detection (e.g., lysosomal localization [bib_ref] A Naphthalimide-Based Lysosome-Targeting Fluorescent Probe for the Selective Detection and Imaging of..., Qian [/bib_ref]. A further improvement of the above examples constitutes the probe developed by Chang and co-workers. The team reported the development of MITO-CC, a FRET-based probe with two-photon excitability that can selectively detect ONOO − in mitochondria [bib_ref] Selective Visualization of the Endogenous Peroxynitrite in an Inflamed Mouse Model by..., Cheng [/bib_ref]. MITO-CC emits at the NIR spectrum in the absence of ONOO − but at the UV-vis range upon its detection, as the FRET acceptor falls off and decomposes [fig_ref] Figure 4: Peroxynitrite-selective tools and their mechanism of detection [/fig_ref]. The authors demonstrated that the probe could detect endogenous ONOO − in both live cells and a mouse inflammation model. The probe consists of a FRET donor and an acceptor as well as a linker, which is also the group responsible for sending the probe into mitochondria. In the absence of ONOO − , the probe emits at 650 nm due to energy transfer from the donor to the acceptor. However, upon reaction with ONOO − , the FRET acceptor is eliminated and decomposed, and the remaining donor fluoresces at 450 nm. (B) GFP variants sfGFP(66Thy) and pnGFP-Ultra with non-natural amino acids. Genetic code expansions on GFP using non-natural amino acids (shown in the boxes) led to the generation of GFPs that are non-fluorescent. The non-natural amino acids are equipped with ONOOsensitive groups, which upon reaction are cleaved off and the resulting GFP is then fluorescent.
Enriching the tools available for the selective detection of ONOO − with genetically encoded reporters is equally challenging, as in the case of HOX-sensing reporters. A way to overcome the limitations related to protein manipulations permitted by traditional approaches (i.e., mutations) is by genetic code expansion [bib_ref] Designer Proteins: Applications of Genetic Code Expansion in Cell Biology, Davis [/bib_ref] , as exemplified by two studies. In an attempt to develop fluorescent protein reporters for ONOO − , two independent teams employed genetic code expansion to create GFPs with non-canonical amino acids, namely sfGFP(66Thy) and pnGFP-Ultra, respectively [bib_ref] Genetically Encoding Thyronine for Fluorescent Detection of Peroxynitrite, Li [/bib_ref] [bib_ref] A High-Performance Genetically Encoded Fluorescent Biosensor for Imaging Physiological Peroxynitrite, Chen [/bib_ref]. The non-canonical amino acids used were tyrosine mimics, and they were chosen to contain peroxynitrite-labile moieties. Genetic code expansion enabled them to replace the tyrosine that is part of GFP's chromophore with their non-canonical amino acids so that the protein is not fluorescent. Upon The probe consists of a FRET donor and an acceptor as well as a linker, which is also the group responsible for sending the probe into mitochondria. In the absence of ONOO − , the probe emits at 650 nm due to energy transfer from the donor to the acceptor. However, upon reaction with ONOO − , the FRET acceptor is eliminated and decomposed, and the remaining donor fluoresces at 450 nm. (B) GFP variants sfGFP(66Thy) and pnGFP-Ultra with non-natural amino acids. Genetic code expansions on GFP using non-natural amino acids (shown in the boxes) led to the generation of GFPs that are non-fluorescent. The non-natural amino acids are equipped with ONOO − sensitive groups, which upon reaction are cleaved off and the resulting GFP is then fluorescent.
## Hydrogen peroxide (h 2 o 2 )
H 2 O 2 is a non-radical ROS that is mainly produced by superoxide dismutase (SOD) through the dismutation of O 2
- − . Despite its low reactivity, H 2 O 2 is a key player in oxidative stress since it can not only modify cysteines on certain proteins, but also because it is a precursor of the highly reactive hydroxyl radical. Out of a large variety of tools available for H 2 O 2 detection, herein we discuss two examples, one small molecule probe and one protein reporter, that offer exceptional advantages over previous tools. It should be noted that the low reactive nature of H 2 O 2 means that tool selectivity cannot be high, and thus, additional control experiments should be performed to dissect the contribution of other ROS.
As discussed earlier, ROS probes could be designed to target some specific organelles (i.e., mitochondria or lysosome). While useful, this approach does not allow the indiscriminate characterization of which organelles or cellular processes could, for instance, induce ROS generation. Hamachi and co-workers reported a novel approach to fill this gap. The authors carefully designed Hyp-L, a small molecule probe that, upon reacting with H 2 O 2 , forms an intermediate that rapidly reacts with proximal proteins, forming covalent adducts [fig_ref] Figure 5: Tools for detecting H2O2 [/fig_ref] [bib_ref] Imaging and Profiling of Proteins under Oxidative Conditions in Cells and Tissues..., Zhu [/bib_ref]. Hyp-L was equipped with a dye (for which a commercial antibody exists) to allow imaging and, importantly, pull-down experiments of the probe-tagged proteins. With this H 2 O 2 -responsive protein labeling, they were able to profile the H 2 O 2 -surrounding proteome and show that activated RAW264.7 macrophages, autophagosomes, as well as phagosomes, endosomes and lysosomes, are rich in H 2 O 2 .
authors carefully designed Hyp-L, a small molecule probe that, upon reacting with H2O2, forms an intermediate that rapidly reacts with proximal proteins, forming covalent adducts [fig_ref] Figure 5: Tools for detecting H2O2 [/fig_ref] [bib_ref] Imaging and Profiling of Proteins under Oxidative Conditions in Cells and Tissues..., Zhu [/bib_ref]. Hyp-L was equipped with a dye (for which a commercial antibody exists) to allow imaging and, importantly, pull-down experiments of the probe-tagged proteins. With this H2O2-responsive protein labeling, they were able to profile the H2O2surrounding proteome and show that activated RAW264.7 macrophages, autophagosomes, as well as phagosomes, endosomes and lysosomes, are rich in H2O2. Two significant benefits of ratiometric imaging are the monitoring of the distribution of the probe that is responsive to a certain stimulus (and hence, it does not emit light in its absence or presence) and allowing accurate quantitative analysis. In the field of protein reporters for ROS, this has been achieved either by the use of a FRET pair or ratiometric probes. Both types, however, have some limitations, namely, the size of the construct (for FRET pair probes), degradation or time lag of maturation of one of the FRET pair proteins, or the use of two excitation lasers (for ratiometric probes). To enable ratiometric imaging free of the aforementioned limitations, Hisabori and co-workers developed a ROS reporter that uses the excitation state intramolecular proton transfer (ESIPT) mechanism [bib_ref] Real-Time Monitoring of the in Vivo Redox State Transition Using the Ratiometric..., Sugiura [/bib_ref]. ESIPTbased reporters are constituted by one fluorescent protein that, despite being excited with a single wavelength, emits light on two different wavelengths, depending on the stimulus. Following several rounds of mutations, the authors successfully engineered a GFP sensor (termed FROG/B) which constitutes the first example of an ESIPT-based protein reporter for detecting ROS [fig_ref] Figure 5: Tools for detecting H2O2 [/fig_ref]. The authors used FROG/B to visualize redox changes in cyanobacterial cells.
## Mapping protein oxidative damage
While characterizing which ROS is elevated, at which organelles and to what extent, it is often challenging to obtain the full picture of events. Apart from potential issues associated with probe selectivity, distribution or stability, this is because of the intrinsic reactivity of ROS, which renders them difficult to be detected before reacting with a wealth of surrounding biomolecules (unless extremely high probe concentration is used). Assessing oxidative stress by looking downstream of ROS production (e.g., protein oxidative damage) largely overcomes these issues. Understanding the oxidative damage of proteins may also offer additional insight regarding protein function (e.g., inhibition [bib_ref] Reversible Inactivation of Protein-Tyrosine Phosphatase 1B in A431 Cells Stimulated with Epidermal..., Lee [/bib_ref] or activation [bib_ref] Methionine Oxidation Activates Pyruvate Kinase M2 to Promote Pancreatic Cancer Metastasis, He [/bib_ref] upon ROS-mediated oxidation). In this section, we present recent applications of small molecule probes for visualizing or identifying oxidized proteins. For the latter, we focus on probes for mass spectrometry (MS)-based chemical proteomics [bib_ref] Activity-Based Protein Profiling: From Enzyme Chemistry to Proteomic Chemistry, Cravatt [/bib_ref] , the general workflow of which is shown in [fig_ref] Figure 6: Chemical proteomics-based identification of oxidized proteins [/fig_ref].
it is often challenging to obtain the full picture of events. Apart from potential issues associated with probe selectivity, distribution or stability, this is because of the intrinsic reactivity of ROS, which renders them difficult to be detected before reacting with a wealth of surrounding biomolecules (unless extremely high probe concentration is used). Assessing oxidative stress by looking downstream of ROS production (e.g., protein oxidative damage) largely overcomes these issues. Understanding the oxidative damage of proteins may also offer additional insight regarding protein function (e.g., inhibition [bib_ref] Reversible Inactivation of Protein-Tyrosine Phosphatase 1B in A431 Cells Stimulated with Epidermal..., Lee [/bib_ref] or activation [bib_ref] Methionine Oxidation Activates Pyruvate Kinase M2 to Promote Pancreatic Cancer Metastasis, He [/bib_ref] upon ROS-mediated oxidation). In this section, we present recent applications of small molecule probes for visualizing or identifying oxidized proteins. For the latter, we focus on probes for mass spectrometry (MS)-based chemical proteomics [bib_ref] Activity-Based Protein Profiling: From Enzyme Chemistry to Proteomic Chemistry, Cravatt [/bib_ref] , the general workflow of which is shown in [fig_ref] Figure 6: Chemical proteomics-based identification of oxidized proteins [/fig_ref]. and oxidized (indicated as dark gray. Oxidized residue is depicted as a red star) proteins is treated with a probe specific for an oxidized amino acid. Upon reaction with the target protein, a covalent complex is formed. Following cell lysis, the probe-modified proteins are biotinylated (using click chemistry) and enriched with streptavidin beads. Finally, digestion and analysis by LC-MS/MS [fig_ref] Figure 6: Chemical proteomics-based identification of oxidized proteins [/fig_ref]. Chemical proteomics-based identification of oxidized proteins. (A) General workflow of chemical proteomics. The proteome of interest consisting of non-oxidized (indicated as light gray) and oxidized (indicated as dark gray. Oxidized residue is depicted as a red star) proteins is treated with a probe specific for an oxidized amino acid. Upon reaction with the target protein, a covalent complex is formed. Following cell lysis, the probe-modified proteins are biotinylated (using click chemistry) and enriched with streptavidin beads. Finally, digestion and analysis by LC-MS/MS leads to target identification. (B) Tagging of proteins with small molecule probes. Upon oxidation with ROS, amino acid residues are modified. The structures of the probes targeting different amino acids, suitable either for imaging (red structures) or chemical proteomics, are shown in boxes. Gray arrows show which probes react with which amino acids.
## Oxidized methionine
Methionine is a sulfur-containing amino acid and, as such, it can first be oxidized to sulfoxide and subsequently to sulfone. Oxidized methionine has been shown to play a pivotal role in strengthening protein-protein interactions [bib_ref] Direct Redox Regulation of F-Actin Assembly and Disassembly by Mical, Hung [/bib_ref] or promoting protein aggregation [bib_ref] Design of Anti-and pro-Aggregation Variants to Assess the Effects of Methionine Oxidation..., Wolschner [/bib_ref] , and so it is essential to develop methods for identifying proteins with such oxidative forms. Unfortunately, no probes have been reported so far to target either of these forms. In order to bypass the need for oxidized methionine-targeted probes, Chang and co-workers recently developed the first ever selective probe targeting non-oxidized methionine (ReACT, [fig_ref] Figure 6: Chemical proteomics-based identification of oxidized proteins [/fig_ref] , which had previously not been possible due to the weak nucleophilicity of methionine. The idea behind their approach was that oxidized methionine-containing proteins could be indirectly identified if the pool of probe-bound proteins between control (no ROS) and sample (with ROS) were compared (since ReACT can facilitate the MS-based identification of methionine-containing proteins). Application of this workflow recently demonstrated the power of this tool, as it revealed that pancreatic cancer metastasis is driven by methionine oxidation-mediated activation of pyruvate kinase M2 [bib_ref] Methionine Oxidation Activates Pyruvate Kinase M2 to Promote Pancreatic Cancer Metastasis, He [/bib_ref].
## Oxidized cysteine
Owing to its high reactivity among amino acids, cysteines (catalytic or not) react rapidly with ROS and become oxidized. Upon oxidation, the thiol group can form disulfide bonds, become nitrosylated, or be converted to sulfenic, sulfinic and sulfonic acids. Herein, we focus only on probes that directly recognize certain modifications (i.e., sulfenic and sulfinic acids), but it should be noted that it is possible to extract information for a certain cysteine modification for which no specific probe exists, as discussed for methionine. For instance, Zahedi and co-workers reported a workflow for specifically isolating proteins that had undergone S-nitrosylation [bib_ref] Proteome-Wide Detection of S-Nitrosylation Targets and Motifs Using Bioorthogonal Cleavable-Linker-Based Enrichment and..., Mnatsakanyan [/bib_ref].
Most reported probes for both the imaging and profiling of sulfenic or sulfinic acidmodified proteins have been developed by the Carrol lab. Due to cysteine's reactivity, the difficulty here is to develop probes that react with one oxidized form but not with cysteine. Following a comparative analysis of several newly synthesized probes, Carrol and coworkers designed a probe, termed BTD, which exhibited great reactivity, as demonstrated by the identification of over 700 sulfenic acid-modified proteins against a colon cancer cell proteome [bib_ref] Diverse Redoxome Reactivity Profiles of Carbon Nucleophiles, Gupta [/bib_ref]. They subsequently developed two new probes: (1) the mitochondriatargeted WYneN, which exhibited a 10-fold increase in kinetics [bib_ref] Wittig Reagents for Chemoselective Sulfenic Acid Ligation Enables Global Site Stoichiometry Analysis..., Shi [/bib_ref] , and (2) a fluorogenic probe, namely CysOx, for live cell imaging [bib_ref] Reaction-Based Fluorogenic Probes for Detecting Protein Cysteine Oxidation in Living Cells, Ferreira [/bib_ref] [fig_ref] Figure 6: Chemical proteomics-based identification of oxidized proteins [/fig_ref]. Similarly, the Carrol team also developed probes for sulfinic acid-modified proteins [bib_ref] A Chemical Approach for the Detection of Protein Sulfinylation, Lo Conte [/bib_ref] [bib_ref] Chemical Proteomics Reveals New Targets of Cysteine Sulfinic Acid Reductase, Akter [/bib_ref] , with their latest one used to exhibit higher selectivity over non-oxidized cysteine (DiaAlk, [fig_ref] Figure 6: Chemical proteomics-based identification of oxidized proteins [/fig_ref]. The application of this probe in different cell lines revealed several interesting findings, such as the fact that the extent of S-sulfinylation of certain proteins is cell line-dependent (despite having been exposed to the same concentration of H 2 O 2 ) [bib_ref] Chemical Proteomics Reveals New Targets of Cysteine Sulfinic Acid Reductase, Akter [/bib_ref].
## Protein carbonylation
Protein carbonylation is a biomarker of oxidative stress, since it is a modification that can take place on various amino acids as well as on the protein backbone. This is because carbonylation can result from three different mechanisms: (1) the direct oxidation of amino acids or protein backbone, (2) the insertion of oxidized lipids (through lipid peroxidation) onto amino acids, and (3) the formation of advanced glycation end products [bib_ref] Oxidative Stress Induced Carbonylation in Human Plasma, Madian [/bib_ref]. Due to the reactivity of the carbonyl groups, several probes have been developed over the years. A notable example is the carbonyl-targeted probe developed by Wang and co-workers. As opposed to the commonly used hydrazine as the recognition element, the team developed m-APA, an aniline probe with high reaction kinetics and selectivity [fig_ref] Figure 6: Chemical proteomics-based identification of oxidized proteins [/fig_ref] [bib_ref] Quantitative Profiling of Protein Carbonylations in Ferroptosis by an Aniline-Derived Probe, Chen [/bib_ref]. Chemical proteomics studies in HT1080 cells revealed that during ferroptosis, several carbonylated residues resulted from oxidized lipids insertion (rather by direct oxidation from ROS), and that oxidized voltage-dependent anion-selective channel protein 2 (VDAC2) may be responsible for mediating ferroptosis. Lastly, a highly sensitive carbonyl-target probe for imaging was developed by Sever and co-workers (TFCH, [fig_ref] Figure 6: Chemical proteomics-based identification of oxidized proteins [/fig_ref] [bib_ref] A Novel Fluorogenic Assay for the Detection of Nephrotoxin-Induced Oxidative Stress in..., Mukherjee [/bib_ref]. The probe was successfully used to monitor protein carbonylation in live cells and in renal tissues during oxidative stress mediated by nephrotoxicity.
## Oxidative stress in ageing and age-related neurodegenerative diseases
In numerous diseases, the amount of ROS is increased and poses a threat to cells and organisms. Oxidative damage has been suggested as both an effect and a mediator of ageing and a range of age-related neurodegenerative pathological conditions [bib_ref] Oxidative Stress in Neurodegenerative Diseases, Chen [/bib_ref]. Nutraceuticals treatment of model organisms has remarkably contributed to elucidating the physiological mechanisms associated with biological ageing. For example, data from the nematode Caenorhabditis elegans conclude that the effects of polyphenolic phytochemical phlorizin (phloridzin) diminish the intracellular ROS levels (measured by H 2 DCF-DA). This decrease in ROS activates a DAF-16-induced stress response and autophagy, subsequently leading to lifespan extension. Moreover, supplementation with phloridzin reverses the inactivation of dopaminergic neurons observed in worm models of PD [bib_ref] Anti-Oxidant and Anti-Aging Effects of Phlorizin Are Mediated by DAF-16-Induced Stress Response..., Park [/bib_ref]. Other compounds, such as the natural product cryptotanshinone [bib_ref] Cryptotanshinone Alleviates Oxidative Stress and Reduces the Level of Abnormally Aggregated Protein..., Cui [/bib_ref] or the glucagon-like peptide-1 receptor (GLP-1R) agonist Exendin-4, have been tested in Caenorhabditis elegans models of AD, concluding that those drugs can decrease oxidative stress and protein aggregation and improve the disease phenotype. In the following paragraphs we will provide exam-ples of how oxidative stress is implicated in disease manifestations in three age-related neurodegenerative disorders, namely PD, AD and HD.
PD is a progressive age-related neurodegenerative condition associated with a loss of dopaminergic neurons in the substantia nigra pars compacta of the brain. The symptomatology gradually increases upon onset of the disease and is mainly related to motor problems. Although the specific molecular mechanisms underlying the loss of dopaminergic neurons is not fully understood, oxidative stress is a well-established trigger of dopaminergic neurotoxicity [bib_ref] Oxidative Stress and Parkinson's Disease, Blesa [/bib_ref].
In this line, Soto-Rojas and co-workers determined ROS (through detection of the fluorescence signal of 2,7-dichloro dihydrofluorescein diacetate [DCFH-DA]) and lipid peroxidation (by evaluating the formation of lipid-soluble fluorescent compounds) in the brain of rats with α-synucleinopathy. They found that these two parameters of oxidative stress, as well as mitochondrial complex I dysfunction, positively correlate with neurodegeneration in different brain areas with α-synucleinopathy, such as the substantia nigra pars compacta, the striatum, the hippocampus and the olfactory bulb. Based on in silico studies, they suggested the involvement of peroxisome proliferator-activated receptors (PPAR) alpha (PPAR-α) and gamma (PPAR-γ) in this process [bib_ref] Oxidative Stress and Mitochondrial Complex I Dysfunction Correlate with Neurodegeneration in an..., Morales-Martínez [/bib_ref] , since PPARs activation induces fatty acid oxidation (which, in turn, generates ROS) [bib_ref] Altered Constitutive Expression of Fatty Acid-Metabolizing Enzymes in Mice Lacking the Peroxisome..., Aoyama [/bib_ref]. However, this mechanism needs to be further explored. Two additional recent studies further highlighted the interplay between ROS and PD. Bardien and co-workers sought to understand the pathophysiological mechanisms associated with familiar autosomal dominant PD. More specifically, they studied the p.G849D variant in the neurexin 2α (NRXN2) gene, which has been described to co-segregate with PD. Using a cellular model carrying the neurexin 2α p.G849D variant, the authors found elevation of H 2 O 2 levels (using a commercially available assay), accompanied by neuronal death [bib_ref] Neurexin 2 p.G849D Variant, Implicated in Parkinson's Disease, Increases Reactive Oxygen Species,..., Cuttler [/bib_ref]. Saiki and co-workers wished to shed light on the mechanism that promotes lysosomal retrograde transport, a regulator of autophagy, in PD. The authors initially demonstrated that JIP4 is upstream of the TRPML1-ALG2 pathway, which is known to promote retrograde transport. Subsequently, it was shown that this pathway can be activated in a ROS-dependent manner, since treatment with H 2 O 2 induces the phosphorylation of JIP4. These results highlight that oxidative stress could, in part, regulate autophagy in the context of PD via the JIP4-TRPML1-ALG2 pathway [bib_ref] Oxidative Stress-Induced Phosphorylation of JIP4 Regulates Lysosomal Positioning in Coordination with TRPML1..., Sasazawa [/bib_ref].
AD is a chronic, progressive neurological disorder that is associated with intraneuronal filamentous inclusions, called neurofibrillary tangles, mostly composed by tau protein aggregates, and extracellular senile amyloid plaques, which are predominantly due to the aggregation of misfolded Aβ peptide. Among neurodegenerative diseases, AD is the most common cause of dementia. Similar to PD, alterations in oxidative stress and mitochondrial dynamics and function are known to play a role in the development of the disease. Studies using in vitro models of AD have shown that treatment with the ROS scavenger N-acetyl-L-cysteine (NAC) reduces Aβ deposition through the activation of the PI3K/Akt/GLUT1 pathway, while it ameliorates the impaired glucose uptake phenotype. In this case, the authors measured intracellular ROS levels by flow cytometry [bib_ref] Scavenging Reactive Oxygen Species Decreases Amyloid-β Levels via Activation of PI3K/Akt/GLUT1 Pathway..., Peng [/bib_ref].
HD is an autosomal dominant neurodegenerative disorder caused by CAG triplet expansions in the Huntingtin gene, encoding an elongated poly-glutamine stretch in the Huntingtin protein. Patients with HD exhibit motor, psychiatric and cognitive deterioration and oxidative stress, although it is not clear whether it is the cause or the consequence of disease progression [bib_ref] Nicotinamide Adenine Dinucleotide Phosphate Oxidases Are Everywhere in Brain Disease, but Not..., Villegas [/bib_ref]. Ellederová and co-workers found higher oxidative stress in fibroblasts from a minipig model of HD, compared with wild-type minipig fibroblasts. The authors measured ROS (by using CellROX deep red reagent), lipid peroxidation (with Image-iT) and membrane permeability (with Calcein-AM). These elevated levels of oxidative stress correlated with the overexpression of mitochondrial superoxide dismutase 2 (SOD2) and the NEIL3 gene (encoding DNA glycosylase). SOD2 is expressed in the mitochondria, where it controls the cell cycle as a response to oxidative stress, whereas glycosylase enzyme repairs in the replication-associated repair of oxidized DNA. Taken together, these data highlight the role of oxidative stress and DNA damage in HD [bib_ref] Age-Related Oxidative Changes in Primary Porcine Fibroblasts Expressing Mutated Huntingtin, Smatlikova [/bib_ref].
Southwell and co-workers overexpressed progerin, a protein that causes premature aging, in murine neurons with HD, and observed ageing-related phenotypes. By treating neurons with H 2 O 2 and quantifying ROS (with CM-H 2 DCFDA), the authors concluded that biological ageing increases the sensitivity of neurons with HD to exogenous oxidative stress [bib_ref] The Interaction of Aging and Cellular Stress Contributes to Pathogenesis in Mouse..., Machiela [/bib_ref] [bib_ref] Biological Aging and the Cellular Pathogenesis of Huntington's Disease, Machiela [/bib_ref]. This sensitivity may highlight the interplay between ROS and protein aggregation, as shown for other neurodegenerative diseases.
In the same line, Oliveira and co-workers used MitoParaquat (MitoPQ) to enhance mitochondrial superoxide production. They observed features of PD in MitoPQ-treated zebrafish and an induction of mutant huntingtin aggregation without increasing cell death in a human cell model of HD, upon treatment with MitoPQ [bib_ref] Mitochondrial Superoxide Generation Induces a Parkinsonian Phenotype in Zebrafish and Huntingtin Aggregation..., Pinho [/bib_ref]. The forkhead box O1 (FOXO1) transcription factor has been shown to be involved in the development of AD and HD. Importantly, upon increased ROS, FOXO1 is activated through the AMPK pathway to promote autophagy, which is a well-established mechanism for the clearance of abnormal proteins and organelles [bib_ref] Cross-Talking Pathways of Forkhead Box O1 (FOXO1) Are Involved in the Pathogenesis..., Liu [/bib_ref].
# Conclusions
When maintained at low levels, ROS serve as signaling molecules and play a positive role in organism survival, by participating in the synthesis of some cellular structures or by being secreted by macrophages as a host defense system against pathogens. Nevertheless, when levels exceed a certain threshold (for instance, due to lower antioxidant defense or higher production), ROS become detrimental to the cell and the organism [bib_ref] Oxidative Stress: Harms and Benefits for Human Health, Pizzino [/bib_ref]. The negative consequences of high ROS levels during oxidative stress are due to the oxidative damage to proteins, nucleic acids and lipids, as well as due to aberrant redox signaling [bib_ref] Guidelines for Measuring Reactive Oxygen Species and Oxidative Damage in Cells and..., Murphy [/bib_ref]. Apart from its documented role in ageing [bib_ref] The Role of Mitochondrial ROS in the Aging Brain, Stefanatos [/bib_ref] , oxidative stress is implicated in a wide variety of age-related diseases such as neurodegenerative disorders [bib_ref] Targeting Oxidative Stress in Disease: Promise and Limitations of Antioxidant Therapy, Forman [/bib_ref].
The prevalence of neurodegenerative disorders is rising and, by 2040, will become the second leading cause of mortality [bib_ref] Neurodegenerative Disease: Brain Windfall, Gammon [/bib_ref]. This group of diseases cause a loss of neurons in the central nervous system and result in defects in memory, movement, cognition or a combination thereof. A common characteristic in many neurodegenerative diseases is the presence of disordered proteins that form toxic aggregates. Several studies have demonstrated the correlation between ROS and protein aggregation for several of these diseases. While it is unclear whether ROS promote aggregation (as in the case of prion protein [bib_ref] Design of Anti-and pro-Aggregation Variants to Assess the Effects of Methionine Oxidation..., Wolschner [/bib_ref] and α-synuclein [bib_ref] Increased Oxidative Stress Exacerbates α-Synuclein Aggregation In Vivo, Scudamore [/bib_ref] or ROS are produced by aggregates (as in the case of Aβ [bib_ref] Redox Proteomics and Amyloid β-Peptide: Insights into Alzheimer Disease, Butterfield [/bib_ref] , these studies certainly place oxidative stress as an important player and call for further insight into its role in neurodegeneration [fig_ref] Figure 7: Graphical summary of the association of oxidative stress in neurodegenerative diseases [/fig_ref].
Delineating the role of oxidative stress in any condition is not only critical, but also challenging. Due to its importance, several tools for evaluating oxidative stress, including small molecules and genetically encoded protein reporters, have been developed over the years. Perhaps the most commonly used approach is the direct quantification of ROS levels. However, methods that have been traditionally employed need to be used with caution due to the intrinsic properties of ROS (short-lived and similar chemical properties), as was recently highlighted [bib_ref] Guidelines for Measuring Reactive Oxygen Species and Oxidative Damage in Cells and..., Murphy [/bib_ref]. Herein, we provided an overview of small molecules and protein reporters recently developed to detect different ROS. Apart from exhibiting improved selectivity, several of these tools have additional desirable characteristics, particularly important for in vivo imaging, such as two-photon excitability, ratiometric imaging, organelle-specific localization and others. combination thereof. A common characteristic in many neurodegenerative diseases is the presence of disordered proteins that form toxic aggregates. Several studies have demonstrated the correlation between ROS and protein aggregation for several of these diseases. While it is unclear whether ROS promote aggregation (as in the case of prion protein [bib_ref] Design of Anti-and pro-Aggregation Variants to Assess the Effects of Methionine Oxidation..., Wolschner [/bib_ref] and α-synuclein [bib_ref] Increased Oxidative Stress Exacerbates α-Synuclein Aggregation In Vivo, Scudamore [/bib_ref] or ROS are produced by aggregates (as in the case of Aβ [bib_ref] Redox Proteomics and Amyloid β-Peptide: Insights into Alzheimer Disease, Butterfield [/bib_ref] , these studies certainly place oxidative stress as an important player and call for further insight into its role in neurodegeneration [fig_ref] Figure 7: Graphical summary of the association of oxidative stress in neurodegenerative diseases [/fig_ref]. Neurons in the brain can suffer deleterious modifications due to high levels of ROS. These ROS can be mitochondrial-generated, for instance due to mitochondrial dysfunction, and further contribute to increasing the mitochondrial damage, since ROS can also oxidize DNA, RNA and lipids. Furthermore, protein aggregation is partly caused by ROS, and subsequently, promotes oxidative stress. All these hallmarks of oxidative stress are well-established and may play a role in many age-related neurodegenerative diseases, such as Alzheimer's, Parkinson's and Huntington's diseases.
Delineating the role of oxidative stress in any condition is not only critical, but also challenging. Due to its importance, several tools for evaluating oxidative stress, including small molecules and genetically encoded protein reporters, have been developed over the years. Perhaps the most commonly used approach is the direct quantification of ROS levels. However, methods that have been traditionally employed need to be used with caution due to the intrinsic properties of ROS (short-lived and similar chemical properties), as was recently highlighted [bib_ref] Guidelines for Measuring Reactive Oxygen Species and Oxidative Damage in Cells and..., Murphy [/bib_ref]. Herein, we provided an overview of small molecules and protein reporters recently developed to detect different ROS. Apart from exhibiting improved selectivity, several of these tools have additional desirable characteristics, particularly important for in vivo imaging, such as two-photon excitability, ratiometric imaging, organelle-specific localization and others.
While advantageous, the high reaction rates of ROS mean that, unless high concentrations are used, an amount of ROS will evade detection. An approach to bypass this issue is to use methods to define the oxidative damage. Although many oxidative modifications are reversible, they are certainly more stable than free ROS. With regards to proteins, oxidative damage could significantly alter the function and properties of protein. ROS-mediated modifications can take place on eight different amino acid residues as well as on the protein backbone, and this can occur either through direct reaction with ROS or by other oxidized biomolecules (e.g., lipids). Small molecule probes targeting specific oxidized residues have been developed and applied for imaging or protein identification. The latter application (using chemical proteomics) has offered unprecedented Neurons in the brain can suffer deleterious modifications due to high levels of ROS. These ROS can be mitochondrial-generated, for instance due to mitochondrial dysfunction, and further contribute to increasing the mitochondrial damage, since ROS can also oxidize DNA, RNA and lipids. Furthermore, protein aggregation is partly caused by ROS, and subsequently, promotes oxidative stress. All these hallmarks of oxidative stress are well-established and may play a role in many age-related neurodegenerative diseases, such as Alzheimer's, Parkinson's and Huntington's diseases.
While advantageous, the high reaction rates of ROS mean that, unless high concentrations are used, an amount of ROS will evade detection. An approach to bypass this issue is to use methods to define the oxidative damage. Although many oxidative modifications are reversible, they are certainly more stable than free ROS. With regards to proteins, oxidative damage could significantly alter the function and properties of protein. ROS-mediated modifications can take place on eight different amino acid residues as well as on the protein backbone, and this can occur either through direct reaction with ROS or by other oxidized biomolecules (e.g., lipids). Small molecule probes targeting specific oxidized residues have been developed and applied for imaging or protein identification. The latter application (using chemical proteomics) has offered unprecedented information, as in the case of the oxidized pyruvate kinase M2 (which was shown to promote cancer) [bib_ref] Methionine Oxidation Activates Pyruvate Kinase M2 to Promote Pancreatic Cancer Metastasis, He [/bib_ref] or the oxidized human prion protein (where aggregation is induced) [bib_ref] Design of Anti-and pro-Aggregation Variants to Assess the Effects of Methionine Oxidation..., Wolschner [/bib_ref]. The combination of probes with different reactivity (cysteine, sulfenic and sulfinic acids) on Caenorhabditis elegans revealed that translation, growth signaling and stress response pathways are redox-sensitive [bib_ref] Global Profiling of Distinct Cysteine Redox Forms Reveals Wide-Ranging Redox Regulation in..., Meng [/bib_ref].
The emerging role of oxidative stress in neurodegenerative diseases, in combination with the lack of cure for any of these diseases, creates the need for further investigation of this interplay. On the one hand, some evidence suggests that ROS may, to some degree, regulate protein aggregation. On the other hand, protein aggregation seems to be accompanied by an increase in ROS levels, which in turn may impair cell defense mechanisms against protein aggregation (e.g., degradation machinery and heat shock response). In either case, all the above indicate that shedding light on this relationship could provide novel therapeutic approaches in the future. While most methods presented herein were not applied on neurodegenerative diseases, we hope this review will inform the community about these tools and inspire further research to unravel the role of oxidative stress in age-related neurodegenerative pathologies.
[fig] Figure 1: Development of oxidative stress. [/fig]
[fig] Figure 2: Figure 2. Interplay between protein aggregates and oxidative damage. (A) Increased protein aggregation in several neurodegenerative diseases is associated with increased ROS levels. This relationship is bidirectional; excessive ROS production is found to induce protein aggregation and vice versa. Two examples of this vicious cycle are shown: (1) a decrease in the mitochondrial levels of Superoxide dismutase 2 (SOD2) gives rise to elevated ROS, which in turn promotes Α-synuclein aggregation (right panel) and (2) the accumulation of Αβ oligomers induces lipid peroxidation, an oxidative damage that further results in protein carbonylation (left panel). (B) Proposed mechanisms of ROS-induced aggregation of prion protein. At physiological conditions, part of the prion protein exists in the α-fold form (depicted as a helix in the middle). However, in disease, its scrapie isoform forms β-sheets, which eventually drive fibrillation. The transition from α-fold to β-sheet is proposed to be mediated by ROS, with two different mechanisms: (1) by methionine oxidation (depicted as red circles), which destabilizes the α-fold form (left side) or (2) by ROS-mediated cleavage-induced by copper-of the protective N-terminus of prion, and subsequently, β-sheet [/fig]
[fig] Antioxidants 2023 ,: 12, x FOR PEER REVIEW 8 of 22 [/fig]
[fig] Figure 3: Tools for the visualization of hypohalous acids. (A) Structure and turn-on mechanism of the methylene blue-based probe (FDOCl-18) upon sensing HOCl. The methylene blue (top structure) is conjugated to naphthalene (indicated as a green star) through the HOCl-sensing group (highlighted by the gray oval, left structure). The methylene blue is non-fluorescent while the sensing group is attached. Upon sensing HOCl, a deformylation mechanism takes place that leads to the liberation of the methylene blue, which is now fluorescent. (B) Sensing mechanism of the organellespecific turn-on probe for HOCl. The fluorescence properties of a two-photon dye (acedan) were quenched by HOCl-labile group (gray circle, left structure). The probe was further equipped with organelle specific-targeted moieties to drive it either to mitochondria (MITO-TP) or to lysosomes (LYSO-TP). (C) Structure of the HOBr ratiometric probe RhSN-mito. In the absence of HOBr, the probe emits at 530 nm. Reaction of the sensing moiety (gray oval, top structure) with HOBr leads to the formation of an additional ring that alters the fluorescence properties of the probe, which now emits at 663 nm. (D) The NemR-cpYFP construct (Hypocrates) for detecting hypohalous acids. The YFP was integrated onto a flexible loop of the NemR. While the YFP exhibits the high fluorescence in non-oxidation conditions, the intensity decreases when the thiol of the cysteine senses the hypohalous acid. [/fig]
[fig] Figure 4: Peroxynitrite-selective tools and their mechanism of detection. (A) Structure and mechanism of the small molecule FRET probe MITO-CC. [/fig]
[fig] Figure 5: Tools for detecting H2O2. (A) Structure and mechanism of H2O2-responsive protein labeling using Hyp-L. A fluorescently labeled (indicated as a green star) small molecule probe with a H2O2-labile group (gray circle) forms a reactive intermediate upon oxidation, which is in turn inserted onto proximal proteins. The proteins can be visualized by fluorescent microscopy or subjected to proteomics workflow for identification. (B) Mechanism of the excitation state intramolecular proton transfer (ESIPT)-based protein reporter. A GFP was engineered to develop an ESIPTbased protein reporter (FROG/B). The final variant is excited by a single wavelength, but the emission wavelength can change depending on the presence or absence of ROS. [/fig]
[fig] Figure 6: Chemical proteomics-based identification of oxidized proteins. (A) General workflow of chemical proteomics. The proteome of interest consisting of non-oxidized (indicated as light gray) [/fig]
[fig] Figure 7: Graphical summary of the association of oxidative stress in neurodegenerative diseases. [/fig]
[table] Table 1: Summary of tools reported in Section 4. [/table]
|
# Introduction
Myocardial bridge (MB) is defined as an epicardial segment of the coronary artery which takes an intramyocardial course, resulting in systolic compression. [bib_ref] Myocardial bridging: a review with emphasis on electrocardiographic findings, Rovai [/bib_ref] It was first recognized at autopsy by Reyman in 1737, first described angiographically by Porstmann and Iwig, 2 and first surgically managed with myotomy by Binet et al. [bib_ref] Myocardial bridge compressing the anterior inter-ventricular artery, Binet [/bib_ref] We report a case of a very long segment left anterior descending artery (LAD) MB requiring a corresponding extensive surgical myotomy.
## Case report
A 55-year-old woman presented to the emergency room with chest pain. Her past medical history was significant for type A aortic dissection repaired with a Dacron vascular graft from the sinotubular junction to the proximal aortic arch 8 years ago, persistent chest pain on medical therapy with known LAD MB, ex-tobacco use (quit 10 years ago), migraines, and gastroesophageal reflux. Her serial electrocardiograms and cardiac enzymes were normal with mildly elevated beta-natriuretic peptide of 290 pg/mL. Computed tomographic angiography (CTA) of the chest was negative for aortic dissection or pulmonary embolism. Transthoracic echocardiogram showed normal biventricular function and size, normal tricuspid aortic valves, and trivial mitral and moderate tricuspid regurgitation. She subsequently underwent right and left cardiac catheterization revealing a LAD with a long intramyocardial segment. Its proximal portion supplied a large bifurcating diagonal branch. There were two separate regions of total systolic compression yet relatively preserved diastolic caliber. Fractional flow reserve (FFR) on the muscle bridge was 0.85 at rest and 0.75 with adenosine, consistent with hemodynamic significant stenosis (Supplemental Video 1, with a snapshot of the same in [fig_ref] Figure 1: Snapshot of the preoperative coronary angiogram at systole showing compression of the... [/fig_ref].
As her New York Heart Association (NYHA) Class III symptoms persisted despite optimal medical therapy comprising beta-channel and calcium channel blockers, surgical management involving dividing the MB was her next and only option.
She underwent redo sternotomy, unroofing of LAD MB, and tricuspid repair with an annuloplasty ring. Dense adhesions made it difficult to localize the distal LAD. Once localized, the MB over the LAD was carefully and slowly divided. The mid LAD coursed under the right ventricular (RV) endocardium and the unroofing created a small 5-mm opening to the RV parallel to the LAD. This opening was repaired with direct suture from the adventitia to the endocardium. The LAD was freed all the way up to the takeoff of its large diagonal branch. The length of the divided intramyocardial segment of LAD measured 100 mm. The tricuspid valve showed annular dilation and was repaired by a 28-mm Carpentier-Edwards (CE) classic annuloplasty ring. Post-pump echocardiography showed preserved biventricular function and no mitral or residual tricuspid regurgitation.
Her postoperative course was uneventful. Postoperative cardiac catherization (Supplemental Video 2, with a snapshot of the same in [fig_ref] Figure 2: Snapshot of the postoperative coronary angiogram showing resolution of any compression of... [/fig_ref] was performed on postoperative day 6 and confirmed resolution of the MB, with an FFR of 0.87 in the distal LAD. She was discharged home a day later.
# Discussion
Myocardial bridging (MB) is considered a benign inborn coronary abnormality. On autopsy, MB has been reported in as many as 40%-80% of cases, 4 angiographically in 0.5%-16.0%. It is more frequent in females, involving mostly the LAD (77.2%), then the left circumflex artery (40%), and the right coronary artery (36%). [bib_ref] Incidence, clinical characteristics, and 4-year follow-up of patients with isolated myocardial bridge:..., Cicek [/bib_ref] The average length of the MB is 21.85 ± 16.10 mm (range: 5-70 mm) with a muscle thickness above the artery of 3.744 ± 1.48 mm. [bib_ref] Myocardial bridges: a prospective forensic autopsy study, Micic-Labudovic [/bib_ref] Our case involved an MB that was 10 cm in length, the longest that it was ever reported.
Although generally benign, MB has been associated with coronary spasm, myocardial infarction, unstable angina, supraventricular and ventricular arrhythmia, syncope, myocardial stunning, 7 ventricular septal rupture, 8 transient ventricular dysfunction, or sudden cardiac arrest or death. [bib_ref] Prevalence and threeyear follow-up of patients with isolated myocardial bridge in the..., Aksakal [/bib_ref] MB of coronary arteries is commonly noted on chest CTA; only approximately one-third of these show systolic compression. When symptomatic, stable angina is the usual presentation. Exercise stress tests often show nonspecific signs of ischemia and do not distinguish between MB and other causes of myocardial ischemia.
In patients with an MB in the LAD, the percentage of arterial compression is related directly to the burden of its proximal atherosclerotic plaque, particularly in patients who otherwise have low coronary artery risk factors. [bib_ref] Functional versus anatomic assessment of myocardial bridging by intravascular ultrasound: impact of..., Yamada [/bib_ref] MB in the LAD is an independent risk factor for more than 50% coronary artery stenosis in proximal LAD, with or without hypertension, [bib_ref] Evaluation of association of myocardial bridge in the left anterior descending coronary..., Tian [/bib_ref] suggestive of an obstructive nature of a LAD MB. Aside from increased propensity for atherosclerosis, diastolic compression and endothelial shear-stress-related vasospasm are the other mechanisms of myocardial ischemia. [bib_ref] Myocardial bridging: contemporary understanding of pathophysiology with implications for diagnostic and therapeutic..., Corban [/bib_ref] Coronary angiography, intracoronary Doppler, multislice CTA, positron emission tomography (PET) scan, and contrast stress echocardiography are the modalities to study the significance of an MB. [bib_ref] Myocardial bridging: contemporary understanding of pathophysiology with implications for diagnostic and therapeutic..., Corban [/bib_ref] In our case, we diagnosed the MB of the LAD angiographically, which showed its systolic narrowing, as well as a reduced FFR. An FFR < 0.75 suggests a significant flow limitation. For a symptomatic patient with MB and an abnormal but nonsignificant FFR (>0.80), intravenous administration of dobutamine can lead to higher pressure gradients and reproduction of anginal symptoms, reflecting a clinically significant MB. [bib_ref] Importance of diastolic fractional flow reserve and dobutamine challenge in physiologic assessment..., Escaned [/bib_ref] First-line treatment of symptomatic MB is medicalnamely, beta blockers. Nitrates should be avoided because symptoms may worsen. [bib_ref] Intramyocardial bridges: anatomo-pathological characteristics, diagnosis, and therapeutic strategies, Zanini [/bib_ref] Coronary stenting is reserved for refractory patients who are deemed too high risks for surgery, as stent fracture and coronary aneurysm have been reported. [bib_ref] Coronary aneurysm and stent fracture following stenting of a myocardial bridge, Lu [/bib_ref] In a retrospective review of 31 patients who underwent surgical myotomy for significant MB, all postoperative patients became symptom free with an improvement in NYHA class from I-III to I-II. [bib_ref] Surgical treatment of myocardial bridging: report of 31 cases, Wu [/bib_ref] Our patient was very symptomatic and was treated medically for an extended period of time. The surgery with the division of a very deep and long MB is of high risk particularly when the RV is punctured intraoperatively; however, the RV hole was easily repaired. Our patient underwent postoperative angiography with FFR (0.87 in the distal LAD), which confirmed a technically good result. Our case is unique as it represented the longest LAD unroofing that has ever been reported.
# Conclusion
Although very common and most often benign, MB may be associated with significant anginal symptoms being the usual presentation. Surgical myotomy of the MB is an effective therapy in alleviating symptoms.
## Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
# Ethical approval
Our institution does not require ethical approval for reporting individual cases or case series.
# Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
## Informed consent
Written informed consent was obtained from the patient(s) for their anonymized information to be published in this article.
# Supplemental material
Supplemental material for this article is available online.
[fig] Figure 1: Snapshot of the preoperative coronary angiogram at systole showing compression of the proximal and mid-segments of the LAD. [/fig]
[fig] Figure 2: Snapshot of the postoperative coronary angiogram showing resolution of any compression of the LAD. [/fig]
|
Establishing the characteristics of an effective pharmacogenetic test for clozapine-induced agranulocytosis
# Introduction
Although there are 22 FDA-approved antipsychotics used to treat schizophrenia, around 30% of schizophrenia patients do not respond to drugs other than clozapine. [bib_ref] Treatment-resistant schizophrenia -the role of clozapine, Meltzer [/bib_ref] Clozapine has superior efficacy for positive symptoms in these treatment-resistant patients 2 and may improve negative symptoms. [bib_ref] Clozapine: a distinct, poorly understood and under-used molecule, Joober [/bib_ref] Furthermore, clozapine reduces suicidal behaviour especially when compared with first generation antipsychotics and overall mortality at population level. [bib_ref] Clozapine treatment for suicidality in schizophrenia: international suicide prevention trial (InterSePT), Meltzer [/bib_ref] [bib_ref] Antipsychotics and mortality in first-onset schizophrenia: prospective Finnish register study with 5-year..., Kiviniemi [/bib_ref] [bib_ref] Mortality, attempted suicide, re-hospitalisation and prescription refill for clozapine and other antipsychotics..., Weitoft [/bib_ref] [bib_ref] Medication and suicide risk in schizophrenia: a nested case-control study, Reutfors [/bib_ref] [bib_ref] 11-year follow-up of mortality in patients with schizophrenia: a population-based cohort study..., Tiihonen [/bib_ref] Despite its proven efficacy, the clinical use of clozapine is limited by the risk of agranulocytosis, a rare but potentially fatal adverse drug reaction, characterized by the acute loss of neutrophils in circulating blood. Agranulocytosis is defined as an absolute neutrophil count (ANC) of less than 500 cells mm − 3 blood. Shortly after clozapine was introduced in Europe in the 1970s, it was withdrawn from the market when 17 cases of agranulocytosis were reported in Finland, of which 8 were fatal. [bib_ref] Agranulocytosis during treatment with chlozapine, Idänpään-Heikkilä [/bib_ref] In 1990, clozapine was reintroduced after its superiority over chlorpromazine for the treatment of refractory schizophrenia was shown. [bib_ref] Clozapine for the treatment-resistant schizophrenic: a double-blind comparison with chlorpromazine, Kane [/bib_ref] However, its use was restricted in most Western countries to treatment of refractory patients, that is, patients who have not improved on at least two different antipsychotics. [bib_ref] Clozapine: balancing safety with superior antipsychotic efficacy, Meltzer [/bib_ref] [bib_ref] Clozapine: a review of clinical practice guidelines and prescribing trends, Warnez [/bib_ref] To prevent agranulocytosis by detecting a fall in ANC, patients treated with clozapine are subject to compulsory haematological monitoring. In Europe, the full white blood cell count and ANC are monitored weekly for the first 18 weeks of treatment and every 4 weeks thereafter for the duration of the treatment.If at any time during treatment the white blood cell count falls below 3000 cells mm − 3 or the ANC below 1500 cells mm − 3 , clozapine should be discontinued immediately and these patients should not be treated with clozapine again except in a controlled setting. [bib_ref] Clozapine-induced leukopenia: arguments for rechallenge, Bogers [/bib_ref] [bib_ref] Restarting clozapine after neutropenia: evaluating the possibilities and practicalities, Whiskey [/bib_ref] [bib_ref] When can patients with potentially life-threatening adverse effects be rechallenged with clozapine?..., Manu [/bib_ref] Although the obligatory monitoring has the benefit of regular contact with a health-care professional, it is an invasive procedure and can be a burden for the patient. Moreover, some patients decline to take clozapine because of the monitoring requirement. [bib_ref] Reasons for non-prescription of clozapine in treatmentresistant schizophrenia, Swinton [/bib_ref] As agranulocytosis can develop within 2-5 days, even weekly monitoring cannot guarantee timely detection in all cases. [bib_ref] Mechanisms of clozapine-induced agranulocytosis, Gerson [/bib_ref] The incidence of agranulocytosis induced by clozapine varies between 0.38 and 0.8%, with approximately 80% of cases occurring within the first 18 weeks. [bib_ref] Reducing clozapine-related morbidity and mortality: 5 years of experience with the Clozaril..., Honigfeld [/bib_ref] [bib_ref] Clozapineinduced agranulocytosis-incidence and risk factors in the United States, Alvir [/bib_ref] [bib_ref] Clozapine-induced agranulocytosis in Finland, 1982-2007: long-term monitoring of patients is still warranted, Lahdelma [/bib_ref] [bib_ref] Active monitoring of 12,760 clozapine recipients in the UK and Ireland. Beyond..., Munro [/bib_ref] The incidence of agranulocytosis decreases from 0.7% in the first year, to 0.07% or lower in the second year of treatment. [bib_ref] Neutropenia and agranulocytosis in patients receiving clozapine in the UK and Ireland, Atkin [/bib_ref] Few cases occur later in the course of treatment, but the risk does not fully disappear. In 2-4% of patients, agranulocytosis is fatal, which corresponds to an overall mortality rate of about 1-3 in 10 000 patients on clozapine. [bib_ref] Beyond white blood cell monitoring: screening in the initial phase of clozapine..., Cohen [/bib_ref] However, most patients recover completely from agranulocytosis with no haematological consequences. [bib_ref] Active monitoring of 12,760 clozapine recipients in the UK and Ireland. Beyond..., Munro [/bib_ref] [bib_ref] Neutropenia and agranulocytosis in patients receiving clozapine in the UK and Ireland, Atkin [/bib_ref] [bib_ref] Clozapine-induced agranulocytosis, Mendelowitz [/bib_ref] In spite of its therapeutic advantages with respect to its efficacy in treatment-resistant schizophrenia, clozapine is underused, mainly owing to the risk of severe adverse events, primarily agranulocytosis and the mandatory haematological monitoring. [bib_ref] Making clozapine safer: current perspectives on improving its tolerability, Nair [/bib_ref] Around 30% of schizophrenia patients meet the indications for clozapine treatment, but the market share of clozapine, which is now a generic drug, was less than 5% in 2010 in the US. [bib_ref] Clozapine: balancing safety with superior antipsychotic efficacy, Meltzer [/bib_ref] A pharmacogenetic test for clozapine-induced agranulocytosis could greatly improve the burden of haematological monitoring if the monitoring requirements could be made less onerous, or be time-limited, for the majority of patients with a low genetic risk for agranulocytosis. Not only would this make clozapine treatment more acceptable for the patient, it would also save considerable health-care resources. On the other hand, the patients who are at a higher risk of developing agranulocytosis could be monitored more frequently or, if the risk is very high, not exposed to clozapine at all. [bib_ref] Treatment-resistant schizophrenia -the role of clozapine, Meltzer [/bib_ref] Pharmacogenetic research of clozapine-induced agranulocytosis has focused on candidate genes in case-control studies. Several associations with human leukocyte antigen (HLA) alleles have been reported, as well as associations with the tumour necrosis factor and N-ribosyldihydronicotin-amide quinone oxidoreductase 2 (NQO2) genes. [bib_ref] Genetics of antipsychotic-induced side effects and agranulocytosis, Chowdhury [/bib_ref] [bib_ref] Clozapine-induced agranulocytosis and its genetic determinants, Opgen-Rhein [/bib_ref] However, few of these findings have been replicated, and the majority of these pharmacogenetic studies suffered from typical candidate gene study issues, namely small sample sizes and inadequate correction for multiple testing. The most promising finding was that the HLA-DQB1 6672G4C polymorphism was associated with clozapine-induced agranulocytosis, with an odds ratio of 16.9. 31 A pharmacogenetic test based on this polymorphism has been marketed, but owing to low sensitivity (21.5%), it failed to be a commercial or clinical success. [bib_ref] Genetics of antipsychotic-induced side effects and agranulocytosis, Chowdhury [/bib_ref] [bib_ref] Opening Pandora's box in the UK: a hypothetical pharmacogenetic test for clozapine, Spencer [/bib_ref] In the first genome-wide association study, amino acid changes in HLA-DQB1 (126Q) and HLA-B (158T) were associated with clozapine-induced agranulocytosis with more modest odds ratios of 0.19 and 3.11, respectively. [bib_ref] Clozapine-induced agranulocytosis is associated with rare HLA-DQB1 and HLA-B alleles, Goldstein [/bib_ref] Here, we investigate the required properties of a clinically useful pharmacogenetic test that could stratify clozapine users with regards to their agranulocytosis risk as described above.
# Methods
We assume that the genetic test divides patients into two groups with different levels of agranulocytosis risk, and that the low-risk (LR) group contains a higher proportion of patients than the high-risk (HR) group. When comparing the outcome of a pharmacogenetic agranulocytosis test with the actual agranulocytosis status, the following scenarios can occur [fig_ref] Table 1: Classification table comparing test outcome with true agranulocytosis status [/fig_ref] :
- True positive: A patient who does develop agranulocytosis is correctly identified as HR. This scenario has probability a. - False positive: A patient who does not develop agranulocytosis is wrongly identified as HR. This scenario has probability b. - False negative: A patient who does develop agranulocytosis is wrongly identified as LR. This scenario has probability c. - True negative: A patient who does not develop agranulocytosis is correctly identified as LR. This scenario has probability d.
If the incidence of clozapine-induced agranulocytosis is a known parameter k, the agranulocytosis risk regardless of test outcome (a+c) equals k and the probability of not getting agranulocytosis (b+d) is 1-k.
Assuming that k is known, two of the following parameters need to be fixed to calculate the probabilities in each cell of [fig_ref] Table 1: Classification table comparing test outcome with true agranulocytosis status [/fig_ref] :
- The proportion of patients in the HR group (x) or the proportion of patients in the LR group (1-x). - The sensitivity of the test, which is the proportion of correctly classified agranulocytosis cases. In [fig_ref] Table 1: Classification table comparing test outcome with true agranulocytosis status [/fig_ref] ,
[formula] Sensitivity ¼ a aþc ¼ a k - [/formula]
The specificity of the test, which is the proportion of correctly classified agranulocytosis-free patients. In [fig_ref] Table 1: Classification table comparing test outcome with true agranulocytosis status [/fig_ref] ,
[formula] Specif icity ¼ d bþd ¼ d 1 -k [/formula]
The proportion of patients in each risk group (x, 1-x) is relevant to this study, as we want to justify a more lenient monitoring schedule for the LR group. The larger the LR group, the more patients on clozapine will benefit from monitoring regime changes. Assuming a single locus test, the size of the two risk groups depends on the allele frequency of the test marker. Hence, we use the proportion of patients in the HR group (x) and test sensitivity (s) to study the cell probabilities in [fig_ref] Table 1: Classification table comparing test outcome with true agranulocytosis status [/fig_ref].
Of primary interest is the agranulocytosis risk in the LR group, as these are the patients for whom the haematological monitoring rules could be relaxed, and this outcome corresponds to the complement of the negative predictive value (NPV), being the proportion of test negative or LR patients who do not develop agranulocytosis. Therefore, we investigate the relationship between agranulocytosis risk in the LR group, test sensitivity and the size of the HR group. The agranulocytosis risk in the LR group is given by
[formula] PðA LR j Þ ¼ P A \ LR ð Þ P LR ð Þ ¼ c c þ d ¼ 1 -s ð Þk 1 -x ¼ 1 -NPV [/formula]
where A stands for developing agranulocytosis. Thus, the agranulocytosis risk in the LR group decreases as test sensitivity increases or as the HR group becomes smaller, which corresponds to the LR group getting larger. As a secondary outcome, we study the agranulocytosis risk in the HR group, which corresponds to the positive predictive value (PPV) or the proportion of test positive patients who are true agranulocytosis cases, and is given by
[formula] PðA HR j Þ ¼ P A \ HR ð Þ P HR ð Þ ¼ a a þ b ¼ sk x ¼ PPV [/formula]
In the HR group, the agranulocytosis risk increases when sensitivity rises or the proportion of patients in the HR group decreases. By definition, the agranulocytosis risk in the HR group must be larger than in the LR group, so
[formula] P A HR j Þ > P A LR j Þ ð ð [/formula]
or expressed in terms of sensitivity and size of the HR group
[formula] sk x > 1 -s ð Þk 1 -x [/formula]
which reduces to s4x. In other words, sensitivity must be larger than the proportion of patients in the HR group.
We explore the relationship between different parameters, focusing on sensitivity (s), and the proportion of patients assigned to the HR group (x), in the context of a genetic test to predict the risk of clozapine-induced agranulocytosis. Furthermore, we develop guidelines for a test that divides the population into a LR and HR group, assuming the total risk of developing clozapine-induced agranulocytosis (k) is 0.8%.We also assess how the pharmacogenetic test based on the HLA-DQB1 6672G4C polymorphism performs under this framework. [bib_ref] Candidate gene analysis identifies a polymorphism in HLA-DQB1 associated with clozapine-induced agranulocytosis, Athanasiou [/bib_ref]
# Results
The key parameters for a clinically effective test, that is, a test that minimizes the agranulocytosis risk in the LR group, are high sensitivity and to a lesser extent a small proportion of patients assigned to the HR group. shows that to obtain a low agranulocytosis risk in the LR group (solid lines), with a concomitant high risk in the HR group (dotted lines), a test should be highly sensitive. Also, for a given sensitivity, a smaller HR group corresponds to lower agranulocytosis risk in the LR group and a higher agranulocytosis risk in the HR group and [fig_ref] Table 2: Hypothetical tests with 80% sensitivity for different proportions of patients in the... [/fig_ref]. The lower the sensitivity of a test is, the smaller the difference between the agranulocytosis risks in both groups. When the sensitivity is equal to the size of the HR group, the risk in the two groups are the same and equal to the overall agranulocytosis risk of 0.8%. A test with sensitivity close to the proportion of patients in the HR group would thus be irrelevant.
Exploration of the relationship between agranulocytosis risk and the proportion of patients in the HR group confirms that a smaller HR group gives rise to a lower agranulocytosis risk in the LR group [fig_ref] Figure 2: Agranulocytosis risk in LR and HR groups by proportion of patients that... [/fig_ref]. The risk in the HR group increases steeply when the size of that group is close to zero. For a given size of the HR group, high sensitivity leads to a low agranulocytosis risk in the LR group and a high risk in the HR group [fig_ref] Figure 2: Agranulocytosis risk in LR and HR groups by proportion of patients that... [/fig_ref] and [fig_ref] Table 3: Hypothetical tests with 10% of patients in the HR group for [/fig_ref]. As in Pharmacogenetic test for clozapine-induced agranulocytosis M Verbelen et al , the risk curves in [fig_ref] Figure 2: Agranulocytosis risk in LR and HR groups by proportion of patients that... [/fig_ref] meet at 0.8% agranulocytosis risk when the proportion of patients in the HR group is equal to the sensitivity (except for 100% sensitivity where the agranulocytosis risk in the LR group is zero as all agranulocytosis cases are detected by the test). A simultaneous assessment of sensitivity and HR group size shows that test sensitivity controls the reduction in agranulocytosis risk seen in the LR group, with high sensitivity leading to low agranulocytosis risk [fig_ref] Figure 3: Agranulocytosis risk in [/fig_ref]. High sensitivity implies that most agranulocytosis cases are identified by the genetic test and classified as HR. By consequence, nearly all patients in the LR group do not develop agranulocytosis, and hence the risk in that . Agranulocytosis risk in LR and HR groups by sensitivity, for different HR group sizes between 5 and 50%. The ■ and ▲ indicate the agranulocytosis risk in the LR and HR groups, respectively, for a test with 80% sensitivity and 10% of patients in the HR group. group is low. For example, if we need the agranulocytosis risk in the LR group to be half the population risk, (that is, ⩽ 0.4%), the test sensitivity must be at least 50.2%. To achieve a stratification where the LR group is at one-fifth of the average agranulocytosis risk, sensitivity greater than 80.1% is required. A small HR group contributes to a low agranulocytosis risk in the LR group by preventing true LR patients from being wrongly classified as HR. A large number of true LR patients maximizes the denominator of the agranulocytosis risk in the LR group, and thus minimizes the risk itself. The agranulocytosis risk in the HR group depends largely on the size of the HR group [fig_ref] Figure 3: Agranulocytosis risk in [/fig_ref]. In a smaller HR group, the ratio of true HR patients versus patients incorrectly classified as HR is larger, and so the agranulocytosis risk in the HR group is larger. High sensitivity increases the number of true HR patients and in that way leads to a high agranulocytosis risk in this group, but even in the ideal scenario of maximum sensitivity, the proportion of true HR patients is limited to 0.8%.
High sensitivity and a small HR group lead to an effective test with the small group of HR patients requiring more frequent monitoring, whereas the majority of patients are assigned to the LR group, which has substantially reduced agranulocytosis risk and could therefore be monitored less frequently.
To comment on the clinical utility of a pharmacogenetic test, an agranulocytosis risk that is acceptable without monitoring must be determined. We propose that an agranulocytosis risk of 0.13% is acceptable, because this corresponds to the risk conferred by the antipsychotic chlorpromazine which does not have mandatory monitoring in the UK. [bib_ref] Haematological toxicity of drugs used in psychiatry, Flanagan [/bib_ref] To achieve this, the sensitivity of the test must be at least 83.9%. We examine four hypothetical pharmacogenetic tests and how the outcomes affect haematological monitoring of patients [fig_ref] Table 4: Four hypothetical pharmacogenetic tests for clozapine-induced agranulocytosis and their clinical impact Test... [/fig_ref].
- Test A is only clinically relevant for the small proportion of HR patients, but owing to its low sensitivity, the agranulocytosis risk in the LR group is too high to reduce monitoring. - Despite a higher sensitivity than test A, test B has no definite impact on the treatment of either risk group. - The characteristics of test C result in an agranulocytosis risk in the LR group that is low enough to stop or reduce haematological monitoring for 90% of patients. This test has the largest clinical impact. - Although the sensitivity of test D is higher than that of test C, the larger size of the HR group in test D implies that fewer patients would benefit from this test.
The pharmacogenetic agranulocytosis test using the HLA-DQB1 6672G4C polymorphism has a sensitivity of 21.5% and specificity of 98.4%. [bib_ref] Candidate gene analysis identifies a polymorphism in HLA-DQB1 associated with clozapine-induced agranulocytosis, Athanasiou [/bib_ref] On the basis of those values and assuming an overall agranulocytosis risk of 0.8%, the test classifies 1.76% of patients as HR, with a high agranulocytosis risk of 9.66%. Conversely, the agranulocytosis risk in the LR group is 0.64%. This is a relative risk of 0.8 or a 20% reduction in risk compared with the agranulocytosis risk without genetic stratification, and exceeds our maximum acceptable agranulocytosis risk of 0.13%. Hence, the HLA-DQB1
6672G4C-based genetic test has limited use in stratifying patients in order to reduce haematological monitoring requirements for a subset of patients.
# Discussion
We have established a framework for assessing the utility of a genetic test for clozapine-induced agranulocytosis and explored the characteristics of tests that would reduce agranulocytosis risk to a level that does not require regular haematological monitoring. In particular, we show that high sensitivity is essential and that a small proportion of patients classified as HR further decreases the agranulocytosis risk in the LR group.
High sensitivity is a self-evident characteristic of a clinically useful test, but the finding that a small HR group is favourable might seem counterintuitive. One could reason that to be sure the LR group contains no agranulocytosis cases, the HR group should include all patients who are at the slightest risk of developing agranulocytosis, and that the HR group should thus be large. However, as the number of true agranulocytosis cases is very low, a large HR group would mainly contain false-positive patients. Instead of a large HR group, a useful test relies on high sensitivity to correctly classify the agranulocytosis cases as HR. A small HR group implies few false positives and many true negatives, which in turn minimizes the agranulocytosis risk in the LR group.
Instead of a single genetic locus, a pharmacogenetic test could also be based on polygenic risk scores, built from combining risk conferred by many genetic loci to identify patients at high risk. In that case, the threshold defining LR and HR groups can be varied, and the effectiveness of a test is typically measured by the area under a receiver operating characteristic curve. [bib_ref] Receiver-operating characteristic analysis for evaluating diagnostic tests and predictive models, Zou [/bib_ref] When the threshold is moved to increase test sensitivity, the size of the HR group will increase as well. It is not straightforward to predict the resulting change-increase or decrease-in agranulocytosis risks in the LR and HR groups, because these depend on the distribution of the polygenic risk scores. Once an appropriate polygenic score threshold has been fixed, a test can be translated easily to the framework developed here.
Pharmacogenetic tests for abacavir, carbamazepine and purine analogues in clinical use have sensitivity, specificity and NPV close to 100% [fig_ref] Table 5: Characteristics of pharmacogenetic tests that predict adverse drug reactions Abacavir HLA-B*5701The HLA-B*5701... [/fig_ref]. [bib_ref] HLA-B* 5701 screening for hypersensitivity to abacavir, Mallal [/bib_ref] [bib_ref] HLA-B* 1502 and risk of Stevens-Johnson syndrome and toxic epidermal necrolysis: US..., Ferrell [/bib_ref] [bib_ref] Imputation of TPMT defective alleles for the identification of patients with highrisk..., Almoguera [/bib_ref] However, the PPVs of these tests vary between 7.7 and 80.6%. These low PPVs are acceptable, as there are alternative treatments available for patients who test positive. In contrast, clozapine is reserved for treatment of refractory schizophrenia and no alternative drug is available. A high PPV would ensure that few patients are unnecessarily excluded from treatment, but a high NPV and consequently low agranulocytosis risk is also important to justify a reduced monitoring schedule for patients in the LR group.
No genetic test for clozapine-induced agranulocytosis currently exists. The proposed test of HLA-DQB1 6672G4C has high specificity, but low sensitivity fails to reduce the agranulocytosis risk in the LR group sufficiently that monitoring could be reduced or ceased. Candidate gene studies have failed to identify a strong, replicated genetic variant that substantially increases risk of Pharmacogenetic test for clozapine-induced agranulocytosis M Verbelen et al clozapine-induced agranulocytosis. [bib_ref] Genetics of antipsychotic-induced side effects and agranulocytosis, Chowdhury [/bib_ref] [bib_ref] Clozapine-induced agranulocytosis and its genetic determinants, Opgen-Rhein [/bib_ref] The first genome-wide association study of clozapine-induced agranulocytosis detected significant associations at two HLA amino acids; 33 at least one further study is in progress,and combined analysis of such studies may identify associated genetic variants that can be rapidly translated to clinical practice. Abbreviations: HLA, human leukocyte antigen; NPV, negative predictive value; PPV, positive predictive value; SNP, single nucleotide polymorphism.
Pharmacogenetic test for clozapine-induced agranulocytosis M Verbelen et al
[fig] Figure 2: Agranulocytosis risk in LR and HR groups by proportion of patients that classified as HR, for different sensitivity values between 20 and 100%. [/fig]
[fig] Figure 3: Agranulocytosis risk in (a) LR group and (b) HR group by sensitivity and proportion of patients in the HR group. In both panels, darker colours represent the desired outcomes of low risk in the LR group and high risk in the HR group. The letters indicate the position of hypothetical tests A, B, C and D; HLA indicates the position of the HLA-DQB1 6672G4C-based test. [/fig]
[table] Table 1: Classification table comparing test outcome with true agranulocytosis status [/table]
[table] Table 2: Hypothetical tests with 80% sensitivity for different proportions of patients in the HR group, showing agranulocytosis risks in the LR group (1-NPV), the HR group (PPV) and specificity of the test [/table]
[table] Table 3: Hypothetical tests with 10% of patients in the HR group for [/table]
[table] Table 4: Four hypothetical pharmacogenetic tests for clozapine-induced agranulocytosis and their clinical impact Test Sensitivity Size of HR group P(A9LR) (1-NPV) P(A9HR) (PPV) Specificity Clinical impact Abbreviations: HR, high risk; LR, low risk; NPV, negative predictive value; PPV, positive predictive value. [/table]
[table] Table 5: Characteristics of pharmacogenetic tests that predict adverse drug reactions Abacavir HLA-B*5701The HLA-B*5701 allele is associated with hypersensitivity reaction. Carriers should avoid abacavir. The HLA-B*1502 allele is associated with SJS-TEN in Han Chinese. Carriers should avoid carbamazepine. SNPs to identify patients with zero wildtype alleles, as they are at high risk of myelotoxicity. [/table]
|
Cross-neutralization of SARS-CoV-2 B.1.1.7 and P.1 variants in vaccinated, convalescent and P.1 infected
# Introduction
More than one year after the declaration of Coronavirus disease 2019 (COVID-19) as a pandemic, 1 severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) is still spreading around the world causing serious public health concerns. A number of effective vaccines are being administered at an unprecedented pace, decreasing the incidence and severe consequences of SARS-CoV-2 infection. However, massive and prolonged worldwide replication let SARS-CoV-2 rapidly explore its genetic space and new variants of concern (VOCs) eventually emerged by the end of 2020. Briefly, VOC indicates a variant for which there is evidence of increased transmissibility and possibly more severe disease as well as decreased neutralization by antibodies elicited by natural infection or vaccination with the original lineage, reduced effectiveness of treatments and diagnostic detection failures. The most relevant variants emerged in United Kingdom (known as 20I/501Y.V1, VOC 202,012/01, B.1.1.7, or alpha), South Africa (known as 20H/501Y.V2, B.1.351, or beta) and Brazil (known as P.1, or gamma) during 2020 but started to spread all around the world between December 2020 and January 2021.These variants possess different mutations on the receptor-binding domain (RBD) of the spike protein responsible for binding to the ACE2 receptor on the human cell surface. Since the RBD is a major target for neutralizing antibodies and all the available vaccines have been produced based on the "original" spike protein, the efficacy of vaccine-induced immune response is being questioned. Furthermore, convalescent people previously infected and recovered from SARS-CoV-2 infection may no longer be protected against SARS-CoV-2 reinfection. [bib_ref] Is recurrence possible in coronavirus disease 2019 (COVID-19)? Case series and systematic..., Gidari [/bib_ref] In Italy, the SARS-CoV-2 lineage B.1, clade 20A.EU1 circulating across Europe was first identified in March 2020 and remained dominant up to November 2020. [bib_ref] Emergence and spread of sars-cov-2 lineages b.1.1.7 and p.1 in Italy, Di Giallonardo [/bib_ref] By the end of February 2021, surveillance data of Umbria region, Italy, revealed the SARS-CoV-2 variants B.1.1.7 and P.1 as accounting for 36.2% and 51.1% of the total cases analyzed, respectively.The vaccination campaign started by the end of December 2020 with the administration of BNT162b2 (Comirnaty® -BioNTech / Pfizer) COVID-19 mRNA vaccine, firstly to healthcare workers. The original aim of this study was to evaluate the dynamics and duration of SARS-CoV-2 neutralizing activity of BNT162b2 -elicited antibody. The subsequent emergence of the different VOCs allowed us to expand the study design and analyze the cross-neutralization pattern with the different variants by using sera from subjects vaccinated with BNT162b2 or infected with different virus lineages. In particular, samples from patients infected with the SARS-CoV-2 P.1 variant have been included in the study together with the ones having a history of SARS-CoV-2 infection from June to October 2020 and candidates as hyper-immune plasma donors.
# Materials and methods
## Design, setting and participants
This research was approved by the Ethics Committee of the Umbria Region (protocol number 20,686/21/OV) and it was conducted in accordance with the Declaration of Helsinki. All participants provided written informed consent. The study used a longitudinal cohort design, with three separate cohorts: healthcare workers of Perugia Hospital vaccinated with BNT162b2 (Comirnaty® -BioNTech/Pfizer) COVID-19 mRNA vaccine; patients with a history of SARS-CoV-2 infection from June to October 2020 and candidates as hyper-immune plasma donors; patients with documented SARS-CoV-2 P1 variant infection.
Samples of vaccinated patients were withdrawn after 14-21 days from the second dose. Furthermore, the subgroup of candidate plasma donors with high neutralizing antibodies (NT-Abs) titers ( ≥1:160) was also analyzed separately to establish if the high titer would be confirmed for the B.1.1.7 and P.1 strains.
## Sars-cov-2 strains and vero e6 cell cultures
All experiments were performed using three SARS-CoV-2 strains isolated in our Biosafety Level 3 (BSL3) virology laboratory at Santa Maria della Misericordia Hospital, Perugia, Italy, as previously described. [bib_ref] SARS-CoV-2 survival on surfaces and the effect of UV-C light, Gidari [/bib_ref] Briefly, the transport medium (UTM) of a nasopharyngeal swab was incubated with a 1:1 nystatin (10,0 0 0 U/mL) and penicillin-streptomycin (10,0 0 0 U/mL) mixture for 1 h at 4 °C to remove bacterial/fungal contamination. The suspension was centrifuged at 400 × g for 10 min, and the supernatant was inoculated on an African green monkey kidney clone E6 (Vero E6) cells monolayer maintained in Eagle's minimum essential medium (MEM) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin at 37 °C with 5% CO 2 . The viral titer in the supernatant was determined by Half-maximal Tissue Culture Infectious Dose (TCID50) endpoint dilution assay 7 and stock aliquots were stored at −80 °C. Whole-genome sequencing of multiple isolates was used to identify a SARS-CoV-2 genome belonging to clade 20A.EU1 (lineage B.1) and clustered with viruses circulating in Italy in spring 2020, a SARS-CoV-2 genome belonging to clade B.1.1.7 (better known as alpha variant), and a SARS-CoV-2 genome belonging to clade P.1 (also known as gamma variant). [bib_ref] Molecular tracing of SARS-CoV-2 in Italy in the first three months of..., Lai [/bib_ref] The SARS-CoV-2 clade 20A.EU1 (lineage B.1) strain was isolated in May 2020 from a symptomatic patient during the first wave of infections. SARS-CoV-2 clade B.1.1.7 and P.1 were isolated on January 2021 during the third wave. Single virus stock aliquots were thawed immediately before each experiment and discarded after use.
## Sars-cov-2 neutralization test
Neutralizing antibodies (NT-Abs) titers against SARS-CoV-2 were evaluated using flat-bottom tissue culture 96-well microtiter plates as previously published. [bib_ref] Prevalence of SARS-CoV-2 specific neutralising antibodies in blood donors from the, Percivalle [/bib_ref] One day prior to the experiment, Vero E6 cells (2.5 × 10 4 /well) were cultivated in MEM + 10% FBS in a 96-well plate incubated at 37 °C + 5% CO 2 . Serum samples of each study cohort were heat-inactivated for 30 min at 56 °C before to be two-fold serially diluted from 1:10 to 1:640 with MEM + 2% FBS and seeded into 96-well microtiter plates. Each sample was tested in duplicate and mixed with an equal volume of medium containing 50 TCID50 of SARS-CoV-2 selected strains and the plates were incubated for 30 min at 37 °C. Sera with known neutralization titer and medium were used as positive and negative control, respectively. Subsequently, the mixtures were transferred to Vero E6 cell containing plates and incubated for 72 h at 37 °C + 5% CO 2 . Then, the medium was removed and the cells were fixed and stained with 0.25% crystal violet and 10% formalin for 30 min at room temperature. The plates were washed to remove excess staining and the absorbance at 595 nm was measured using a microtiter plate reader (Multiskan Fc Photometer, Thermo Fisher Scientific). The neutralizing titer was determined as the maximum dilution showing reduction ≥ 90% of CPE respect to the virus control.
## Sanger sequencing and variant assignation
Viral RNA was extracted from 150 μl of viral stock using the ZR Viral RNA Kit (Zymo Research), according to the manufacturer's protocol. The cDNA was generated by random hexamer-driven reverse transcription using 10 μl of heat-denaturated RNA extract, 660 μM dNTPs, 6 μl 5X ImProm-II TM Reaction Buffer, 50 ng hexanucleotides, 1.5 mM MgCl 2 , 20 U RNasin® Plus RNase Inhibitor and 1 U of ImProm-II TM Reverse Transcriptase (Promega) in a final volume of 30 μl. Reactions were run for 30 min at 37 °C followed by enzyme inactivation for 5 min at 80 °C. The cDNA was used as the template to amplify a spike region of about 20 0 0 nucleotides spanning the mutations involved in variant designation. To design conserved primers, the alignment of reference genomes available at Gisaid site was used ( https://www.gisaid.org/ ). The PCR mixture included 3 μl cDNA, 5 μl 5X Q5 Reaction buffer (NEB), 4 pmol P1004-Fwd (5 -AATTAGAGAAAACAACAGAGT-3 ) and P976-Rev (5 -AAATTTGTGGGTATGGCAATAGAGTTA-3 ), 150 μM dNTPs and 0.5 U Q5 Hot Start High-Fidelity DNA Polymerase (NEB) in a final volume of 25 μl. Reactions were performed with an initial denaturation step at 98 °C for 3 min followed by 40 cycles each including 30 s at 58 °C, 1 min and 15 s at 72 °C and 10 s at 98 °C and a final step at 72 °C for 5 min. Bidirectional DNA sequencing reactions were performed using the BrilliantDye TM Terminator Kit v1.1 (Nimagen) with four different primers spanning the spike region of interest. Briefly, 3 μl of PCR products, diluted at final concentration of 1-3 ng/μl, were mixed with 3.2 pmol/μl of each sequencing primer, 0.5 μl of BrilliantDyeTM Terminator Ready Reaction Sequencing and 2 μl of 5x Sequencing Buffer in a final volume of 10 μl. The reactions were denatured at 96 °C for 1 min followed by 25 cycles at 50 °C for 5 s, 60 °C for 4 min and 96 °C for 10 s. Sequencing reactions were treated with X-Terminator® Purification kit (Applied Biosystems) and then resolved by capillary electrophoresis with the 3130 XL Genetic Analyzer (Applied
# Statistical analysis
Statistical analysis was performed using Graphpad 8.3. Data were tested for normality using the Kolmogorov-Smirnov test and were presented as mean with the respective standard deviation (SD) or median with interquartile range (IQR), as appropriate. Data with normal distribution were analysed with one-way repeated measures analysis of variance (ANOVA) and Bonferroni's multiplecomparison test. For nonparametric variables, the Friedman test and Dunn's multiple comparison test were performed. The relationship between NT-Abs titre and time passed since diagnosis was evaluated by Spearman r correlation test. A p-value < 0.05 was considered significant.
# Results
## Design, setting and participants
As shown in [fig_ref] Figure 1: Patients selection flowchart [/fig_ref] A-C, three cohorts of subjects were enrolled in this study.
## Vaccinated healthcare workers
This cohort was composed of 103 healthcare workers vaccinated with the BNT162b2 COVID-19 mRNA vaccine [fig_ref] Figure 1: Patients selection flowchart [/fig_ref].
The mean age was 44.7 ± 10.1 years and 18.1% were male. All subjects were tested for SARS-CoV-2 antibody before vaccination. Ninety subjects had no SARS-CoV-2 infection, either before or within 21 days from completed vaccination.
Three subjects had asymptomatic SARS-CoV-2 infection before vaccination. Six subjects had mild or asymptomatic SARS-CoV-2 infection between the first and second vaccine dose. Three subjects were lost to follow-up. Another participant had an anaphylactic reaction to the first dose of the vaccine and did not undergo the second one. Consequently, these 13 patients were analyzed separately.
For subjects completing the two-dose vaccination schedule, serum samples were drawn in median 16 days (IQR 15-18 days) after the second dose.
The distribution of the 90 vaccinated healthcare workers according to NT-Abs titers is shown in [fig_ref] Figure 2: Distribution of neutralizing antibodies [/fig_ref] A, E. In detail, the median NT-Abs titer was 1:80 (1:40-1:80) when serum was tested with 20A.EU1 and B.1.1.7 strains. However, when the same sera were tested with P.1 strain the median NT-Abs titer was 1:20 (1:10-1:40). The median NT-Abs titer was significantly higher for 20A.EU1/B.1.1.7 than for P.1 ( p < 0.0 0 01). The mean titer was 3.3fold higher for 20A.EU1 than P.1 strain.
All the three subjects with previous SARS-CoV-2 infection had an NT-Abs titer of 1:640 for 20A.EU1 and B.1.1.7, and 1:160 to 1:640 for the P.1 variant.
Among patients with SARS-CoV-2 infection occurring between the first and the second vaccine dose, 5/6 were from P.1 strain and 1/6 strain was not identified. Four of the five patients with P.1 infection were analyzed among the P.1 patients' cohort, one was lost to follow-up.
Only one participant developed SARS-CoV-2 infection after two weeks from completed vaccination: she was a young woman with an NT-Abs titer of 1:20 for 20A.EU1 and B.1.1.7 and 1:10 for P.1. She contracted the B.1.1.7 variant and complained mild cough and rhinorrhea without fever.
## Candidates as hyper-immune plasma donors
We postulated that patients with a history of SARS-CoV-2 infection acquired from June to October 2020 were infected with SARS-CoV-2 lineage 20A.EU1, the highly predominant variant in Italy at that time. [bib_ref] Emergence and spread of sars-cov-2 lineages b.1.1.7 and p.1 in Italy, Di Giallonardo [/bib_ref] We enrolled 90 convalescent patients as candidates as hyper-immune plasma donors [fig_ref] Figure 1: Patients selection flowchart [/fig_ref].
Patients were enrolled in a median of 67 days (IQR 43-87, range 22-246) after diagnosis of SARS-CoV-2 infection.
The distribution of NT-Abs titers is shown in [fig_ref] Figure 2: Distribution of neutralizing antibodies [/fig_ref] B, F. In detail, the median NT-Abs titer was 1:160 (1:80-1:320), 1:80 (1:80-1:160) and 1:20 (1:10-1:40) when serum was tested with 20A.EU1, B.1.1.7, and P.1 strains, respectively. The median NT-Abs titer was significantly higher for 20A.EU1 than for B.1.1.7 (mean titer 1.6fold higher) and P.1 (mean titer 6.7-fold higher) ( p = 0.0 0 02 and p < 0.0 0 01 respectively). Furthermore, the median NT-Abs titer was significantly higher for B.1.1.7 than for P.1 (mean titer 4.2-fold higher) ( p < 0.0 0 01).
A subgroup of 64 patients had a high NT-Abs titer ( ≥1:160) and, consequently were considered eligible as hyper-immune plasma donors. This subgroup had a median NT-Abs titer of 1:240 (1:160-1:320), 1:160 (1:80-1:160) and 1:30 (1:20-1:40) when serum was tested with 20A.EU1, B.1.1.7, and P.1 strains, respectively [fig_ref] Figure 2: Distribution of neutralizing antibodies [/fig_ref]. The median NT-Abs titer was significantly higher for 20A.EU1 than for B.1.1.7 (mean titer 1.8-fold higher) and P.1 (mean titer 8.5-fold
## Sars-cov-2 p.1 variant infection cohort
Twenty-two patients with ascertained SARS-CoV-2 P.1 infection acquired from 21st November 2020 to 8th February 2021 in the Umbria region were enrolled, including 5 healthcare workers who already had a dose of BNT162b2 vaccine [fig_ref] Figure 1: Patients selection flowchart [/fig_ref]. Two patients were tested at multiple time points, one 2 and the other 3 times. Each measure was considered separately, therefore, a total of 25 samples were considered for the analysis. The mean age was 61.1 ± 19.8 years and 13 (59.1%) were male. Patients were enrolled a median of 21.0 days (IQR 15.3-34.0, range 12-94) after diagnosis of SARS-CoV-2 infection.
The distribution of the 25 sera according to NT-Abs titers is shown in [fig_ref] Figure 2: Distribution of neutralizing antibodies [/fig_ref] In detail, the median NT-Abs titer was 1:10 (IQR < 1:10-1:30, range < 1:10-1:80), 1:10 (IQR 1:5-1:20, range < 1:10-1:80) and 1:80 (IQR 1:40-1:320, range < 1:10-1:640) when serum was tested with 20A.EU1, B.1.1.7, and P.1 strains, respectively. The median NT-Abs titer was significantly higher for P.1 than for 20A.EU1 (mean titer 12.2-fold higher) and B.1.1.7 (mean titer 10.9fold higher) ( p < 0.0 0 01). The median NT-Abs titer was not significantly different for 20A.EU1 and B.1. No correlation between NT-Abs titre and time passed since diagnosis was found for any of the three lineages.
# Discussion
The emergence of VOCs could significantly complicate SARS-CoV-2 pandemic. Starting from January 2021, in the Umbria region, a small area in the center of Italy, SARS-CoV-2 B.1.1.7 and P.1 variants widely spread causing an important increase of cases during the third wave. In particular, the prevalence of P.1 was disproportionally higher, reaching 36.2% as compared to 4.1% of the whole Italy, while the prevalence of B.1.1.7 was in line with the rest of the country (51.1% versus 54.0%, respectively).It must be noted that P.1 outbreaks involved healthcare workers and thus may have been easier to detect within surveillance programs.
Few studies have so far investigated how different mutations/variants could impact the immunologic response both of vaccinated and convalescent people. Due to different complexity of the assays, neutralizing antibody titration has been used far more often than analysis of cell mediated response and most studies have used pseudoviruses instead of live virus. However, neutralizing antibodies are of primary importance as a key component of immune memory providing sterilizing immunity. Even substerilizing NT-Abs titers would limit infection and COVID-19 severity. [bib_ref] Immunological memory to SARS-CoV-2 assessed for up to 8 months after infection, Jennifer [/bib_ref] In this study, we only evaluated the role of neutralizing antibodies but, also CD4 + and CD8 + T cell responses play important roles in COVID-19 course modulating disease severity and contrasting viral replication. Tarke et al. demonstrated that T cell responses elicited by either SARS-CoV-2 natural infection or vaccination with mRNA vaccines were not affected by the B.1.1.7, B.1.351, P.1 and CAL.20C variants. [bib_ref] Negligible impact of SARS-CoV-2 variants on CD4 + and CD8 + T..., Tarke [/bib_ref] However, SARS-CoV-2 specific CD4 + T cells and CD8 + T cells showed an earlier decline with a half-life of 3-5 months compared to IgG to the spike protein that was relatively stable over 6 months. [bib_ref] Immunological memory to SARS-CoV-2 assessed for up to 8 months after infection, Jennifer [/bib_ref] Sub-sterilizing immunity and T cell responses probably contribute together to limit COVID-19 severity in patients with reinfection or infection after vaccination due to VOCs.
Neutralizing activity of BNT162b2 -elicited sera against B.1.1.7 variant was widely tested. All studies concluded that neutralizing activity to this VOC is largely preserved. Some authors demonstrated roughly equivalent NT-Abs activity to B.1.1.7 and the wildtype strain, [bib_ref] Multiple SARS-CoV-2 variants escape neutralization by vaccine-induced humoral immunity, Garcia-Beltran Wilfredo [/bib_ref] [bib_ref] Sensitivity of infectious SARS-CoV-2 B1.1.7 and B.1.351 variants to neutralizing antibodies, Planas [/bib_ref] [bib_ref] Neutralizing Activity of BNT162b2-Elicited Serum, Liu [/bib_ref] while others detected less efficient neutralization against the B.1.1.7 variant. [bib_ref] Suthar Mehul S. Neutralizing antibodies against SARS-CoV-2 variants after infection and vaccination, Edara Venkata Viswanadh [/bib_ref] [bib_ref] mRNA vaccine-elicited antibodies to SARS-CoV-2 and circulating variants, Wang [/bib_ref] [bib_ref] Neutralization of SARS-CoV-2 lineage B.1.1.7 pseudovirus by BNT162b2 vaccine-elicited human sera, Muik [/bib_ref] [bib_ref] Neutralization of SARS-CoV-2 spike 69/70 deletion, E484K and N501Y variants by BNT162b2..., Xie [/bib_ref] In particular, different studies found a reduction of B.1.1.7 neutralizing titer from 1.9 to 3.3-fold respect to wild-type strain. [bib_ref] Antibody evasion by the P.1 strain of SARS-CoV-2, Dejnirattisai [/bib_ref] [bib_ref] Antibody evasion by the P.1 strain of SARS-CoV-2, Dejnirattisai [/bib_ref] Chen et al. studied the in vitro impact of different SARS-CoV-2 spike protein mutations on serum neutralizing activity by introducing individual point mutations in the spike gene into an infectious complementary DNA clone of the 2019n-CoV/USA_WA1/2020. They found similarly reduced inhibitory activity against viruses containing the E484K spike mutation, such as P.1 and B.1.351. [bib_ref] Resistance of SARS-CoV-2 variants to neutralization by monoclonal and serum-derived polyclonal antibodies, E Chen Rita [/bib_ref] Other studies reported the relative resistance of P.1 to neutralization by multiple therapeutic monoclonal antibodies, convalescent plasma, and sera from vaccine recipients with respect to the wild-type strain. Conversely, Liu et al., studying the neutralizing activity of BNT162b2 -elicited serum, did not find a significant difference in neutralization activity across wild-type, B.1.1.7 and P.1 spike proteins, while neutralization of B.1.351-spike protein was still robust but lower, 14 a finding corroborated by other studies. Overall, data on P.1 neutralization by antibody elicited by natural or artificial exposure to the wild type spike protein remain equivocal, likely due to the use of different methods and the lack of reference isolates at this time.
In our study, the lower neutralizing activity of BNT162b2elicited sera on the P.1 variant may explain the five cases of P.1 variant infection in vaccine recipients.
Our vaccine recipients with previous SARS-CoV-2 infection had a neutralizing activity significantly higher for all the three strains compared to the uninfected vaccinees. These results, despite the low number of samples, are in line with the published data. Studies on SARS-CoV-2 convalescent patients who previously contracted wild-type SARS-CoV-2 infection demonstrated neutralizing activity to B.1.1.7 and P.1 VOCs but with significantly lower titers compared with the homologous virus (2-4.5-fold and 3.1fold, respectively). However, few studies did not detect a reduction of activity on B.1.1.7, highlighting limited consistency of literature data also with convalescent sera. Similar to our findings, Colier et al., demonstrated that vaccinated patient serum has higher neutralizing activity against B.1.1.7 compared to convalescent ones (about 3.6-fold of difference). [bib_ref] Sensitivity of SARS-CoV-2 B1.1.7 to mRNA vaccine-elicited antibodies, Collier Dami [/bib_ref] Indeed, in our study, vaccine sera neutralized both 20A.EU1 and B.1.1.7 strains with the same efficacy. We separately analyzed a subgroup of high NT-Abs titer convalescent patients (NT-Abs ≥1:160) and we found that these sera could not be considered as hyperimmune to B.1.1.7 and P.1 VOCs because they were significantly less effective in neutralizing these variants. Thus, the emergence of VOCs can abrogate this therapeutic option when the SARS-CoV-2 lineage of donors and recipient do not match.
To the best of our knowledge, this is the first study that evaluated neutralizing activity of sera from P.1 variant infected patients to other SARS-CoV-2 lineages. As expected, we found an important reduction of NT-Abs against 20A.EU1 and B.1.1.7 strains (12.2 and 10.9-fold, respectively). Considering that patients with a previous wild-type infection had a titer reduction of 6.7 and 1.58-fold on P.1 and B.1.1.7, it appears that patients with a previous P.1 infection are less protected from further SARS-CoV-2 reinfections from other variants.
# Conclusions
The impact of VOCs on serum neutralization activity and protection from SARS-CoV-2 (re)infection needs to be further clarified. Various studies have shown different results and come to different conclusions. Our data corroborate the concept that B.1.1.7 and P.1 are less efficiently neutralized by convalescent sera from subjects infected by the original virus. However, BNT162b2 vaccineelicited human sera have an equivalent neutralization potency on the B.1.1.7 but lower on the P.1 variant. Convalescent P.1 patients are less protected from other SARS-CoV-2 strains.
# Ethical statements
The Vero E6 cell line was kindly provided by Istituto Zooprofilattico Sperimentale di Brescia, Brescia, Italy.
This research was approved by the Ethics Committee of the Umbria Region, protocol numbers 18,344/20/OV and 20,686/21/OV.
## Disclaimers
Nothing to declare.
[fig] Figure 1: Patients selection flowchart. Coronavirus disease 2019, COVID-19; Severe Acute Respiratory Syndrome Coronavirus 2, SARS-CoV-2; Neutralizing antibodies, NT-Abs. Biosystems). Chromatograms were assembled and edited with the DNAStar 7.1.0 SeqMan module and imported in the clade assignment tool ( https://clades.nextstrain.org/ ) to determine the variant. [/fig]
[fig] Figure 2: Distribution of neutralizing antibodies (NT-Abs) titers of BNT162b2 -vaccinated healthcare workers ( N = 90) (A, E), convalescent candidates as plasma donors ( N = 90) (B, F), high titers plasma donors ( N = 64) (C, G) and SARS-CoV-2 P1 infected patients ( N = 25) (D, H). NT-Abs titers against SARS-CoV-2 were evaluated using flat-bottom tissue culture 96-well microtiter plate serum dilution assay. Serums have been tested against 20A.EU1, B.1.1.7 and P.1 strains isolated from symptomatic patients with Coronavirus disease 2019 (COVID-19). Panels A, B, C and D show NT-Abs titers distribution for the three strains. Panels E, F, G and H show how NT-Abs serum titers of each patient change for the three variants. Data were presented as median with interquartile range (IQR). The Friedman test and Dunn's multiple comparison test were performed. A p-value < 0.05 was considered significant. * Significant if compared to 20A.EU1 strain. # Significant if compared to P.1 strain.higher) ( p < 0.0 0 01) and it was higher for B.1.1.7 than for P.1 (mean titer 4.3-fold mean titer) ( p < 0.0 0 01). [/fig]
|
The potential role of peak nasal inspiratory flow to evaluate active sinonasal inflammation and disease severity
Although the pathophysiology of nasal polyposis is incompletely understood, rhinologists have seldom studied it with rhinomanometry or peak nasal inspiratory flow (PNIF) due to technical limitations and the perception that polyp size might impair reproducibility and the usefulness of recordings. The objective of this study is to assess how measures of rhinomanometry and PNIF relate to disease activity. Nineteen patients with polyps, 15 patients with chronic sinusitis without polyps and 11 negative controls were evaluated with active anterior rhinomanometry and PNIF. Sinusitis and polyp patients were re-evaluated after medical treatment. Polyp patients had the highest median Lund-Mackay score (14) and a median Johansen score of 1. PNIF and its variation after treatment were also lowest in this group (median 90 L/min before and after treatment; median variation of 0 L/min). Nasal resistance was similar between groups, and only correlated with Johansen score (Spearman = 0.517, p = 0.048) after treatment. Our study suggests that evaluating polyp patients using rhinomanometry and PNIF may provide useful and reproducible data. Several findings considered together suggest that polyp size is not the main determinant of nasal functional changes in these patients, warranting further studies to verify whether PNIF changes reflect sinus inflammation or merely airway obstruction.
Nasal polyposis (NP) is a chronic nasal inflammatory disease and is considered an endotype of chronic rhinosinusitis (CRS) [bib_ref] European position paper on rhinosinusitis and nasal polyps 2012, Fokkens [/bib_ref] [bib_ref] Endotype-driven treatment in chronic upper airway diseases, De Greve [/bib_ref] [bib_ref] Positioning the principles of precision medicine in care pathways for allergic rhinitis..., Hellings [/bib_ref]. The prevalence of NP is reported to be 1-4% of the general population [bib_ref] European position paper on rhinosinusitis and nasal polyps 2012, Fokkens [/bib_ref] [bib_ref] Diagnosis and treatment of nasal polyps, Lund [/bib_ref] [bib_ref] Prevalence of nasal polyps in adults: The Skovde population-based study, Johansson [/bib_ref] , reaching as high as 32% in cadaver studies [bib_ref] Origin of nasal polyps: An endoscopic autopsy study, Larsen [/bib_ref]. The average incidence of symptomatic NP has been estimated to be around 0.86 and 0.39 patients/1,000 individuals per year in male and female patients, respectively [bib_ref] The estimated incidence of symptomatic nasal polyps, Larsen [/bib_ref]. Medical or surgical treatment significantly improve quality of life but surgery may be required in up to 50% of patients 8 and 8-25% of those may require revision surgery [bib_ref] European position paper on rhinosinusitis and nasal polyps 2012, Fokkens [/bib_ref] [bib_ref] Polypectomy compared with ethmoidectomy in the treatment of nasal polyposis, Devars Du Mayne [/bib_ref]. The socio-economical cost of CRS parallels those of diabetes mellitus, chronic obstructive pulmonary disease and coronary artery disease [bib_ref] The health and productivity cost burden of the "top 10" physical and..., Goetzel [/bib_ref]. Poorly controlled CRS is estimated to cost around 2000€ per year, per patient [bib_ref] Cost analysis of regular and filgrastim treatment in patients with refractory chronic..., Van Agthoven [/bib_ref] , based on direct health services use, and indirect costs of absenteism and presenteism. The worsening of associated co-morbidities (ex: asthma, sleep apnea, etc.) also represents an economical burden.
Rhinologists agree that research to determine the most effective treatment plan is a priorityand should be encouraged by proper authorities [bib_ref] Rhinitis, not to sniff at, Fokkens [/bib_ref] , with emphasis on the principles of precision medicine and patient centered care [bib_ref] Positioning the principles of precision medicine in care pathways for allergic rhinitis..., Hellings [/bib_ref]. These explicitly recommend the use of relevant clinical biomarkers to elaborate personalized treatment plans.
The central role of epithelial cell lesion in the pathogenesis of NP has long been recognized [bib_ref] Pathogenesis of nasal polyps, Tos [/bib_ref] and mucosal inflammation has been extensively studied and established as the cause rather than the consequence of NP in experimental models [bib_ref] Polyp pathogenesis-A histopathological study in experimental otitis media, Caye-Thomasen [/bib_ref] [bib_ref] The relationship of nasal polyps, infection, and inflammation, Norlander [/bib_ref]. However, little was investigated on the effect of nasal ventilation in NP. Consequently, no nasal ventilation variables are considered as possible disease biomarkers in NP. Epithelial metaplasia and ciliary dysfunction were noted with increased local nasal airflow [bib_ref] Do chronic changes in nasal airflow have any physiological or pathological effect..., Boyce [/bib_ref]. Known mesenchimal cell mechano-transduction mechanisms [bib_ref] Primary cilia-mediated mechanotransduction in human mesenchymal stem cells, Hoey [/bib_ref] as well as analogies between endothelial cell transmembrane proteins and shear stress lesion with NP nitric oxide and epithelial disaggregation mechanisms suggest that ventilation www.nature.com/scientificreports/ pathogenesis. Recently, computerized fluid dynamic models reinforced this idea by demonstrating increased positive pressure areas where polyps were found [bib_ref] Investigation of flow characteristics in regions of nasal polypoid change, Chung [/bib_ref]. In short, several studies suggest that nasal ventilation may indeed influence the behaviour of NP. There are several techniques to objectively evaluate nasal ventilation, of which rhinomanometry and peak nasal inspiratory flow (PNIF) are the best studied. Several guidelines, panel reports and consensus help researchers and clinicians to apply these techniques in a standardized and comparable manner [bib_ref] Diagnostic tools in rhinology EAACI position paper, Scadding [/bib_ref] [bib_ref] The new agreement of the international RIGA consensus conference on nasal airway..., Vogt [/bib_ref]. A recognized limitation is the inability to obtain measures in a setting of severe obstruction, and polyps may also limit the reliability of results because of reduced reproducibility due to variable size between individuals and throughout time, related to disease behaviour or treatment effect, as reported previously [bib_ref] What happens with untreated nasal polyps over time? A 13-year prospective study, Oscarsson [/bib_ref]. The future in nasal functional evaluation is moving towards the use of computerized analysis of nasal anatomy based on fluid dynamics models [bib_ref] Investigation of flow characteristics in regions of nasal polypoid change, Chung [/bib_ref] , or, possibly new methods for non-invasive evaluations such as those based on thermistors [bib_ref] Thermistor at a distance: Unobtrusive measurement of breathing, Fei [/bib_ref] [bib_ref] Dynamic 3D print of the breathing function, Duong [/bib_ref] , that have been adapted from other areas. Such methods may provide very refined analysis of the patients' breathing function without the need for artificial ventilation patterns or patient compliance. However, this will also come at the cost of equipment, time, dedicated personnel and specific training to do so. Moreover, they still lack clinical implementation and validation. PNIF, on the other hand, is a well-established method to evaluate nasal function objectively that can be performed by the otolaryngology surgeon during any consultation. This will bear minimal costs and the results will be available immediately. For these reasons PNIF may prove to be an indispensable tool in rhinology practice, as it has been on clinical research [bib_ref] Measurements of nasal airflow and patency: A critical review with emphasis on..., Ottaviano [/bib_ref].
In summary, NP is a socially relevant disease and new disease biomarkers must be identified to help manage treatment in a personalized manner. Few studies have reported results from PNIF and/or rhinomanometry in NP patient cohorts [bib_ref] What happens with untreated nasal polyps over time? A 13-year prospective study, Oscarsson [/bib_ref] [bib_ref] Pre-and postoperative evaluation of patients with nasal obstruction undergoing endoscopic sinus surgery, Sipila [/bib_ref] [bib_ref] Nasal polyps. A clinical study of endoscopic score systems and epidemiology in..., Johansson [/bib_ref] [bib_ref] Impact of baseline nasal polyp size and previous surgery on efficacy of..., Djupesland [/bib_ref] [bib_ref] The evaluation of pattern and quality of sleep in patients with chronic..., Uz [/bib_ref] and, to the best of our knowledge, none have specifically tried to measure the effect of polyp size on those techniques. In light of the previous considerations, our aim is to evaluate if polyp size alters, or not, PNIF and rhinomanometry results, which has implications in the validity of testing those techniques as NP disease biomarkers in the future.
We hypothesize that polyp size does not significantly change the results from nasal functional studies. To test that hypothesis, PNIF and rhinomanometry measurements were compared against polyp size. Additionally, the same measurements were recorded in patients with CRS or without nasal mucosa disease. Comparing those groups against NP patients, will help to answer the question of whether the presence of polyps is inevitably associated with poorer nasal ventilation.
# Methods
Study design. This is a cross-sectional and prospective study. Three patient groups were considered for recruitment: negative controls (NC), without nasal mucosa disease, positive controls, with CRS (without polyps) and patients with NP. This project respected all national and European regulations and legislation concerning good medical practice, clinical research, bioethics and data protection. The study protocol was submitted to and approved by the Ethics Committees from NOVA Medical School and Centro Hospitalar do Oeste. All patients gave their written and informed consent to participate in the study. Given the exploratory nature of the study and scarce information in the literature, a convenience sample size was decided based on examples of published studies [bib_ref] Comparison of decongestive capacity of xylometazoline and pseudoephedrine with rhinomanometry and MRI, Caenen [/bib_ref] [bib_ref] What is the impact of positive airway pressure in nasal polyposis? An..., Balsalobre [/bib_ref]. The study was developed in the context of a PhD project and funded by a prize from the NOVARTIS|Excellence in Medicine program and a grant from NOVA Saúde (Universidade Nova de Lisboa Rectorate).
Setting. The study takes place in an Otolaryngology and Head and Neck Surgery clinic in Hospital de Torres Vedras-Centro Hospitalar do Oeste. All eligible patients who met the inclusion criteria were sequentially enrolled after informed written consent was obtained. The main author was responsible for recruiting and evaluating (clinical, endoscopic, PNIF and rhinomanometry measurements) every study participant.
Patient recruitment and follow-up. Patients were evaluated and recruited to each group according to the diagnostic criteria published in EP3OS 1 . Those patients who met the following exclusion criteria were not enrolled in the study: patients unable to provide written informed consent,with congenital or pos-traumatic anatomic abnormalities,proven or suspected nasal neoplasia; choanal, antro-choanal or isolated single nasal polyps; isolated maxillary or sphenoid sinusitis. All patients had a baseline evaluation with PNIF and rhinomanometry and were medically treated according to the recommendations in the EP3OS document. Patients with CRS were provided with systemic and nasal topical corticosteroid and long-term chlarithromycin. Patients with NP were provided with the same and also leukotriene antagonists. Patients in the CRS and NP were re-evaluated 5-8 weeks after beginning medical treatment.
Nasal functional evaluations. Active anterior rhinomanometry was performed using a NR6 manometer from GM Instruments ® with the research software (4-phase curve analysis) and PNIF was performed with a Youlten debitometer, according to published consensus recommendations [bib_ref] Diagnostic tools in rhinology EAACI position paper, Scadding [/bib_ref]. A graphic depiction of the equipment and recording set-ups may be found in these guidelines [bib_ref] Diagnostic tools in rhinology EAACI position paper, Scadding [/bib_ref]. A twenty minutes acclimatization period was respected and all evaluations were performed before and after nasal decongestion with topical xylomethazoline spray (Vibrocil Actilong ® )-a puff in each nostril, followed by another 5 min after, and measurements ten minutes after the second puff. All recordings took place in the same room at an average temperature between 20° and 25° and although complete atmospherical standardization would be ideal, it is not strictly necessary Study variables and analysis. Age, gender and smoking status were registered as base population characteristics. The extension of sinusitis was measured using the Lund-Mackay score from computed tomographic Scientific RepoRtS | (2020) 10:12674 | https://doi.org/10.1038/s41598-020-69693-6
www.nature.com/scientificreports/ scans of the sinuses (each side and total). Polyp size was recorded considering Johansen's scale [bib_ref] The effect of budesonide (Rhinocort) in the treatment of small and medium-sized..., Vendelo Johansen [/bib_ref] , modified according to Kramer et al. [bib_ref] Nasal polyposis: Eosinophils and interleukin-5, Kramer [/bib_ref] , in order to distinguish large (polyp size 3) from obliterative lesions (polyp size 4): 0-no polyps,1-polyps not reaching the middle turbinate lower border,2-polyp not reaching the inferior turbinate lower border; 3-beyond the inferior turbinate lower border; 4-complete obstruction of the nasal airway. PNIF values were recorded before and after decongestion and the difference in the best of these values, before and after treatment was calculated (ΔPNIF). From the rhinomanometry evaluations we recorded resistance at 150 Pa from the left (Re) and right (Rd) nasal cavities and whole nose (Rt), as well as Broms angle (Ve, Vd and Vt) in both inspiratory and expiratory phases. A descriptive statistics study was performed for all variables. Continuous variables are presented with the average and standard deviation or median and interquartile range, as appropriate, and categorical variables are presented as proportions. Parametric or non-parametric independent or related samples tests were used with Bonferroni correction, according to whether the distribution of the data in each variable was normal or not. The Wilcoxon test was used to compare measurements before and after decongestion and the Chi-squared or Fisher test were used to compare categorical variables. A p-value < 0.05 was considered statistically significant and nasal cavities were compared independently from the patient (i.e. as if they were "different separate individuals"). Effect sizes for comparisons were calculated and interpreted according to Cohen criteria, as described by Ethical approval. All procedures performed in this study involving human participants were in accordance with the ethical standards of both institutional research committees (NOVA Medical School and Centro Hospitalar do Oeste) and with the 1964 Helsinki declaration and its later amendments and comparable ethical standards.
## Informed consent. informed consent was obtained from all individual participants included in the study.
Human research and informed consent. The study was approved by the institutional review board and namely by the ethics committee from NOVA Medical School and Centro Hospitalar do Oeste. All study participants were volunteers and were only included after providing their free and informed consent by writing.
# Results
Patient recruitment. Eleven healthy patients were evaluated, as well as 16 CRS patients (one lost for follow-up after beginning treatment and another recruited only after treatment) and 19 NP patients (two lost for follow-up)-see study flowchart in [fig_ref] Figure 1: Patient recruitment flowchart [/fig_ref]. When data is missing (due to absent records or inability to measure functional variables), n values for each analysis may differ from N values for each group.
Population characteristics. The individual descriptive characteristics for all patients can be found in [fig_ref] Table 1: Patient base characteristics [/fig_ref]. Significant differences were found regarding age, smoking status and Lund-Mackay scores. Age difference is significant between groups (p = 0.016) as well as smoking status (p = 0.027), due to the characteristics of the NP group. The proportion of smokers between genders was not significantly different (8/22 females vs 7/24 males, p = 0.755). Lund-Mackay scores are significantly different between groups, although, despite the observed trend for higher values in the NP group, statistical significance could not be reached when comparing CRS to NP patients (p values between 0.07 and 0.17).
Polyp size. Average polyp sizes and their changes with treatment are described in . One patient had bilateral score 4 polyposis in the first evaluation. Three patients showed complete regression of their polyps and an average change of 0.5 in score in each side was observed after treatment, which was significant (right, left and total: p = 0.020, p = 0.023 and p = 0.009, respectively).
## Pnif.
A preliminary analysis concluded that there was no significant difference between PNIF values recorded before and after decongestion. With that in mind, we opted to discuss and present baseline results before decongestion, except where noted otherwise . The change in PNIF after treatment is shown as ΔPNIF. A significant difference between NC patients and CRS or NP patients was evident and significant (p = 0.010). After treatment, sinusitis and polyp patients showed significantly different average PNIF recordings (p = 0.043). Although other comparisons could not reach statistical significance, the data suggest a clinically significant trend for ΔPNIF to be lower in NP patients (35 L/min in sinusitis vs 0 L/min in polyps). Furthermore, PNIF change after treatment was significant in CRS patients (p = 0.048), but not NP patients (p = 0.674).
Rhinomanometry. All the recordings from rhinomanometry produced a large volume of data that was submitted to an exploratory analysis. Resistance and flow parameters improved significantly after decongestion in the control group. However, no significant changes were found from the effect of decongestion or the respiratory cycle phase in sinusitis and polyp patients. Also, a multiple variable, mixed-effect model analysis reinforced this impression (unpublished data). To abbreviate and avoid redundancy, rhinomanometry measurements are presented in [fig_ref] Table 4: Rhinomanometry results-Resistances [/fig_ref] and concern the pre-decongested and inspiratory phase results. Adequate readings were impossible in one CRS patient post-treatment (6.7%) and some patients with large polyps (sizes 3 or 4): 3/19 pre-treatment (15.8%) and 2/17 post-treatment (11.8%).
No statistically significant differences were found between groups and in the same patient after treatment. www.nature.com/scientificreports/ show for their scheduled appointments and could not be summoned to return for evaluation. One patient was recruited only after treatment as the first evaluation was before the beginning of the project. Lund-Mackay (right) 0 (0-1) 5.5 (4-6) 7 (5-7.5) p = 0.000
Lund-Mackay (left) 0 (0-2) 5 (4-6) 7 (6-7) p = 0.000
Lund-Mackay (total) 0 (0-3) 10 (8-12) 14 (11.5-15) p = 0.000 . Modified Johansen scale polyp sizes. Results are shown as median (P 25 -P 75 ). Three patients showed complete polyp regression after treatment (score result = 0). All post-treatment changes were statistically significant (see text).
## Polyp size right left total
Pre-treatment (n = 19) . PNIF values (L/min). Results are shown as median [P . ΔPNIF values in the CRS group only concern 14 individuals, given losses during follow-up (see [fig_ref] Figure 1: Patient recruitment flowchart [/fig_ref]. PNIF shows a trend to be smaller and vary less in NP patients. See text for intergroup comparisons. NC negative controls, CRS chronic rhinosinusitis patients, NP nasal polyps patients, N/A not applicable, ΔPNIF change in PNIF after treatment. www.nature.com/scientificreports/ (Spearman = − 0.49, p = 0.046). All correlations that were found are weak and [fig_ref] Figure 2: Several graphs illustrate the effect of polyp size on nasal ventilation measures... [/fig_ref] illustrates that differences may have reduced clinical significance, especially after decongestion.
# Discussion
Overall, findings from our study suggest that PNIF and rhinomanometry can provide objective measures of nasal ventilation in almost all sinusitis and polyp (except patients with severe size 4 polyps). Published studies have presented PNIF and rhinomanometry values to compare polyp behaviour over time or after treatment [bib_ref] What happens with untreated nasal polyps over time? A 13-year prospective study, Oscarsson [/bib_ref] [bib_ref] Pre-and postoperative evaluation of patients with nasal obstruction undergoing endoscopic sinus surgery, Sipila [/bib_ref] [bib_ref] Nasal polyps. A clinical study of endoscopic score systems and epidemiology in..., Johansson [/bib_ref] [bib_ref] Impact of baseline nasal polyp size and previous surgery on efficacy of..., Djupesland [/bib_ref] [bib_ref] The evaluation of pattern and quality of sleep in patients with chronic..., Uz [/bib_ref].
In our study we have also tried to assess whether polyp size could be a confounding factor that influences those results. When comparing and correlating polyp size against the several measurements and moments, we could only find a few weak correlations with rhinomanometry results. This suggests that differences found with these www.nature.com/scientificreports/ methods may potentially reflect disease pathophysiology rather than the mechanical effect of polyp size. As such, PNIF and rhinomanometry should definitely be considered in future studies investigating NP pathophysiology. Although sample size may limit statistical power, we found significant differences and relevant effect sizes that are backed up by the apparently consistent and clinically relevant trend in favour of PNIF being able to discriminate patients without polyps from patients with polyps better than rhinomanometry. PNIF should thus be studied as a potential nasal ventilation biomarker of disease severity, particularly considering that it can be easily measured in any otolaryngology outpatient visit. Below follows a point by point discussion of our work. The distribution of patient characteristics in our sample is similar to that of others in the literature in the multiple aspects we discuss next. There is a slight male predominance in the NP group, although not the 2:1 reported elsewhere [bib_ref] Commentary on gender differences in prevalence, treatment, and quality of life of..., Ference [/bib_ref]. Studies in caucasian and asian populations agree that polyps tend to present in older individuals [bib_ref] Age-related increased prevalence of asthma and nasal polyps in chronic rhinosinusitis and..., Cho [/bib_ref] , peaking around 50-59 years of age 7 . It has also been reported that in patients with sinusitis, those with polyps tend to be older [bib_ref] Incidence and associated premorbid diagnoses of patients with chronic rhinosinusitis, Tan [/bib_ref] [bib_ref] Rinossinusite crónica: Correlação entre a clínica e o score lund-mackay, Breda [/bib_ref]. Our sample shows those same trends. The incidence (but not the gender proportion) of smokers (32.6%) follows that of the Portuguese population. In a manner similar to our patients, other studies have also reported a higher prevalence of smokers in sinusitis groups [bib_ref] Chronic rhinosinusitis without nasal polyps, Cho [/bib_ref] , as well as a lower proportion of NP patients who smoke, both in national and international cohorts [bib_ref] Rinossinusite crónica: Correlação entre a clínica e o score lund-mackay, Breda [/bib_ref] [bib_ref] The relationship between smoking and nasal polyposis, Erbek [/bib_ref] [bib_ref] Relative frequencies of symptoms and risk factors among patients with chronic rhinosinusitis..., Bohman [/bib_ref]. Although Lund-Mackay scores aren't used in clinical decision, they represent the extent of sinus inflammation (even though the score may be different from 0 in any given healthy individual). Increased scores have been associated with more symptoms and the presence of nasal polyps [bib_ref] Rinossinusite crónica: Correlação entre a clínica e o score lund-mackay, Breda [/bib_ref] [bib_ref] Lund-Mackay score is predictive of bleeding in ethmoidectomy for nasal polyposis, Mortuaire [/bib_ref] and the same was found in our co-hort. To assess disease severity, 3-or 4-point scales have often been used with good inter-observer consistency in the practice of rhinology [bib_ref] Reproducibility of the three-dimensional endoscopic staging system for nasal polyposis, Sousa [/bib_ref]. These scales are simple to apply and easy to compare in different studies. Although polyp size may vary over time and does not predict treatment success, patients often tend to present with small polyps regardless of when they present for evaluation [bib_ref] What happens with untreated nasal polyps over time? A 13-year prospective study, Oscarsson [/bib_ref] [bib_ref] Nasal polyps. A clinical study of endoscopic score systems and epidemiology in..., Johansson [/bib_ref] [bib_ref] Impact of baseline nasal polyp size and previous surgery on efficacy of..., Djupesland [/bib_ref]. Our patients also had predominantly small polyps.
With the exception of a score 4 polyp patient, all other individuals in our study could perform a reliable PNIF maneuver, despite the fact that around 7.34% unobstructed caucasians aren't able to do so [bib_ref] 4-Phase-rhinomanometry (4PR)-Basics and practice, Vogt [/bib_ref]. Rhinomanometry at 150 Pa was impossible only in patients with very large polyps (15.8% pre-and 11.8% post-treatment) and also one CRS patient (6.7%). These results demonstrate that these methods can effectively measure nasal ventilation in most patients with NP.
Considering each side of the nose independently in the analysis is important because the same patient may have different polyp sizes. This is possible only because they are functionally separated (no septal perforations). This way, a broader range of nasal airways in different nasal cycle states is also represented in readings before decongestion. Such an analysis effectively doubles sample size in unilateral comparisons (22 nasal cavities in the control group, 32 in sinusitis patients and 38 in polyp patients). We have also considered polyp patients according to global Johansen scores distributed into three categories: 0 represents no polyps, applicable only after successful medical treatment (3 patients); 2 represents small intrameatal polyps (8 individuals before treatment and 10 after treatment); and > 2 represents patients with larger polyps that impinge on the nasal airway corridors reducing their cross-sectional area (11 patients before treatment and 4 patients after treatment). Despite these adjustments, no functional variable correlated with polyp size unilaterally before or after treatment. No statistical differences were found with rhinomanometry (unilateral or global measurements), despite a trend for poorer results in NP patients. While sample size could be the cause, we need to consider the possibility that there is an inherently small magnitude of differences between patients. This may have also compromised unilateral correlations with polyp size. Global variables, however, did show some differences. PNIF in our study appears to be lower and change less after treatment in NP patients, in a magnitude similar to the reports by other authors [bib_ref] What happens with untreated nasal polyps over time? A 13-year prospective study, Oscarsson [/bib_ref] [bib_ref] Nasal polyps. A clinical study of endoscopic score systems and epidemiology in..., Johansson [/bib_ref] [bib_ref] Impact of baseline nasal polyp size and previous surgery on efficacy of..., Djupesland [/bib_ref]. This effect may be independent of polyp size as there were no consistent or relevant correlations with polyp scores. Other studies have reported a correlation of polyp size and PNIF 30 , but that finding was merely a measure of treatment effect. We have analyzed PNIF separately before and after decongestion. Overall differences before and after topical decongestion were not significant, and the comparisons between groups displayed the same significances and followed the same trends. We opted to present pre-decongestion results as they are easier to obtain and compare between studies and also because they represent the natural functional state of nasal function in the patients' everyday life. PNIF results negatively correlated with Lund-Mackay scores before treatment. Lower PNIF values have been associated with polyps, even when asymptomatic [bib_ref] Prevalence of nasal polyps in adults: The Skovde population-based study, Johansson [/bib_ref] and an increased probability of requiring treatment over time [bib_ref] What happens with untreated nasal polyps over time? A 13-year prospective study, Oscarsson [/bib_ref]. In our study, PNIF was similar between sinusitis and polyp patients before treatment. However, significant changes were noted after treatment in CRS patients, but not polyp patients. This happened despite the changes in polyp size mentioned above and in contrast to the absence of significant changes in manometry results. Rhinomanometry results are similar to other reports in the literature [bib_ref] Pre-and postoperative evaluation of patients with nasal obstruction undergoing endoscopic sinus surgery, Sipila [/bib_ref] [bib_ref] The evaluation of pattern and quality of sleep in patients with chronic..., Uz [/bib_ref] , which reinforces the idea that there are naturally small differences between groups. Interestingly, if we apply to this and other studies the logarithmic class transformation proposed by , all groups will be ranked in class 1 for nasal obstruction (less severe), clearly pointing to a difference in the pathophysiology of symptoms in patients with sinonasal inflammation. This may be related to mucosal edema and inflammation induced neuropathic mechanisms [bib_ref] Pathophysiology of nasal congestion, Naclerio [/bib_ref]. The same mechanism may also possibly explain the negative correlation between Lund-Mackay scores and PNIF that disappears after treatment. The correlation between Lund-Mackay and polyp size, however, is already recognized [bib_ref] The Lund-Mackay staging system for chronic rhinosinusitis: How is it used and..., Hopkins [/bib_ref]. Correlations between rhinomanometry and polyp size were only present after treatment and readily disappeared with decongestion. Decongestion has been shown to significantly alter polyp size [bib_ref] Do topical nasal decongestants affect polyps, Johansson [/bib_ref] , and ventilation recordings in patients with allergic rhinitis [bib_ref] Decongestion test in patients with allergic rhinitis: Functional evaluation of nasal airflow, Ciprandi [/bib_ref]. We didn't design our study to answer specifically why decongestion had such a small effect in manometric readings in our sinusitis and polyp patients. It's hard to find answers in the literature because there are little available studies that discuss such findings in similar populations or study designs. At this time, we can only hypothesize that some resistance to the decongestant effect may be caused by mucus, epithelial barrier or endothelial changes due to the effect of inflammation or tissue remodeling. www.nature.com/scientificreports/ Broms method allows calculation of nasal airway resistance at the point where the rhinomanometric curve intersects a circle with radius 200 (crossing both axes at 200 Pa and 200 cm 3 s −1 ) and the angle from a line that connects such point to the origin of the graph will be lower the higher the resistance in that nasal cavity. Broms angles were presented because they complement or may substitute nasal resistance values when they cannot be measured in a given patient. Broms angle was also the only dimension that we could measure when considering the different phases of respiration-inspiration and expiration. While inspiratory values behaved similarly to nasal resistance values, expiratory values never correlated with polyp size and there were no significant differences in multiple comparisons. Given that airflow patterns in the nasal cavity are different when comparing inspiration to expiration, we must consider the hypothesis that the difference found may represent a baseline pathophysiological difference. We propose that in future studies researchers make an effort to discriminate inspiratory from expiratory pathophysiology.
In conclusion, PNIF and rhinomanometry provide measurable results in NP and polyp size doesn't appear to significantly influence recordings, especially if pre-and post-decongestion readings are obtained. The results from this study are consistent with other reports in the literature. It appears differences between sinusitis patients with or without polyps may be naturally small and PNIF may be a more useful clinical biomarker of disease severity. Despite the limitations in our study, its results should encourage us and other groups to continue investigating the clinical usefulness of PNIF and rhinomanometry, as well as their contribution to the study of NP pathophysiology.
## Data availability
The datasets generated during and/or analysed during the current study are available on reasonable request from either the corresponding author or from the Research Center of the host institution (Centro Hospitalar do Oeste) via e-mail: [email protected].
Received: 23 September 2019; Accepted: 15 July 2020
[fig] Figure 1: Patient recruitment flowchart. NC negative controls, CRS chronic rhinosinusitis patients, NP nasal polyps patients, PNIF peak nasal inspiratory flow, N/A not applicable. Patients lost for follow-up failed to [/fig]
[fig] Figure 2: Several graphs illustrate the effect of polyp size on nasal ventilation measures in polyp patients after medical treatment. Broms angles were recorded in ° and nasal resistance as Pa cm −3 s. Box-plots are a graphical depiction of the dispersion measures calculated from the recordings obtained from our patients in that given moment and setting-Medians are represented by the thick bars and the lower and upper extremities of the error-bars in the box-plots are P 25 and P 75 , respectively. In some patients, the values recorded fall outside the calculated P 25 -P 75 interval and these are represented as outlier circles in (a,c,d,e). These outliers stress the point that some patients will show particularly good or bad manometry results despite the presence of small polyps (Global Johansen score of 2). Correlations were mentioned in the text, although no statistical significance is found when comparing different global Johansen score categories. Decongestion may further abate differences from polyp size. Sample sizes are 0 (n = 3), 2 (n = 10) and > 2 (n = 4). (a,c,e) report pre-decongestion recordings, while (b,d,f) report post-decongestion recordings.Scientific RepoRtS | (2020) 10:12674 | https://doi.org/10.1038/s41598-020-69693-6 [/fig]
[table] Table 1: Patient base characteristics. Sex is shown as proportion, age as average ± standard deviation, smokers as percentage of smoking patients and Lund-Mackay results as median (P 25 -P 75 ). NC negative controls, CRS chronic rhinosinusitis patients, NP nasal polyps patients, F/M female/male. Statistically significant differences were found in all variables, except sex. See text for inter/intragroup comparisons. [/table]
[table] Table 4: Rhinomanometry results-Resistances (Pa cm −3 s ) and Broms angles (°). Results are shown as median [P25 -P 75 ]. In some cells, n values are lower than the group size because adequate readings could not be obtained from the patient (large polyposis) or because patients were lost for follow-up (see text). NC negative controls, CRS chronic rhinosinusitis patients, NP nasal polyps patients, N/A not applicable, ΔPNIF change in PNIF after treatment. No statistical significant differences or trends were noted and hence p-values are omitted for the sake of clarity (see text). [/table]
|
Atypical presentation of complex regional pain syndrome: neuropathic itching - A case report -
KSPSComplex regional pain syndrome (CRPS) is a painful disorder, characterized by pain following an inciting injury and a variable convergence of signs and symptoms of sensory, vasomotor, sudomotor, trophic, and motor dysfunction[1]. CRPS appears to be caused by various components of inflammatory factors, autoimmune factors, neuronal plasticity, and autoimmune dysregulation. Inexplicable pain, limb edema, skin color changes, and elevated skin temperature can manifest after trauma and surgery due to an exaggerated inflammatory response[2]. Although neurogenic inflammation is a normal process after nerve injury, tissue injury in Background: In some patients with neuropathic pain (NP), such as complex regional pain syndrome (CRPS), itching rather than pain is the main symptom making diagnosis and treatment difficult.Case: We report a case of a 23-year-old male with a history of hypoxic brain damage who presented with pruritus of the left foot and ankle. His left foot was fractured, and he underwent surgery 6 months previously. After the operation and cast application, he developed uncontrolled pruritus, swelling, sweating, and flushing of the left foot skin with limping. On examination, he showed well-known features of CRPS without pain. He was diagnosed with an atypical CRPS with neuropathic itching (NI). With treatment modalities used for NP and CRPS, his pruritus subsided gradually, and the his ankle mobility improved.Conclusions: Unexplained itching can be the main symptom in some CRPS patients. Treatment according to NP can improve symptoms of NI in CRPS patients.
ly observed in both these conditions [bib_ref] Itch and neuropathic itch, Hachisuka [/bib_ref]. In NP, the hypersensitive sensory states are termed allodynia and hyperalgesia whereas the corresponding terms used in NI are alloknesis and hyperknesis. Over the past few years, studies have attempted to describe the pathophysiology of itch and its neuropathic origin. Therefore, in the light of this research, our understanding of pruritus suggests that the skin is not the only reason for the sensation of itch. Damage to the nervous system may interfere with the pruritic afferent pathways allowing painful and noxious sensations to be perceived as pruritic [bib_ref] Itch and neuropathic itch, Hachisuka [/bib_ref].
Although the concept of NI has been emerging, few studies have covered its pathophysiology. Unfortunately, common treatments for pruritus are often not effective for pruritus of such origins. Furthermore, NI affects a patient's quality of life to an extent similar to NP. Herein, we discuss the atypical presentation of NI in a patient with CRPS and the treatment modalities used to improve his symptoms.
## Case report
We obtained written informed consent from the patient after surgery to publish this report.
A 23-year-old male presented with uncontrolled, chronic pruritus of his left foot and ankle. When he first visited the clinic, he was limping and had reduced range of motion, swelling, and sweating and flushing of the skin over the affected area. Interestingly, he was mainly concerned with the severe itching sensation and did not complain of pain. Six months before his first visit to our clinic, he had undergone open reduction and internal fixation of his left foot due to recurrent left first metatarsal fracture. To start with, the initial injury was a left great toe fracture, prompting a cast application. From the time of injury, the patient only complained of itching, rather than pain, over the affected foot. Due to the accompanying itch, he had kicked his foot against a hard wall in frustration and developed a comminuted toe fracture. Therefore, an additional surgical correction was performed. A plaster cast was applied for 3 months after the surgical correction due to the complex nature of the recurrent metatarsal fracture. Immediately after removal of the plaster cast, he developed uncontrolled pruritus and discomfort, skin color change, and swelling over the surgical site, which lasted for more than 3 months.
This patient had a history of congenital heart disease, with a functional single ventricle with transposition of the great arteries and pulmonary atresia, and had undergone three open heart surgeries by 5 years of age. At 11 days after birth, he underwent central shunt operation as his first cardiac surgery. Two weeks postoperatively, he experienced a sudden cardiac arrest for unknown reasons, and was revived. However, this had resulted in the development of ischemic brain damage [fig_ref] KSPSFigure 1: Magnetic resonance imaging of the brain shows multiple chronic cerebral infarctions at... [/fig_ref]. The follow-up brain magnetic resonance imaging had shown multiple chronic cerebral infarctions at both the parieto-occipital lobes and left anterior watershed zone. It also showed thinning of the posterior portion of the corpus callosum and multifocal small lacunar infarctions in the left cerebellum. The ischemic stroke caused mental retardation and repeated seizures for years afterwards.
We and a consultant dermatologist did not find skin lesions could cause itching except for edema and change in skin color at the affected area. There was no change in the medication administered during the period when the patient started complaining of symptoms. In addition, itching caused by infection could be ruled out with the physical examination and normal laboratory findings. Physical examination revealed edema, change and asymmetry in skin color, sweating, and decreased range of motion at the affected foot and ankle. Moreover, he complained of severe itching when the affected foot and ankle were rubbed using a soft brush (alloknesis), and pinprick-evoked itch was also observed (hyperknesis).
Simple X-ray showed osteopenia at the affected foot and ankle. The temperature difference between both feet was up to 2.42°C in thermography. We found sympathetic postganglionic sudomotor dysfunction or sweat gland abnormality in the left distal leg and foot by quantitative sudomotor axon reflex test (QSART) test. In addition, a 3-phase bone scan test found diffusely increased perfusion and blood pool and delayed bone uptake in the left ankle and foot suggesting CRPS [bib_ref] Bone scintigraphy in patients with pain. Korean, Shin [/bib_ref] [fig_ref] Figure 2: The three-phase bone scan shows diffusely increased perfusion, blood pool, and delayed... [/fig_ref]. Unfortunately, the energy dispersive X-ray analysis test and electrodiagnostic study were not performed because the patient did not cooperate. Although the patient did not complain of pain at all, these findings with his the symptoms and signs over the affected foot and ankle partially met the diagnostic criteria for CRPS as per the International Association for the Study of Pain (Budapest criteria) [bib_ref] Proposed new diagnostic criteria for complex regional pain syndrome, Harden [/bib_ref] , and met the criteria for CRPS-NOS (not otherwise specified) [bib_ref] Validation of proposed diagnostic criteria (the "Budapest Criteria") for complex regional pain..., Harden [/bib_ref] with itching.
We initiated his treatment carefully, with sub-anesthetic doses of intravenous (IV) ketamine therapy (0.1-0.3 mg/kg) administered in the outpatient pain clinic, along with prescribed oral doses of pregabalin (75 mg, twice a day) and
## Posterior 1458k
## Post iv 3 hrs 38 min
www.anesth-pain-med.org naproxen (500 mg, twice a day) to control his symptoms. One month later, his gait had improved, and the edema and profuse sweating had resolved. Although he experienced a few side effects, such as urinary retention and generalized edema, he could relatively tolerate them. With the same treatment every month, the pruritus improved gradually and completely subsided by the sixth month, along with the skin discoloration. Subsequently, we discontinued the IV ketamine therapy and prescribed reduced doses of pregabalin only. Therefore, the patient was successfully treated with medications used for the treatment of NP.
# Discussion
There could be three explanations for the questionable symptoms observed in this case. First, delayed removal of the plaster cast may have triggered the itching and skin color change. It could also have caused muscle atrophy, deep vein thrombosis, joint stiffness due to prolonged immobilization, gait abnormalities, and calf muscle weakness [bib_ref] Cast and splint immobilization: complications, Halanski [/bib_ref]. This could explain some of his symptoms, but they do not usually last for more than 1 month; if they do, physicians are required to look for other causes for these symptoms. Another possibility of itching was the atypical presentation of CRPS. Fracture of an extremity is a common inciting event of CRPS type 1. According to a previous report, 7% of patients with a single fracture of the wrist, scaphoid, ankle, or metatarsal bone develop CRPS type 1 [bib_ref] Complex regional pain syndrome, current concepts and treatment options, Urits [/bib_ref]. With the obvious inciting events, the patient showed skin color changes, edema, and abnormal sudomotor activity, without the presence of other conditions that could account for the dysfunction. Except for the fact that the continuous pain and allodynia were substituted by continuous severe itching and alloknesis, results of the imaging test, QSART test, and 3-phase bone scan and physical evidence pointed to an atypical presentation of CRPS-NOS with itch. His successful treatment with medications used for NP supports this diagnosis. Lastly, due to his cerebrovascular disease, it is possible that the abnormal pruritus may have been caused by a lesion in the afferent sensory pathway, rather than by a cutaneous lesion or a peripheral stimulus. An unusual itch with a neurologic origin is called an NI and is defined as an itch caused by a pathology located at any point along the afferent pathway of the nervous system [bib_ref] Itch and neuropathic itch, Hachisuka [/bib_ref].
About 30% of patients with NI have a peripheral neuropathic cause, including post-herpetic itch, brachioradial pruritus, notalgia paresthetica, trigeminal trophic syndrome, and itch caused by burns or keloids [bib_ref] Clinical presentation, management, and pathophysiology of neuropathic itch, Steinhoff [/bib_ref]. It has been reported that 15% of itches arise from central nervous system disorders [bib_ref] The neurology of itch, Dhand [/bib_ref]. Spinal cord disorders including inflammatory transverse myelitis, neoplasms, cavernous hemangiomas, and post-traumatic Brown-Sequard syndrome have also been implicated for the same [bib_ref] Clinical presentation, management, and pathophysiology of neuropathic itch, Steinhoff [/bib_ref] [bib_ref] Disease mechanisms in neuropathic itch, Binder [/bib_ref]. Ischemic stroke of the subcortical area or brainstem is the most common central cause of NI [bib_ref] The neurology of itch, Dhand [/bib_ref]. As shown in [fig_ref] KSPSFigure 1: Magnetic resonance imaging of the brain shows multiple chronic cerebral infarctions at... [/fig_ref] , the patient was previously diagnosed with multiple hypoxic ischemic encephalopathy. Since NI can occur in patients with central nervous system disorders, we also cannot rule out the possibility of NI due to cerebrovascular disease.
The diagnostic criteria for NI has yet to be definitively established. Currently, its diagnosis is primarily based on clinical characteristics specific to NI syndromes, with history taking as an important process to exclude dermatologic or systemic causes, as well as detect neuropathic causes [bib_ref] Clinical presentation, management, and pathophysiology of neuropathic itch, Steinhoff [/bib_ref]. Sensory tests, electrodiagnostic tests, autonomic function tests, imaging studies, and skin biopsies may also help detect and localize potential neuropathic causes [bib_ref] Clinical presentation, management, and pathophysiology of neuropathic itch, Steinhoff [/bib_ref]. In the present case, due to the patient's non-cooperation, we could not perform an electrodiagnostic test and skin biopsy; however, results from the sensory test, 3-phase bone scan, and QSART test suggested that his itching was neuropathic in origin.
Similar to the circumstances of NI diagnosis, there is also no specific treatment for this condition [bib_ref] Clinical presentation, management, and pathophysiology of neuropathic itch, Steinhoff [/bib_ref]. Generally, management of NI begins with non-pharmacological measures used for itch, followed by other therapies in a stepwise approach, wherein the choice of treatment is usually based on the size and localization of the pruritus [bib_ref] Neuropathic itch: diagnosis and management, Stumpf [/bib_ref] [bib_ref] Expectations about the effectiveness of pain-and itch-relieving medication administered via different routes, Peerdeman [/bib_ref]. In particular, treatment of itching caused by neuropathy is often based on NP management [bib_ref] Clinical presentation, management, and pathophysiology of neuropathic itch, Steinhoff [/bib_ref]. Moreover, preventing and treating secondary scratch-induced skin lesions is also important to reduce further itching and infection. Since antihistamines, corticosteroids, and most pain medications were largely ineffective in treating NI, we decided to systematically administer inhibitors of neuronal excitability, and thus the patient was treated using medications for CRPS with NP. More specifically, monthly IV sub-anesthetic ketamine doses with daily oral pregabalin and naproxen were administered to the patient for 6 months. Naproxen was prescribed in this case to reduce the release of inflammatory mediators that may affect NP and/or NI. Although he experienced a few side effects, such as urinary retention and generalized edema, the pruritus was successfully treated.
In conclusion, when a patient complains of unexplained itching not resolved by general itching treatments, a probable neuropathy should also be considered among its various KSPS causes. Careful history taking and evaluation of symptoms, with some neurologic and radiologic tests to detect neuropathic causes, can help identify NI. Although there is no established treatment for NI, treatment according to NP can improve these symptoms.
## Conflicts of interest
No potential conflict of interest relevant to this article was reported.
# Data availability statement
All data generated or analyzed during this study are included in this published article.
[fig] KSPSFigure 1: Magnetic resonance imaging of the brain shows multiple chronic cerebral infarctions at both the parieto-occipital lobes and left anterior watershed zone and thinning of the posterior portion of the corpus callosum (A). Multifocal small lacunar infarctions are also seen in the left cerebellum (B). [/fig]
[fig] Figure 2: The three-phase bone scan shows diffusely increased perfusion, blood pool, and delayed bone uptake in the left foot and ankle.99m Tc-DPD: 99m Tc-3, 3-diphosphono-1, 2-propanodicarboxylic acid, IV: intravenous. PLANTAR RT (1 min) LT PLANTAR RT (delay) [/fig]
|
Emergency Management of Medical Wastewater in Hospitals Specializing in Infectious Diseases: A Case Study of Huoshenshan Hospital, Wuhan, China
# Introduction
The novel coronavirus disease (COVID-19) first appeared in Wuhan, China, in December 2020, before spreading rapidly worldwide [bib_ref] SARS-CoV-2 in hospital wastewater during outbreak of COVID-19: A review on detection,..., Achak [/bib_ref]. By the end of December, 2021, there were approximately 280 million cases of COVID-19 cases worldwide, with a mortality rate exceeding 2%. The social, economic, and environmental impacts of COVID-19 have to date far exceeded those of the severe acute respiratory syndrome (SARS) epidemic in 2003 [bib_ref] Greater China, and the Pathologies of Globalization and Transition, Delisle [/bib_ref]. Hospital wastewater contains hazardous elements and a wide variety of microbial pathogens and viruses [bib_ref] SARS-CoV-2 in hospital wastewater during outbreak of COVID-19: A review on detection,..., Achak [/bib_ref]. Therefore, this type of wastewater poses a significant risk to human and environment health . A recent study showed that the release of partially treated or raw medical wastewater into surface waterways poses a risk of the spread of disease through the "fecal-oral" transmission route [bib_ref] Impacts of COVID-19 pandemic on the wastewater pathway into surface water: A..., Bandala [/bib_ref]. Particularly within the context of the COVID-19 pandemic, appropriate management of medical wastewater is of great significance for minimizing the risks posed to human and environment health [bib_ref] SARS-CoV-2 in hospital wastewater during outbreak of COVID-19: A review on detection,..., Achak [/bib_ref] [bib_ref] Disinfection technology of hospital wastes and wastewater: Suggestions for disinfection strategy during..., Wang [/bib_ref].
President Xi in China responded to the outbreak of COVID-19 in Wuhan by stating: "If Wuhan beats the pandemic, then Hubei province will succeed, and if Hubei defeats (3) the legislative foundation for wastewater emergency treatment; and (4) information disclosure and public opinions. The results of the present study can provide a reference for the emergency regulation and management of hospital wastewater bearing pathogens.
## Methods and data resource
The Chinese government decided to rapidly construct two temporary field hospitals during the COVID-19 pandemic to ease the burden on the local hospitals. The Wuhan Huoshenshan Hospital was constructed in just nine days, three days less than the time required to construct the Leishenshan Hospital. The Huoshenshan Hospital was also the first field hospital globally to be constructed to care for COVID-19 patients. Considering the global spread of COVID-19 and the serious shortage of medical supplies, and the importance of medical wastewater management in many countries and region, the lessons and experience of Huoshenshan Hospital in China are expected to provide valuable information for other countries and regions to assist in combating the COVID-19 outbreak. Therefore, the present study chose Huoshenshan Hospital as a case study.
The present study combined qualitative and quantitative information on the management of medical wastewater by Huoshenshan Hospital emergency management measures of medical wastewater were systemically summarized. Information was obtained from three categories of data resources; (1) major media sources; (2) official information; and (3) published studies. Information relating to the construction of the Huoshenshan Hospital and the three-tier environmental management of medical wastewater were collected mainly from major media sources (e.g., Xinhua News, Beijing Youth Daily, China Environmental News, Global Times, etc.). Information was obtained from a total of 14 media sources, accounting for 29% of the total references. Information on the relevant regulations and policies for the emergency treatment of medical wastewater and environmental monitoring data were obtained from official report and statistical data , including the Wuhan Bureau of Ecological Environment, and websites of related authorities (e.g., the CCTV international online and the Ministry of Ecology and Environment). Recent published studies were analyzed to obtain information on the health and environmental risks of medical wastewater, other studies' management system, and further highlighted the necessity and effectiveness of emergency measure of medical wastewater for the case study.
As an important management lesson, such as public participation regarding the case study, its basic information was obtained via questionnaire survey. The questionnaire was structured to understand the extent of information disclosure and to facilitate public participation, There were eight questions in the questionnaire, which could be categorized into three parts: (1) a basic characterization of the respondent (i.e., ages, regions, education level and occupation, Q1 to Q4; (2) the behavior of the respondents in relation to public participation, Q5; (3) the means of public participation employed by the respondent and its context within the case study, Q6 to Q8. Online questionnaires were randomly distributed in May 2020 via a WeChat group and a Q chat group among friends, students, and colleagues. In total, data information was analyzed from 212 individuals of different ages, regions, educational levels, and occupations using Microsoft Excel (Microsoft Corporation. Redmond, WA, USA).
# Results and discussion
The present study systemically summarized the process of construction of the Huoshenshan Hospital and its successful environmental management of medical wastewater based on the three categories of information (major media sources, official information, and published studies). Sections 3.1-3.5 outline the lessons learnt.
## Construction progress of the huoshenshan hospital
The design of the hospital was initiated just three days after Wuhan went into lockdown. Construction of the hospital, as well as the laying down of high-density polyethylene (HDPE) anti-seepage film for wastewater, was completed by noon of the 29 January 2020.
All medical supporting facilities were installed by the 1 February 2020, and the hospital was officially handed over to China's Army Joint Service on the following day. [fig_ref] Figure 1: Timeline of progress in the construction progress of the Huoshenshan Hospital, Wuhan,... [/fig_ref] shows a timeline of progress of the construction of the Huoshenshan Hospital. The hospital was situated in a sanatorium area near Zhiyin Lake in southwestern Wuhan, covered an area of~34,000 square meters, and had a capacity of 1000 beds. The hospital utilized modern information technology including a 5th Generation (5G) Mobile Network, Artificial Intelligence (AI), and Internet of Things (IoT) to achieve intelligent security, remote medical consultation, intelligent reviews of medical records, and "contactless" operation and maintenance. For example, 5G coverage and wired broadband were utilized to connect to a remote consultation platform that allowed experienced doctors in other provinces around China to conduct remote consultations with patients in Wuhan.
based on the three categories of information (major media sources, official information, and published studies). Sections 3.1-3.5 outline the lessons learnt.
## Construction progress of the huoshenshan hospital
The design of the hospital was initiated just three days after Wuhan went into lockdown. Construction of the hospital, as well as the laying down of high-density polyethylene (HDPE) anti-seepage film for wastewater, was completed by noon of the 29 January 2020. All medical supporting facilities were installed by the 1 February 2020, and the hospital was officially handed over to China's Army Joint Service on the following day. [fig_ref] Figure 1: Timeline of progress in the construction progress of the Huoshenshan Hospital, Wuhan,... [/fig_ref] shows a timeline of progress of the construction of the Huoshenshan Hospital. The hospital was situated in a sanatorium area near Zhiyin Lake in southwestern Wuhan, covered an area of ~34,000 square meters, and had a capacity of 1000 beds. The hospital utilized modern information technology including a 5th Generation (5G) Mobile Network, Artificial Intelligence (AI), and Internet of Things (IoT) to achieve intelligent security, remote medical consultation, intelligent reviews of medical records, and "contactless" operation and maintenance. For example, 5G coverage and wired broadband were utilized to connect to a remote consultation platform that allowed experienced doctors in other provinces around China to conduct remote consultations with patients in Wuhan.
## Development of relevant regulations and policies
As shown in [fig_ref] Table 1: A summary of regulations relating to management of medical wastewater in China [/fig_ref] , the design of wastewater treatment facilities at the Huoshenshan Hospital during the pandemic was in strict accordance with existing specifications (HJ 2029-2013) and standards . In particular, the design of the wastewater treatment facilities was in adherence to the Standards for the Discharge of Water Pollutants from Medical Organizations. These standards regulate medical wastewater discharge to limit the release of toxins and pathogens into the environment, such as fecal coliform bacteria, total residual chlorine, etc.. The legislative branch in China responded quickly to the COVID-19 pandemic by improving the legal structure of the public emergency services. In particular, the regulations relating to wastewater discharged from hospitals were updated to consider the risk posed by infectious diseases. Prior to the COVID-19 pandemic, medical wastewater in China was discharged directly into the municipal sewer system, with medical wastewater allowed to mix with municipal wastewater, and this mixture of wastewater treated at a municipal wastewater treatment plant, following a regulation and management system similar to those practiced in some developed countries, such as Australia and Spain [bib_ref] What have we learned from worldwide experiences on the management and treatment..., Verlicchi [/bib_ref]. However, the COVID-19 pandemic motivated the Ministry of Ecology and Environment to apply the Internet/AI technology (mentioned in Section 3.1) for the construction of an online communication platform. This platform allowed remote communication among relevant experts, front-line operators, and ecological and environmental management personnel. This improved remote communication and allowed experts groups to provide online technical guidance for the management of medical wastewater of the online communication platform based on Internet/AI technology, including its disinfection and disposal and equipment operation. Meanwhile, the online platform facilitated the development and prompt release of the Notice, the Technical Program, and the Monitoring Program upon the completion of the Huoshenshan Hospital based on the experience of medical wastewater regulation and management obtained since the SARS outbreak. These legal documents proposed the legislative requirements for the regulation and management of medical wastewater at the source. These management aspects included the classification, collection, storage, and treatment of wastewater, as well as guidance for emergency monitoring of variables, such as residual chlorine and biological toxicity. The documents also considered measures for the protection from and control of pathogen-bearing wastewater. The requirements specified the characteristics of COVID-19 indicators for use in monitoring and disinfection, as well as sterilization treatment criteria, to achieve the effective management of the medical wastewater produced by the Huoshenshan Hospital within the shortest time frame. [fig_ref] Figure 2: A conceptual diagram illustrating the three-tier system for the management of medical... [/fig_ref] shows the three-tier system for the management of medical wastewater. The system facilitates the systematic treatment of wastewater, as well as the management of routes of disease transmission during outbreaks of infectious diseases. The system stipulates strict separation of medical wastewater and collection at the source, specialized treatment at different levels, and discharge to specified standards. Identify the responsible monitoring parties, and monitoring data and strengthen monitoring, reporting, and reviewing [fig_ref] Figure 2: A conceptual diagram illustrating the three-tier system for the management of medical... [/fig_ref] shows the three-tier system for the management of medical wastewater. The system facilitates the systematic treatment of wastewater, as well as the management of routes of disease transmission during outbreaks of infectious diseases. The system stipulates strict separation of medical wastewater and collection at the source, specialized treatment at different levels, and discharge to specified standards.
## Three-tier prevention and control system
## Three-tier prevention and control system
## Tier-one prevention and source control (source collection)
Medical wastewater carrying SARS-CoV-2 increases the risks of continuous disease transmission and spread of the viruses [bib_ref] COVID-19: Mitigating transmission via wastewater plumbing systems, Gormley [/bib_ref]. Therefore, higher standards need to be adopted for the treatment of wastewater generated by hospitals treating infectious diseases than those for ordinary hospitals. The tier-one prevention and control system should act to prevent and control the source. Wastewater generated at the Huoshenshan Hospital was physically isolated and collected in a fully enclosed manner to prevent exposure of wastewater to the ambient air to the soil [bib_ref] Discussion on the safety design of Huoshenshan hospital drainage system, Li [/bib_ref]. A layer of high-density polyethylene (HDPE) film was installed into the foundation of the hospital to prevent seeping and to ensure physical isolation of the aboveground structures from groundwater and soil. Domestic sewage and medical wastewater from the infectious disease ward were collected separately, and medical wastewater was delivered separately, in sealed storage tanks.
## Tier-two protection and control (disposal system)
The wastewater in the sealed storage tanks at the hospital was then transported from the hospital through a special pipeline into the medical wastewater treatment system. This treatment system was a tier-two prevention and control wastewater disposal system. The successful design and construction of this system were viewed as a case study in emergency management of medical wastewater. Two sets of equipment required for the treatment of wastewater during this period were installed, with one set maintained as standby to increase the resilience of the system. A single device was able to treat 800-1000 tons of wastewater per day. As shown in [fig_ref] Figure 3: Treatment process of medical wastewater for Huoshenshan Hospital [/fig_ref] , the tier-two protection and control system incorporated multiple processes, including pre-disinfection, a septic tank, a conditioning tank, biological and chemical treatment (degradation of chemical oxygen demand (COD) and ammonia nitrogen (ANH 3 -N)), sedimentation, and (secondary) disinfection. The experience gained in medical wastewater treatment during the SARS pandemic, in combination with reference "HJ 2029-2013", allowed the effective treatment of fecal coliforms and SARS-CoV by disinfection, and the Hospital implemented a safe, effective, and cost-effective two-level intensified disinfection system [bib_ref] Disinfection technology of hospital wastes and wastewater: Suggestions for disinfection strategy during..., Wang [/bib_ref]. A pre-disinfectionseptic tank-disinfection tank (secondary disinfection), lasting for 5 h, far exceeds the national standard of 1.5 h. This standard has been incorporated into the nationally implemented Design Guidelines for Emergency Treatment Facilities for COVID-meeting standard (GB18466-2005), the sludge was sealed and transported to a hazardous waste disposal site and disposed through incineration as hazardous waste [bib_ref] Discussion on the safety design of Huoshenshan hospital drainage system, Li [/bib_ref]. The wastewater treatment facilities at the Huoshenshan Hospital were equipped with online monitoring stations for dynamic tracking of the management and control of wastewater quality. Eco-environment division personnel conducted daily on-site inspection of pollution control facilities, during which they assessed the discharge and monitor- Anaerobic degradation in the septic tank generally went through acidogenesis and methanogenesis, during which organic wastewater and sewage with high concentrations of COD and ANH 3 -N were treated. This process produced a low quantity of surplus sludge, which was easily concentrated and dewatered, during which the sludge was also disinfected. The automatic mechanical screen removed debris to avoid being negatively affected, such as tissue paper products. The activated sludge method is a commonly applied biological treatment process, which allows the removal of organic pollutants in colloidal and dissolved forms in wastewater. Degraded organic matter and non-degraded pollutants existed in the sludge and were separated from the water through solid-liquid separation. The wastewater was disinfected to reach the specified pretreatment standard. [fig_ref] Table 2: Medical wastewater associated with standard, influent, and effluent quality [/fig_ref] shows the design effluent quality of medical wastewater. The moving bed biofilm reactor (MBBR) process utilizes the traditional fluidized bed and contact oxidation process to effectively eliminate organic pollutants. The operation of the process was also simple, with resistance to shock loads and low costs [bib_ref] Ozonation for source treatment of pharmaceuticals in hospital wastewater-Ozone lifetime and required..., Hansen [/bib_ref]. Therefore, MBBR was selected as the biological and chemical treatment technology used in the hospital. The sludge generated from the wastewater treatment was concentrated and dewatered, following which it was transported to a centralized location for disposal. The odor (e.g., ammonia, chlorine, hydrogen sulfide gas) from the wastewater treatment station was collected, disinfected, and deodorize to finally meet standard GB18466-2005, following which it was discharged [bib_ref] Sewage Treatment Project Design of Wuhan Huoshenshan and Leishenshan Hospitals, Peng [/bib_ref]. The sludge storage tank was disinfected with lime and bleach, and the sludge was removed using sealed centrifugal sludgers and mobile filter trucks. After meeting standard (GB18466-2005), the sludge was sealed and transported to a hazardous waste disposal site and disposed through incineration as hazardous waste [bib_ref] Discussion on the safety design of Huoshenshan hospital drainage system, Li [/bib_ref].
The wastewater treatment facilities at the Huoshenshan Hospital were equipped with online monitoring stations for dynamic tracking of the management and control of wastewater quality. Eco-environment division personnel conducted daily on-site inspection of pollution control facilities, during which they assessed the discharge and monitoring data [bib_ref] Discussion on the safety design of Huoshenshan hospital drainage system, Li [/bib_ref]. When discharge did not meet the specified standards, a sequential investigation was conducted to identify and solve the problems. The strict monitoring and control measures implemented achieved a certain degree of success. The results of monitoring of hospital wastewater quality showed that the levels of the characteristic indicators, such as residual chlorine and fecal coliform, were generally in conformance with the limits set in the Standards for the Discharge of Water Pollutants from Medical Organizations (GB 18466-2005) [fig_ref] Table 2: Medical wastewater associated with standard, influent, and effluent quality [/fig_ref] from 1-21 February 2020. During the hospital operation stage from 2 February to the close of the hospital, the levels of characteristic indicators of the effluent at the outlet of a contact disinfection tank were maintained within the standard range, e.g., a specified standard (GB 18466-2005) for the total residual chlorine [fig_ref] Table 2: Medical wastewater associated with standard, influent, and effluent quality [/fig_ref].
## Tier-three protection and control (discharge and further disposal of wastewater)
The treated wastewater was tested by the online monitoring station to ensure that the water quality of wastewater met the standards (GB18466-2005) and technical specifications (HJ 2019-2013) shown in [fig_ref] Table 2: Medical wastewater associated with standard, influent, and effluent quality [/fig_ref]. This treated medical wastewater was then discharged into the municipal sewer system, finally reaching the municipal wastewater treatment plant (Shiyang Wastewater Treatment Plant). This is the tier-three prevention and control measure. The municipal wastewater treatment plants further treated the wastewater by means of the activated sludge method to ensure that the produced treated effluent reached the standard for pollutants discharged by urban wastewater treatment plants .
[formula] ≤100 MPN.L −1 ≤3.0 × 10 8 MPN.L −1 ≤5000 MPN.L −1 ≤100 2 Intestinal bacteria ND - - - 3 [/formula]
Tubercule bacilli ND --4 pH value 6-9 6-9 6-9 7.9 ± 0.7
[formula] 5 COD ≤60 mg L −1 ≤350 mg L −1 ≤250 mg L −1 40.66 ± 5.23 6 BOD ≤20 mg L −1 ≤150 mg L −1 ≤100 mg L −1 - 7 SS ≤20 mg L −1 ≤120 mg L −1 ≤60 mg L −1 - 8 ANH 3 -N ≤15 mg L −1 ≤30 mg L −1 - 1.53 ± 0.38 9 [/formula]
The total residual chlorine 6.5~10 mg L −1 --6.5~10 mg L −1
## 10
Animal and vegetable oil ≤5 mg L −1 ≤50 mg L −1 ≤20 mg L −1 -11 SARS-CoV-2 ND -ND -Note: data, reference [bib_ref] Discussion on the safety design of Huoshenshan hospital drainage system, Li [/bib_ref] ; BOD-biological oxygen demand; COD-chemical oxygen demand; SS-suspended solids; sign -, vailable; ND, cannot be detected.
## Measures for the management of environmental monitoring
The Ministry of Ecology and Environment focused on the environmental monitoring of designated hospitals in Wuhan during the pandemic and swiftly formulated a Monitoring Program on 30 January 2021 [fig_ref] Table 1: A summary of regulations relating to management of medical wastewater in China [/fig_ref]. This program specified the emergency monitoring of characteristic pollutants for pandemic prevention and control, including for residual chlorine and biological toxicity.
The Wuhan Environmental Monitoring Center used a field investigation to rapidly draft the Emergency Monitoring Program for COVID-19 Outbreak in Wuhan (hereinafter referred to as the "Wuhan Monitoring Program") [bib_ref] Design and Operation Management of Wastewater Treatment Station of COVID-19 Specialized Hospital..., Zeng [/bib_ref]. The center also mobilized relevant enterprises and monitoring units to provide real-time dynamic monitoring of the medical wastewater produced by 63 designated hospitals in the city, including the Huoshenshan Hospital, according to the program. Monitoring of the medical wastewater of the designated hospitals (including Huoshenshan Hospital) conformed to the requirements of the Wuhan Monitoring Program. The monitoring program followed the three-tier accountability system of data monitoring, recording, and storage. Additionally, the Environmental Emergency Monitoring Information Briefing conducted the pandemic in Wuhan was prepared and submitted to the higher-level monitoring authority for verification and record keeping. Environmental monitoring of medical wastewater quality during the epidemic also promoted the development of ecological and environmental big data platforms. A three-tier accountability system for monitoring the data collected from wastewater units was developed. General environmental monitoring stations were established, and the legislative, scientific, and ethical foundations of the environmental monitoring system were constructed.
## Importance of environmental governance
## Information disclosure and public participation
With the development of internet technology, Chinese netizens have been playing an increasingly substantial role in public opinion supervision and management of environmental protection [bib_ref] The impacts of governmental performance assessment policy and citizen participation on improving..., Wu [/bib_ref]. The 46th Statistical Report on the Development of Internet in China by the China Internet Network Information Center (CNNIC)found that as of June 2020. Internet users in China has reached 940 million. This represents an Internet penetration rate of 67.0%. Mobile users accounted for 99.2% of the total Internet users, thereby providing a solid foundation for receiving and publishing information. The Ministry of Environmental Protection promulgated the Measures for Public Participation in Environmental Protection [fig_ref] Table 3: Laws, measures, and regulations regarding the disclosure of environmental safety information to... [/fig_ref] designated hospitals (including Huoshenshan Hospital) conformed to the requirements of the Wuhan Monitoring Program. The monitoring program followed the three-tier accountability system of data monitoring, recording, and storage. Additionally, the Environmental Emergency Monitoring Information Briefing conducted the pandemic in Wuhan was prepared and submitted to the higher-level monitoring authority for verification and record keeping. Environmental monitoring of medical wastewater quality during the epidemic also promoted the development of ecological and environmental big data platforms. A three-tier accountability system for monitoring the data collected from wastewater units was developed. General environmental monitoring stations were established, and the legislative, scientific, and ethical foundations of the environmental monitoring system were constructed. . A three-tier accountability system for monitoring hospital wastewater treatment units. . A three-tier accountability system for monitoring hospital wastewater treatment units. To protect the rights and interests of citizens, legal persons, and other organizations to obtain information of environmental pollution sources using legal means and to issue guidance on public participation in environmental protection To protect the rights of citizens, legal persons, and other organizations in obtaining environmental information, as well as participating in and supervising environmental protection issues; develop avenues for public participation and promote the legal and orderly development of public participation in environmental protection Information disclosure by government agencies is an important means of providing information to the general public. The dissemination of emergency information (information disclosure) online has become an active means of expressing public opinion, further boosting public participation [bib_ref] Temporal and spatial evolution of online public sentiment on emergencies, Li [/bib_ref]. During the COVID-19 pandemic, the Ministry of Ecology and Environment disseminated environmental information relating to the emergency to the public in accordance with the laws [bib_ref] The Scientific and Technolo Department of Environ Characterization and Treatment of Medical..., Basturka [/bib_ref]. This information included details of the medical wastewater generated by the Huoshenshan Hospital and provided accurate and timely information to the public and social organizations by various forms of media and through the official website [bib_ref] Design and Operation Management of Wastewater Treatment Station of COVID-19 Specialized Hospital..., Zeng [/bib_ref] [bib_ref] Headlines: Quiet and busy-Lunar New Year's Eve in Wuhan, Ma [/bib_ref]. This disclosure of information ensured that information on the process of the construction of the medical wastewater treatment facilities and medical wastewater treatment data of the Huoshenshan Hospital were open and transparent.
## Survey on public participation within the development and operation of the huoshenshan hospital
The release of information on the COVID-19 pandemic has been timely, accurate, open, and transparent. Netizens received not only the open and transparent release of pandemic information but were also active supporters and protectors of this policy. Information on the construction of the Huoshenshan Hospital was released online by government and state media [bib_ref] Headlines: Quiet and busy-Lunar New Year's Eve in Wuhan, Ma [/bib_ref] , and public participation became an important driver for the orderly progression of the construction of the hospital and its associated wastewater treatment management measures. The present study conducted a questionnaire survey to understand the breadth and depth of information disclosure on the hospital and its medical wastewater management that was disclosed, as well as the extent of public participation.
As shown in [fig_ref] Figure 5: Characterization of survey respondents [/fig_ref] , most respondents were from Hubei Province, the nearest province to the case study. The respondents were representative of a wide range of ages, educational backgrounds, and occupations, although the majority of respondents were students between 18-30 years with a bachelor degree. the widely disseminated information on the management of medical wastewater produced by the case hospital [fig_ref] Figure 6: Survey on public interest in information released for the construction and operation... [/fig_ref]. This openness of information ensured enthusiasm and a degree of participation by the public within the construction of the hospital and the management of its wastewater to an extent. Concurrently, the requirement for broad public participation in the Measures for Environmental Information Disclosure (Trial) has been fulfilled. The survey asked the respondents to specify the primary way they used to remain informed and to express their opinion on the construction of wastewater facilities and the management of wastewater at the Huoshenshan Hospital. Most respondents (67%) indicated that they participated through following the news and comments on CCTV reports [fig_ref] Figure 7: Survey on how the general public exercised their rights in public opinion... [/fig_ref]. The live broadcast of CCTV attracted 50 million "cloud supervisors" during the peak viewing times. Overseas netizens were also able to simultaneously watched the broadcast through YouTube and other foreign video websites [bib_ref] Beijing: CCTV International Online, Zhi [/bib_ref]. The survey results revealed the enthusiasm, positivity, and high levels of awareness of the public regarding construction and operation of the hospital. The survey results also showed that connectivity of the current society allowed the general public to receive transparent information through a variety of avenues. This allowed public opinion to supervise the construction and operation of the hospital through a variety of real-time, dynamic, and prompt approaches, including publishing comments on the official websites of relevant agencies, Weibo, and CCTV news websites/live broadcast platforms, contacting government officials through mails and hotlines, and commenting at WeChat official accounts. The findings showed that most of the population (88%) had paid close attention to the development of the Huoshenshan Hospital [fig_ref] Figure 6: Survey on public interest in information released for the construction and operation... [/fig_ref]. The respondents indicated that they had obtained information on the hospital through multiple channels, including TV news, social media, such as Weibo and WeChat, and Apps providing short videos to mobile users. This informed the audience of information disclosure. Most of the public was able to actively utilize the well-developed Internet and a variety of web tools to receive the widely disseminated information on the management of medical wastewater produced by the case hospital [fig_ref] Figure 6: Survey on public interest in information released for the construction and operation... [/fig_ref]. This openness of information ensured enthusiasm and a degree of participation by the public within the construction of the hospital and the management of its wastewater to an extent. Concurrently, the requirement for broad public participation in the Measures for Environmental Information Disclosure (Trial) has been fulfilled. The survey asked the respondents to specify the primary way they used to remain informed and to express their opinion on the construction of wastewater facilities and the management of wastewater at the Huoshenshan Hospital. Most respondents (67%) indicated that they participated through following the news and comments on CCTV reports [fig_ref] Figure 7: Survey on how the general public exercised their rights in public opinion... [/fig_ref]. The live broadcast of CCTV attracted 50 million "cloud supervisors" during the peak viewing times. Overseas netizens were also able to simultaneously watched The survey asked the respondents to specify the primary way they used to remain informed and to express their opinion on the construction of wastewater facilities and the management of wastewater at the Huoshenshan Hospital. Most respondents (67%) indi-cated that they participated through following the news and comments on CCTV reports [fig_ref] Figure 7: Survey on how the general public exercised their rights in public opinion... [/fig_ref]. The live broadcast of CCTV attracted 50 million "cloud supervisors" during the peak viewing times. Overseas netizens were also able to simultaneously watched the broadcast through YouTube and other foreign video websites [bib_ref] Beijing: CCTV International Online, Zhi [/bib_ref]. The survey results revealed the enthusiasm, positivity, and high levels of awareness of the public regarding construction and operation of the hospital. The survey results also showed that connectivity of the current society allowed the general public to receive transparent information through a variety of avenues. This allowed public opinion to supervise the construction and operation of the hospital through a variety of real-time, dynamic, and prompt approaches, including publishing comments on the official websites of relevant agencies, Weibo, and CCTV news websites/live broadcast platforms, contacting government officials through mails and hotlines, and commenting at WeChat official accounts.
## Comparison with developed countries
Disclosure of environmental information Disclosure represents a new tool for environmental governance in the era of big data and information [bib_ref] The developing trends and driving factors of environmental information disclosure in China, Li [/bib_ref]. Developed countries, such as the European Union (EU) member states, the United States (U.S.), and Japan, have actively promoted public participation and have established related legal systems to guarantee and emphasize the importance of public participation and social supervision [bib_ref] Current Situation, Characteristics and Promotion of Public Participation and Social Supervision on..., Liu [/bib_ref]. For example, the U.S. passed the National Environmental Policy Act as early as 1969, which represented the first established the status and rights of the public to participate in environmental impact assessment. The Basic Environment Law of Japan enacted in 1993 specified the avenues, procedures, and means of public participation. The Environmental Impact Assessment Law of 2003 of China defined public participation in environmental impact assessment of construction projects. Some recent studies have shown the positive effect of public participation in and supervision of environmental management for the mitigation of environmental degradation [bib_ref] Public participation modes in China's environmental impact assessment process: An analytical framework..., Yao [/bib_ref]. Although public legislation for public participation of China has been gradually improving, there remains a demand for legislative and executive guarantees for public participation and environmental supervision for specific industries. To some extent, these gaps in legislation have impacted the extent of public participation. For example, there remains no legislation related to public participation in the management of medical wastewater. Therefore, the case study on the construction of the Huoshenshan Hospital and the management of its wastewater can act as a reference for facilitating the participation of netizens in supervision of environmental management though public opinion through a variety of network platforms and information tools. The results of the present study showed a rapid progression in the process compared to that during the SARS epidemic period. Achieving the positive effect of public opinion supervision of environmental management through public opinion has two requirements. First, the disclosure and transparency of information needs to be improved. Secondly, public awareness should be enhanced and enthusiasm and motivation of netizens and non-governmental organizations (NGOs) should be promoted to ultimately mobilize the public to take part in environmental management.
## Limitation and further study
The present study had certain limitations, particularly regarding the survey on the
## Comparison with developed countries
Disclosure of environmental information Disclosure represents a new tool for environmental governance in the era of big data and information [bib_ref] The developing trends and driving factors of environmental information disclosure in China, Li [/bib_ref]. Developed countries, such as the European Union (EU) member states, the United States (U.S.), and Japan, have actively promoted public participation and have established related legal systems to guarantee and emphasize the importance of public participation and social supervision [bib_ref] Current Situation, Characteristics and Promotion of Public Participation and Social Supervision on..., Liu [/bib_ref]. For example, the U.S. passed the National Environmental Policy Act as early as 1969, which represented the first established the status and rights of the public to participate in environmental impact assessment. The Basic Environment Law of Japan enacted in 1993 specified the avenues, procedures, and means of public participation. The Environmental Impact Assessment Law of 2003 of China defined public participation in environmental impact assessment of construction projects. Some recent studies have shown the positive effect of public participation in and supervision of environmental management for the mitigation of environmental degradation [bib_ref] Public participation modes in China's environmental impact assessment process: An analytical framework..., Yao [/bib_ref]. Although public legislation for public participation of China has been gradually improving, there remains a demand for legislative and executive guarantees for public participation and environmental supervision for specific industries. To some extent, these gaps in legislation have impacted the extent of public participation. For example, there remains no legislation related to public participation in the management of medical wastewater. Therefore, the case study on the construction of the Huoshenshan Hospital and the management of its wastewater can act as a reference for facilitating the participation of netizens in supervision of environmental management though public opinion through a variety of network platforms and information tools. The results of the present study showed a rapid progression in the process compared to that during the SARS epidemic period. Achieving the positive effect of public opinion supervision of environmental management through public opinion has two requirements. First, the disclosure and transparency of information needs to be improved. Secondly, public awareness should be enhanced and enthusiasm and motivation of netizens and non-governmental organizations (NGOs) should be promoted to ultimately mobilize the public to take part in environmental management.
## Limitation and further study
The present study had certain limitations, particularly regarding the survey on the public participation based on the questionnaire method. First, only 14% of respondents were not students. This possibly has impacted the representativeness of the designed respondents to some extent. Second, the questions in the survey mainly focused on the information disclosure and the means of public participation, thereby possibly limiting the accuracy of information on public participation. More accurate information on public participation can be obtained in a future study by increasing the range of occupations among respondents, adding problems designed in the questionnaire, such as public opinion on the quality of information received (timeliness, detail, ease of access, understandability, comprehension, etc.) and the levels of public confidence regarding any actions taken in response to their comments/suggestions.
# Conclusions
The present study aimed to provide a roadmap for the emergency management of medical wastewater in the future. Effective management measures for the medical wastewater originating from hospitals treating infectious diseases, using the Huoshenshan Hospital as a case study, were systematically summarized based on three types of data resources, namely media sources, official information, and published articles. The main findings were as follows:
(1) The successful and rapid construction of a modern emergency hospital with modern information technology, including 5G, AI, and IoT, occurred concurrently with the construction of effective wastewater management equipment, including an underground HDPE film layer for the isolation of source's wastewater. (2) The three-tier protection and control system was implemented for the emergency treatment of wastewater. The system was implemented by classification and collection of the medical wastewater from the ward, implementation of hospital wastewater treatment facilities, and municipal sewer system. Tier one focused on the prevention of pollution prevention at the source; tier two concentrated on the disinfection of the wastewater and the biological and chemical treatment of organic pollutants, to ensure that discharged wastewater met the water quality standards. (3) The management of monitoring aimed to ensure effective operation of monitoring equipment, standardization of monitoring process, and the provision of timely, reliable, verifiable, and recordable monitoring data. The three-tier accountability system within the data monitoring process guaranteed accuracy of monitoring data and also promoted the construction and development of big data platforms. (4) Medical Wastewater prevention and control technologies and measures were effectively implemented through the emergency legislative guarantee for the collection, monitoring, treatment, and discharge of the medical wastewater, as well as the modern information techniques. (5) The results of the questionnaire survey confirmed that the most of the population follow updates on the Huoshenshan Hospital through online tools, such as TV news, Weibo, WeChat, and video-sharing apps for mobile users. The public participated in online reviews and surveys of medical wastewater treatment facilities and management issues. Disclosure of environmental information should be strengthened, and public participation in industry-specific legislation should be improved, to promote the supervision of the management and control of medical wastewater through public opinion.
# Data availability statement:
The data presented in this study mainly derived from major media sources, office information and published studies shown in reference lists, which are openly available. Additionally, the data supporting public participation regarding the case study was obtained via questionnaire survey conducted by authors team, all relative results have been shown in Section 3.5.3.
[fig] Figure 1: Timeline of progress in the construction progress of the Huoshenshan Hospital, Wuhan, China, between January and February, 2020. Images were taken from the official website of the CITIC General Institute of Architectural Design & Research Co., with permission. [/fig]
[fig] Figure 2: A conceptual diagram illustrating the three-tier system for the management of medical wastewater.3.3.1. Tier-One Prevention and Source Control (Source Collection) [/fig]
[fig] Figure 3: Treatment process of medical wastewater for Huoshenshan Hospital. [/fig]
[fig] 17, Figure 5: Int. J. Environ. Res. Public Health 2022, 19, x 12 of Characterization of survey respondents. [/fig]
[fig] Figure 6: Survey on public interest in information released for the construction and operation of Huoshenshan Hospital. [/fig]
[fig] Figure 5: Characterization of survey respondents. [/fig]
[fig] 17, Figure 7: Int. J. Environ. Res. Public Health 2022, 19, x 13 of Survey on how the general public exercised their rights in public opinion supervision. [/fig]
[fig] Figure 7: Survey on how the general public exercised their rights in public opinion supervision. [/fig]
[table] Table 1: A summary of regulations relating to management of medical wastewater in China. [/table]
[table] Table 2: Medical wastewater associated with standard, influent, and effluent quality. [/table]
[table] Table 3: Laws, measures, and regulations regarding the disclosure of environmental safety information to the public in China. [/table]
|
Chinese Consensus Report on Family-Based Helicobacter pylori Infection Control and Management (2021 Edition)
# Introduction
Helicobacter pylori has infected half of the world population and is a major health threat for Chinese families and society. It is also a heavy economic and healthcare burden for the country due mostly to the high infection rate and healthcare costs that are attributed to its related gastrointestinal (GI) and extra-GI diseases. [bib_ref] Fifth Chinese national consensus report on the management of Helicobacter pylori infection, Liu [/bib_ref] [bib_ref] Consensus on eradication of Helicobacter pylori and prevention and control of gastric..., Du [/bib_ref] [bib_ref] Recent progress in Helicobacter pylori treatment, Hu [/bib_ref] [bib_ref] Cost-effectiveness analysis of screen-and-treat strategy in asymptomatic Chinese for preventing Helicobacter pylori-associated..., Chen [/bib_ref] A portion of infected people will develop various degrees of GI diseases, such as dyspepsia (5%-10%), chronic gastritis (90%), peptic ulcers (15%-20%) and gastric malignancies (1%); H. pylori infection is also closely associated with a number of extra-GI diseases, such as irondeficiency anaemia, idiopathic thrombocytopenic purpura, autoimmune diseases, cardiovascular and cerebrovascular diseases, etc. Eradication of H. pylori is recommended to reduce the infection rate and occurrence of related diseases. [bib_ref] Fifth Chinese national consensus report on the management of Helicobacter pylori infection, Liu [/bib_ref] [bib_ref] Consensus on eradication of Helicobacter pylori and prevention and control of gastric..., Du [/bib_ref] [bib_ref] Recent progress in Helicobacter pylori treatment, Hu [/bib_ref] [bib_ref] Cost-effectiveness analysis of screen-and-treat strategy in asymptomatic Chinese for preventing Helicobacter pylori-associated..., Chen [/bib_ref] Over the past three to four decades, the general H. pylori infection rate in China has slowly declined Guidelines due to the continued intervention, education, improved sanitary condition and drinking water quality (online supplemental figures 1-4). [bib_ref] Time trends and other sources of variation in Helicobacter pylori infection in..., Li [/bib_ref] [bib_ref] Changing trends of disease burden of gastric cancer in China from 1990..., Zhang [/bib_ref] [bib_ref] China's drinking water sanitation from 2007 to 2018: a systematic review, Wang [/bib_ref] One meta-analysis in 2020, which includes 670 572 participants from 26 provinces of mainland China, found that during 1983-2018, H. pylori prevalence declined by 0.9% annually. The overall prevalence was .7% in 2006-2018. Infection rate varies greatly among different geographical areas and is much higher in rural areas. [bib_ref] Time trends and other sources of variation in Helicobacter pylori infection in..., Li [/bib_ref] In addition, the age-standardised gastric cancer (GC) incidence rate has decreased from 37.56 per 100 000 in 1990 to 30.64 per 100 000 in 2019, with an estimated annual percentage change of −0.41, but GC incidence has increased from 317.34 thousand in 1990 to 612.82 thousand in 2019. The cause for increased GC incidence appears due to an increase in the middle to elderly population (≥40 years of age), along with the growth of the general population (online supplemental figures 2 and 3). It is estimated that in the next 25 years, the numbers of new GC cases and deaths will continue increase, while the rates of incidence and death should steadily decline. [bib_ref] Changing trends of disease burden of gastric cancer in China from 1990..., Zhang [/bib_ref] During this period, China's social and family structure have also profoundly changed (online supplemental figure 5),with an increase in the total population and family number of the nation from 1.13 billion and 278.6 million in 1990 to 1.41 billion and 494.1 million, respectively, in 2021. The traditional multiplegeneration family structure has shrank toward fewer generation and smaller size, with a decrease in average family size from 4.05 person/family in 1990 to 2.62 person/family in 2021,probably due to urbanisation, industrialisation and, previously, the 'onechild policy'. All these are important factors related to H. pylori infection.
Since 2017, four major consensus reports for H. pylori infection control and related disease prevention have been published in China. These include the 'Fifth Chinese National Consensus Report on the Management of Helicobacter pylori Infection', 1 the 'Consensus on Chronic Gastritis in China (2017, Shanghai)', [bib_ref] Consensus on chronic gastritis in China (2017, Shanghai), Fang [/bib_ref] 'National Integrated Traditional Chinese and Western Medicine Management of Helicobacter pylorirelated Diseases' [bib_ref] National integrated traditional Chinese and Western medicine management of Helicobacter pylori-related diseases, Hu [/bib_ref] and the 'Consensus on Eradication of Helicobacter pylori and Prevention and Control of Gastric Cancer in China (2019, Shanghai)'. [bib_ref] Consensus on eradication of Helicobacter pylori and prevention and control of gastric..., Du [/bib_ref] In addition, important international consensus reports have been published to guide the management of H. pylori infection and related disease prevention at the global level. [bib_ref] Kyoto global consensus report on Helicobacter pylori gastritis, Sugano [/bib_ref] [bib_ref] Management of Helicobacter pylori infection-the Maastricht V/Florence Consensus Report, Malfertheiner [/bib_ref] [bib_ref] Helicobacter pylori management in ASEAN: The Bangkok consensus report, Mahachai [/bib_ref] [bib_ref] The Toronto consensus for the treatment of Helicobacter pylori infection in adults, Fallone [/bib_ref] [bib_ref] Guidelines for the management of Helicobacter pylori infection in Japan, Kato [/bib_ref] [bib_ref] Screening and eradication of Helicobacter pylori for gastric cancer prevention: the Taipei..., Liou [/bib_ref] [bib_ref] Houston consensus conference on testing for Helicobacter pylori infection in the United..., El-Serag [/bib_ref] However, no corresponding guidelines and strategies have been designated for the management of H. pylori infection in the general public and among family members to block its transmission and development of related diseases.
H. pylori infection is largely a family-based disease, with the advances in clinical practice and public awareness, the detrimental effects of family-based H. pylori infection have been increasingly recognised, [bib_ref] H pylori: treatment for the patient only or the whole family?, Sari [/bib_ref] [bib_ref] Focusing on whole family based-Helicobacter pylori infection management and clinical research to..., Ding [/bib_ref] [bib_ref] Global whole family based-Helicobacter pylori eradication strategy to prevent its related diseases..., Ding [/bib_ref] which requires attention from physicians and public health administrative officials. Traditionally, the 'test-and-treat' and 'screen-and-treat' strategies are available for various infected populations, and they are critical in guiding global H. pylori eradication programmes over the past decades. However, clinical practice has found the need for refinement on these strategies, [bib_ref] H pylori: treatment for the patient only or the whole family?, Sari [/bib_ref] [bib_ref] Focusing on whole family based-Helicobacter pylori infection management and clinical research to..., Ding [/bib_ref] [bib_ref] Global whole family based-Helicobacter pylori eradication strategy to prevent its related diseases..., Ding [/bib_ref] [bib_ref] Helicobacter pylori treatment for gastric cancer prevention, Malfertheiner [/bib_ref] as the daily practice has met challenges and found that it is hard to differentiate a low-prevalence area from many highly infected communities in China, and the treatment processes are also affected by patient adherence, selection of treatment population and cost-benefit estimations. Furthermore, it is difficult to control infection from source without the infected family members being engaged; therefore, development of novel comprehensive strategies is desired to solve these problems and to facilitate bacterial eradication. Population-wide screening and eradication of H. pylori to prevent GC in highly infected areas have recently been proposed by several consensus reports and supported by large-scale investigations. The notion of 'family-based H. pylori infection control and management' was also recently introduced in China to block its intrafamilial transmission, and it appears to be an effective, practical and promising strategy to reduce H. pylori infection and to prevent related diseases. [bib_ref] Focusing on whole family based-Helicobacter pylori infection management and clinical research to..., Ding [/bib_ref] [bib_ref] Global whole family based-Helicobacter pylori eradication strategy to prevent its related diseases..., Ding [/bib_ref] The National Clinical Research Centre for Digestive Diseases (Shanghai); Gastrointestinal Early Cancer Prevention & Treatment Alliance of China; Helicobacter pylori Study Group of Chinese Society of Gastroenterology, Chinese Medical Association (CMA); and Chinese Alliance for Helicobacter pylori Study therefore decided to hold a joint meeting to discuss and adopt the notion of familybased H. pylori infection control and management as a practical strategy to curb H. pylori intrafamilial transmission and development of related diseases in order to reduce its infection rates, related disease and GC burden for the nation in the coming decades.
The consensus suggestions will be instrumental to prevent H. pylori spread among family members and subsequently reduce related disease, GC incidence and medical expenditure. It will also be helpful to optimise household living style and to improve public awareness and health for Chinese residents and society. Experience from this practice would be beneficial not only for China, but also valuable as a reference for international communities that have high infection rates and related disease burdens.
# Methods
Preparation and construction of consensus-related clinical questions (CQs) and statements followed a search and systematic review of relevant documents; 57 experts from 41 major universities and institutions in 20 provinces and regions of Mainland China were invited to review evidence and to modify statements. Key questions related to familybased H. pylori infection control and management were searched and drafted, and 23 original statement items were formed. The consensus drafting process refers to the population, intervention, comparator and outcome principle, [bib_ref] Development of a framework to identify research gaps from systematic reviews, Robinson [/bib_ref] and we also referred to the international consensus formulation processes. [bib_ref] Kyoto global consensus report on Helicobacter pylori gastritis, Sugano [/bib_ref] [bib_ref] Management of Helicobacter pylori infection-the Maastricht V/Florence Consensus Report, Malfertheiner [/bib_ref] [bib_ref] Helicobacter pylori management in ASEAN: The Bangkok consensus report, Mahachai [/bib_ref] Evaluation of quality of evidence and strength of recommendations: grading of recommendations assessment, development and evaluation system is used to evaluate the quality of evidence and strength of recommendations. [bib_ref] Clinical guidelines Committee of the American College of physicians. the development of..., Qaseem [/bib_ref] Quality of evidence was divided into four levels: high, moderate, low and very low. Strength of recommendations was divided into two levels: strong recommendation (the benefit is significantly greater than the risk or vice versa) and conditional recommendation (benefit is greater than the risk or vice versa). Quality of evidence is only one of the factors that determine the strength of recommendation, and low-quality evidence may also get strong recommendations. Experts provide recommendations based on both the available evidence and their personal opinion and experience.
Consensus reaching process: Delphi method was applied to reach a consensus on relevant CQs and statements. The constructed CQs and statements were first sent to all experts by email; feedbacks from experts were incorporated into the statements. After two rounds of consultation and modification of statements, the initial consensus statements were reached. On 30 January 2021, an online meeting was organised using the 'Zoom Cloud Meetings' platform (because the COVID-19 pandemic prevented a face-to-face meeting), which included 35 available experts from 27 institutions in 18 provinces to discuss and make necessary changes on the statements. Experts discussed the original CQs and 23 statements on family-based H. pylori infection transmission, control and management, and reached consensus on 16 statements with modifications.
Experts voted anonymously through the Chinese online 'Questionnaire Star' platform, and electronic ballots were automatically calculated; this process was monitored by the meeting secretary. Voting opinions were divided into six levels: (1) agree strongly; (2) agree, with minor reservations;
(3) agree, with larger reservations; (4) disagree, with larger reservations; (5) disagree, with minor reservations; and (6) disagree completely. Voting results of 1+2 ≥80% were considered to reach consensus.
The voting results were later finalised, and the contents were prepared for publication. Three versions of the consensus report were drafted for different audiences: (1) a Chinese version of the report 31 was published by Chinese Journal of Digestion, which is the official journal of the Chinese Society of Gastroenterology, CMA and circulated mostly in mainland China; (2) a popular science version of the report 32 which was modified from the professional version and published in Chinese by Health World, the CMA-affiliated popular science magazine, and circulated in mainland China aimed at the general audience; and (3) this English language version of the report, which is aimed at international communities.
# Results
The 16 statements that reached consensus are summarised in table 1 and detailed as follows; 7 statements that failed to reach consensus are presented in online supplemental table 1. [bib_ref] Intrafamilial clustering of Helicobacter pylori infection, Drumm [/bib_ref] [bib_ref] Helicobacter pylori infection in spouses of patients with duodenal ulcers and comparison..., Georgopoulos [/bib_ref] [bib_ref] Role of infected parents in transmission of Helicobacter pylori to their children, Rothenbacher [/bib_ref] [bib_ref] Gastroenteritis and transmission of Helicobacter pylori infection in households, Perry [/bib_ref] [bib_ref] Concordance of Helicobacter pylori infection among children in extended-family homes, Garg [/bib_ref] [bib_ref] Evidence of intra-familial transmission of Helicobacter pylori by PCR-based RAPD fingerprinting in..., Nahar [/bib_ref] [bib_ref] Analysis of intra-familial transmission of Helicobacter pylori in Japanese families, Osaki [/bib_ref] Regarding the role and infection status of parents in the transmission of H. pylori to children, both clinical surveys and molecular biology studies such as random amplified polymorphic DNA fingerprinting analysis have indicated that the infected parents, especially mothers, play a key role in H. pylori transmission within a household. When parents are infected by H. pylori, the infection rate of their children increases significantly; spread also occurs between spouses and among siblings. [bib_ref] Intrafamilial clustering of Helicobacter pylori infection, Drumm [/bib_ref] [bib_ref] Helicobacter pylori infection in spouses of patients with duodenal ulcers and comparison..., Georgopoulos [/bib_ref] [bib_ref] Role of infected parents in transmission of Helicobacter pylori to their children, Rothenbacher [/bib_ref] [bib_ref] Gastroenteritis and transmission of Helicobacter pylori infection in households, Perry [/bib_ref] [bib_ref] Concordance of Helicobacter pylori infection among children in extended-family homes, Garg [/bib_ref] [bib_ref] Evidence of intra-familial transmission of Helicobacter pylori by PCR-based RAPD fingerprinting in..., Nahar [/bib_ref] [bib_ref] Analysis of intra-familial transmission of Helicobacter pylori in Japanese families, Osaki [/bib_ref] These results suggest that intrafamilial transmission is an important cause of H. pylori infection, but the risk of infection among family members varies from family to family. Factors such as living habits, race, hygiene conditions, socioeconomic status and family size all contribute to transmission. Bacterial spread among family Guidelines members also varies greatly between different geographical regions. [bib_ref] Evidence of intra-familial transmission of Helicobacter pylori by PCR-based RAPD fingerprinting in..., Nahar [/bib_ref] All H. pylori-infected patients will develop histological gastritis, and clinical manifestations after H. pylori infection are very different. In some infected people, H. pylori infection does not have any symptoms or signs, and it is only discovered during medical examinations or routine checkups. A portion of infected people will follow Correa's cascade and develop chronic non-atrophic gastritis, atrophic gastritis, intestinal metaplasia, dysplasia and GC after years or decades of slow development. CQ2. What are the transmission routes for H. pylori spread within family units or among family members? Statement 2: H. pylori is transmitted by oral route, and intrafamilial transmission is one of the major sources of infection.
Evidence quality: high. Recommendation strength: strong recommendation 78.9%, conditional recommendation 21.1%.
Consensus level: 94.3%.
Comments: Mounting evidence has demonstrated that H. pylori is transmitted through oral-oral, fecal-oral routes and water sources [fig_ref] Table 2: Common Helicobacter pylori transmission routes and preventive measures [/fig_ref]. H. pylori can be isolated and cultured from vomitus, as well as occasionally from saliva and cathartic stools of infected people, and the bacterium can be detected in dental plaque, cavities, saliva specimens and dental pulp using in vitro molecular detection methods such as PCR and western blots. [bib_ref] Helicobacter pylori colonization of the oral cavity: A milestone discovery, Yee [/bib_ref] [bib_ref] Distribution of Helicobacter pylori and Periodontopathic Bacterial Species in the Oral Cavity, Kadota [/bib_ref] [bib_ref] Molecular detection of Helicobacter pylori based on the presence of cagA and..., Flores-Treviño [/bib_ref] [bib_ref] The presence of Helicobacter pylori in dental plaque of children and their..., Tsami [/bib_ref] In addition, H. pylori can also be detected in a variety of animals, and isolated and cultured in the gastric mucosa of sheep, 46 goats and cows. [bib_ref] Genotyping and antibiotic resistance properties of Helicobacter pylori strains isolated from human..., Ranjbar [/bib_ref] Studies have also found that H. pylori can survive for a certain period in food stuffs, such as milk, juice, vegetables, ready-to-eat foods and different kinds of meats. [bib_ref] Helicobacter pylori: A foodborne pathogen?, Quaglia [/bib_ref] Investigations in Latin America using molecular techniques or bacterial culture methods have found that H. pylori is also present in drinking water, fresh water, well water, estuarine water, seawater and marine products. [bib_ref] Specific detection of cultivable Helicobacter pylori cells from wastewater treatment plants, Moreno [/bib_ref] [bib_ref] Detection of Helicobacter pylori in the coastal waters of Georgia, Puerto Rico..., Holman [/bib_ref] [bib_ref] Novel detection of Helicobacter pylori in fish: A possible public health concern, Abdel-Moein [/bib_ref] [bib_ref] Rural health alert: Helicobacter pylori in well water, Reavis [/bib_ref] [bib_ref] Identification of viable Helicobacter pylori in drinking water supplies by cultural and..., Santiago [/bib_ref] Although the specific routes of transmission remain to be confirmed, these results strongly indicate that they are one of the major causes of infection among family members. H. pylori infection is thus speculated as a foodborne or waterborne disease, with humans and animals the likely reservoirs. Due to its contagious nature, H. pylori infection is usually spread among family members, and the infected strains among members can be completely same or very similar after mutation. Infected individuals may also carry strains of different origins, suggesting that there is also a possibility of exogenous infection. [bib_ref] Role of infected parents in transmission of Helicobacter pylori to their children, Rothenbacher [/bib_ref] Further investigations are warranted to clarify the proportion and importance of both infection routes. Evidence quality: moderate. Strength of recommendation: strong recommendation 73.6%, conditional recommendation 26.4%.
Consensus level: 92.1%. Comments: Individuals infected by H. pylori usually will not clear by itself without proper treatment; family members infected by H. pylori are always a potential source of infection and have the possibility for continued transmission. [bib_ref] Intrafamilial clustering of Helicobacter pylori infection, Drumm [/bib_ref] [bib_ref] Helicobacter pylori infection in spouses of patients with duodenal ulcers and comparison..., Georgopoulos [/bib_ref] [bib_ref] Role of infected parents in transmission of Helicobacter pylori to their children, Rothenbacher [/bib_ref] [bib_ref] Gastroenteritis and transmission of Helicobacter pylori infection in households, Perry [/bib_ref] [bib_ref] Concordance of Helicobacter pylori infection among children in extended-family homes, Garg [/bib_ref] [bib_ref] Evidence of intra-familial transmission of Helicobacter pylori by PCR-based RAPD fingerprinting in..., Nahar [/bib_ref] [bib_ref] Analysis of intra-familial transmission of Helicobacter pylori in Japanese families, Osaki [/bib_ref] When living in the same household with one or more H. pylori-infected family members, other members such as spouses and children are exposed with increased infection risk. Although H. pylori may not necessarily infect every member of a household, the likelihood of infection depends on close contact, living habits and hygiene conditions. [bib_ref] Evidence of intra-familial transmission of Helicobacter pylori by PCR-based RAPD fingerprinting in..., Nahar [/bib_ref] Therefore, education of family members about good hygiene practices, eating habits, awareness of H. pylori and its detrimental effects, handwashing, and avoidance of raw or contaminated food and water are all important to help prevent cross-contamination and reduce the risk of infection [fig_ref] Table 3: Helicobacter pylori infection control and management strategies Strategies Characteristics Applications and limitations... [/fig_ref].
For Chinese families, a common problem that deserves close attention is the use of serving/public chopsticks and spoons. As Chinese families have this historical tradition, and some of them usually share foods in the same dish or bowl, sometimes using the same utensils, which are sources of H. pylori crosscontamination. Therefore, changing habit and promoting separate meals/serving for individual family members are desirable, especially during family gatherings. and adolescents (10-19 years of age), and infection is mainly transmitted by parents, especially by mothers. [bib_ref] Intrafamilial clustering of Helicobacter pylori infection, Drumm [/bib_ref] Common routes of transmission among family members include sharing utensils and food, chewing food before feeding children, kissing, drinking contaminated water and poor hygiene habits. One 2006 survey of childhood H. pylori infection conducted in Shanghai, China, included 1119 healthy school children and adolescents; the results showed that infection rate among asymptomatic children in the 7-year-old age group was 30.9%, with an average annual increase of 3.2%, reaching the adulthood infection level at the age of 12. In 2014, a retrospective survey in China that included 1634 paediatric inpatients undergoing endoscopy found that among children aged under 3, 4-6, 7-10 and 11-18 years groups, H. pylori infection rates were 24.6%, 27.2%, 32.9%, and 34.8%, respectively. [bib_ref] Association between Helicobacter pylori infection and pathological changes in the gastric mucosa..., Yu [/bib_ref] In recent years, due to improved living conditions and healthcare, the overall H. pylori infection rate has declined. One H. pylori infection survey in 2011 conducted in three major Chinese cities (Beijing, Guangzhou and Chengdu) included 3491 children; the results showed that in children aged 1-3, 4-6, 7-9, 10-12, 13-15 and 16-18 years, the infection rates were 0.6%-4.9%, 5.6%-9.7%, 3.9%-7.1%, 8.6%-12.1%, 6.2%-17.2% and 13.0%-33.0%, respectively. [bib_ref] Prevalence and risk factors of Helicobacter pylori infection in asymptomatic Chinese children:..., Ding [/bib_ref] However, for children and adolescents living in rural areas, due to limited access to healthcare and poor sanitation conditions, the infection rates are expected to be higher than those in cities. [bib_ref] Time trends and other sources of variation in Helicobacter pylori infection in..., Li [/bib_ref] Further investigations are required to determine rural infection conditions. These results demonstrate that H. pylori infection rate in children increases with age. Infection is mainly acquired during childhood and adolescent stages, [bib_ref] Association between Helicobacter pylori infection and pathological changes in the gastric mucosa..., Yu [/bib_ref] [bib_ref] Prevalence and risk factors of Helicobacter pylori infection in asymptomatic Chinese children:..., Ding [/bib_ref] and it can also be acquired during adulthood, although the chances are slightly less. In addition, family living style, living environment, education level, socioeconomic status and family size are factors that affect its spread.
## Cq4. when and
## Cq5. should all h. pylori-infected adult family members be treated to eliminate the infection?
Statement 5: For all H. pylori-infected adult family members in a household, eradication should be considered.
Evidence quality: moderate. Recommendation strength: strong recommendation 52.6%, conditional recommendation 47.4%.
Consensus level: 81.5%. Comments: For infected adult family members, H. pylori eradication is recommended following guidelines from the 'Fifth Chinese National Consensus Report on the Management of Helicobacter pylori Infection', unless there are competing considerations. [bib_ref] Fifth Chinese national consensus report on the management of Helicobacter pylori infection, Liu [/bib_ref] The 2015 'Kyoto Global Consensus Report on Helicobacter pylori Gastritis' also indicated that H. pylori gastritis is an infectious disease. Eradicating H. pylori can eliminate the source of infection, reduce infection rate of the population and prevent serious complications. [bib_ref] Kyoto global consensus report on Helicobacter pylori gastritis, Sugano [/bib_ref] Other guidelines and consensus suggestions including the 2017 'Management of Helicobacter pylori Infection-The Maastricht V/Florence Consensus Report', consensus of the USA, Japan and Asia-Pacific region, [bib_ref] Kyoto global consensus report on Helicobacter pylori gastritis, Sugano [/bib_ref] [bib_ref] Management of Helicobacter pylori infection-the Maastricht V/Florence Consensus Report, Malfertheiner [/bib_ref] [bib_ref] Helicobacter pylori management in ASEAN: The Bangkok consensus report, Mahachai [/bib_ref] [bib_ref] The Toronto consensus for the treatment of Helicobacter pylori infection in adults, Fallone [/bib_ref] [bib_ref] Guidelines for the management of Helicobacter pylori infection in Japan, Kato [/bib_ref] [bib_ref] Screening and eradication of Helicobacter pylori for gastric cancer prevention: the Taipei..., Liou [/bib_ref] [bib_ref] Houston consensus conference on testing for Helicobacter pylori infection in the United..., El-Serag [/bib_ref] and results from several large-scale clinical observations [bib_ref] Effects of Helicobacter pylori treatment on gastric cancer incidence and mortality in..., Li [/bib_ref] [bib_ref] Helicobacter pylori eradication to prevent gastric cancer in a high-risk region of..., Wong [/bib_ref] [bib_ref] The benefit of mass eradication of Helicobacter pylori infection: a community-based study..., Lee [/bib_ref] [bib_ref] Mass eradication of Helicobacter pylori to reduce gastric cancer incidence and mortality:..., Chiang [/bib_ref] have all recommended eradicating H. pylori before the occurrence of gastric mucosal atrophy and intestinal metaplasia to reduce the risk of GC. Therefore, unless there are competing considerations, eradicating H. pylori infection is recommended for all adults within the family.
The phenomenon of family cluster infections can also partially explain the fact that gastric mucosal precancerous lesions or GC are present in one or more infected family members at different times of their lives, indicating that H. pylori infection may play a critical role in disease progression. [bib_ref] Increased prevalence of precancerous changes in relatives of gastric cancer patients: critical..., El-Omar [/bib_ref] Evidence quality: moderate. Strength of recommendation: strong recommendation 50.0%, conditional recommendation 50.0%.
Consensus level: 86.8%. Comments: Because H. pylori infection usually results in mild or no gastric mucosal lesions in paediatric patients, few studies have been performed on the relationship between H. pylori infection and gastric mucosal lesions in children. Recent reports have noted that gastric mucosal atrophy and intestinal metaplasia are present in H. pylori-infected children, even in those who are very young. [bib_ref] Helicobacter pylori infection in children and adolescents, Okuda [/bib_ref] Earlier small studies indicated that atrophy was present in 0%-72% of the samples studied. [bib_ref] Does gastric atrophy exist in children?, Dimitrov [/bib_ref] However, recent surveys of 524 children in China 63 and 131 children in Japan 70 showed that the incidence of atrophic gastritis in H. pylori-infected children was 4.4% and 10.7%, respectively. In stage II and III atrophic gastritis of H. pylori-positive Japanese children, the incidence of intestinal metaplasia is 4.6% in both the antrum and body of the stomach. [bib_ref] Association between gastric atrophy and Helicobacter pylori infection in Japanese children: a..., Kato [/bib_ref] A small Mexican study in 2014 on gastric biopsies from 82 children with chronic gastritis found that 8.5% of the samples had gastric mucosal atrophy, and 6.1% had intestinal metaplasia; and among the 36 H. pylori-infected children, six (16.7%) had atrophic gastritis. [bib_ref] Antral atrophy, intestinal metaplasia, and preneoplastic markers in Mexican children with Helicobacter..., Villarreal-Calderon [/bib_ref] In Tunisia, an H. pylori-prevalent area, one study in 2009 found that patients with gastric mucosal atrophy accounted for 9.3% (32/345 cases) of the enrolled children and 14.5% (32/221 cases) of the enrolled patients with chronic gastritis; among these 32 children with atrophic gastritis, 30 were infected by H. pylori. [bib_ref] Gastric atrophy and Helicobacter pylori infection in children, Boukthir [/bib_ref] However, research results are inconsistent between different regions and countries. For example, in a 2012 study of 96 H. Guidelines pylori-infected children in Brazil, gastric mucosal atrophy was not found, and gastric mucosal atrophy was relatively rare in French and Austrian children infected by H. pylori. [bib_ref] Gastric phenotype in children with Helicobacter pylori infection undergoing upper endoscopy, Hoepler [/bib_ref] [bib_ref] In French children, primary gastritis is more frequent than Helicobacter pylori gastritis, Kalach [/bib_ref] [bib_ref] Upper gastrointestinal histopathological findings in children and adolescents with nonulcer dyspepsia with..., Carvalho [/bib_ref] Although further research is needed, these data suggest that gastric mucosal atrophy and intestinal metaplasia in children and adolescents may present in a similar way to adult infection and are probably more common in H. pylori highly infected areas than they were previously thought. Therefore, active intervention based on infection and disease status is required. It is also necessary to explore further on its natural course, consequences, potential carcinogenic risks and factors that lead to atrophy and metaplasia during H. pylori infection.
CQ7. Within a household, should all H. pylori-infected children be treated to eliminate the infection? In 2015, the Gastroenterology Group of the Chinese Paediatric Association, CMA, published a guideline on the management of H. pylori infection in children.The consensus recommends that H. pylori must be eradicated in children with peptic ulcer and patients with gastric mucosa-associated lymphoid tissue (MALT) lymphoma. H. pylori eradication can be offered for children with chronic gastritis, those with a family history of GC or unexplained refractory iron-deficiency anaemia, those planning to take long-term non-steroidal antiinflammatory drugs (NSAIDs, including low-dose aspirin), and those whose guardians or older siblings strongly demand it. The indications to detect H. pylori infection include these conditions and children with a family history of GC in firstdegree relatives, but H. pylori testing is not recommended as a routine assay.For North America and Europe, the joint ESPGHAN/ NASPGHAN guidelines for the management of H. pylori in children and adolescents (Update 2016) recommended against a test-and-treat strategy for H. pylori infection in children but recommended testing and treating H. pylori in children with gastric or duodenal ulcer diseases. [bib_ref] Joint ESPGHAN/NASPGHAN guidelines for the management of Helicobacter pylori in children and..., Jones [/bib_ref] As these recommendations are primarily based on and for the settings of North America and Europe, where H. pylori infection rates are low and decreasing, along with the related disease burden, the guideline may therefore not apply to other areas of the world where the infection rate and related disease burden are high, such as Asia, Latin America and many developing countries.
In 2020, the Japanese Society of Paediatric Gastroenterology, Hepatology and Nutrition published a revised guideline managing H. pylori infection in children. [bib_ref] The updated JSPGHAN guidelines for the management of Helicobacter pylori infection in..., Kato [/bib_ref] The guideline recommends that H. pylori must be eradicated in children with gastric or duodenal ulcers, histological gastric mucosal atrophy, gastric MALT lymphoma, protein-losing gastroenteropathy, iron-deficiency anaemia and chronic idiopathic thrombocytopenic purpura. The report also recommends using at least two H. pylori tests, such as two non-invasive tests (eg, breath test and stool antigen test) or a biopsy-based and a non-invasive test (eg, breath test) for more accurate diagnosis of active infection.
This guideline 79 also recommends that for children aged 5 years or older with active infections, H. pylori eradication therapy can be offered when reinfection possibility is considered. Consideration of H. pylori eradication therapy is recommended for children who have active gastritis, undergo gastroscopy due to abdominal symptoms, have a family history of GC in first-degree or second-degree relatives, and in whom active H. pylori infection has been found. However, H. pylori eradication therapy is not recommended for chronic primary urticaria. A test-and-treat approach is not recommended for asymptomatic children for preventing GC or preventing reinfection in adults in families where H. pylori infection has been eradicated. Compared with its previous versions, the updated recommendations have further clarified the indications for H. pylori detection and eradication in children.
Regarding the methods for diagnosis and confirmation of H. pylori eradication in children and adolescents, commonly applied 13 C-urea breath test (UBT), stool antigen test and serum antibody test, as well as invasive methods including rapid urease test, histological examination and culture, have been recommended. These methods have different accuracy and application limitations, and can be offered based on the indications mentioned in the aforementioned guidelines. [bib_ref] Upper gastrointestinal histopathological findings in children and adolescents with nonulcer dyspepsia with..., Carvalho [/bib_ref] [bib_ref] Joint ESPGHAN/NASPGHAN guidelines for the management of Helicobacter pylori in children and..., Jones [/bib_ref] At least two tests are recommended for more accurate diagnosis. [bib_ref] The updated JSPGHAN guidelines for the management of Helicobacter pylori infection in..., Kato [/bib_ref] Eradication therapies for children and adolescents are slightly different from adult regimens, and body weight, dose calculation-based triple therapies are recommended, including proton pump inhibitor (PPI) plus two antibiotics, such as amoxicillin and clarithromycin or metronidazole; the course is 7-14 days in Japan 79 to 10-14 days in China, North America and Europe regimens. Bismuth is not approved in Japan 79 but is recommended in the consensus reports from China, North America and Europe. To achieve maximal therapeutic effects, consultation with parents or guardians before eradication therapy is mandatory.
CQ8. Should all H. pylori-infected elderly family members be treated for the infection? Comments: H. pylori infection rate is higher in elderly population (60-65 years or older). Eradication of H. pylori can improve GI symptoms in this group of patients, and, to a certain extent, prevent, delay or even partially reverse gastric mucosal atrophy and intestinal metaplasia, as well as reducing the incidence of GC. [bib_ref] Helicobacter pylori therapy for the prevention of metachronous gastric cancer, Choi [/bib_ref] [bib_ref] Treatment of Helicobacter pylori in special patient populations, Nguyen [/bib_ref] [bib_ref] Benefit/risk assessment and issues related to antibiotic use of Helicobacter pylori eradication..., Wang [/bib_ref] [bib_ref] Primary Helicobacter pylori resistance in elderly patients over 20 years: A Bulgarian..., Boyanova [/bib_ref] Studies have shown that the elderly population has no significantly increased resistance to antibiotics commonly used to eradicate H. pylori such as amoxicillin and furazolidone, but resistance to quinolone (levofloxacin) and clarithromycin has increased. Eradication treatment should be considered for this population unless competing considerations exist.
The elderly population often has one or more disease conditions at the time of treatment, such as heart, lung, cerebrovascular and kidney diseases, or uses long-term NSAIDs. Therefore, caution should be exercised and risk-benefit assessment should be carried out before H. pylori eradication. Individual or personalised treatment plans should be selected based on the patient's drug intake history, physiopathological condition, disease status and adverse drug reactions. [bib_ref] Helicobacter pylori therapy for the prevention of metachronous gastric cancer, Choi [/bib_ref] [bib_ref] Treatment of Helicobacter pylori in special patient populations, Nguyen [/bib_ref] [bib_ref] Benefit/risk assessment and issues related to antibiotic use of Helicobacter pylori eradication..., Wang [/bib_ref] [bib_ref] Primary Helicobacter pylori resistance in elderly patients over 20 years: A Bulgarian..., Boyanova [/bib_ref] Discussing treatment regimens with family members or carers is recommended. Proper instruction and education about eradication processes before and during treatment can improve adherence and render the individualised treatment safe and effective. (1) test and treat, which is recommended for young patients with uninvestigated dyspepsia but is unsuitable for older patients and those with alarm symptoms; and (2) screen and treat, which is for patients who have a family history of GC, have alarm symptoms or live in GC-prevalent areas. [bib_ref] Fifth Chinese national consensus report on the management of Helicobacter pylori infection, Liu [/bib_ref] In areas where the infection rate is low, the benefits of population-based screen-and-treat strategy are limited due to declining H. pylori infection rates and related disease burden such as peptic ulcers, precancerous lesions and GC. For areas where the infection rate and related disease burden are high, this strategy is beneficial, cost-effective and worth population-wide screening and intervention (table 2). [bib_ref] Fifth Chinese national consensus report on the management of Helicobacter pylori infection, Liu [/bib_ref] H. pylori is transmitted orally, mostly among family members during childhood and adolescents. Prevention and control of intrafamilial transmission among family members is therefore a practical approach to block the transmission chain and to reduce the infection rate in a society; this approach has been suggested to reduce the infection rate and related diseases. [bib_ref] Focusing on whole family based-Helicobacter pylori infection management and clinical research to..., Ding [/bib_ref] The notion of family-based H. pylori infection control and management was also recently introduced as a third important strategy to block H. pylori transmission and infection in China 19 20 31 32 (figure 1), and the strategy appears practical, easily manageable, with good family member adherence, and does not require differentiating high-infection and low-infection areas. Although large-scale household-based clinical trials to prove its efficacy are currently lacking, meta-analyses and scattered small reports have begun to demonstrate that whole family-based cotreatment approach is a superior strategy to regimens treating only infected individuals in reducing H. pylori reinfection rates. [bib_ref] H pylori: treatment for the patient only or the whole family?, Sari [/bib_ref] This is in line with the concept of mass screening and eradication of H. pylori in populations to reduce its related disease burden. [bib_ref] Screening and eradication of Helicobacter pylori for gastric cancer prevention: the Taipei..., Liou [/bib_ref] It is also necessary to investigate further and to formulate detailed strategies to control household H. pylori infection, ideally by randomised controlled trials (RCTs).
CQ10. Is concurrent treatment of H. pylori-infected family members helpful to reduce the chance of reinfection among family members? Statement 10: Concurrent treatment of H. pylori-infected family members is helpful to reduce the chance of reinfection after its eradication.
Evidence quality: moderate.
Recommendation strength: strong recommendation 65.7%, conditional recommendation 34.3%.
Consensus level: 81.5%. Comments: One critical issue after H. pylori eradication is its reinfection and recurrence; study results from different regions have yielded slightly different outcomes, ranging from 0% to 12%. One Chinese study in 2019 showed that recurrence rates of 1 and 3 years in adults after H. pylori eradication (including recrudescence within 1 year and reinfection over 1 year after eradication) were 1.75% and 4.61%, respectively. Low income and poor hygiene conditions are independent risk factors responsible for recurrence. [bib_ref] Recurrence of Helicobacter pylori infection: incidence and influential factors, Xue [/bib_ref] A systematic review in 2017 showed that the global annual H. pylori recurrence, reinfection and recrudescence rates were 4.3%, 3.1% and 2.2%, respectively. [bib_ref] Systematic review with meta-analysis: the global recurrence rate of Helicobacter pylori, Hu [/bib_ref] However, these results vary by regions and countries. For example, a Korean study in 2013 showed that the longterm (37.1 months) average reinfection rate was 10.9%, and the annual reinfection rate was 3.5% after H. pylori eradication. [bib_ref] Long-term follow-up Helicobacter pylori reinfection rate and its associated factors in Korea, Kim [/bib_ref] The results may also differ between developing and developed countries [bib_ref] Helicobacter pylori recurrence in developed and developing countries: meta-analysis of 13C-urea breath..., Niv [/bib_ref] and between urban and rural areas because residents may experience different healthcare accessibilities, Guidelines living environments and sanitation conditions. [bib_ref] Recurrence of Helicobacter pylori infection: incidence and influential factors, Xue [/bib_ref] However, very few studies have compared the reinfection rate in H. pyloriinfected households versus the general infection conditions in a community.
H. pylori infection among family members can occur before, during or after treatment. Family members living with H. pyloriinfected individuals are exposed to an increased infection risk due to factors such as similar living habits, sharing food or food wares, intimate contact and chewing food before feeding children. This could also be the reason that some patients are reinfected after eradication. [bib_ref] Reinfection after successful eradication of Helicobacter pylori in three different populations in..., Bruce [/bib_ref] Concurrent treatment of all H. pylori-infected family members could theoretically eliminate its cross-infection and spread within the household because the familial source of infection is removed. [bib_ref] Global whole family based-Helicobacter pylori eradication strategy to prevent its related diseases..., Ding [/bib_ref] Recent studies have begun to show that family-based cotreatment of H. pylori infection could partially reduce recurrence rates and increase eradication rates compared with the approach of treating only infected individuals. [bib_ref] H pylori: treatment for the patient only or the whole family?, Sari [/bib_ref] A small, single-centre, non-RCT observation in China [bib_ref] Role of comprehensive treatment in Helicobacter pylori eradication, Jiang [/bib_ref] in 2014 showed that the 24-month cumulative recurrence rate in the single-infected patient treatment group was 19.7%, and in the family cotreatment group, it was only 7.4%. These results are in line with a previous report by Sari et al in Turkey [bib_ref] H pylori: treatment for the patient only or the whole family?, Sari [/bib_ref] in 2008. However, in countries or regions where H. pylori infection rates are low, no difference exists between these two treatment approaches. For example, one study in the UK 94 in 2004 found no difference between the two treatment groups, even after 62.2 months of follow-up on 50 families. One German study [bib_ref] Prevalence of H. pylori-infection in family members of H. pylori positive and..., Knippig [/bib_ref] in 2002 that was conducted with a 2-year follow-up of 108 cases of H. pylori-eradicated patients found no reinfection; even some family members were H. pylori-positive. In 2002, Gisbert et al in Spain 96 followed up 120 cases of H. pylori-eradicated patients and found 6.8% of the patients whose spouses were H. pyloripositive were reinfected per year. However, molecular biology study showed that bacterial strains of the reinfected patients were different from those of their partners, suggesting that reinfected H. pylori strains are from different sources.
In 2021, a meta-analysis demonstrated that whole-family H. pylori eradication is a superior strategy to the single-infected patient treatment approach. [bib_ref] Whole family-based Helicobacter pylori eradication is a superior strategy to single-infected patient..., Zhao [/bib_ref] The overall recurrence rate at 1-year follow-up in the pooled data was 2.7% for the wholefamily treatment group, and 10.5% for single-infected treatment patient group. Studies in this area currently lack large-scale, multicentre RCTs, and further stratified investigations are warranted.
CQ11. Should H. pylori be screened and treated among family members living in the same household with patients who have GC or gastric mucosal precancerous lesions? and H. pylori-infected family members of patients with GC have increased risk of developing gastric mucosal precancerous lesions and GC. Although H. pylori infection-related diseases are more frequently present in adulthood, infections are mostly acquired during childhood and adolescents; the chronic gastritis progresses into atrophic gastritis, intestinal metaplasia, dysplasia and GC after years of slow development. Despite other factors such as genetics, chemical factors, diet, living habits and age are also closely related to histological gastric mucosal atrophy and intestinal metaplasia; H. pylori is the most important known cause of these precancerous lesions and GC, especially for patients who carry the CagA-positive genotype strains. As the high-risk group of individuals for GC, family members of patients with GC should be tested for H. pylori, and those who are H. pylori-positive should receive eradication therapy. For family members with gastric mucosa precancerous lesions, H. pylori infection status should be tested and eradicated if it is positive. In addition, regular endoscopic examination is recommended for early GC surveillance in these patients. Patients with peptic ulcers usually have higher H. pylori infection rates (90.4%), and family members living with the patient are also exposed to a higher risk of infection. For H. pylori-infected adult patients in the family, eradication therapy is recommended unless competing considerations exist. Recent studies on the genotyping of infected persons also show that type I H. pylori strains (CagA and VacA positive) are the major type of H. pylori infection in Chinese residents, and it induces severer inflammation, gastric mucosal lesions and are more prevalent in patients with GC than type II (CagA and VacA negative) strains. This is also supported by population-based studies and confirmed by basic science research experiments. Future investigations in this area are of great importance for preventing gastric mucosal lesion and GC in first-degree relatives of patients with GC. The concept of eradicating H. pylori infection in whole family unit is also reflected in existing international consensus. For example, in the 2018 'Helicobacter pylori Management in ASEAN: the Bangkok Consensus Report', 14 statement 4 recommends that 'eradication of H. pylori reduces the risk of GC and family members of patients with GC should be screened and treated'. The 2020 'Screening and Eradication of Helicobacter pylori for Gastric Cancer Prevention: the Taipei Global Consensus' also recommended 'screening and eradication of H. pylori for gastric cancer prevention in populations with a high incidence or high risk of gastric cancer' (statement [bib_ref] Time trends and other sources of variation in Helicobacter pylori infection in..., Li [/bib_ref] , and that this approach should be included in the national health insurance plan (statement 12). [bib_ref] Screening and eradication of Helicobacter pylori for gastric cancer prevention: the Taipei..., Liou [/bib_ref] The 2016 'Houston Consensus Conference on Testing for Helicobacter pylori Infection in the United States' also included several family member-related testing and treatment statements. [bib_ref] Houston consensus conference on testing for Helicobacter pylori infection in the United..., El-Serag [/bib_ref] Additionally, a number of clinical observations and international consensuses have clearly highlighted that eradication of H. pylori can relieve the inflammatory level, the risk of gastric mucosal atrophy, intestinal metaplasia, dysplasia and even the incidence of GC. [bib_ref] Effects of Helicobacter pylori treatment on gastric cancer incidence and mortality in..., Li [/bib_ref] Young generations would benefit most from this strategy because eliminating H. pylori infection can reduce the risk of precancerous lesions, along with GC in later life. It also reduces the chance of transmission to their children. [bib_ref] Screening and eradication of Helicobacter pylori for gastric cancer prevention: the Taipei..., Liou [/bib_ref] Even at the stage of atrophic gastritis and intestinal metaplasia, eradicating H. pylori is also beneficial in preventing the deterioration of gastric mucosal lesions, and reducing the occurrence of metachronous cancer in patients with GC after endoscopic resection treatment. CQ12. What are the suitable treatment regimens for Chinese residents to eradicate H. pylori among family members? Comments: In 2017, the Helicobacter pylori Study Group of Chinese Society of Gastroenterology, CMA, published its 'Fifth National Consensus Report on the Treatment of Helicobacter pylori Infection', which proposed seven different combinations of bismuth and PPI-containing quadruple treatment regimens (BQT) as the main empiric therapy to eradicate H. pylori. The course of treatment is 10 or 14 days (PPI+bismuth+two antibiotics). These regimens are currently commonly applied in clinical practice with high eradication rates 1 (table 4). They are also suitable for family-based H. pylori eradication in the household. Infected family members should be offered a quadruple regimen with known high curative rate locally and low adverse drug reaction rate as recommended. [bib_ref] Fifth Chinese national consensus report on the management of Helicobacter pylori infection, Liu [/bib_ref] Bismuth does not have drug resistance and is safe and convenient in short-term application. Therefore, unless there are competing considerations, the consensus recommends the empirical bismuth quadruple regimen to treat H. pylori infection. Its current eradication rate is above 85% and, in some instances, above 90% (table 4), However, the actual eradication rates also have variations, depending on different regimens used and geographical location. One large-scale clinical trial that involves a total of 94 101 subjects in 2016 found that the overall H. pylori eradication rate was only 72.9% in the infection group by using the omeprazole, tetracycline, metronidazole and bismuth citrate 10-day regimen. Gender, body mass index, history of stomach disease, baseline delta over baseline value of 13 C-UBT, missed medication doses, smoking and drinking were independent predictors of eradication failure. [bib_ref] A large randomised controlled intervention trial to prevent gastric cancer by eradication..., Pan [/bib_ref] While in other studies, where eradication therapy that uses different regimens such as esomeprazole, amoxicillin, clarithromycin and bismuth citrate for 14 days, the eradication rates are around to or above 85%. [bib_ref] Comparison the cost-efficacy of furazolidonebased versus clarithromycin-based quadruple therapy in initial treatment..., Yi [/bib_ref] In addition to China, the BQT therapies are also recommended as treatment strategies by several international consensus reports, such as the 'Management of Helicobacter pylori Infection-The Maastricht V/Florence Consensus Report', the 'Toronto Consensus for the Treatment of H. pylori Infection in Adults', and 'Helicobacter pylori Management in ASEAN: the Bangkok Consensus Report". [bib_ref] Management of Helicobacter pylori infection-the Maastricht V/Florence Consensus Report, Malfertheiner [/bib_ref] [bib_ref] Helicobacter pylori management in ASEAN: The Bangkok consensus report, Mahachai [/bib_ref] [bib_ref] The Toronto consensus for the treatment of Helicobacter pylori infection in adults, Fallone [/bib_ref] Two concerns for the widespread use of antibiotics are their adverse effects and H. pylori resistance. In 2019, one Chinese study investigated H. pylori resistance rates and found that in a treatment-naïve adult group, the resistance rates for metronidazole, clarithromycin, levofloxacin, amoxicillin, rifampicin and tetracycline were 78.4%, 19.0%, 23.3%, 1.2%, 1.7% and 2.3%, respectively. The previously treated adult group had significantly higher resistance rates for metronidazole (99.2%), clarithromycin (58.3%) and levofloxacin (52.3%). [bib_ref] Characteristics of Helicobacter pylori antibiotic resistance: data from four different populations, Liu [/bib_ref] Recently, double-dose, high-frequency PPI+amoxicillin combination therapy without bismuth has shown promising results in H. pylori eradication. The regimen has been reported to achieve eradication rates similar to quadruple therapy in several individual trials, with eradication rates ranging from 84.7% to 95.3%. [bib_ref] High-dose dual therapy is superior to standard first-line or rescue therapy for..., Yang [/bib_ref] This regimen is relatively simple and easy to implement, with high patient adherence and fewer adverse drug reactions, providing clinicians an alternative option in practice. However, it has limitations: this approach is not suitable for patients who are allergic to amoxicillin, and a history of amoxicillin use may lead to the risk of drug resistance. Drug susceptibility tests are required in cases of repeated eradication failures. [bib_ref] Correction: Eradication efficacy of modified dual therapy compared with bismuth-containing quadruple therapy..., Yang [/bib_ref] In addition to the aforementioned PPI-based regimen, potassium-competitive acid blocker (P-CAB)-based regimen was also introduced recently. 122-125 P-CABs are a novel and heterogeneous class of drugs that competitively block the potassiumbinding site of gastric H + /K + -ATPase. Vonoprazan (VPZ), a P-CAB, has shown a strong inhibitory effect on gastric acid secretion, which lasts longer and is not affected by CYP2C19 gene polymorphism. This provides physicians a new regimen option for eradicating H. pylori. [bib_ref] Systematic review with meta-analysis: the efficacy of vonoprazan-based triple therapy on Helicobacter..., Jung [/bib_ref] [bib_ref] Efficacy of Vonoprazan for Helicobacter pylori eradication, Kiyotoki [/bib_ref] [bib_ref] Seven-day vonoprazan and low-dose amoxicillin dual therapy as first-line Helicobacter pylori treatment:..., Suzuki [/bib_ref] Studies have compared the efficacy of VPZ versus PPIs in the treatment of acid-related disorders and for H. pylori eradication. Optimised VPZ-amoxicillin dual therapy can reliably achieve eradication rates equal to or higher than 95%. 122-125 VPZ-based dual or triple therapy has shown similar or even higher eradication rates than the corresponding PPI-based therapy in intention-to-treat and perprotocol treatments for H. pylori eradication, potentially overcoming the limitations of PPIs. [bib_ref] Systematic review with meta-analysis: the efficacy of vonoprazan-based triple therapy on Helicobacter..., Jung [/bib_ref] Further large-scale, prospective, multicentre randomised trials will be required to verify its efficacy.
CQ13. What is the management strategy in treating H. pylori infection among family members?
## Guidelines
Comments: The nature of H. pylori eradication is similar to the treatment of other commonly found pathogenic bacterial infections, but their characteristics are different. H. pylori eradication rates can be improved by using high-efficiency acidsuppressing PPIs to raise the stomach pH and therefore increase antibiotic bioavailability. [bib_ref] Relative potency of proton-pump inhibitors, Helicobacter pylori therapy cure rates, and meaning..., Graham [/bib_ref] In recent years, with widespread H. pylori eradication in many hospitals and clinics in China, the H. pylori resistance rate to antibiotics has been gradually increasing, which is accompanied with a decline in the empirical eradication rate. Failure of the first-time eradication may lead to the development of bacterial resistance and reduce the range of antibiotic options for future retreatment. Hence, successfully eradicating H. pylori infection in the first treatment is desirable.
H. pylori resistance to antibiotics is usually related to local bacteria resistance patterns and previous antibiotic usage. [bib_ref] Recent progress in Helicobacter pylori treatment, Hu [/bib_ref] Before selecting an empirical treatment regimen, the cure rate, adverse drug reactions, convenience, accessibility, adherence and costs, along with the antibiotic resistance situation of the patient, should be evaluated to achieve the maximal therapeutic effect. It should also avoid misuse or unnecessary use of antibiotics. When treating H. pylori infection in children and adolescents, the safety and benefits of eradication should be assessed carefully based on individual infection history, antibiotic usage and disease status; doses of antibiotics should be adjusted according to body weight before eradication is implemented. [bib_ref] Guidelines for the management of Helicobacter pylori infection in Japan, Kato [/bib_ref] Regarding the efficacy of susceptibility-guided individualised treatment and empirical therapy in eradicating H. pylori infection, a recent Chinese study confirmed that susceptibility-guided individualised therapy and previously proven locally highly effective empiric therapy are highly and equally effective. [bib_ref] Randomised controlled trial: susceptibility-guided therapy versus empiric bismuth quadruple therapy for first-line..., Chen [/bib_ref] However, for patients who have repeatedly failed eradication, individualised therapy guided by drug-sensitivity testing is recommended. In addition, if the patient's history of antibiotic use can be incorporated into the selection of empirical treatment regimen, it will help to increase the first-time H. pylori eradication rate and reduce the chance of developing drug resistance, and only therapies that are known to be highly effective locally should be used empirically. Comments: Non-invasive H. pylori tests include UBTs, serum antibody tests and stool antigen tests. These have been recommended by consensus reports to detect H. pylori infection both in China and internationally. They are currently commonly used in clinical settings and are suitable to detect H. pylori infection among family members. However, each of these methods has its own advantages and restrictions, and they can be selected based on their characteristics and applied either alone or in combination to obtain accurate results and to avoid falsepositive and false-negative results [fig_ref] Table 5: Accuracy of current commonly used Helicobacter pylori diagnosis methods Serum anti-H [/fig_ref]. C-UBT and 14 C-UBT have the advantages of accuracy, relatively high specificity, convenience and being unaffected by patchy distribution of H. pylori in the stomach; their sensitivity and specificity are all above 90%. However, when test results are close to cut-off values, they must be treated with caution or verified by another method or repeated assay. Furthermore, 13 C-UBT and 14 C-UBT results are easily affected by prior use of antibiotics, PPIs and certain Chinese medicines [bib_ref] Fifth Chinese national consensus report on the management of Helicobacter pylori infection, Liu [/bib_ref] [bib_ref] Management of Helicobacter pylori infection-the Maastricht V/Florence Consensus Report, Malfertheiner [/bib_ref] ; residual foods in the stomach, partial gastrectomy, gastric bleeding, gastric tumours, severe gastric mucosal atrophy or low amount H. pylori bacteria numbers in the gastric mucosa also interfere with test results, and false negatives and false positives may occur. Serum antibodies and stool antigen tests are not affected by these factors and are good alternative methods for confirmation, or they can be assayed together for more accurate results. On the occasion of serum tests being positive in untreated patients, they are more likely infected by H. pylori, and treatment options can be considered if active infection is confirmed. In addition, although both 13 C-UBT and 14 C-UBT are considered safe and efficient, for children and pregnant women, 13 C-UBT is the preferred test for these groups of population to avoid the possible radiation exposure. [bib_ref] Helicobacter pylori detection and antimicrobial susceptibility testing, Mégraud [/bib_ref] Serum H. pylori antibody testing has been widely used in hospitals in China with sensitivity and specificity around 72.7%-90.7% and 68.4%-100.0%, [bib_ref] Optimal serological tests for the detection of Helicobacter pylori infection in the..., Xia [/bib_ref] [bib_ref] Evalation of detecting methods of Helicoabter pylori infection histology, serology, 13C-urea breath..., Su [/bib_ref] [bib_ref] Evaluation of three commercial enzyme-linked immunosorbent assay kits for diagnosis of Helicobacter..., Leung [/bib_ref] respectively. H. pylori genotyping antibody testing (for CagA, VacA and UreB) showed higher sensitivity and specificity all above 90%. [bib_ref] Detection of H. pylori antibody profile in serum by protein array, Han [/bib_ref] However, for patients whose H. pylori has been eradicated, due to the longterm presence of serum antibodies, these tests cannot differentiate between past and current infections and are not suitable for follow-up; 13 C-UBT and 14 C-UBT, or stool antigen testing, can make up for these shortcomings. Detection of H. pylori antigen in stools with monoclonal antibodies for the diagnosis of infection is currently available in China ; the sensitivity and specificity are around 73.9%-95% and 86.8%-100%, respectively [fig_ref] Table 5: Accuracy of current commonly used Helicobacter pylori diagnosis methods Serum anti-H [/fig_ref]. Therefore, combined use of multiple methods is of great value to effectively differentiate H. pylori infection status among family members. Commercial products such as H. pylori detection kits for oral plaque or saliva are also simple and rapid, but their sensitivity and specificity need improvements, and they are expected to be the preliminary methods for household family member screening in the future. Other invasive methods for H. pylori detection include rapid urease tests, [bib_ref] Detection of Helicobacter pylori infection by invasive and non-invasive techniques in patients..., Hussein [/bib_ref] [bib_ref] Accuracy of the ultra-rapid urease test for diagnosis of Helicobacter pylori infection, Mcnicholl [/bib_ref] [bib_ref] Validation of diagnostic tests and epidemiology of Helicobacter pylori infection in Bangladesh, Aftab [/bib_ref] histological examination of gastric mucosal biopsies, cultures and molecular techniques. These are available in various clinical settings, but they are inconvenient and require endoscopy to obtain biopsy samples. H. pylori cultures can be used for drug susceptibility testing and bacteriological research, and molecular techniques can be used with specimens such as faeces or gastric mucosal tissues, especially for DNA genotyping of bacterial strains and detection of drug-resistant gene mutations. These methods are applied in different scenarios to detect family member infection.
CQ15. Do we need an H. pylori infection control and management strategy at the general public or community levels to prevent its infection and related disease? Comments: intrafamilial spread is one of the major routes of H. pylori transmission and infection. In addition to the testand-treat and screen-and-treat strategies, the newly introduced concept of family-based H. pylori infection control and management is another important strategy. The strategy is to screen, identify, treat and follow-up all H. pylori-infected family members and is expected to reduce the overall infection burden for the nation in the coming decades. As family members become engaged and motivated, it would be easier to implement H. pylori eradication programmes beginning from family units and to extend them to the general public at the national level. The advantage of this approach (figure 1) is that it proposes not only to treat the visiting patients but also to engage nonvisiting, untreated but infected family members. It also considers the gastric mucosal lesion progression of infected family members. Patients' adherence will be improved with the participation and encouragement from family relatives; therefore, prevention and management of H. pylori infection at public and community levels should include the notion of family-based H. pylori infection control and management to reduce the sources of infection. [bib_ref] Focusing on whole family based-Helicobacter pylori infection management and clinical research to..., Ding [/bib_ref] These approaches will be helpful to block the H. pylori transmission chain and to reduce the likelihood of reinfection. It is currently the most economical, practical and manageable approach to reduce H. pylori infection burden for the nation. At the general public and community levels, this can be achieved through advocacy by government health administrative officials, education programmes from multiple forms of media to increase public awareness about the detrimental effects of H. pylori and to promote healthy lifestyles and habits. [bib_ref] Application of a social media platform as a patient reminder in the..., Luo [/bib_ref] One concern about this strategy is that it might overscreen family members who are not infected. Despite the fact that cost-benefit assessments on large scale, population-wide treatment programmes have yet to be determined, previous results have shown that population-wide screening is cost-effective in GC and related disease prevention. In addition, as noninvasive urease breath tests, serological tests and stool antigen tests are more affordable, accessible and efficient, this strategy has provided a practical solution for whole family-based H. pylori infection prevention and control, which is especially suitable for areas of high H. pylori infection and related disease prevalence. Clinicians including family doctors should also guide visiting patients in clinical practice to avoid unnecessary waste of medical resources and panic about H. pylori while diagnosing and treating infected persons and their family members. [bib_ref] Efficacy, safety, and immunogenicity of an oral recombinant Helicobacter pylori vaccine in..., Zeng [/bib_ref] [bib_ref] How far are we from vaccination against Helicobacter pylori infection?, Stubljar [/bib_ref] [bib_ref] Efficacy, immunogenicity, and safety of a parenteral vaccine against Helicobacter pylori in..., Malfertheiner [/bib_ref] Due to the complexity of H. pylori antigen preparation and the human immune response, it is still at laboratory developmental stage and not yet suitable for clinical application. Therefore, prevention and control of H. pylori infection will still need to use antibiotics to control infection and prevent recurrence. Until an effective vaccine is available, preventing new infections is as important as eradicating existing infections among family members, and both are effective measures for H. pylori infection prevention and control.
## Cq16. will an
# Discussion
H. pylori infection is a family-based, population-wide infectious disease. In addition to the traditional test-and-treat and screenand-treat strategies, this consensus introduced the third novel family-based H. pylori infection control and management strategy for H. pylori infection prevention and control at the national level. The strategy is to screen, identify, treat and follow up on all H. pylori-infected family members (figure 1). It is an extension of the previous consensus to eradicate H. pylori and is able to increase family members' engagement and awareness of infection, prevent or reduce the spread of bacteria within households, and save later medical expenses. It is expected that future practice will confirm the convenience and efficacy of this approach, and provide clinical practice a better option for global H. pylori eradication and prevention.
The advantage of this strategy is that it changed the current clinical practice from only treating H. pylori-infected individuals into actively caring for the infection and disease status of the entire family, including non-visiting, untreated but infected family members, and trying to eliminate the source of infection and reduce the possibility of reinfection within households. Family members who are at high stakes are easily motivated and engaged to receive test and treatment. The strategy also integrates the relevant gastric mucosal lesions and disease progression that was already present in the infected family members, thus subsequently reducing the chance for further deterioration of mucosal inflammation and precancerous lesions. [bib_ref] Focusing on whole family based-Helicobacter pylori infection management and clinical research to..., Ding [/bib_ref] In addition, the practice is also expected to positively influence or improve the family's hygiene habit and prevent other infectious diseases. Family doctors and physicians should also provide family members and the general public relevant information about H. pylori and, at the same time, avoid unnecessary waste of medical recourses.
## Guidelines
Despite the consensus statements and conclusions being well supported by the current available evidence, this strategy has limitations. First, the concept of family-based H. pylori infection control and management is introduced to the field for only several years before reaching current consensus [bib_ref] H pylori: treatment for the patient only or the whole family?, Sari [/bib_ref] [bib_ref] Focusing on whole family based-Helicobacter pylori infection management and clinical research to..., Ding [/bib_ref] [bib_ref] Global whole family based-Helicobacter pylori eradication strategy to prevent its related diseases..., Ding [/bib_ref] ; therefore, experience from this practice is very limited and future refinement is necessary. Second, this consensus is based on and for the setting of China and areas with high infection rates and related disease burden; some of the statements may not be suitable for other countries or areas, as infection condition and treatment strategies vary greatly between different countries or communities. Third, available evidences supporting the strategy on health economics/cost-effectiveness, and several other statements are actually weak, and they are in much need to be improved and strengthened by future investigations. Fourth, massive population-wide eradication-related problems, such as antibiotic resistance, drugs' adverse effects and refractory cases, may increase with the increasing number of patients treated, which deserves serious attention from clinicians, pharmacists and health administrative officials. This version of the consensus will need to be improved and revised based on future clinical practice and evidence-based medicine.
For future perspectives, there are several important areas that need to be improved: first, one critical issue discussed in statement 6 is that if H. pylori might induce gastric mucosal precancerous lesions in children and adolescents, and if the CagA-positive and VacA-positive strains might have different effects compared with CagA-negative and VacA-negative strains in inducing gastric atrophy and intestinal metaplasia in this population, future investigations in these areas are critical to provide evidence to guide future clinical practice and consensus recommendations. Second, another statement that caused much debate and barely passed the consensus is statement 10, which tackles the problem if concurrent treatment of H. pylori-infected family members might reduce the chance of reinfection among family members. As large-scale, randomised controlled clinical trials are actually lacking in these areas, further studies are required for more convincing evidence. Third, future investigations on the molecular mechanisms that H. pylori infection induced gastric precancerous lesion and cancer are critically important to increase our understanding on the carcinogenesis, and to help identify therapeutic targets for intervention and prevention. Fourth, techniques for reliable diagnosis methods for H. pylori infection, confirmation of eradication, and detection drug resistance to guide antibiotics selections are also needed for more convenient and easy access; future development in these areas is of great value to assist accurate detection and susceptibilityguided H. pylori eradication.
In conclusion, the current work summarises previous studies and presents novel consensus recommendations on family-based H. pylori prevention and management. Despite the fact that it just started from China, the practice actually paves the road to explore a global whole family-based H. pylori eradication project; therefore, it has the potential to provide clinical practice a novel avenue for the management of H. pylori infection at a broader level. It is expected to play an important role in reducing H. pylori spread among the Chinese population, increase family members' awareness of infection, improve public health and reduce related diseases and GC burden in the coming decades. In addition to China, the notion and practice could also be a valuable reference and benefit other H. pylori highly infected areas or communities globally.
[fig] - 68 Section 2: Prevention and management of H. pylori infection in children and elderly people within the household CQ6. Does H. pylori infection induce gastric mucosal precancerous lesions in children and adolescents? Statement 6: The relationship between H. pylori infection and gastric mucosal precancerous lesions in children and adolescents needs further investigation. [/fig]
[fig] Figure 1: Flowchart of family-based Helicobacter pylori infection control and management. In clinical settings, visiting patients are questioned for symptoms and signs, and Helicobacter pylori infection status is screened by urease breath tests, serological tests or stool antigen tests. If the patient is H. pylori-positive, their family members are recommended to test for H. pylori using one or more of these methods. Family members usually include parents, spouses, children and others living in the same household. The infected patients and family members are advised to treat the infection based on individual condition and follow-up in 4 weeks. If patients or their family members are H. pylori-negative, routine follow-up and no treatment are required. For patients with endoscopy-confirmed gastric precancerous lesions such as atrophy, intestinal metaplasia and intraepithelial neoplasia, H. pylori infection status should be tested, and if it is positive, eradication therapy should be offered and regular endoscopy surveillance should be performed regardless of H. pylori infection status. Dashed line with arrow indicates interaction and close relationship. [/fig]
[table] Table 1: Summary of the 16 statements Helicobacter pylori infection and transmission among family members in the householdStatement 1: H. pylori is a bacterial pathogen that is transmissible from person to person and especially among family members. [/table]
[table] Table 2: Common Helicobacter pylori transmission routes and preventive measures [/table]
[table] Table 3: Helicobacter pylori infection control and management strategies Strategies Characteristics Applications and limitations 1. Test and treat. Recommended for uninvestigated young patients with dyspeptic symptoms, but not for older patients or persons with alarm symptoms Not suitable for areas with high H. pylori infection rates and high incidences of GC 2. Screen and treat. Recommended for patients with family history of GC and alarm symptoms; not suitable for areas with low H. pylori infection rates Suitable for areas with high H. pylori infection rates and high incidences of GC 3. Family-based control and management. Targeting H. pylori-infected individuals within the family; screening, treating and following up H. pylori-infected family members within the household Application areas are not affected by H. pylori infection rates or incidences of GC. GC, gastric cancer. [/table]
[table] Table 5: Accuracy of current commonly used Helicobacter pylori diagnosis methods Serum anti-H. pylori IgG antibody test 142-144 Serum anti-H. pylori protein array UreB/CagA/VacA 145 HpSA, Helicobacter pylori stool antigen test; N/A, not available; RUT, rapid urease test; UBT, urea breath test. [/table]
|
Does Facial Resemblance Enhance Cooperation?
Facial self-resemblance has been proposed to serve as a kinship cue that facilitates cooperation between kin. In the present study, facial resemblance was manipulated by morphing stimulus faces with the participants' own faces or control faces (resulting in self-resemblant or other-resemblant composite faces). A norming study showed that the perceived degree of kinship was higher for the participants and the self-resemblant composite faces than for actual first-degree relatives. Effects of facial self-resemblance on trust and cooperation were tested in a paradigm that has proven to be sensitive to facial trustworthiness, facial likability, and facial expression. First, participants played a cooperation game in which the composite faces were shown. Then, likability ratings were assessed. In a source memory test, participants were required to identify old and new faces, and were asked to remember whether the faces belonged to cooperators or cheaters in the cooperation game. Old-new recognition was enhanced for self-resemblant faces in comparison to other-resemblant faces. However, facial self-resemblance had no effects on the degree of cooperation in the cooperation game, on the emotional evaluation of the faces as reflected in the likability judgments, and on the expectation that a face belonged to a cooperator rather than to a cheater. Therefore, the present results are clearly inconsistent with the assumption of an evolved kin recognition module built into the human face recognition system.
# Introduction
Theories within evolutionary psychology predict that the human mind comprises mechanisms that serve to facilitate kinrecognition. Distinguishing between close genetic kin and non-kin is essential for cooperation and incest avoidance. Cooperation has long been a puzzle for evolutionary theorists because it often implies accepting costs to help others, which seems inconsistent with Darwin's belief that all organisms strive to increase their own fitness. A major breakthrough in the evolutionary explanation of cooperation came when William Hamilton proposed his theory of inclusive fitness [bib_ref] The evolution of altruistic behavior, Hamilton [/bib_ref] [bib_ref] The genetical evolution of social behaviour. I, Hamilton [/bib_ref]. According to this theory, helping relatives may pay off in evolutionary terms because it implies investing into individuals that share genes with the helper. Therefore, helping behavior that is directed to close kin can support the individual's inclusive fitness. Hamilton's law states that helping is profitable when r.c/b, where r is the degree of kinship between two individuals, c is the cost incurred by the helping individual, and b is the benefit of the individual who receives help. An implication of this formula is that cooperation increases with the degree of kinship [bib_ref] Some neo-Darwinian decision rules for altruism: Weighing cues for inclusive fitness as..., Burnstein [/bib_ref]. However, in order to be able to restrict certain forms of cooperation to close kin, it is necessary to discriminate kin from non-kin. ''If [an individual] could learn to recognize those of his neighbors who really were close relatives and could devote his beneficial actions to them alone an advantage to inclusive fitness would at once appear. Thus a mutation causing such discriminatory behavior itself benefits inclusive fitness and would be selected'' (p. 21f) [bib_ref] The genetical evolution of social behaviour. II, Hamilton [/bib_ref]. Therefore, kin selection theory predicts that people might have evolved kin recognition mechanisms that help them to selectively channel cooperation to close kin.
What psychological mechanism might underlie kin recognition in humans? There is evidence that human kin recognition is based on a number of highly automatic, cue-based kin recognition processes, most of which are also found in other animals [bib_ref] Psychology of human kin recognition: Heuristic cues, erroneous inferences, and their implications, Park [/bib_ref]. It is well established that proximity during childhood is probably used as a kinship cue in humans. Maternal-perinatal association may increase the effects of this variable [bib_ref] The architecture of human kin detection, Lieberman [/bib_ref]. Lieberman and Lobel [bib_ref] Kinship on the Kibbutz: Coresidence duration predicts altruism, personal sexual aversions and..., Lieberman [/bib_ref] investigated the effects of living in close proximity during the first years of life in a Kibbutz. Coresidence duration during childhood was found to be correlated with sexual aversion directed towards opposite-sex peers, and it was also positively correlated with the level of altruism directed towards peers, consistent with other studies showing that coresidence duration has positive effects on cooperation [bib_ref] The architecture of human kin detection, Lieberman [/bib_ref]. These findings suggest that spatial proximity during childhood may be used as a kinship cue to regulate incest avoidance and kin support. Thus, it is well established that coresidence during childhood and maternal association are associated with incest avoidance and cooperation. Also, humans typically learn explicitly and repeatedly who their relatives are. In the presence of such strong and highly reliable cues to kinship, other cues may be ignored [bib_ref] The architecture of human kin detection, Lieberman [/bib_ref].
A second class of kinship recognition mechanisms is based on phenotype matching. There is evidence that both animals and humans use body odor to discriminate between kin and non-kin [bib_ref] The human brain is a detector of chemosensorily transmitted HLA-class I-similarity in..., Pause [/bib_ref] [bib_ref] Human kin recognition by olfactory cues, Porter [/bib_ref] [bib_ref] MHC-correlated mate choice in humans: A review, Havlicek [/bib_ref]. In addition to these mechanisms, it has also been proposed that humans use facial self-resemblance as a kinship cue [bib_ref] Facial resemblance enhances trust, Debruine [/bib_ref] [bib_ref] Trustworthy but not lust-worthy: Context-specific effects of facial resemblance, Debruine [/bib_ref] [bib_ref] Facial resemblance increases the attractiveness of same-sex faces more than other-sex faces, Debruine [/bib_ref]. The closer related two individuals are, the more genes they share, and the more similar they will look. In the present study, we examine the effects of facial self-resemblance on cooperation, liking, and trust. Note that facial resemblance could influence cooperation in two ways. One hypothesis is that we learn the faces of our family members when we grow up. When we encounter similar looking individuals, the emotional reactions towards our family members may affect the emotional evaluation of resembling faces via transference effects [bib_ref] Facial-feature resemblance elicits the transference Effect, Kraus [/bib_ref]. These transference effects may be due to a general tendency to generalize our feelings towards other people and objects to people and objects that resemble them. A second hypothesis is that spotting selfresemblance in a stranger's face might represent a special adaptation that serves to promote kin support. Such an adaptation would only provide an advantage over contextual kinship cues such as verbal communication, co-socialization, childhood coresidence, and maternal-perinatal association when family relationships are unclear (i.e., when paternal uncertainty is high) [bib_ref] Cross-cultural perceptions of facial resemblance between kin, Alvergne [/bib_ref]. Given the comparably low estimates of non-paternity in many human societies [bib_ref] Estimating the prevalence of nonpaternity in Germany, Wolf [/bib_ref] [bib_ref] Low nonpaternity rate in an old Afrikaner family, Greeff [/bib_ref] , it is unclear whether natural selection would favor this mechanism.
Evidence in favor of such a mechanism is mixed. It has been shown that people (especially men) are more willing to help children whose faces resemble their own [bib_ref] Reactions to children's faces: Resemblance affects males more than females, Platek [/bib_ref] [bib_ref] How much paternal resemblance is enough? Sex differences in hypothetical investment decisions..., Platek [/bib_ref] , but the effects are inconsistent [bib_ref] Men do not have a stronger preference than women for self-resemblant child..., Bressan [/bib_ref]. There are also several studies showing that facial self-resemblance facilitates trust and cooperation in socialdilemma games [bib_ref] Facial resemblance enhances trust, Debruine [/bib_ref] [bib_ref] A cue of kinship promotes cooperation for the public good, Krupp [/bib_ref]. The most reliable effects of facial resemblance on positive pro-social attributions have been obtained by DeBruine [bib_ref] Trustworthy but not lust-worthy: Context-specific effects of facial resemblance, Debruine [/bib_ref] [bib_ref] Facial resemblance increases the attractiveness of same-sex faces more than other-sex faces, Debruine [/bib_ref] [bib_ref] Kin recognition: Evidence that humans can perceive both positive and negative relatedness, Krupp [/bib_ref] [bib_ref] Opposite-sex siblings decrease attraction, but not prosocial attributions, to selfresembling opposite-sex faces, Debruine [/bib_ref] in a paradigm that seems to be deliberately designed to detect even small and subtle effects of facial resemblance on trustworthiness judgments. In most studies, the procedure starts by creating a composite face by morphing the faces of 20 individuals. Presumably, this aspect of the procedure serves to eliminate all distinctive features of the face that might influence pro-social attributions. Then different versions of the same composite face are created-one that is morphed with the participant (the self-resemblant face), and others that are morphed with other participants. The different versions of the face are simultaneously presented, and the procedure forces participants to pick the one that seems most trustworthy. Note that the versions are identical, with the only difference being the self-or-otherresemblance of the faces. Although this paradigm has proven useful for obtaining reliable effects of facial self-resemblance on judgments of trustworthiness [bib_ref] Trustworthy but not lust-worthy: Context-specific effects of facial resemblance, Debruine [/bib_ref] [bib_ref] Facial resemblance increases the attractiveness of same-sex faces more than other-sex faces, Debruine [/bib_ref] [bib_ref] Kin recognition: Evidence that humans can perceive both positive and negative relatedness, Krupp [/bib_ref] [bib_ref] Opposite-sex siblings decrease attraction, but not prosocial attributions, to selfresembling opposite-sex faces, Debruine [/bib_ref] , it seems like a highly artificial task that is deliberately designed to detect even minimal effects of facial resemblance. One could argue that participants are forced to rely on facial resemblance when making their trustworthiness judgments because all other variables that could have influenced their behavior have been eliminated. Everyday situations are not at all like that. For instance, in most situations there is plenty of information available that could influence the decision of whether to trust somebody or not. If the effects of facial resemblance on cooperation are due to a kin recognition module that has been selected for its beneficial effects on inclusive fitness, then facial resemblance should have pronounced effects on cooperative behavior even when other information is available. Furthermore, often the task is not to pick a cooperation partner from a group of similar individuals, but rather to decide whether to engage in social cooperation with a certain individual or not.
In the present study, our aim was to test whether the effects of facial resemblance generalize to a situation that provides a more realistic assessment of the effects of facial self-resemblance on cooperation. There were several procedural differences compared to the studies of DeBruine described above. First, the faces are presented sequentially. Furthermore, we morphed the participants' faces with real faces instead of facial composites to reduce the artificiality of the stimulus material. There is some evidence that the facial resemblance effects generalize to this paradigm [bib_ref] Facial resemblance enhances trust, Debruine [/bib_ref] , but generally the results are inconsistent. For instance, there are several experiments that examine whether voters show enhanced trust in a politician whose face has been morphed with their own face [bib_ref] Facial similarity between voters and candidates causes influence, Bailenson [/bib_ref] [bib_ref] Transformed facial similarity as a political cue: A preliminary investigation, Bailenson [/bib_ref]. Only in one of these experiments an effect of facial self-resemblance on political preference was revealed; in all the other experiments interactions were found which could not be reliably replicated. Although the results were interpreted as evidence for the assumption that voters prefer candidates who resemble them, the findings are highly inconsistent, and a more objective interpretation would be that the hypothesis is only weakly supported.
To examine the effects of facial resemblance on cooperation and trust, we applied a paradigm that has been shown to be very sensitive to cues of facial trustworthiness in a previous study [bib_ref] On the flexibility of social source memory: A test of the emotional..., Bell [/bib_ref]. Participants played a cooperation game in which they could invest money into a joint business venture with virtual interactants who either cooperated or cheated. In a surprise test phase, participants saw the faces together with new faces, and were asked to rate the likability of the faces. Next, they classified the faces as old or new. When a face was classified as old, they were asked to indicate whether the face belonged to a cheating or cooperative interactant. This paradigm has proven to be a useful tool in the study of cooperation [bib_ref] On the flexibility of social source memory: A test of the emotional..., Bell [/bib_ref] [bib_ref] Enhanced old-new recognition and source memory for faces of cooperators and defectors..., Bell [/bib_ref]. Participants' behavior in this paradigm is strongly affected by properties of the stimulus faces that are known to affect trust and cooperation. For instance, it is known that facial trustworthiness is strongly associated with the overall positive or negative evaluation of a face [bib_ref] Evaluating faces on trustworthiness: An extension of systems for recognition of emotions..., Todorov [/bib_ref]. Consistent with this finding, the a-priori likableness of the faces (that had been assessed in a norming study) was positively associated with (a) the willingness to invest into the cooperation game, (b) the emotional evaluation of the face as expressed in the likability ratings, and (c) the tendency to guess that a face belonged to a cooperator in the source memory test. All of these dependent variables were also affected by the facial expression of the faces. Smiling, relative to an angry facial expression, was associated with enhanced cooperation and likableness of the faces, and was also associated with benevolent guessing in the source memory test [bib_ref] On the flexibility of social source memory: A test of the emotional..., Bell [/bib_ref]. What is more, likability ratings and guessing biases were also affected by the pre-normed trustworthiness of the faces. These findings prove that the paradigm applied in the present study is highly sensitive to manipulations of facial trustworthiness.
Based on the assumption that facial self-resemblance enhances trust and cooperation, we predicted that these findings should be replicated using facial self-resemblance-manipulated by morphing the participants' own faces with other faces-as the independent variable. Specifically, we predicted that facial self-resemblance would be positively associated with (a) the willingness to invest into the cooperation game, (b) the emotional evaluation of the face as expressed in the likability ratings, and (c) with benevolent guessing in the source memory test. In addition to its effects on a potential kin recognition mechanism, facial self-resemblance may also have beneficial effects on memory simply because it is an extremely well known stimulus feature to which people have been exposed very often during their lifetime. For highly overlearned faces such as one's own face, people may have developed rich representations [bib_ref] Robust representations for faces: Evidence from visual search, Tong [/bib_ref] that may facilitate face identification, face processing, and face recognition [bib_ref] Robust representations for faces: Evidence from visual search, Tong [/bib_ref] [bib_ref] Recognizing one's own face, Kircher [/bib_ref]. Enhanced memory for familiar faces has also been demonstrated for other super-familiar faces of family members, friends, and celebrities [bib_ref] Recognizing familiar and unfamiliar faces, Klatzky [/bib_ref] [bib_ref] Familiarity enhances visual working memory for faces, Jackson [/bib_ref]. Expertise may also explain why people are better at recognizing faces of their own race, sex, and age [bib_ref] Recognizing faces of other ethnic groups: An integration of theories, Sporer [/bib_ref] [bib_ref] Thirty years of investigating the own-race bias in memory for faces: A..., Meissner [/bib_ref] [bib_ref] An own gender bias and the importance of hair in face recognition, Wright [/bib_ref] [bib_ref] Evidence for an own-age bias in face recognition, Anastasi [/bib_ref]. Based on the assumption that super-familiarity facilitates face recognition, superior old-new recognition of self-resembling in comparison to other-resembling faces can be expected.
# Methods
# Ethics statement
The study was approved by the ethics committee of the medical faculty of Heinrich Heine University Düsseldorf. Participants signed an informed consent before participating in the experiment.
## Participants
The original sample consisted of 66 students at Heinrich Heine University Düsseldorf aged 18 to 30 years. Only people without facial hair were allowed to participate in the study. Participants who did not match the inclusion criteria were offered to participate in another study. Participants were matched with yoked partners of the same sex who participated in the experiment on another day. Four women and their yoked partners had to be excluded from the data analysis because two women recognized their own face, one was interrupted by a fire alarm, and one failed to attend the second session. The remaining sample consisted of 58 white adults (40 female, 18 male; mean age = 21.36, SD of age = 2.96). 82.76% of the participants had at least one sibling.
# Materials
A total of 80 male and 80 female white stimulus faces with neutral expressions were taken from face databases [bib_ref] FACES-A database of facial expressions in young, middle-aged, and older women and..., Ebner [/bib_ref] [bib_ref] The Karolinska directed emotinal faces -KDEF: Department of Clinical Neuroscience, Psychology section, Lundqvist [/bib_ref] [bib_ref] Presentation and validation of the Radboud Faces Database, Langner [/bib_ref]. All photographs had the same size (4006499 pixel grayscales). Each photograph showed a face on a white background. To avoid morphing artifacts, only faces without facial hair were selected. Half of the male and female faces were assigned to two sets of 40 male and 40 female photographs each (henceforth Set A and B). All faces were edited to have the same brightness.
The photographs of the participants' faces were taken approximately 1 week before the proper experiment. Light conditions and the size of the faces were held constant. All participants had a neutral facial expression when photographed. Participants were asked to remove their glasses, piercings, and makeup, which could have interfered with the morphing process. Scars and birth marks were removed digitally (using Photoshop CS5). Participants were told that the photograph would be used to show it to other participants in the cooperation game.
Each participant's face was morphed with all persons of the same sex in either Set A or B. One participant and his or her yoked partner formed a pair. One member of each pair was randomly assigned to either Set A or B; the other member was assigned to the remaining set. Each participant's face (i.e., eye brows, eyes, nose, and mouth) was cut out, edited to have the same brightness as the stimulus face, and pasted into the shape of the stimulus face. Then the participant's face was morphed with the stimulus face using Abrosoft FantaMorph Deluxe 5.2.5 (see [fig_ref] Figure 1: Example morph. [/fig_ref] , which had been used in a previous study [bib_ref] Westermarck, Freud, and the incest taboo: Does familial resemblance activate sexual attraction?, Fraley [/bib_ref]. Similar to previous studies [bib_ref] Facial resemblance enhances trust, Debruine [/bib_ref] [bib_ref] A cue of kinship promotes cooperation for the public good, Krupp [/bib_ref] [bib_ref] Facial similarity between voters and candidates causes influence, Bailenson [/bib_ref] [bib_ref] Transformed facial similarity as a political cue: A preliminary investigation, Bailenson [/bib_ref] , the participants' faces and the stimulus faces were blended in a 40:60 ratio to create the selfresembling face (i.e., a morph consisted of 40% of the participant's face, and 60% of the stimulus face). We morphed only the inner face of the participants to ensure that superficial similarities (e.g., in the participants' hairstyle or clothes) between the stimulus faces and the participants could not influence the results, and to avoid artifacts of the morphing procedure.
## Norming study
Method. To ensure that we successfully manipulated selfresemblance, we conducted a norming study in which 25 students who did not participate in the proper experiment rated the similarity and the perceived degree of kinship between the participants' faces and the morphed stimulus faces. For comparison, we selected photographs of first degree relatives from the KinFace database [bib_ref] Understanding kin relationships in a photo, Xia [/bib_ref] [bib_ref] Kinship verification through transfer learning, Xia [/bib_ref] , which provides 200 photographs of celebrities in young adult age and the same number of facial photographs of first-degree relatives (parent or child) photographed at approximately the same age. To rule out that knowledge about the celebrities influenced the results, we selected only faces of white parent-child pairs that were not recognized by four independent raters (half of which were female). Sixteen father-son pairs, and four mother-daughter pairs were selected based on these criteria. These twenty parent-child face pairs were compared with sixteen male and four female participant-morph face pairs that were randomly selected from the whole pool of available participant-morph face pairs by a computer program.
All faces had the same size (2286265 pixel grayscales). The faces were edited so that only the inner face (eye brows, eyes, nose, and mouth) was shown on a white background. For each participant in the norming study, half of the twenty faces of the KinFace database (8 male and 2 female faces), and half of the twenty faces of the participants (8 male and 2 female faces) were presented together with matching faces (i.e., with first-degree relatives from the KinFace database, or with self-resemblant faces). The other half of the faces were presented with non-matching faces from the same sources.
First, participants were asked to indicate whether they knew one of the faces. Then, participants were asked to judge the resemblance of the faces on a scale ranging from 1 (''very dissimilar'') to 6 (''very similar''). Finally, participants were asked to rate the perceived degree of kinship of the two faces on a scale ranging from 1 (''not related'') to 6 (''closely related''). Nine participants indicated to know at least one of the shown faces. To rule out that knowledge about the faces influenced the results, these nine participants were excluded from further data analysis.
Results. The results are shown in [fig_ref] Figure 2: Results of the norming study [/fig_ref]. A 262 MANOVA showed that the facial resemblance ratings differed as a function of match, F(1,15) = 53.72, p,.01, g 2 = .78, and database, F(1,15) = 30.54, p,.01, g 2 = .67. Most importantly, there was also a significant interaction between the two variables, F(1,15) = 14.78, p,.01, g 2 = .50. Matching parent-child pairs were perceived as being more similar than non-matching face pairs taken from the KinFace database, t(15) = 4.26, p,.01, g 2 = .55. However, the difference in facial resemblance between matching and mismatching face pairs was even more pronounced for participant-morph pairs used in the present study, t(15) = 7.21, p,.01, g 2 = .78.
It seems especially interesting whether people used facial resemblance to estimate the degree of kinship between the individuals depicted. The kinship ratings were significantly affected by match, F(1,15) = 50.31, p,.01, g 2 = .77, and database, F(1,15) = 41.99, p,.01, g 2 = .74. These main effects were qualified by a significant interaction between match and database, F(1,15) = 11.21, p,.01, g 2 = .43. Consistent with previous results showing that people can detect kinship among others based on facial resemblance [bib_ref] Cross-cultural perceptions of facial resemblance between kin, Alvergne [/bib_ref] [bib_ref] Perceptual factors affecting the ability to assess facial resemblance between parents and..., Kaminski [/bib_ref] [bib_ref] Kin recognition and the perceived facial similarity of children, Maloney [/bib_ref] , first-degree relatives of the KinFace database were correctly judged to be more closely related than mismatching pairs of faces from the same source, t(15) = 3.43, p,.01, g 2 = .44. However, the difference between matching and mismatching face pairs was even more pronounced for participant-morph face pairs, t(15) = 6.95, p,.01, g 2 = .76. Matching participant-morph face pairs were judged to be more closely related than parent-child pairs from the KinFace database that were actually closely related, t(15) = 6.08, p,.01, g 2 = .71.
The norming study clearly shows that facial similarity was successfully manipulated. Even more importantly, facial resemblance had pronounced effects on the perceived degree of kinship between the participants of the proper experiment and the morphs. Participants and morphs were perceived to be even more closely related than actual first-degree relatives. One might criticize that the estimation of the resemblance and perceived kinship of actual first-degree relatives might be underestimated due to paternal uncertainty. However, recent studies show that the prevalence of non-paternity in Western societies is so low (about 1%) that it can be considered irrelevant for the comparison reported above [bib_ref] Estimating the prevalence of nonpaternity in Germany, Wolf [/bib_ref]. Thus, if a kinship recognition mechanism exists that is sensitive to facial resemblance, one would predict that it should be activated by the manipulation of facial resemblance used in the present study.
## Procedure
Participants attended two sessions. In the first session, which took about ten minutes, the participants were informed that their photograph would be shown to other participants in a socialcooperation game in which they would also take part in the second session. The second session was scheduled to occur about one week later and lasted about 45 minutes. The participants played the social-cooperation game and took part in a surprise source monitoring test.
Social-cooperation game. In the first part of the experiment, participants played the social-cooperation game successfully used by Bell et al. [bib_ref] On the flexibility of social source memory: A test of the emotional..., Bell [/bib_ref] to examine the effects of facial likability and facial expression on trust and cooperation. In the social-dilemma game, participants saw the faces of 40 interactants, all of which had the same sex as the participant. Twenty self-resembling faces (i.e., morphs of the participant) and 20 other-resembling faces (i.e., morphs of the same-sex yoked partner) were shown. Half of each type of interaction partner cheated, and the other half cooperated. The faces were randomly assigned to these conditions. Morphed photographs of yoked partners were used as control faces (instead of the original photographs) to avoid confounding self-resemblance with morphing. Otherwise, it would not have been possible to distinguish the effects of self-resemblance from effects of morphing. For instance, averaged faces are known to be perceived as more attractive [bib_ref] Attractive faces are only average, Langlois [/bib_ref]. Therefore, it is possible that morphed faces are associated with enhanced cooperation per se. Therefore, morphed pictures were used in both the experimental and the control condition to control for effects of morphing.
Each trial started with the presentation of a black silhouette on the left side of the screen representing the participant, and the face of the interactant on the right side of the screen [fig_ref] Figure 3: Screenshot of the social-cooperation game [/fig_ref]. Above the participant's silhouette, his or her current account balance was shown. At the beginning of the game, the account balance was 550 cents. The participant was required to decide whether to invest 15 On the left side, the participant's face (eye brows, eyes, nose mouth) is shown that has been pasted into the shape of the stimulus person's face. On the right side, an example for a stimulus face is shown. In the center, the face morph is shown (morph ratio = 40:60). Both faces are used for illustration purposes only (i.e., because of copyright restrictions, the faces do not correspond to faces actually used in the present study). Written consent (as outlined in the PLoS consent form) was obtained from both individuals before publishing these photos. doi:10.1371/journal.pone.0047809.g001 or 30 cents into a joint business venture with the interactant. Once the investment was selected and confirmed, the selected amount was shown in an arrow appearing on the upper right corner of the participant's silhouette. After 600 ms, the arrow moved towards the center of the screen (within 600 ms). Next, the interactant's investment was shown in an arrow appearing on the upper left corner of the interactant's face. Whether the participant made or lost money depended on whether the interactant cooperated or cheated. If the interactant was a cooperator, he or she reciprocated and invested as much money as the participant did (15 or 30 cents). If the interactant was a cheater, he or she refused to reciprocate and invested no money. The arrow with the interactant's investment moved to the center of the screen, where the sum of investments was shown. A bonus (M of the sum of investments) was added, and the total sum was shown. In a last step, the total sum was divided between the interactants. Both received half of the total sum, regardless of what they had invested. The corresponding amounts were presented in two arrows at the center of the screen that moved towards the interaction partners at the left and right side of the screen, respectively. When the participant played with a cooperator, both interactants made money (5 or 10 cents, depending on the participant's investment). When the participant played with a cheater, the cheater benefitted at the expense of the participant. The participant had a loss that was as large as the gain in the cooperator condition (5 or 10 cents, depending on the participant's investment). On both sides, the amount of gain or loss (i.e., the difference between each interactant's initial investment and his or her share of the total sum) was shown. At the end of the experiment, the amount scored was paid out.
Source memory test. When the cooperation game was completed, participants received the instructions for the test phase. They saw a sequence of 80 faces, half of which had been encountered in the cooperation game, and half were new. Half of the new faces were self-resembling faces, and the other half resembled the yoked control partner. First, participants were asked to rate the likability of each face on a scale ranging from 1 (''not likable at all'') to 6 (''extremely likable''). Then, they were asked to indicate whether the face was old or new. When a face was classified as old, participants were asked to specify whether the face belonged to a cheater or to a cooperator. Then the next face was presented. It was randomly determined whether a face was used as the face of an interactant in the cooperation game or as a new face in the old-new face recognition test.
In a debriefing interview after the source monitoring test, participants were asked whether they noticed anything about the stimulus faces. Two of the participants reported that they had recognized their own face, as one would expect when manipulating facial self-resemblance near threshold. The two participants and their yoked partners were excluded from further analysis. Note that it is standard to manipulate facial resemblance near threshold [bib_ref] Facial resemblance enhances trust, Debruine [/bib_ref] [bib_ref] Transformed facial similarity as a political cue: A preliminary investigation, Bailenson [/bib_ref].
## Design
A 262 within-subjects-design was used with facial resemblance (self vs. other) and interactant behavior (cheating vs. cooperative) as within-subject factors. The dependent variables were the cooperation-game investments, the likability ratings, old-new face recognition performance, source memory performance, and the amount of benevolent guessing in the source memory test. Given a sample size of N = 58, 80 responses in the source memory test and a = .05, the power to detect a difference between the source guessing parameters for self-resembling and other-resembling interactants with an effect size of w = 0.06 (which is in the order of magnitude of the effects on guessing observed by Bell et al. [bib_ref] On the flexibility of social source memory: A test of the emotional..., Bell [/bib_ref] was reasonably large (1-b = .98). The same applies to the general linear model within-subject comparisons. An effect of size f = .25 could be detected with a probability of 1-b = .95, assuming an average population correlation between the levels of the behavioral history repeated-measures variable of r = .6. The power calculation was conducted using GNPower [bib_ref] G*Power 3: A flexible statistical power analysis program for the social, behavioral,..., Faul [/bib_ref].
# Results
The results of the present experiment are displayed on the right side of [fig_ref] Figure 4: Cooperation-game investments, likability ratings, and the amount of benevolent guessing in the... [/fig_ref]. For reasons of comparison, we display the results of Experiment 1 of Bell et al. [bib_ref] On the flexibility of social source memory: A test of the emotional..., Bell [/bib_ref] on the left side of [fig_ref] Figure 4: Cooperation-game investments, likability ratings, and the amount of benevolent guessing in the... [/fig_ref]. In that experiment, we had manipulated facial likability by presenting faces of high or low facial likability that were selected based on consensus judgments. As can be clearly seen, all three dependent variables displayed in [fig_ref] Figure 4: Cooperation-game investments, likability ratings, and the amount of benevolent guessing in the... [/fig_ref] were significantly affected by facial likability in the previous study. These dependent variables were also used in the present study. We can thus be sure that the present paradigm is highly sensitive to manipulations of facial cues to trustworthiness.
## Game investments
The upper panel of [fig_ref] Figure 4: Cooperation-game investments, likability ratings, and the amount of benevolent guessing in the... [/fig_ref] shows the mean investments in the cooperation game. On the left side, the effects of facial likability (obtained in the previous study) are shown. As can be seen, facial likability had a pronounced effect on the participants' decision to invest into the cooperation game. Given that investing only makes sense when cooperation is anticipated, this finding is consistent with the idea that facial likability enhances trust. In contrast, as the right side of [fig_ref] Figure 4: Cooperation-game investments, likability ratings, and the amount of benevolent guessing in the... [/fig_ref] shows, there was no effect of self-resemblance on the investments, F(1,57) = 0.03, p = .87, g 2 ,.01. This finding is clearly inconsistent with the assumption that facial self-resemblance facilitates trust and cooperation. A trial-based analysis [fig_ref] Figure 5: Mean cooperation-game investments for self-and other-resembling interactants in each of the 20... [/fig_ref] showed that the game investments did not differ as a function of self-resemblance throughout the experiment. Thus, there was no difference between self-and other-resemblant faces even in the initial trials.
## Test-phase likability ratings
The center panel of [fig_ref] Figure 4: Cooperation-game investments, likability ratings, and the amount of benevolent guessing in the... [/fig_ref] shows facial likability ratings. Again, the left side shows the results of the previous study [bib_ref] On the flexibility of social source memory: A test of the emotional..., Bell [/bib_ref] in which faces with high a-priori likability received higher likability ratings than faces with low a-priori likability, consistent with the idea that there is a high level of agreement between individuals about the features that make a face likable or unlikable [bib_ref] Evaluating faces on trustworthiness: An extension of systems for recognition of emotions..., Todorov [/bib_ref]. In contrast, in the present experiment, facial self-resemblance had no effect on the emotional evaluation of the faces as expressed in the likability ratings, F(1,57) = 1.51, p = .22, g 2 = .03. The likability ratings were descriptively lower for cheaters than for cooperators, F(1,57) = 3.28, p = .08, g 2 = .05. There was no interaction between self-resemblance and interactant behavior, F(1,57) = 0.01, p = .93, g 2 ,.01. Thus, the results indicate that there is not always a strong link between self-resemblance and the emotional evaluation of faces.
## Old-new-recognition
There was a main effect of self-resemblance on old-newrecognition in terms of the sensitivity measure of the two-high threshold (2-HT) model P r [bib_ref] Pragmatics of measuring recognition memory -Applications to dementia and amnesia, Snodgrass [/bib_ref] , F(1,57) = 7.48, p,.01, g 2 = .12. Consistent with previous studies showing enhanced recognition of other types of self-resembling (own-race, own-age, and own-sex) faces [bib_ref] Recognizing faces of other ethnic groups: An integration of theories, Sporer [/bib_ref] [bib_ref] Thirty years of investigating the own-race bias in memory for faces: A..., Meissner [/bib_ref] [bib_ref] An own gender bias and the importance of hair in face recognition, Wright [/bib_ref] [bib_ref] Evidence for an own-age bias in face recognition, Anastasi [/bib_ref]
## Measuring source memory
To analyze performance in the source memory test, we used the multinomial 2-HT model of source monitoring proposed by Bayen, Murnane and Erdfelder [bib_ref] Source discrimination, item detection, and multinomial models of source monitoring, Bayen [/bib_ref] and displayed in . This model is often used in source memory research [bib_ref] Multinomial processing tree models. A review of the literature, Erdfelder [/bib_ref] to separately assess the cognitive processes of old-new recognition, source monitoring and guessing which are assumed to underlie the observable classification performance in the source memory task. A huge advantage of this model is that validation studies have shown empirically that manipulations that are known to affect oldnew recognition, source memory, and guessing, are accurately reflected, respectively, in the model parameters representing oldnew recognition, source memory, and guessing [bib_ref] Source discrimination, item detection, and multinomial models of source monitoring, Bayen [/bib_ref] [bib_ref] Influences of source-item contingency and schematic knowledge on source monitoring: Tests of..., Bayen [/bib_ref].
The first processing tree displayed in shows the processes that are assumed to occur when a cooperator face is presented at test. The probability D coop represents the probability of recognizing the face as old. Parameter d coop represents the conditional probability of remembering that the face is that of a cooperative interactant. With the complementary probability 1d coop , it is not remembered that the face belonged to a cooperator. Then, the source has to be guessed. The source guessing parameter g represents the probability of guessing that the face belongs to a cooperative interactant. The complementary probability 1-g reflects the probability that a face is guessed to belong to a cheating interactant. When the face is not recognized as an old face with probability 1-D coop , it may still be guessed that the face is old with probability b. For these faces, it may be guessed that the face belonged to a cooperative interactant with probability g, or it may be guessed that the face belonged to a cheater with probability 1-g. With probability 1-b, the face is guessed to be new. The second tree represents the processing of the cheater faces in an analogous way. The last tree represents the processing of new (distractor) faces.
Please note that for reasons of simplification, only one set of trees is shown in . To analyze the results of the source memory test adequately, we need two sets of these trees, one representing the self-resembling faces and one representing the other-resembling faces. Therefore, we also need two sets of parameters. For instance, there is one parameter representing the probability of guessing that a self-resembling face belongs to a cooperator, g Self , and one parameter representing the probability that an other-resembling face belongs to a cooperator, g Other .
## Source memory and guessing
The model shown in has more free parameters than can be identified based on the data, but it can be made identifiable by imposing appropriate restrictions on the parameters. Bayen et al. [bib_ref] Source discrimination, item detection, and multinomial models of source monitoring, Bayen [/bib_ref] provided a taxonomy of identifiable submodels of the source monitoring model. The most parsimonious identifiable model that still fits our data (Submodel 4 in the taxonomy of Bayen et al. [bib_ref] Source discrimination, item detection, and multinomial models of source monitoring, Bayen [/bib_ref] includes the assumptions that old-new recognition and source memory do not differ between cooperators and cheaters (D coop = D cheat = D new ; d coop = d cheat ), which is consistent with the analysis of the old-new recognition data reported above and previous studies examining memory for cheaters and cooperators in social-dilemma games [bib_ref] Enhanced old-new recognition and source memory for faces of cooperators and defectors..., Bell [/bib_ref]. A discrepancy between these assumptions and our data would be reflected in the goodness-of-fit statistic G 2 , which is asymptotically X 2 distributed. However, the base model that included these assumptions fit the data well, G 2 (4) = 3.12, p = .54, w = 0.03.
First, we tested whether the parameters representing item recognition could be set to be equal between self-resembling and other-resembling faces. As a restriction on the base model this assumption generates one additional degree of freedom. Selfresembling faces were better remembered than other-resembling faces, DG 2 (1) = 6.95, p,.01, w = 0.04. This result confirms the analysis of the old-new recognition data reported above. Next, we tested whether the parameters representing source memory for self-and other-resembling faces could be set to be equal. Source memory did not differ between self-and other-resembling faces, DG 2 (1) = 1.12, p = .29, w = 0.02.
We were most interested in testing whether guessing differed between self-and other-resembling faces. Specifically, previous studies have shown that facial likability, facial trustworthiness, and smiling were associated with an increased tendency towards guessing that a face was associated with cooperative behavior. It has been shown that these guessing biases are a good measure of trust and positive social expectations [bib_ref] On the flexibility of social source memory: A test of the emotional..., Bell [/bib_ref]. Based on the assumption that facial self-resemblance is associated with trust and positive expectations about the future social behavior of interactants, we expected that the guessing parameter g (reflecting the probability of guessing that a face belonged to a cooperative interactant if the true source of the face was not known) should be higher for self-resembling in comparison to other-resembling faces. Again, the left side of the lower panel of [fig_ref] Figure 4: Cooperation-game investments, likability ratings, and the amount of benevolent guessing in the... [/fig_ref] shows the effects of facial likability obtained in the previous study [bib_ref] On the flexibility of social source memory: A test of the emotional..., Bell [/bib_ref]. Clearly, the probability estimate of parameter g was higher for likable than for unlikable faces, which means that participants had a pronounced bias towards guessing that likable faces for which the true source was not known belonged to cooperators, and that unlikable faces belonged to cheaters. This finding is consistent with the idea that likable facial features are associated with positive expectations towards the cooperative behaviors of others [bib_ref] On the flexibility of social source memory: A test of the emotional..., Bell [/bib_ref]. In contrast, as the right side of the lower panel of [fig_ref] Figure 4: Cooperation-game investments, likability ratings, and the amount of benevolent guessing in the... [/fig_ref] shows, the guessing parameter g did not differ at all between self-and other-resembling faces, DG 2 (1) = 0.67, p = .41, w = 0.01. Facial self-resemblance was clearly not associated with the expectation that the face belonged to a cooperator.
## Additional analyses
In two additional analyses, we evaluated whether the participants' sex, or their having siblings modulated the effects of resemblance on any of the dependent variables. A priori, it might have seemed possible that either female or male participants are more sensitive to the facial resemblance manipulation than male or female participants, respectively. It has been previously discussed that women may be more sensitive to facial resemblance in incest avoidance situations because they have to bear more costs when the offspring suffers from genetic defects than men [bib_ref] Effect of facial self-resemblance on the startle response and subjective ratings of..., Lass-Hennemann [/bib_ref]. Furthermore, it has been suggested that men are more sensitive to the self-resemblance of children's faces due to an adaptation that has evolved as a solution to the problem of paternal uncertainty [bib_ref] Estimating the prevalence of nonpaternity in Germany, Wolf [/bib_ref] [bib_ref] Low nonpaternity rate in an old Afrikaner family, Greeff [/bib_ref]. No gender effects are to be expected with respect to the effects of self-resemblance on cooperation, but a priori it seemed possible that sex-specific adaptations support the detection of facial self-resemblance and might therefore influence the results. Furthermore, it has been previously suggested that having siblings might enhance kin recognition in some situations (the effects have been demonstrated in a mating context, but not in an exchangerelevant context) [bib_ref] Opposite-sex siblings decrease attraction, but not prosocial attributions, to selfresembling opposite-sex faces, Debruine [/bib_ref]. However, the analyses revealed that the participants' sex and their having siblings had absolutely no influence on the self-resemblance effect.
The participants' sex had no main effects or interaction effects on any dependent variables, with one exception. There was a main effect of sex on the likability ratings, F(1,56) = 4.66, p,.01, g 2 = .21. Female faces received higher likability ratings by female participants than male faces by male participants. Although this effect should be interpreted with caution because of the comparably small sample of male participants, it might be due to the common finding that female faces are generally perceived as being more likable than male faces. For the interpretation of the present findings, the effect can be considered irrelevant Lower panel: Probability estimate of parameter g, which represents guessing that a face (of which the behavior was not known) belonged to a cooperative person. Thus, a high g probability estimate represents a tendency towards guessing that the face belonged to a cooperator, whereas a low g probability estimate represents a tendency towards guessing that a face belonged to a cheater. The error bars represent the standard errors of the means. doi:10.1371/journal.pone.0047809.g004 because there was no interaction between sex and facial resemblance.
Having siblings had no main effects or interaction effects on any of the dependent variables. However, there was a main effect of having siblings on the guessing parameter g in the source memory test, DG 2 (2) = 9.37, p,.01, w = 0.04. Participants with siblings had a higher tendency of guessing that a face (of which the behavior was not known) belonged to a cooperative interactant than participants without siblings. This finding might suggest that people who grew up with siblings have more positive expectations towards the social environment. However, the effect should be interpreted with caution because of the small number of participants without siblings in our study. For the central research question of our study this main effect is irrelevant because there was no interaction with facial resemblance.
# Discussion
The data pattern found in the present study is clear and consistent. Consistent with previous studies showing enhanced recognition of self-resembling (i.e., own-race, own-sex, and own- . The source memory model adapted from Bayen et al. [bib_ref] Source discrimination, item detection, and multinomial models of source monitoring, Bayen [/bib_ref]. The rounded rectangles on the left side represent the stimulus category (cooperative, cheating, new). The rectangles on the right side represent the participants' responses to the faces in the source memory test. The letters along the links represent the probabilities with which certain cognitive states occur (D = probability of correctly recognizing a face as old or new; d = probability of correctly identifying the source of a face as cheating or cooperative; g = probability of guessing that a face belonged to a cooperative interactant; b = probability of guessing that a face was old). doi:10.1371/journal.pone.0047809.g006 age) faces [bib_ref] Recognizing faces of other ethnic groups: An integration of theories, Sporer [/bib_ref] [bib_ref] Thirty years of investigating the own-race bias in memory for faces: A..., Meissner [/bib_ref] [bib_ref] An own gender bias and the importance of hair in face recognition, Wright [/bib_ref] [bib_ref] Evidence for an own-age bias in face recognition, Anastasi [/bib_ref] , facial self-resemblance had beneficial effects on old-new face recognition. This finding is most likely due to the fact that one's own face is a highly familiar stimulus with a particularly rich representation embedded in a large network [bib_ref] Robust representations for faces: Evidence from visual search, Tong [/bib_ref] , which may have beneficial effects on encoding and retrieval even when the participants are not aware of it. However, the hypotheses that were based on evolutionary theories of kin recognition were clearly not supported by the present study. Based on the assumption that facial self-resemblance is a kinship cue that stimulates kin support, we expected that facial self-resemblance should be associated with enhanced cooperation, likability, and trust. However, these predictions were clearly refuted by the results of the present study. (1) Facial self-resemblance did not affect the willingness to engage in cooperation. Participants did not invest more money into the cooperation game when the interactants' faces resembled their own face. [bib_ref] The genetical evolution of social behaviour. I, Hamilton [/bib_ref] The emotional evaluation of the stimulus faces was not affected by facial selfresemblance in that self-resembling faces were rated as likable as other-resembling faces. (3) Facial resemblance did not affect trust in the sense of having confidence that the other person has cooperative intentions. The probability of guessing that the face belonged to a cooperator rather than to a cheater (when the true source of the face was not known) did not differ between self-and other-resembling faces. This guessing bias has previously been proven to be a good measure of positive emotional expectations in social contexts [bib_ref] On the flexibility of social source memory: A test of the emotional..., Bell [/bib_ref] [bib_ref] Look who's talking! Facial appearance can bias source monitoring, Nash [/bib_ref]. In summary, our findings imply that selfresemblance does not lead to positive expectancies regarding the pro-social behavior of others, and does not trigger pro-social behavior towards others.
Before discussing the consequences of the present findings for theories about kin detection and facial resemblance, we discuss several methodological issues, and explain why we think that the present null results cannot be attributed to methodological problems. First, the null results cannot be attributed to low statistical power. Using 40 facial photographs of 58 participants should guarantee a sufficient reliability of the dependent measures and a sufficient statistical power to detect even small to medium effects. Obviously, statistical reasons cannot explain the discrepancy to previous studies such as that of DeBruine [bib_ref] Facial resemblance enhances trust, Debruine [/bib_ref] in which significant effects were obtained using eight facial photographs of 24 participants.
Second, when interpreting null effects, one faces the problem that this outcome could be either due to the absence of an effect of the independent variable or due to an insensitivity of the method used to measure the effect. To avoid this problem, we applied a paradigm that has been shown to be very sensitive to cues of trustworthiness in previous studies [bib_ref] On the flexibility of social source memory: A test of the emotional..., Bell [/bib_ref] [bib_ref] Enhanced old-new recognition and source memory for faces of cooperators and defectors..., Bell [/bib_ref]. For instance, high facial likability was associated with high investments in the social cooperation game, high likability ratings, and a tendency towards guessing that the person was associated with cooperative behavior (as shown in [fig_ref] Figure 4: Cooperation-game investments, likability ratings, and the amount of benevolent guessing in the... [/fig_ref]. Likewise, the facial expression of the faces (smiling vs. anger), and facial trustworthiness had pronounced effects on these measures. Therefore, it seems noticeable that facial self-resemblance did not even lead to a consistent tendency towards cooperation, trust, and sympathy.
There is thus good evidence that the paradigm applied in the present study is sensitive to manipulations of facial trustworthiness. This rules out several alternative interpretations of the null effect. For instance, a potential concern is that our paradigm might have been insensitive to the effects of facial self-resemblance because participants did not believe to interact with real persons in the cooperation game. However, this hypothesis certainly does not explain the discrepancy to other studies in which effects of selfresemblance have been obtained in highly artificial laboratory situations [bib_ref] Trustworthy but not lust-worthy: Context-specific effects of facial resemblance, Debruine [/bib_ref] [bib_ref] Facial resemblance increases the attractiveness of same-sex faces more than other-sex faces, Debruine [/bib_ref] [bib_ref] Kin recognition: Evidence that humans can perceive both positive and negative relatedness, Krupp [/bib_ref] [bib_ref] Opposite-sex siblings decrease attraction, but not prosocial attributions, to selfresembling opposite-sex faces, Debruine [/bib_ref] , and is inconsistent with previous evidence showing that the paradigm used here provides sensitive measures for detecting effects of facial trustworthiness [bib_ref] On the flexibility of social source memory: A test of the emotional..., Bell [/bib_ref].
Finally, one might be tempted to argue that the present null results could be explained by assuming that the facial phenotype matching mechanism is sensitive to other facial features than those manipulated in the present study. However, this interpretation is seriously challenged by the results of the norming study showing that the facial features manipulated in the present study clearly affected kinship judgments. The morphed stimulus faces and the faces of participants were rated as being even more closely related than actual first-degree relatives, which shows that self-resemblance was successfully manipulated. Given the results of the norming study, it seems quite plausible that a kin recognition mechanism should be activated by the manipulation used in the present study if it existed. Note that a popular hypothesis in Evolutionary Psychology is that people have evolved mechanisms that allow them to make accurate judgments about kinship based on facial features [bib_ref] Cross-cultural perceptions of facial resemblance between kin, Alvergne [/bib_ref] , consistent with the results of the present norming study. Furthermore, it has been speculated that the ability to recognize kin relationships between third parties might be a byproduct of the capacity to detect one's own kin and relies on the same kin detection mechanism [bib_ref] Cross-cultural perceptions of facial resemblance between kin, Alvergne [/bib_ref]. In contrast to these speculations, the present results show an interesting dissociation between the effects of facial resemblance on judgments of kin relationships in the norming study, and its lack of effect on trust and cooperation in the experiment. It is unclear why a kin recognition mechanism that is based on facial self-resemblance should be sensitive to other facial features than a kin recognition mechanism that is based on the resemblance between other people's faces.
Thus, the present study raises doubts about the robustness of the facial self-resemblance effect on trust and cooperation. It is noticeable that the paradigm used in the present study was shown to be highly sensitive to the effects of other cues of facial trustworthiness [bib_ref] On the flexibility of social source memory: A test of the emotional..., Bell [/bib_ref]. Furthermore, as our norming study shows, we used a strong manipulation of facial resemblance. Nevertheless, no effect on trust and cooperation was obtained. From this pattern of results, we conclude that the effects of facial resemblance on trust and cooperation are small and negligible in comparison to the large effects of other cues to facial likability and facial trustworthiness. Therefore, it seems questionable whether it is necessary to postulate a highly specific kin recognition module built into the face recognition system to explain effects of facial self-resemblance that seem to be only reliable in highly artificial laboratory tasks. In everyday life, there seem to be much more potent influences on cooperation that may mask an effect of facial resemblance. Therefore, we think that the fitness benefits associated with such a subtle and unreliable facial resemblance effect are too small to be selected for. Hence, it is questionable that the effects of facial resemblance on cooperation represents a highly specific adaptation built into the face processing system.
A possible interpretation of the overall pattern of results is that people do indeed show incest avoidance and enhanced cooperation towards close kin. This may have a subtle influence on the responses to kin-resembling faces that can be detected in strictly controlled tests. When seeing a face that resembles a close family member, people may show a transference effect, which is the wellknown observation that we are influenced by the emotions towards our friends and relatives when we learn about new persons that resemble them or have something in common with them [bib_ref] Facial-feature resemblance elicits the transference Effect, Kraus [/bib_ref] [bib_ref] Childhood physical and emotional abuse by a parent: Transference effects in adult..., Berenson [/bib_ref]. This transference effect might be due to a general tendency to generalize our feelings and reactions. However, this effect may be so small and negligible in comparison to other influences on facial evaluation and cooperation that it can be only detected in highly artificial laboratory situations and is likely to be masked by much stronger influences on cooperative behavior in everyday life. If correct, then it is unlikely that the influence of self-resemblance on cooperation is strong enough to represent a highly specialized adaptation that has been selected for its beneficial effects on inclusive fitness.
Another possible interpretation of the present results would be to argue that the kinship recognition mechanism does not globally affect cooperation, but has rather restricted effects on cooperative behaviors. Our paradigm has been shown to accurately reflect the effects of variables (facial likability, smiling) that are commonly known to affect social perception and social behavior. However, it is possible that the beneficial effects of kinship on cooperation are not mediated by likability and trust. Given that cooperation with kin pays off in evolutionary terms even when the favors are not returned, one might cooperate with close relatives even when one does not expect that they will reciprocate. Therefore, one may expect to find more pronounced effects of kinship cues on measures of pure altruism (e.g., in a dictator game) than in the present paradigm that may be more closely linked to reciprocal altruism. Another interesting hypothesis is that people might only rely on phenotype matching in environments in which childhood coresidence and association to the mother are no reliable indicators of kinship (for example, when illegitimate children are frequent) [bib_ref] Cross-cultural perceptions of facial resemblance between kin, Alvergne [/bib_ref]. In any case, the present results clearly show facial resemblance does not generally lead to detectable effects on all measures of trust and cooperation, as was claimed previously [bib_ref] Facial resemblance enhances trust, Debruine [/bib_ref] [bib_ref] Trustworthy but not lust-worthy: Context-specific effects of facial resemblance, Debruine [/bib_ref]. The positive influence of facial self-resemblance on cooperation seems to be much more subtle and restricted in scope than previously thought.
[fig] Figure 1: Example morph. [/fig]
[fig] Figure 2: Results of the norming study. Mean ratings for mismatching and matching parent-child face pairs (taken from the KinFace database) and participant-morph face pairs. Upper panel: Mean facial resemblance ratings on a scale ranging from 1 (''very dissimilar'') to 6 (''very similar''). Lower panel: Mean ratings of the perceived degree of kinship on a scale ranging from 1 (''not related'') to 6 (''closely related''). The error bars represent the standard errors of the means. doi:10.1371/journal.pone.0047809.g002 [/fig]
[fig] Figure 3: Screenshot of the social-cooperation game. On the left side, the participant is represented by a black silhouette. On the right side, a self-resembling or other-resembling face is shown representing the interactant. The number in the upper arrows refer to the investments. In this example, the interactant is cooperative and invests the same amount of money as the participant (30 cents). In the center of the screen, the sum of the investments and the bonus are shown. The numbers in the lower arrows refer to the participant's and the interactant's share of the total sum. doi:10.1371/journal.pone.0047809.g003 [/fig]
[fig] Figure 4: Cooperation-game investments, likability ratings, and the amount of benevolent guessing in the source memory test as a function of facial likability in Bell et al. 's (in press) Experiment 1 (for comparison only) and as a function of facial self-resemblance in the present experiment. Upper panel: Mean investments into the cooperation game. Participants could choose to invest 15 or 30 cents. Middle panel: Mean test phase likability ratings on a scale ranging from 1 (''not likable at all'') to 6 (''extremely likable''. [/fig]
[fig] Figure 5: Mean cooperation-game investments for self-and other-resembling interactants in each of the 20 trials of the cooperation game. doi:10.1371/journal.pone.0047809.g005 [/fig]
|
Protocol: Adaptive Implementation of Effective Programs Trial (ADEPT): cluster randomized SMART trial comparing a standard versus enhanced implementation strategy to improve outcomes of a mood disorders program
Background: Despite the availability of psychosocial evidence-based practices (EBPs), treatment and outcomes for persons with mental disorders remain suboptimal. Replicating Effective Programs (REP), an effective implementation strategy, still resulted in less than half of sites using an EBP. The primary aim of this cluster randomized trial is to determine, among sites not initially responding to REP, the effect of adaptive implementation strategies that begin with an External Facilitator (EF) or with an External Facilitator plus an Internal Facilitator (IF) on improved EBP use and patient outcomes in 12 months. Methods/Design: This study employs a sequential multiple assignment randomized trial (SMART) design to build an adaptive implementation strategy. The EBP to be implemented is life goals (LG) for patients with mood disorders across 80 community-based outpatient clinics (N = 1,600 patients) from different U.S. regions. Sites not initially responding to REP (defined as <50% patients receiving ≥3 EBP sessions) will be randomized to receive additional support from an EF or both EF/IF. Additionally, sites randomized to EF and still not responsive will be randomized to continue with EF alone or to receive EF/IF. The EF provides technical expertise in adapting LG in routine practice, whereas the on-site IF has direct reporting relationships to site leadership to support LG use in routine practice. The primary outcome is mental health-related quality of life; secondary outcomes include receipt of LG sessions, mood symptoms, implementation costs, and organizational change. Discussion: This study design will determine whether an off-site EF alone versus the addition of an on-site IF improves EBP uptake and patient outcomes among sites that do not respond initially to REP. It will also examine the value of delaying the provision of EF/IF for sites that continue to not respond despite EF.Trial registration: ClinicalTrials.gov identifier: NCT02151331
# Background
It can take years to translate evidence-based practices (EBPs) into community-based settings [bib_ref] Implementation research in mental health services: an emerging science with conceptual, methodological,..., Proctor [/bib_ref]. This researchto-practice gap is especially pronounced for psychosocial EBPs for mood disorders (depression, bipolar disorders), which represent the top ten causes of disability according to the World Health Organization. Mood disorders are associated with significant functional impairment, high medical costs, and preventable mortality [bib_ref] Global mortality, disability, and the contribution of risk factors: Global Burden of..., Murray [/bib_ref]. For many patients, pharmacotherapy is not enough to improve outcomes, with psychosocial treatments in addition to pharmacotherapy recommended [bib_ref] Canadian Network for Mood and Anxiety Treatments (CANMAT) and International Society for..., Yatham [/bib_ref] [bib_ref] Clinical practice guidelines for bipolar disorder from the Department of Veterans Affairs, Bauer [/bib_ref].
Despite the availability of psychosocial EBPs [bib_ref] Implementation research in mental health services: an emerging science with conceptual, methodological,..., Proctor [/bib_ref] [bib_ref] Canadian Network for Mood and Anxiety Treatments (CANMAT) and International Society for..., Yatham [/bib_ref] [bib_ref] Clinical practice guidelines for bipolar disorder from the Department of Veterans Affairs, Bauer [/bib_ref] [bib_ref] Moving treatment research from clinical trials to the real world, Roy-Byrne [/bib_ref] for mood disorders, they rarely get implemented and sustained in community-based practices [bib_ref] Overcoming barriers and creating opportunities to reduce burden of affective disorders: a..., Wells [/bib_ref] [bib_ref] From silos to bridges: meeting the general health care needs of adults..., Horvitz-Lennon [/bib_ref] [bib_ref] National Comorbidity Survey Replication: The epidemiology of major depressive disorder: results from..., Kessler [/bib_ref]. This is primarily due to a lack of available strategies to help providers embed the EBP into routine clinical workflows, and garner support from clinical and administrative leadership on the EBP's added value to the practice [bib_ref] Moving treatment research from clinical trials to the real world, Roy-Byrne [/bib_ref] [bib_ref] Translating evidence-based depression management services to community-based primary care practices, Kilbourne [/bib_ref] [bib_ref] Reinventing evidence-based interventions?, Sullivan [/bib_ref] [bib_ref] Bridging community intervention and mental health services research, Wells [/bib_ref] [bib_ref] Implementation of evidence-based practice and organizational performance, Hovmand [/bib_ref] [bib_ref] The sustainability of evidence-based practices in routine mental health agencies, Swain [/bib_ref] [bib_ref] Mobilizing organizations for health enhancement: theories of organizational change, Steckler [/bib_ref] [bib_ref] The effects of organizational culture and climate on the access to mental..., Glisson [/bib_ref] [bib_ref] Organizational climate partially mediates the effect of culture on work attitudes and..., Aarons [/bib_ref] [bib_ref] The organizational context of children's mental health services, Glisson [/bib_ref] [bib_ref] The cross-level effects of culture and climate in human services teams, Glisson [/bib_ref] [bib_ref] Financial incentives and accountability for integrated medical care in Department of Veterans..., Kilbourne [/bib_ref] [bib_ref] Organizational contexts of primary care approaches for managing problem drinking, Schutte [/bib_ref] [bib_ref] Management of mental disorders in VA primary care practices, Kilbourne [/bib_ref] [bib_ref] Translating scientific opportunity into public health impact: a strategic plan for research..., Insel [/bib_ref]. As a result, outcomes for persons with mental disorders remain suboptimal [bib_ref] Translating evidence-based depression management services to community-based primary care practices, Kilbourne [/bib_ref]. New healthcare initiatives including medical home models designed to improve efficiency and value have not included specific strategies to assist local providers in implementing EBPs in routine care [bib_ref] Fostering accountable health care: moving forward in medicare, Fisher [/bib_ref] [bib_ref] Building a medical neighborhood for the medical home, Fisher [/bib_ref] [bib_ref] A national strategy to put accountable care into practice, Mcclellan [/bib_ref]. Moreover, up to 98% of patients with mood disorders receive care from smaller clinical practices, which may not have the tools to fully implement medical homes [bib_ref] Primary care and behavioral health practice size: the challenge for health care..., Bauer [/bib_ref].
For EBPs to reach these patients, effective implementation strategies are needed [bib_ref] Disseminating innovations in health care, Berwick [/bib_ref] [bib_ref] Wells KB: Fidelity, adherence, and robustness of interventions, Duan [/bib_ref] [bib_ref] A practice change model for quality improvement in primary care practice, Cohen [/bib_ref] [bib_ref] Models, strategies, and tools: theory in implementing evidence-based findings into health care..., Sales [/bib_ref]. Implementation strategies that have been highly specified, operationalized, and previously used to implement EBPs into usual care settings include Replicating Effective Programs (REP) [bib_ref] Stall R: Implementing evidence-based interventions in health care: application of the replicating..., Kilbourne [/bib_ref]. REP primarily focuses on standardization of the EBP implementation into routine care settings through toolkit development and marketing, provider training, and program assistance [fig_ref] Figure 1: Replicating Effective Programs [/fig_ref]. Used successfully to improve the uptake of brief HIV interventions [bib_ref] The Fox Project, Tax [/bib_ref] [bib_ref] Self-efficacy: toward a unifying theory of behavioral change, Bandura [/bib_ref] , REP is a low-intensity intervention with minimal costs for sites. However, when applied to the implementation of a psychosocial EBP for mood disorders, REP resulted in less than half of sites implementing the EBP [bib_ref] Cluster randomized adaptive implementation trial comparing a standard versus enhanced implementation intervention..., Kilbourne [/bib_ref].
Recognizing the need for more intensive implementation strategies that leverage local provider initiative and outside expertise [bib_ref] Research & services partnerships: lessons learned through a national partnership between clinical..., Kirchner [/bib_ref] , REP was enhanced to include Facilitation [bib_ref] Cluster randomized adaptive implementation trial comparing a standard versus enhanced implementation intervention..., Kilbourne [/bib_ref] [bib_ref] Research & services partnerships: lessons learned through a national partnership between clinical..., Kirchner [/bib_ref]. Two Facilitation roles evolved: an External Facilitator (EF), employed from outside the local site, and an Internal Facilitator (IF), employed at each of the local sites. EFs are situated at a central location and provide technical expertise and program support in implementing the EBP at the local site. In contrast, IFs have a direct reporting relationship to site leadership and have protected time to support providers in implementing EBPs by helping them align the EBP activities with the priorities of the clinic providers and local leadership.
A recent randomized controlled trial [bib_ref] Cluster randomized adaptive implementation trial comparing a standard versus enhanced implementation intervention..., Kilbourne [/bib_ref] [bib_ref] Re-engaging veterans with serious mental illness into care: preliminary results from a..., Kilbourne [/bib_ref] of added an External Facilitator (REP + EF) versus REP alone found that among 88 sites not responding to REP, sites randomized to receive REP + EF compared to sites continuing with REP were more likely to adopt the EBP within 6 months (defined as percentage of patients receiving care management: 56% vs. 28%) [bib_ref] Re-engaging veterans with serious mental illness into care: preliminary results from a..., Kilbourne [/bib_ref]. This suggests that while REP + EF led to increased uptake, a more intensive, locally oriented implementation strategy (e.g., Internal Facilitation) might be needed for EBP uptake in sites that do not fully implement the EBP. In a different study, REP in combination with EF and IF (REP + EF/IF) versus standard REP alone improved adoption and fidelity to a mood disorder EBP [bib_ref] Enhanced fidelity to treatment for bipolar disorder: results from a randomized controlled..., Waxmonsky [/bib_ref] [bib_ref] Long-term clinical outcomes from a randomized controlled trial of two implementation strategies..., Kilbourne [/bib_ref] , underscoring the promise of REP + EF/IF. However, because IF requires additional time commitment from sites, and not all sites may need IF, an adaptive implementation strategy approach is a more practical design. In contrast to measuring implementation non-response (or correlates of it) and not using it to guide improved implementation and outcomes, in adaptive implementation strategies, implementation interventions are augmented in direct response to limited adoption of EBPs among specific sites based on circumstances that may not be observable at baseline. This study addresses key scientific questions that need to be answered in order to develop an effective adaptive implementation strategy. Notably, among sites not responding to REP (i.e., not adopting the EBP), is it best to augment with REP + EF/IF or with REP + EF, or is it best to delay the provision of REP + EF/ IF for sites that continue not to respond despite an initial REP + EF augmentation?
## Aims and objectives
The overarching goal of this study is to build the most effective adaptive implementation strategy involving two cutting-edge implementation strategies (REP and Facilitation) to improve practice-level uptake of EBPs and patient outcomes. The EBP, life goals (LG), is an evidence-based psychosocial treatment delivered in six individual or group sessions that was shown in seven randomized controlled trials to improve mental and physical health outcomes among patients with mood disorders [bib_ref] Long-term clinical outcomes from a randomized controlled trial of two implementation strategies..., Kilbourne [/bib_ref] [bib_ref] Improving medical and psychiatric outcomes among individuals with bipolar disorder: a randomized..., Kilbourne [/bib_ref] [bib_ref] Long-term effectiveness and cost of a systematic care program for bipolar disorder, Simon [/bib_ref] [bib_ref] Life Goals Collaborative Care for patients with bipolar disorder and cardiovascular disease..., Kilbourne [/bib_ref] [bib_ref] Pilot randomized trial of a cross-diagnosis collaborative care program for patients with..., Kilbourne [/bib_ref] [bib_ref] Cooperative Studies Program 430 Study Team: Collaborative care for bipolar disorder: part..., Bauer [/bib_ref]. We will use a sequential multiple assignment randomized trial (SMART) [bib_ref] A "SMART" design for building individualized treatment sequences, Lei [/bib_ref] [bib_ref] Designing a pilot sequential multiple assignment randomized trial for developing an adaptive..., Almirall [/bib_ref] design to build the adaptive intervention in which data on early response or non-response to REP will be used to determine the next implementation strategy.
## Primary study aim
The primary aim of this study is to determine in a SMART implementation trial, among patients with mood disorders in sites that do not exhibit response to REP alone after 6 months (i.e., <50% patients receive ≥3 LG sessions), the effect of adding an External and Internal Facilitator (REP + EF/IF) versus REP + EF on patient-level changes in mental health-related quality of life (MH-QOL; primary outcome), receipt of LG sessions, mood symptoms, implementation costs and organizational change (secondary outcomes) from month 6 to month 24.
## Secondary aim 1
The first secondary aim is to determine, among REP + EF sites that continue to exhibit non-response after an additional 6 months, the effect of continuing REP + EF vs. REP + EF/IF on patient-level changes in the primary and secondary outcomes from month 12 to month 24.
## Secondary aim 2
The second secondary aim is to estimate the site-level costs of REP + EF/IF compared to REP + EF.
## Secondary aim 3
The third secondary aim is to describe the implementation of EF and EF/IF at the site level, including interaction between the two roles and the specific strategies EFs and IFs use to facilitate LG uptake across different sites.
The result of this current SMART study will be an optimized, adaptive implementation strategy that that could be applied to healthcare systems to improve outcomes for patients with mental disorders.
# Methods
This cluster randomized SMART implementation trial [fig_ref] Figure 2: SMART trial design of REP combined with External [/fig_ref] involves community-based clinics (sites) from Michigan (MI) and Colorado (CO) that care for persons with mood disorders (depression or bipolar disorders). This study was reviewed and approved the local institutional review boards.
## Setting
Eighty community-based mental health or primary care clinics from the two states will be recruited to participate in the study based on lists of primary care and community mental health programs available from state organizations. Since the EBP to be implemented (LG) has been shown to be effective in improving outcomes across different settings (primary care, community mental health) [bib_ref] Improving medical and psychiatric outcomes among individuals with bipolar disorder: a randomized..., Kilbourne [/bib_ref] [bib_ref] Long-term effectiveness and cost of a systematic care program for bipolar disorder, Simon [/bib_ref] [bib_ref] Life Goals Collaborative Care for patients with bipolar disorder and cardiovascular disease..., Kilbourne [/bib_ref] [bib_ref] Cooperative Studies Program 430 Study Team: Collaborative care for bipolar disorder: part..., Bauer [/bib_ref] [bib_ref] Randomized controlled trial to assess reduction of cardiovascular disease risk in patients..., Kilbourne [/bib_ref] , a diverse array of sites will be recruited to maximize study generalizability.
## Sites
Site inclusion criteria include the following:
1. Community-based mental health or primary care clinic located in Michigan or Colorado with at least 100 unique patients diagnosed with or treated for mood disorders. 2. Availability of a bachelor's-or master's-level healthcare provider with a mental health background and experience with implementing individual or group sessions (core modality of LG) who can be trained to provide LG to up to 20 adult patients with mood disorders in the clinic in a 1-year period. 3. Availability of an employee at the site with direct reporting authority to the leadership of the site or parent practice organization who could serve as potential internal facilitator.
## Study design flow
Primary care or mental health outpatient clinics confirming eligibility using a standard organizational assessment and agreeing to participate will initially be offered REP for 6 months in order to implement LG [fig_ref] Figure 2: SMART trial design of REP combined with External [/fig_ref]. Based on previous research [bib_ref] Implementation of collaborative depression management at community-based primary care clinics: an evaluation, Bauer [/bib_ref] and preliminary data, it is expected that after 6 months of REP, at least 80% of sites will be non-responsive to REP. The primary focus of the randomized comparisons in this study are sites that are initially non-responsive to REP, defined based on our preliminary studies as <50% of previously identified patients receiving at least three LG sessions (≥3 out of 6) to achieve minimum clinically significant results [bib_ref] A brief measure for assessing generalized anxiety disorder: the GAD-7, Spitzer [/bib_ref] [bib_ref] The Internal State Scale: replication of its discriminating abilities in a multisite,..., Bauer [/bib_ref] [bib_ref] A manic-depressive symptom self-report in optical scanable format, Glick [/bib_ref] [bib_ref] Three questions can detect hazardous drinkers, Gordon [/bib_ref] [bib_ref] The World Health Organization Composite International Diagnostic Interview short-form (CIDI-SF), Kessler [/bib_ref]. Although not part of the primary randomized comparisons, sites that are responsive at 6 months will continue to be followed and outcomes will be assessed.
## Patients within sites
The unit of analysis is LG eligible patients within sites who are diagnosed with or treated for mood disorders.
Patients are identified during the first 6 months as part of REP, i.e., prior to site-level randomization (see step 1 of REP below). Patient-level primary and secondary outcomes will be assessed by independent evaluators (i.e., study associates who are not aware of the assignment to REP + EF or REP + EF/IF) across all sites prior to identifying site response/non-response status at month 6, prior to identifying site response/non-response status at month 12, and at months 18 and 24.
## Site randomization
The unit of intervention (randomization) is the site. Sites that are not responding to REP at month 6 will be randomized 1:1 by the study data analyst to receive additional External Facilitation (REP + EF) or External plus Internal Facilitation (REP + EF/IF). After another 6 months (at month 12), (i) REP + EF sites that are still non-responsive (defined as <50% patients receiving ≥3 life goals sessions) will be randomized 1:1 to either continue REP + EF or augmentation with IF (REP + EF/IF) for an additional 12 months, and (ii) intervention will be discontinued for sites that are responsive. The month 6 randomization will be stratified by state, practice type (primary care or mental health site) [bib_ref] The method of minimization for allocation to clinical trials. A review, Scott [/bib_ref] [bib_ref] Minimization: a new method of assigning patients to treatment and control groups, Taves [/bib_ref] , and by siteaverage baseline MH-QOL (e.g., low (<40) vs. high (≥40)) at baseline. The month 12 randomization will be stratified by state, practice type, and site-average MH-QOL at month 12. This will ensure that intervention groups are balanced for site variables that may correlate highly with outcomes. The study analyst will generate the stratified permuted-block random allocation lists (blocks of size 2, 4, and 6) using a computer program such as PROC PLAN in SAS. Sites are expected to become eligible and accrue in groups of 6-12 (staggered entry). A site is considered randomized once the study analyst informs the study coordinator of each site's random assignment.
## Evidence-based practice to be implemented
LG [bib_ref] Long-term clinical outcomes from a randomized controlled trial of two implementation strategies..., Kilbourne [/bib_ref] [bib_ref] Improving medical and psychiatric outcomes among individuals with bipolar disorder: a randomized..., Kilbourne [/bib_ref] [bib_ref] Long-term effectiveness and cost of a systematic care program for bipolar disorder, Simon [/bib_ref] [bib_ref] Life Goals Collaborative Care for patients with bipolar disorder and cardiovascular disease..., Kilbourne [/bib_ref] [bib_ref] Pilot randomized trial of a cross-diagnosis collaborative care program for patients with..., Kilbourne [/bib_ref] [bib_ref] Cooperative Studies Program 430 Study Team: Collaborative care for bipolar disorder: part..., Bauer [/bib_ref] [bib_ref] Randomized controlled trial to assess reduction of cardiovascular disease risk in patients..., Kilbourne [/bib_ref] ] is a psychosocial intervention for mood disorders delivered in six individual or group sessions [fig_ref] Table 1: Components of the life goals programSymptom profileCreate a personal symptom profile for... [/fig_ref]. Based on social cognitive theory [bib_ref] Service delivery in older patients with bipolar disorder: a review and development..., Kilbourne [/bib_ref] [bib_ref] The primacy of self-regulation in health promotion, Bandura [/bib_ref] , LG encourages active discussions focused on individuals' personal goals that are aligned with healthy behavior change and symptom management strategies.
LG was chosen to be the EBP to implement because mood disorders are common in both community-based primary care and mental health clinics [bib_ref] The descriptive epidemiology of commonly occurring mental disorders in the United States, Kessler [/bib_ref] [bib_ref] The prevalence and disability of bipolar spectrum disorders in the US population:..., Judd [/bib_ref] , are considered to be high-priority populations based on input from community partners, and because LG was shown to improve outcomes including mental health-related quality of life in this group [bib_ref] Improving medical and psychiatric outcomes among individuals with bipolar disorder: a randomized..., Kilbourne [/bib_ref] [bib_ref] Long-term effectiveness and cost of a systematic care program for bipolar disorder, Simon [/bib_ref] [bib_ref] Life Goals Collaborative Care for patients with bipolar disorder and cardiovascular disease..., Kilbourne [/bib_ref] [bib_ref] Cooperative Studies Program 430 Study Team: Collaborative care for bipolar disorder: part..., Bauer [/bib_ref] [bib_ref] Randomized controlled trial to assess reduction of cardiovascular disease risk in patients..., Kilbourne [/bib_ref]. Compared to usual care,
LG improved outcomes among a cross-diagnosis sample of community-based outpatients with mood disorders [bib_ref] Improving medical and psychiatric outcomes among individuals with bipolar disorder: a randomized..., Kilbourne [/bib_ref] [bib_ref] Long-term effectiveness and cost of a systematic care program for bipolar disorder, Simon [/bib_ref] [bib_ref] Life Goals Collaborative Care for patients with bipolar disorder and cardiovascular disease..., Kilbourne [/bib_ref] [bib_ref] Pilot randomized trial of a cross-diagnosis collaborative care program for patients with..., Kilbourne [/bib_ref] [bib_ref] Cooperative Studies Program 430 Study Team: Collaborative care for bipolar disorder: part..., Bauer [/bib_ref] [bib_ref] Randomized controlled trial to assess reduction of cardiovascular disease risk in patients..., Kilbourne [/bib_ref] , notably a four-point increase in mental and physical health-related quality of life scores based on the SF-12 (e.g., Cohen's D = .36) [bib_ref] Improving medical and psychiatric outcomes among individuals with bipolar disorder: a randomized..., Kilbourne [/bib_ref] [bib_ref] Long-term effectiveness and cost of a systematic care program for bipolar disorder, Simon [/bib_ref] [bib_ref] Cooperative Studies Program 430 Study Team: Collaborative care for bipolar disorder: part..., Bauer [/bib_ref].
LG has been shown to be equally effective in patients with co-occurring substance use and medical comorbidities [bib_ref] Long-term effectiveness and cost of a systematic care program for bipolar disorder, Simon [/bib_ref] [bib_ref] Life Goals Collaborative Care for patients with bipolar disorder and cardiovascular disease..., Kilbourne [/bib_ref] [bib_ref] Cooperative Studies Program 430 Study Team: Collaborative care for bipolar disorder: part..., Bauer [/bib_ref] [bib_ref] Randomized controlled trial to assess reduction of cardiovascular disease risk in patients..., Kilbourne [/bib_ref] [bib_ref] Is the collaborative chronic care model effective for patients with bipolar disorder..., Kilbourne [/bib_ref]. Community-based providers helped to adapt LG [bib_ref] Long-term effectiveness and cost of a systematic care program for bipolar disorder, Simon [/bib_ref] [bib_ref] Life Goals Collaborative Care for patients with bipolar disorder and cardiovascular disease..., Kilbourne [/bib_ref] [bib_ref] Pilot randomized trial of a cross-diagnosis collaborative care program for patients with..., Kilbourne [/bib_ref] , but as with many psychosocial EBPs, have not been widely implemented in smaller practices [bib_ref] Public-academic partnerships: evidence-based implementation: the role of sustained community-based practice and research..., Kilbourne [/bib_ref].
## Implementation strategies rep
All sites will receive REP ( project [bib_ref] Stall R: Implementing evidence-based interventions in health care: application of the replicating..., Kilbourne [/bib_ref] [bib_ref] Public-academic partnerships: evidence-based implementation: the role of sustained community-based practice and research..., Kilbourne [/bib_ref] [bib_ref] Replicating effective programs: HIV/AIDS prevention technology transfer, Neumann [/bib_ref] and includes initial marketing of the LG program by study investigators, package dissemination and training by an off-site trainer, as-needed program support in using LG (program assistance), and LG uptake monitoring for up to 6 months [bib_ref] Stall R: Implementing evidence-based interventions in health care: application of the replicating..., Kilbourne [/bib_ref] [bib_ref] Long-term clinical outcomes from a randomized controlled trial of two implementation strategies..., Kilbourne [/bib_ref]. The theories underlying REP [bib_ref] Orientation and training: preparing agency administrators and staff to replicate an HIV..., Adams [/bib_ref] [bib_ref] The role of technical assistance in the replication of effective HIV interventions, O'donnell [/bib_ref] include Rogers' Diffusion of Innovationsand Social Learning Theory. For this study, REP includes the following components:
Step 1: marketing and dissemination of LG package, in-service, and patient selection Regional in-services will first be provided by study investigators who will give an overview of LG including the evidence, and details on how to implement LG in their setting. Each site will designate at least one provider with a mental health background to implement LG ("LG provider"), and the LG package will be disseminated by the REP trainer to these providers. The LG REP package includes all of the components needed to implement LG, including the LG provider manual; a protocol for identifying patients to enroll in LG;
LG session scripts and focus points covered in each session in a semi-directed fashion; the registry template for tracking enrolled patients' progress in LG sessionbased personal goals, symptoms, and health behavior change; scripts for follow-up calls, patient workbooks, and an implementation manual describing logistics (e.g., identifying rooms if group sessions are used, identifying patients for LG, medical record templates for LG sessions, billing codes). Because LG effectiveness was already demonstrated in patients with mood disorders and that the primary goal of the study is to promote LG implementation, LG providers will identify and offer LG to patients at their site with mood disorders based on guidance from the REP trainer [bib_ref] Long-term clinical outcomes from a randomized controlled trial of two implementation strategies..., Kilbourne [/bib_ref] [bib_ref] Public-academic partnerships: evidence-based implementation: the role of sustained community-based practice and research..., Kilbourne [/bib_ref]. The trainer will work with LG providers to identify up to 30 patients within a 3-month period who are appropriate for the LG program using the following criteria that are included in the registry template:
1. adults 21 years or older with a diagnosis and current documentation of antidepressant or mood stabilizer for a mood disorder (depression or bipolar disorder) based on medical record review 2. not currently enrolled in residential treatment 3. can understand English and have no terminal illness or cognitive impairment that precludes participation in outpatient psychosocial treatment based on confirmation by the treating clinician.
Assuming a 33% refusal or ineligibility rate [bib_ref] Life Goals Collaborative Care for patients with bipolar disorder and cardiovascular disease..., Kilbourne [/bib_ref] , as well as a potential 12-month attrition rate of around 20% from prior studies [bib_ref] Life Goals Collaborative Care for patients with bipolar disorder and cardiovascular disease..., Kilbourne [/bib_ref] , LG providers will be expected to identify and initially offer LG to 30 patients, assuming 20 will initially participate, with 16 to complete 12-month follow-ups.
Step 2: REP LG training and provider competency
LG providers will undergo a 1-day training program that has been provided to over 200 clinicians nationally [bib_ref] Long-term clinical outcomes from a randomized controlled trial of two implementation strategies..., Kilbourne [/bib_ref] [bib_ref] Cooperative Studies Program 430 Study Team: Collaborative care for bipolar disorder: part..., Bauer [/bib_ref] [bib_ref] Design and implementation of a randomized trial evaluating systematic care for bipolar..., Simon [/bib_ref]. The LG trainer will first provide an orientation to the evidence behind LG and core elements and then a step-by-step walk-through of LG components. The trainer will then demonstrate each LG session and follow-up contact procedures, as well as the standardized criteria for identifying patients who are appropriate for LG in routine practice.
LG providers will break into groups and practice each component. After identifying eligible patients, LG providers will make an initial contact to each patient, introduce the LG program, and schedule in-person group or individual sessions. Individual sessions typically last 50 min each, while group sessions typically last around 90 min to allow for additional discussions between group members. Participants can make up sessions over the phone if they are unable to make in-person sessions. In the sessions, the LG provider will encourage active discussions that progressively have participants identify a personal symptom profile and triggers to the mood symptom or episode and develop an activity plan for identifying warning signs of symptoms and an activity plan for adopting a specific Randomize to non-responding sites
Step 1: Market LG, disseminate LG package. (a) Marketing (pre-implementation): hold pre-implementation meetings with site representatives to describe the benefits of LG and identify potential LG providers and internal facilitators (IF). (b) Dissemination: regional in-services to disseminate
LG implementation guide to providers, schedule trainings.
## √ √ √
Step 2: Train site providers in LG. Conduct regionalized 8-h training for LG providers covering LG session content and delivery via LG website as well as patient tracking and monitoring over time.
## √ √ √
Step 3: As-needed program assistance and LG uptake monitoring via secure web-based reporting sheets. Ad hoc program support available by study program support assistant.
## √ √ √
## Rep + external facilitator (rep + ef)
Step 1: Initiation and benchmarking: EF contacts each LG provider, works with LG provider to identify potential barriers and facilitators to LG uptake from the LG provider's perspective, and sets measurable goals to uptake.
## √ √
Step 2: Coaching: EF makes calls on a biweekly basis to site's LG providers to develop rapport and provide specific guidance on overcoming barriers in implementing LG components by aligning LG's strengths with LG's available influence at the site. If needed, EF refers LG provider to study LG program support assistant.
## √ √
Step 3: Public recognition of "bright spots" (high-performing sites): EF provides state-specific report on sites' progress and disseminates LG provider success stories.
## √ √
## Rep + external and internal facilitator (rep + ef/if)
Step 1: Initiation and benchmarking: EF contacts each LG provider and holds call with LG provider and IF to give background on LG, review potential barriers and facilitators to LG uptake, and set measurable goals to LG uptake.
## √
Step 2: Leveraging: IF meets with LG provider biweekly, identifies local site priorities per leadership input, identifies other LG program champions, and helps LG provider summarize and describe added value of LG to leadership and site providers (e.g., consistency with other initiatives, support from leadership).
## √
Step 3: Coaching: IF, EF, and LG hold biweekly calls to develop rapport; EF provides guidance to LG on overcoming specific barriers to LG uptake by aligning LG's strengths with LG's available influence at the site (EF), and IF aligns goals of LG provider with existing site priorities, based on feedback from site leadership. If needed, EF refers LG provider to study LG program support assistant.
## √
Step 4: Ongoing marketing: IF, leadership, and LG provider summarize progress and develop business and sustainability plans.
√ health behavior to mitigate symptoms and promote wellness. Participants will be given a workbook with exercises on behavior change goals, symptom assessments, and coping strategies. The LG provider will also make individual contacts to patients after the end of the sessions on a regular basis to review symptoms and behavior change.
LG providers are trained to handle patients with elevated symptoms or suicidal ideation as part of their clinical responsibilities.
Step 3: As-needed REP program assistance and LG monitoring
The final phase of REP consists of as-needed program assistance provided to LG clinicians who contact the LG program support specialist. In addition, each LG provider is sent a standard monitoring form on a biweekly basis for up to 6 months to have them record the number of patients approached and number receiving each LG session. Monitoring forms will be used to assess non-response across sites and will be corroborated based on patient self-reported LG use from the follow-up surveys (see data collection below). Sites will receive feedback reports from the program support assistant that includes performance on uptake measures (i.e., how many enrolled, how many sessions) in comparison to other sites. Also, sites will receive periodic (quarterly) newsletters about the study progress.
## Facilitation
The Facilitation implementation strategy will consist of EF and IF. Based on the Promoting Action on Research Implementation in Health Services (PARiHS) Framework [bib_ref] A guide for applying a revised version of the PARIHS framework for..., Stetler [/bib_ref] [bib_ref] Evaluating the successful implementation of evidence into practice using the PARiHS framework:..., Kitson [/bib_ref] [bib_ref] Enabling the implementation of evidence based practice: a conceptual framework, Kitson [/bib_ref] [bib_ref] Getting evidence into practice: the role and function of facilitation, Harvey [/bib_ref] [bib_ref] Role of "external facilitation" in implementation of research findings: a qualitative evaluation..., Stetler [/bib_ref] , facilitation is defined as the process of interactive problem solving and support that occurs in the context of a recognized need for improvement and a supportive interpersonal relationship [bib_ref] Research & services partnerships: lessons learned through a national partnership between clinical..., Kirchner [/bib_ref] [bib_ref] Role of "external facilitation" in implementation of research findings: a qualitative evaluation..., Stetler [/bib_ref].
LG providers from sites that are not responsive to REP in 6 months after initiating LG at their site will be randomized to receive REP + EF or REP + EF/IF. Sites receiving REP + EF will be contacted by the EF, who resides off site. The EF will contact individual providers from each site on a regular basis and conduct 1-hour phone calls to provide guidance on implementing LG components. Notably, the EF will be trained to identify individual strengths of each LG provider and help leverage those strengths with available influence the LG provider has at the site. The EF will also set measurable objectives in implementing LG (e.g., number of patients completing at least one group session), review implementation progress, and where appropriate, refer the provider to the study program support assistant for specific guidance on LG use.
LG providers from sites that are not responsive to REP initially and randomized to receive REP + EF/IF will also be contacted by the EF, and the study team will also contact the site to identify an IF at the time of randomization. The IF will be an individual supervising the LG provider in some capacity. In this arm, LG providers will have regular calls with the EF and IF, as well as meet with IF one on one on a regular basis. As with the REP + EF arm, EFs will identify the individual strengths of each LG provider and help leverage those strengths with available influence the LG provider has at the site. In addition, the IF at each site will meet with the LG provider and will use their internal knowledge of the site to help the LG provider identify opportunities to align LG program goals with existing site priorities. Sites randomized to receive IF will receive up to $5,500 to cover IF-related time.
## Ensuring fidelity to implementation strategies
Fidelity monitoring will be used to assess whether each site is receiving the core components of each implementation strategy (REP, EF, and IF) and to ensure that there is no contamination across roles. Data from LG provider logs and EF/IF activities completed by study staff will be used to ascertain fidelity within each 6-month period of implementation strategy exposure using established checklists. All sites will get the same LG REP package, and LG training will be conducted by the study trainer who will hold regional trainings in each state. The EF will be trained by study investigators based on a 2-day training program developed for national roll-out of both the REP and EF/IF programs.
Fidelity to REP is defined based on number of sites receiving an LG package, number of providers completing the 1-day LG training program, and number of completed lists of patients receiving LG. Fidelity to the EF role will be defined using the following criteria based on the components outlined in: 1) number of completed calls by the EF with LG providers at each site; 2) number of documented barriers, facilitators, and specific measurable goals to LG uptake; and 3) documentation of LG provider strengths and available opportunities to influence site activities and overcome barriers. Fidelity to the IF role is defined as 1) number of meetings with the LG provider and EF, 2) number of meetings IF and LG provider have with site leadership, 3) number of documented opportunities to leverage LG uptake with existing site priority goals, and 4) development of a strategic plan to implement LG by the IF.
## Primary outcomes and measures
Study staff members (not site employees) will collect patient and provider data to ensure consistency of outcomes data collection over time. Outcomes assessors will be blinded to implementation condition. The assessment package previously implemented by study investigators was informed by the RE-AIM framework for evaluating implementation of EBPs and includes measures used in routine clinical care [bib_ref] Evaluating the public health impact of health promotion interventions: the RE-AIM framework, Glasgow [/bib_ref] [bib_ref] Implementation outcomes of evidence-based quality improvement for depression in VA community based..., Fortney [/bib_ref]. Key measures include patient-level outcomes, LG fidelity, organizational factors, and REP, EF, and IF activities [bib_ref] Cooperative Studies Program 430 Study Team: Collaborative care for bipolar disorder: part..., Bauer [/bib_ref] [bib_ref] VA Cooperative Study #430 Team: Medical comorbidity and health-related quality of life..., Fenn [/bib_ref] [bib_ref] Therapeutic drug and cardiovascular disease risk monitoring in patients with bipolar disorder, Kilbourne [/bib_ref].
## Patient data collection
Lists of patients and contact information that were identified by the LG providers during pre-randomization will be sent to study staff members who will conduct independent clinical assessments. As the study is focused on implementation processes for an evidence-based treatment that does not involve randomizing at the patient level, local institutional review boards (IRBs) considered the protocol to not fall under research, and no patient informed consent is required for the clinical assessments. Staff members will contact patients and describe the purpose of the phone-based clinical assessment, which will be conducted at baseline, 6, 12, 18, and 24 months later.
The clinical assessment will include previously established measures used in routine care, including the SF-12 for MH-QOL [bib_ref] A 12-item short-form health survey: construction of scales and preliminary tests of..., Ware [/bib_ref] (primary outcome), the Patient Health Questionnaire (PHQ-9) for mood symptoms [bib_ref] The PHQ-9: validity of a brief depression severity measure, Kroenke [/bib_ref] [bib_ref] Validation and utility of a self-report version of PRIME-MD: the PHQ primary..., Spitzer [/bib_ref] , the World Health Organization Disability Assessment Scale (WHO-DAS) for functional impairment [bib_ref] Global "burden of disease"-study for psychiatric disorders, Ustun [/bib_ref] [bib_ref] On the development and psychometric testing of the WHO screening instrument to..., Rehm [/bib_ref] , GAD-7 for anxiety symptoms [bib_ref] A brief measure for assessing generalized anxiety disorder: the GAD-7, Spitzer [/bib_ref] , and the Internal State Scale for manic symptoms [bib_ref] The Internal State Scale: replication of its discriminating abilities in a multisite,..., Bauer [/bib_ref] [bib_ref] A manic-depressive symptom self-report in optical scanable format, Glick [/bib_ref] (secondary and exploratory outcomes). Additional information [bib_ref] Long-term clinical outcomes from a randomized controlled trial of two implementation strategies..., Kilbourne [/bib_ref] [bib_ref] Three questions can detect hazardous drinkers, Gordon [/bib_ref] [bib_ref] The World Health Organization Composite International Diagnostic Interview short-form (CIDI-SF), Kessler [/bib_ref] [bib_ref] Questionnaires from the National Health Interview Survey, Chyba [/bib_ref] [bib_ref] Algorithm for assessing patients' adherence to oral hypoglycemic medication, Kilbourne [/bib_ref] including participant demographics will also be ascertained from the clinical assessment during each follow-up period.
## Provider and organizational surveys
Study staff members will survey providers from the participating sites, including the site clinical director and LG providers on organizational factors that might impact implementation and outcomes. As the study's principal institutional review board (University of Michigan Medical School Institutional Review Board -IRBMED) considered this study to fall under quality improvement activities, no informed consent is required from providers or clinic directors. Clinic directors will be contacted prior to the initiation of the REP in-service to complete an initial site organizational survey [bib_ref] Enhanced fidelity to treatment for bipolar disorder: results from a randomized controlled..., Waxmonsky [/bib_ref] [bib_ref] Long-term clinical outcomes from a randomized controlled trial of two implementation strategies..., Kilbourne [/bib_ref] that includes questions on resources, staff turnover, and integrated care [bib_ref] Reynolds CF 3rd: Organizational factors and depression management in community-based primary care..., Post [/bib_ref]. A longitudinal organizational assessment will also be given to the clinic directors and LG providers prior to the initiation of REP, then again at 12 and 24 months later to assess changes in organizational features that might be impacted by REP and Facilitation. The assessment includes two established questionnaires focused on organizational capacity to implement EBPs. The Implementation Leadership Scale (ILS) assesses the degree of organizational support for EBPs, and the Implementation Climate Scale (ICS) assesses staff's impressions on expectations and support for effective EBP implementation through their policies, procedure, and behaviors [bib_ref] The Implementation Leadership Scale (ILS): development of a brief measure of unit..., Aarons [/bib_ref].
## Life goals fidelity
Because this study is designed to assess real-world implementation, minimally invasive measures to assess LG fidelity will be based on provider logs sent via a secure web-based form to study staff on a weekly basis. Study staff members will also collect confirmatory data from patient surveys on receipt of LG sessions at the patient level. The fidelity monitor calculates a total score based on number of sessions completed by each patient and the percentage completing five of six sessions. Average patient completion of sessions of ≥75% was associated with improved mental health-related quality of life [bib_ref] Cooperative Studies Program 430 Study Team: Collaborative care for bipolar disorder: part..., Bauer [/bib_ref].
## Secondary aims: facilitation implementation and costs
Data on the implementation of EF and EF/IF will be collected by study staff members using previously established assessments for monitoring implementation activities [bib_ref] Cluster randomized adaptive implementation trial comparing a standard versus enhanced implementation intervention..., Kilbourne [/bib_ref] [bib_ref] Enhanced fidelity to treatment for bipolar disorder: results from a randomized controlled..., Waxmonsky [/bib_ref] [bib_ref] Leading from the middle: replication of a re-engagement program for veterans with..., Goodrich [/bib_ref]. Study staff members will interview LG providers, EF, IF, and site clinic leaders regarding barriers and facilitators to LG, REP, and EF/IF strategies, interactions between the EF and IFs, and specific strategies EFs and IFs use to facilitate LG uptake across different sites.
Implementation strategy costs will be ascertained by the study staff members based on study staff logs of REP activities, and LG provider and EF and IF logs based on a standardized list of activity categories [fig_ref] Table 3: Summary of specific implementation strategy activities [/fig_ref]. EF and IF time is likely to account for the vast majority of costs associated with the implementation strategies, including site employees (LG provider time on training and LG implementation, IF time at meetings) as well as project staff (REP package development, training/program assistance, EF Facilitator time in site follow-ups). All costs will be multiplied by personnel wage rates including fringe. Patient-level costs will also be estimated from self-reported utilization survey data on inpatient, ER, and outpatient use. Costs will be assigned using Current Procedural Terminology (CPT) codes, and a relative value unit (RVU) weight will allow us to use the Medicare Fee Schedule to calculate a standardized cost in US dollars for each service, adjusted for annual levels of inflation using the consumer price index.
## Analyses
Intent to treat analyses will be performed. The primary analysis will compare interventions in non-responding sites beginning with REP + EF/IF versus interventions beginning with REP + EF on longitudinal patient-level change in number of LG sessions received, SF-12 mental health-related quality of life scores, and PHQ-9 scores. This analysis is a two-sample comparison of cells A + B vs. C + D + E [fig_ref] Figure 2: SMART trial design of REP combined with External [/fig_ref]. For this analysis, the longitudinal outcomes will be measured at months 6 (prerandomization), 12, 18, and 24. The primary contrast is the between groups difference in change from month 6 to month 18. The follow-up contrast at month 24 will also be examined in this and all subsequent analyses.
The primary aim analysis for health-related quality of life (the primary longitudinal outcome) will use linear mixed models (LMM;, also known as random effects models. The unit of analysis is the individual patient within a site (recall that approximately 20 individuals will be identified prior to randomization). LMMs use all available measurements, allowing individuals to have an unequal number of longitudinal observations and producing unbiased parameter estimates as long as unobserved values are missing at random. The analysis will fit a threelevel (repeated measures for each individual clustered within site) LMM with fixed effects for the intercept, time, group, and a group-by-time interaction term, where group is an indicator of REP + EF/IF vs. REP + EF. The LMM will include random effects for site and time, an unstructured within-person correlation structure for the residual errors, and it will adjust for state and type of practice (primary care or mental health site). LMMs similar to the above will be conducted for the secondary patient-level outcomes: change in number of LG sessions received, functional impairment, and mood symptoms.
Secondary aim analyses will be conducted to determine whether continuing REP + EF vs. augmenting with This secondary analysis is a comparison of cells B vs. C [fig_ref] Figure 2: SMART trial design of REP combined with External [/fig_ref]. Outcomes will be examined using an LMM similar to that described above, except (a) including only the subset of sites that do not respond at month 12 to REP + EF, (b) using monthly longitudinal outcomes from month 12 to month 24, and (c) the LMM will use dummy indicators for time (i.e., time-saturated model since there are only three measurement times for each longitudinal outcome). The longitudinal course of discontinuing REP + EF (or REP + EF/IF) will also be examined at month 12 among sites that are responsive at month 12.
Other secondary aim analyses include comparison of EF and EF + IF costs over time using a standardized list of REP and EF/IF activities [fig_ref] Table 3: Summary of specific implementation strategy activities [/fig_ref] , as well as assessment of organizational change over time [bib_ref] Current directions in mediation analysis, Mackinnon [/bib_ref]. Additional exploratory analyses will also compare the different adaptive interventions embedded in the SMART design [fig_ref] Figure 2: SMART trial design of REP combined with External [/fig_ref] in terms of changes in longitudinal outcomes using methods based on Robins and colleagues [bib_ref] Dynamic regime marginal structural mean models for estimation of optimal dynamic treatment..., Orellana [/bib_ref] [bib_ref] Causal models for estimating the effects of weight gain on mortality, Robins [/bib_ref].
Missing values may occur in outcomes due to dropout or inability to reach patients for follow-up (anticipated 10% attrition). A thorough investigation of mechanisms for missing data will be carried out and will be dealt with using multiple imputation procedures. In stability analyses, data will be analyzed with and without the multiple imputation strategy. Any discrepancies will be reported and carefully examined.
## Sample size and power
The estimated sample size for this study is based on a comparison of between groups (REP + EF/IF versus REP + EF) in changes in the most conservative effect size estimate (mental health-related quality of life scores) between month 6 and month 18 (Cohen's D = .23). This is a twosample comparison of patients within sites in cells A + B + C versus D + E [fig_ref] Figure 2: SMART trial design of REP combined with External [/fig_ref]. To account for the between-site variation induced by the within-site correlation in quality of life outcomes, we inflate the variance term in the standard sample size formula by 1 + (n − 1) × ICC, where ICC is the site inter-class correlation coefficient for the SF-12 mental health-related quality of life component score. Based on our previous randomized controlled implementation trial [bib_ref] Cluster randomized adaptive implementation trial comparing a standard versus enhanced implementation intervention..., Kilbourne [/bib_ref] [bib_ref] Enhanced fidelity to treatment for bipolar disorder: results from a randomized controlled..., Waxmonsky [/bib_ref] , the site ICC was estimated at .01. Using a two-sided, two-sample t test, a Type I error rate of 5%, and with 60 sites (30 sites randomized to REP + EF/IF versus 30 sites randomized to REP + EF) and 16 patients per site after accounting for attrition, we will have 94% power to detect clinically significant changes in quality of life scores, assuming an ICC = .01. Based on previous data, with SD = 8.35 for the MH-QOL component score, this effect size corresponds to being able to detect a clinically meaningful difference of at least 4 units in MH-QOL score.
## Trial status
Sites will be identified and participation confirmed by October 2014. Site training and REP procedures will begin in the fall of 2014. A timeline of implementation activities is provided in [fig_ref] Table 3: Summary of specific implementation strategy activities [/fig_ref].
# Discussion
To date, this will be one of the largest randomized controlled trials to develop a site-level adaptive implementation strategy. This is also one of the first studies to test the augmentation of an established implementation strategy (REP) using REP + EF/IF or REP + EF alone among sites that exhibit non-response after 6 months. This study will determine whether augmentation of the implementation strategy is needed (e.g., REP + EF/IF) or whether in some circumstances withholding augmentation may result in a delayed implementation effect of REP + EF alone among non-responsive sites. Second, this study will utilize a novel SMART design developed by the study investigators [bib_ref] Designing a pilot sequential multiple assignment randomized trial for developing an adaptive..., Almirall [/bib_ref] to accomplish this implementation trial. SMARTs allow efficient comparison of the overall impact of receiving different intervention augmentation strategies over time and incremental costs of one intervention strategy over another in improved outcomes among non-responsive sites. In addition, this design allows the potential to determine the added value of implementation strategies applied to real-world treatment settings.
There is a paucity of research on effective implementation strategy that improves EBP uptake and ultimately patient outcomes, notably in smaller, community-based, safety net practices, which serve a substantial proportion of patients with mood disorders. There have been few rigorous trials of implementation strategy to promote the uptake of EBPs in community-based practices [bib_ref] Implementation research in mental health services: an emerging science with conceptual, methodological,..., Proctor [/bib_ref] [bib_ref] Stall R: Implementing evidence-based interventions in health care: application of the replicating..., Kilbourne [/bib_ref]. Previous implementation studies have focused on highly organized practices such as the VA or staff-model HMOs and mainly involved intensive implementation strategy that might not be feasible to apply to smaller, lower-resourced practices. Among the frameworks that guide implementation efforts (e.g., [bib_ref] Fostering implementation of health services research findings into practice: a consolidated framework..., Damschroder [/bib_ref] , few have been operationalized sufficiently to enable community-based practices to enhance EBP uptake.
Adaptive Implementation of Effective Programs Trial (ADEPT) represents a growing cadre of rigorous yet realworld trials that involve site-level randomization of quality improvement strategies to augment the implementation of evidence-based practices. Specifically, informed consent from participants is not required per the recommendation of the IRB because the focus of ADEPT is on the use of implementation strategies at the site level, above and beyond available resources to disseminate effective programs, in order to assist existing providers in adopting an evidencebased practice. Moreover, the delivery of the evidencebased practice to patients is not altered or controlled at the sites by the study team, and patient clinical assessment includes measures that are used across many of the practice settings as part of routine clinical care. These types of studies have high generalizability to real-world settings because they reflect the pragmatic strategies needed to improve healthcare delivery and foster learning health systems [bib_ref] Proactive tobacco treatment and population-level cessation: a pragmatic randomized clinical trial, Fu [/bib_ref] [bib_ref] Informed consent in randomized quality improvement trials: a critical barrier for learning..., Pletcher [/bib_ref]. ADEPT evaluates the use of an evidence-based intervention (i.e., life goals) that is superior to usual care from the patient perspective while testing an implementation strategy (facilitation) that is beneficial to providers and practices by improving their ability to adopt such evidence-based practices [bib_ref] Informed consent in randomized quality improvement trials: a critical barrier for learning..., Pletcher [/bib_ref]. Designation as a QI trial does not remove IRB regulatory oversight but instead reduces the burden on researchers (costs) and patients (time) to participate in written informed consent and formal survey assessment meetings which do not reflect real-world care but instead introduce potential patient selection biases that undermine patient participation in the intervention [bib_ref] Proactive tobacco treatment and population-level cessation: a pragmatic randomized clinical trial, Fu [/bib_ref] [bib_ref] Informed consent in randomized quality improvement trials: a critical barrier for learning..., Pletcher [/bib_ref].
REP uses key tactical strategies that can promote effective EBP adoption in community-based providers. However, these providers may face multiple organizational barriers at their sites including competing demands or lack of experience in garnering leadership buy-in that are beyond the scope of toolkit dissemination, structured training, or initial performance feedback [bib_ref] From silos to bridges: meeting the general health care needs of adults..., Horvitz-Lennon [/bib_ref] [bib_ref] Implementation of evidence-based practice and organizational performance, Hovmand [/bib_ref] [bib_ref] The sustainability of evidence-based practices in routine mental health agencies, Swain [/bib_ref] [bib_ref] Interventions in organizational and community context: a framework for building evidence on..., Mendel [/bib_ref]. Addressing these barriers may require strategic thinking and multilevel organizational alliances [bib_ref] Emerging models of depression care: multi-level ('6 P') strategies, Pincus [/bib_ref] [bib_ref] Aligning incentives in the treatment of depression in primary care with evidence-based..., Frank [/bib_ref] that engage providers in supporting and owning the implementation [bib_ref] Medicare's flagship test of pay-forperformance did not spur more rapid quality improvement..., Ryan [/bib_ref] [bib_ref] The effect of pay-forperformance in hospitals: lessons for quality improvement, Werner [/bib_ref] [bib_ref] The use of physician financial incentives and feedback to improve pediatric preventive..., Hillman [/bib_ref]. While leadership support in EBP adoption is important [bib_ref] Creating a quality-improvement dialogue: utilizing knowledge from frontline staff, managers, and experts..., Parker [/bib_ref] [bib_ref] Understanding team-based quality improvement for depression in primary care, Rubenstein [/bib_ref] [bib_ref] How can we increase translation of research into practice? Types of evidence..., Glasgow [/bib_ref] , involvement and buy-in from frontline providers are also crucial to EBP sustainability [bib_ref] Leading from the middle: replication of a re-engagement program for veterans with..., Goodrich [/bib_ref] [bib_ref] Creating a quality-improvement dialogue: utilizing knowledge from frontline staff, managers, and experts..., Parker [/bib_ref] [bib_ref] Uncovering middle managers' role in healthcare innovation implementation, Birken [/bib_ref] [bib_ref] Balancing participation and expertise: a comparison of locally and centrally managed health..., Parker [/bib_ref] [bib_ref] Diffusion of innovations in service organizations: systematic review and recommendations, Greenhalgh [/bib_ref] [bib_ref] Revisiting interaction in knowledge translation, Ginsburg [/bib_ref] [bib_ref] Sustainability of a practice-individualized preventive service delivery intervention, Stange [/bib_ref] [bib_ref] From understanding health care provider behavior to improving health care: the QUERI..., Rubenstein [/bib_ref] [bib_ref] Interventions to improve the delivery of preventive services in primary care, Hulscher [/bib_ref]. Hence, REP may need to be augmented to address these organizational barriers to adoption [bib_ref] A randomized controlled trial of CQI teams and academic detailing: can they..., Goldberg [/bib_ref] [bib_ref] A CQI intervention to change the care of depression: a controlled study, Solberg [/bib_ref] and to ultimately show value of the EBP through the triple aim (improving patient outcomes, experience, lowering costs).
Our study will determine the added value of augmenting implementation programs through Facilitation on EBP uptake, sustainability, and patient outcomes, ultimately determining the public health impact of implementation strategies for persons with mental disorders. Akin to stepped treatment for patients, adaptive SMART designs provide a tailored and potentially cost-effective approach to augmenting standard implementation programs based on the specific needs of a site. This approach is needed to better understand how to improve EBP uptake and patient mental health outcomes using Facilitation, whereby REP is augmented given early signs of site non-response. While REP might be sufficient for some sites in adopting EBPs, our preliminary studies and prior research [bib_ref] Transfer of research-based HIV prevention interventions to community service providers: fidelity and..., Kelly [/bib_ref] [bib_ref] Protocol: Adaptive Implementation of Effective Programs Trial (ADEPT): cluster randomized SMART trial..., Kelly [/bib_ref] suggest that the majority will need additional assistance. Moreover, sites initially not responding to REP (i.e., limited adoption of EBPs) were unlikely to do so in the future. IF is more expensive for sites to implement, as the additional customization and relationship building across individual sites due to variations in culture, climate, and capacity can be time consuming [bib_ref] Balancing participation and expertise: a comparison of locally and centrally managed health..., Parker [/bib_ref]. It is also unclear how long Facilitation is needed to achieve its effect [bib_ref] Interventions in organizational and community context: a framework for building evidence on..., Mendel [/bib_ref] [bib_ref] Creating a quality-improvement dialogue: utilizing knowledge from frontline staff, managers, and experts..., Parker [/bib_ref] [bib_ref] From understanding health care provider behavior to improving health care: the QUERI..., Rubenstein [/bib_ref]. Adaptive interventions involving REP and augmentation through Facilitation are needed to determine the added value of REP + EF/IF on improved patient outcomes, how long Facilitation need to be continued to achieve outcomes, and the cost-effectiveness of REP + EF/IF.
While this study has a number of strengths, including the use of a novel implementation study design within community-based practices, there are key limitations that need to be considered. First, there is the potential for self-selection of site early adopters which can limit generalizability. The involvement of sites most willing to participate in studies is inevitable with any implementation design. We have mitigated this potential selection effect by inviting all sites from networks of practices through state associations from across the country. In addition, LG uptake and fidelity measures do not include direct observations of sessions. In order to conduct an implementation trial that reflected real-world considerations, we chose to use indirect methods to measure LG uptake, and cost considerations precluded us from conducting in-person observations. Nonetheless, because the study was considered non-regulated (non-research), there is potential for increased generalizability especially if patient dropout is minimized because they will be assessed as part of routine clinical care. In addition, there is a chance for contamination between the EF and IF implementation strategies, especially if a less engaged IF might promote the EF to become more active in assisting the LG provider in identifying barriers and facilitators to program uptake. Finally, not all sites from each state participated, and those what are enrolled may not be representative of community-based practices nationwide.
# Conclusions
The results of this study will inform the implementation of evidence-based practices across sites in two states, as well as how to conduct practical implementation studies in real-world healthcare settings. This study also sets the stage for determining the added value of more intensive implementation strategies within sites that need additional support to promote the uptake of EBPs. Ultimately, adaptive implementation strategies may produce more relevant, rapid, and generalizable results by more quickly validating or rejecting new implementation strategies, thus enhancing the efficiency and sustainability of implementation research and potentially lead to the rollout of more cost-efficient implementation strategies.
[fig] Figure 1: Replicating Effective Programs (REP) and Internal/External Facilitation (EF/IF). [/fig]
[fig] Figure 2: SMART trial design of REP combined with External (EF, REP + EF) and Internal Facilitation (IF, REP + EF/IF). [/fig]
[table] Table 1: Components of the life goals programSymptom profileCreate a personal symptom profile for a specific mental health condition with a focus on identifying early warning signs and ways to use this knowledge to bolster self-management skills [/table]
[table] Table 2: Description of REP, REP + EF, and REP + EF/IF implementation strategies to enhance the uptake of the life goals (LG) evidence-based practice [/table]
[table] Table 3: Summary of specific implementation strategy activities [/table]
|
Totally Tubeless Outpatient Percutaneous Nephrolithotomy: Initial Case Report
We report the first case of totally tubeless outpatient percutaneous nephrolithotomy (PCNL). Our patient was discharged home safely less than 4 hours following uncomplicated PCNL with no nephrostomy tube, ureteral stent, or urethral catheter. Followup the next day in clinic confirmed that the procedure was successful, as the patient was clinically well and stone free. To our knowledge, this is the first case report of totally tubeless (no nephrostomy, no ureteral stent) PCNL performed on a truly outpatient basis.
# Introduction
Most endourologic procedures can be done on an outpatient basis, yet patients requiring PCNL are routinely admitted to hospital postoperatively. With the advent of tubeless PCNL [bib_ref] Tubeless percutaneous renal surgery, Bellman [/bib_ref] [bib_ref] Safety and efficacy of tubeless percutaneous nephrostolithotomy, Delnay [/bib_ref] [bib_ref] Almost totally tubeless percutaneous nephrolithotomy: further evolution of the technique, Goh [/bib_ref] [bib_ref] Tubeless percutaneous renal surgery: review of first 112 patients, Limb [/bib_ref] and more recently, totally tubeless PCNL [bib_ref] Effectiveness of totally tubeless percutaneous nephrolithotomy in selected patients: a prospective randomized..., Istanbulluoglu [/bib_ref] [bib_ref] Totally tubeless percutaneous nephrolithotomy, Aghamir [/bib_ref] [bib_ref] Tubeless and stentless percutaneous nephrolithotomy, Gupta [/bib_ref] [bib_ref] Totally tubeless percutaneous nephrolithotomy, Crook [/bib_ref] , postoperative length of stay following PCNL has decreased. Singh et al. have reported "ambulatory" PCNL [bib_ref] Ambulatory PCNL" (tubeless PCNL under regional anesthesia)-a preliminary report of 10 cases, Singh [/bib_ref] , however when one looks closely at their report, each patient was discharged home the following day. We recently reported a successful case series of tubeless outpatient PCNL in three cases, all of which involved ureteral stent placement. This case report represents the first known case of totally tubeless (no percutaneous nephrostomy tube or ureteral stent) PCNL performed on an outpatient basis.
## Case report
A generally healthy 58-year-old woman presented with a 3month history of intermittent right flank pain and several episodes of intermittent gross hematuria. On physical examination, she appeared well and her stated age. Pre-operative blood pressure was 100/72 mmHg, and pre-operative heart rate was 78 beats per minute. Urinalysis revealed microhematuria, and urine culture was negative. Pre-operative bloodwork revealed a hemoglobin of 147 g/L and a creatinine of 56 µmol/L. KUB and CT scan ) revealed an 11 × 5 mm right lower pole calculus with no hydronephrosis. At our institution, all patients with lower pole stones greater than 1 cm in maximum diameter are offered shock wave lithotripsy, ureterorenoscopic laser lithotripsy, and PCNL. The patient stated very clearly that she was not interested in passing stone fragments, and she wanted the most definitive procedure that would best ensure complete removal of her stone in a single setting, a PCNL. Informed consent was obtained for PCNL, and she was treated with oral levofloxacin 500 mg daily for seven days prior to the operative procedure.
Intraoperatively, intravenous Ampicillin 1 g and Gentamicin 80 mg were administered. The procedure was performed in the prone position. Flexible cystourethroscopy was normal. Retrograde pyelogram showed the stone in a lower calyx. Percutaneous access was achieved at the tip of an interpolar calyx because its infundibulum was wide and there was a wide angle between it and the lower pole calyx harbouring the stone. The tract was dilated using a 30 French balloon dilator. Using a 24 French rigid nephroscope, the stone was readily identified and removed intact using duckbill graspers. There was negligible bleeding during the procedure and negligible bleeding on inspection of the tract at the end of the procedure. Therefore, decision was made to not leave any indwelling nephrostomy tube or ureteral stent.
## Advances in urology
Total operative time, including cystoscopy and retrograde pyelogram, was 47 minutes.
During the early postoperative period, her vital signs were stable, and there was minimal pain and hematuria. The urethral catheter was removed prior to discharge, and the voided urine was pink in colour. She was discharged home 72 minutes after being discharged from the recovery room and less than four hours after leaving the operating room. At time of discharge, arrangements were made for a follow-up telephone call later that evening and postoperative assessment the following day as an outpatient. Assessment on postoperative day no. 1 included bloodwork, noncontrast CT scan and a clinic appointment with the attending urologist.
The patient had no complaints at telephone followup and clinic follow-up on the first postoperative day. In the first 24 hours following hospital discharge, the patient had very little discomfort, as evidenced by the fact that she took only four acetaminophen 500 mg tablets and absolutely no narcotics. On examination, she appeared well, her vitals signs were stable, and her incision was clean and intact. A urine sample revealed minimal hematuria [fig_ref] Figure 2: Urine sample on postoperative day no [/fig_ref]. Her hemoglobin was 127 g/L, and her creatinine was 73 µmol/L. A CT scan showed a normal kidney with minimal postsurgical changes in the tract but no evidence of any significant hematoma, fluid collections, or residual stones . Stone analysis proved it to be a calcium oxalate stone.
# Discussion
Although generally safe and effective, PCNL can be associated with significant morbidity, including infection, hemorrhage, urinary obstruction, and urinary leakage. To prevent these complications, placement of nephrostomy tubes is standard practice following PCNL in most centers. Although nephrostomy tubes provide postoperative drainage of urine from the collecting system, they can cause discomfort and increase hospital stay.
In attempt to reduce morbidity from nephrostomy tubes, tubeless PCNL emerged, which initially involved using ureteral stents for renal drainage in place of nephrostomy tubes [bib_ref] Tubeless percutaneous renal surgery, Bellman [/bib_ref] [bib_ref] Safety and efficacy of tubeless percutaneous nephrostolithotomy, Delnay [/bib_ref] [bib_ref] Almost totally tubeless percutaneous nephrolithotomy: further evolution of the technique, Goh [/bib_ref] [bib_ref] Tubeless percutaneous renal surgery: review of first 112 patients, Limb [/bib_ref]. These initial studies showed that tubeless PCNL was safe and effective. Since that time, some groups started performing totally tubeless PCNL, questioning the need for any type of drainage following PCNL [bib_ref] Effectiveness of totally tubeless percutaneous nephrolithotomy in selected patients: a prospective randomized..., Istanbulluoglu [/bib_ref] [bib_ref] Totally tubeless percutaneous nephrolithotomy, Aghamir [/bib_ref] [bib_ref] Tubeless and stentless percutaneous nephrolithotomy, Gupta [/bib_ref] [bib_ref] Totally tubeless percutaneous nephrolithotomy, Crook [/bib_ref]. However, such patients were still routinely admitted postoperatively.
We recently reported our experience with outpatient tubeless PCNL in a very small series of patients. These patients had placement of a ureteral stent intraoperatively at the end of their PCNL procedure in order to ensure adequate renal drainage. As a result of the success in our small case series, combined with the straightforward and uncomplicated nature of this particular patient's operation, we elected to attempt and to perform an outpatient PCNL without any nephrostomy tube or ureteral stent. All of our previously published discharge criteriawere met, and we arranged outpatient follow-up on postoperative day no. 1.
[L] C40 W400 Our patient did not have any significant pain postoperatively, nor did she endure any minor or major complications. She experienced minimal hemorrhage as evidenced by her minimal and transient gross hematuria and her postoperative hemoglobin [fig_ref] Table 1: Pre-operative and postoperative clinical parameters [/fig_ref]. CT scan confirmed that she was stone-free. This case report represents the first successful case of totally tubeless PCNL performed on a completely outpatient basis. The role of totally tubeless outpatient PCNL in larger stones has yet to be determined. We propose that this approach could be extended to larger stones and needs to be studied in a larger and more diverse patient population before it is widely employed as standard treatment.
[fig] Figure 2: Urine sample on postoperative day no. 1. [/fig]
[table] Table 1: Pre-operative and postoperative clinical parameters. [/table]
|
Evaluation of Children Caries Risk Factors: A Narrative Review of Nutritional Aspects, Oral Hygiene Habits, and Bacterial Alterations
Citation: Butera, A.; Maiorani, C.; Morandini, A.; Simonini, M.; Morittu, S.; Trombini, J.; Scribante, A.
# Introduction
Dental caries is one of the most common chronic conditions in childhood. About 50% of preschoolers in several countries have at least one carious lesion, and this has a negative impact on the quality of life for both the child and the family. Caries is regarded as a public health problem, the etiology of which is reflected in eating habits, the type of dental brush, and socioeconomic indicators [bib_ref] Impact of oral hygiene on oral health-related quality of life of preschool..., Shaghaghian [/bib_ref] [bib_ref] Parent's and Children's Behavior and Knowledge about Oral Health, Beljan [/bib_ref]. As far as the socio-economic situation is concerned, there is evidence that children from economically vulnerable or low-educated families have a higher prevalence rate of dental caries. This condition may influence parents' or guardians' perception of children's oral health, having less knowledge of factors associated with caries and the need for dental care, as well as less access to health services [bib_ref] Children's toothbrushing frequency: The influence of parents' rationale for brushing, habits and..., Trubey [/bib_ref].
In more economically disadvantaged countries, there is, in fact, a greater incidence of caries in children. The economic crisis and the lack of public health programmes affect the availability of both food and oral hygiene resources. High values of DMFT (Decayed Missing Filled Teeth) in countries such as Israel, Syria, Perù, and Serbia are linked to the lack of free distribution of fluoride toothpastes and toothbrushes in the poorest communities, and, in Syria, there is the economic crisis that has caused an increase in sugar intake and an increase in the barriers that families face for the availability of fresh food and oral hygiene products; studies carried out in Perù also highlight poverty, sugar intake, and a low literacy rate, which all encourage the increase of ECC in children.
Although the key factors causing carious lesions in adults and children are similar, there are unique risk factors regarding the latter, probably because the oral microbial flora and host defense mechanisms are still in development. In addition, the surfaces of freshly eroded teeth may have hypoplastic defects that facilitate the accumulation of bacterial biofilms and an increased risk of caries [bib_ref] Risk factors for dental caries in young children: A systematic review of..., Harris [/bib_ref].
A review of the literature conducted in Canada has highlighted the major risk factors involved: enamel hypoplasia, eating habits, difficulty brushing, caregiver influence, low income, and low level of education. In addition, it would appear that those who do not perform a first examination of the child within 24 months have a higher risk of developing carious lesions, since it would appear to be higher (38%) in the male sex [bib_ref] Determining the prevalence and risk factors for early childhood caries in a..., Schroth [/bib_ref].
The aim, therefore, was to evaluate the main risk factors of caries in children in order to reduce the incidence through managing the oral microbiota, reducing the ingestion of acidic foods and sugars, and motivating oral hygiene at home-to reduce the presence of Streptococci spp. and Lactobacilli spp. In fact, nutrition is implicated in the risk of carious lesions and the erosion of the enamel; the latter disease is on the rise and is caused by food or acidic substances, without bacterial involvement. Incorrect habits, such as eating disorders, together with the acidic pH of some foods, weaken the enamel and consequently make the tooth structure more fragile. In addition, diet components can contribute to the development of enamel defects, such as hypoplasia and fluorosis [bib_ref] Dental Erosion in Patients with Gastroesophageal Reflux Disease (GERD) in a Sample..., Alavi [/bib_ref] [bib_ref] Association between dental erosion and possible risk factors: A hospital-based study in..., Reddy [/bib_ref].
## Caries assessment
From the analysis of the literature, it is known that carious lesions occur due to the demineralization of enamel and dentine by the organic acids formed from bacteria present in dental plaques through the anaerobic metabolism of dietary sugars [bib_ref] Microbiology of Dental Decay and Periodontal Disease, Loesche [/bib_ref].
One of the main barriers available to the body is undoubtedly saturated saliva, which contains calcium and phosphate at pH 7 and promotes the process of remineralization, or deposits minerals in the porous areas where the process of demineralization of enamel or dentine began. If the pH of the oral cavity remains high enough for sufficient time, this process may occur; however, if the acidity persists, the demineralization process progresses, thereby increasing the porosity of the enamel until the formation of a carious lesion [bib_ref] Saliva and dental erosion, Buzalaf [/bib_ref].
This development requires sugar and bacteria, but it is certainly influenced by tooth susceptibility, bacterial profile, saliva quantity and quality, and the time when fermentable food carbohydrates are available for bacterial fermentation. In light of this evidence, and of the multiple factors involved, it is important to be able to keep the oral cavity well cleaned, even in a newborn, with the help of wet lap pads with a physiological solution after every sucking, avoiding spoiled habits such as baby bottles with milk and/or juices or any sweetened substance before going to sleep, thus preventing a serious form of caries, ECC (baby bottle caries). Then, at the appearance of the first dental elements, you can introduce the use of the toothbrush with an ergonomic handle and a non-slip grip, and a round and small head with extra-soft bristles. It is important that the child becomes familiar with the toothbrush. In addition, it is good practice to use a fluorinated toothpaste or toothpaste that contains remineralizing principles, with parent supervision to guide the child in brushing for at least 2 min, 2 times a day; this check is useful at least up to 7-8 years. From 3 years onwards, a fully deciduous dentition should be used for a fluorinated toothpaste with 1000 ppm of fluoride, or always a toothpaste with remineralizing principles, to increase the appearance of the first permanent elements-up to 1450 ppm of fluoride [bib_ref] Promoting parenting strategies to improve tooth brushing in children: Design of a..., De Jong-Lenters [/bib_ref] [bib_ref] Fluoride toothpaste efficacy and safety in children younger than 6 years: A..., Wright [/bib_ref]. As for the intake of fluoride, which is contained in the toothpaste, the American Dental Association recommends the use of a "smear"-sized portion of toothpaste for up to 3 years of age to avoid the risk of fluorosis (problem related to possible ingestion of toothpaste by the child), from which you can then use a pea-sized portion of toothpaste.
# Materials and methods
As caries is the most common non-communicable disease in the world, based on the evidence gathered, it is necessary to implement oral health education programmes in schools (both for children and parents), programmes aimed at the most vulnerable groups, and programmes for the training of health professionals; it is also important to pay attention to the consumption of sugar, a problem related to other systemic diseases such as diabetes, cardiovascular diseases, cancer, and obesity, and thereby seeking to promote oral and nutritional hygiene education programs in schools. Furthermore, having highlighted the relationship between socioeconomic conditions and caries, it is appropriate to integrate policy programmes that promote affordable access to care essentials for dental needs and affordable fluoride toothpastes [bib_ref] Making Cavities History: A Global Policy Consensus for Achieving a Dental Cavity-Free..., Pitts [/bib_ref].
The International Caries Classification and Management System (ICCMS TM ) has provided clear guidance for an appropriate management plan, which can be customized to be preventive and can be adapted to caries risk through clinical examination, risk assessment, and personalised assistance planning. The first step is the assessment of patient risk factors related to caries, namely head and neck radiation, dry mouth, inadequate oral hygiene practices, insufficient exposure to topical fluoride, high frequency of intake of sugary foods and drinks, symptom-driven dental care, socioeconomic status, and mother's caries experience. Subsequently, the intra-oral risks related to caries, such as hyposalivation, PUFA factors, the experience of caries, the presence of plaque, the presence of devices or restorations that favor the accumulation of plaque, and exposed radicular surfaces, are assessed; the evaluation for the presence of caries can then be carried out visually and can possibly be associated with an x-ray examination, using the ICDAS ranking system (International Caries Detection and Assessment System) [bib_ref] Icdas Ii Criteria (International Caries Detection and Assessment System), Dikmen [/bib_ref] and assessing whether these lesions are active or inactive. The ICDAS system uses several codes: the first for healthy teeth (code 0) and the next two for caries limited to enamel, white stain / brown (codes 1 and 2). The following two categories (code 3 and 4) are considered to be caries that extend to exposed dentine-free enamel. The remaining two categories (codes 5 and 6) are considered as caries with exposed dentin. The third step involves the analysis of these first collected data to provide information on the probability of new lesions and for indications of the activity of the lesions highlighted in order to develop plans for the treatment and management of the caries, as shown in [fig_ref] Table 1: Managing a patient's risk factors [/fig_ref].
Numerous tools have been developed for caries risk assessment, such as Cariogram, CAMBRA, ADA, and AAPD. The first uses an algorithm to calculate the percentage of risk combined with a clinical opinion; the second for the calculation of risk uses the instructions and the "caries balance", evaluating the clinical observations, preventive factors, biological and environmental risk factors, and the opinion of the health care professional; the third uses modules that contain clinical observations, preventive factors, and risk factors, as well as the last-mentioned method.
## Focused question
In order to provide a new guideline for health professionals for the assessment of caries risk, research was carried out on the main risk factors involved.
## Elegibility criteria
First, we analyzed the studies published in English in accordance with the following inclusion criteria:
Type of studies. Case-control, cross-sectional, cohort, longitudinal studies, and clinical trials. Type of participants. Participant with caries and ECC. Type of interventions. Case-control, cross-sectional, cohort studies, and clinical trials that have evaluated the possible etiological factors involved in the development of caries.
Outcome type. Each variable included in the studies was taken into account. We included studies that have assessed risk factors for children's caries (primary outcome: dietary factors, oral hygiene, and microflora) and possible preventive factors (secondary outcome). We included in the second phase only those studies that met all the inclusion criteria, that is to say, the analysis of the selected studies according to the exclusion criteria: (I) studies where the authors had not reported at least one of the parameters chosen as outcomes; (II) studies performed on participants with concomitant systemic pathologies/treatments that could have affected outcomes; and (III) studies that have not analyzed the possible risk factors of children's caries.
## Search strategy
The review is based on the research of studies in reference to the PICO model (Population: caries risk in pediatric patients; Intervention: review of nutritional risk factors, oral hygiene, and change in microbiological flora; Comparison: with all the studies in the literature that compared various risks in addition to those examined; Outcome: a proactive approach allows us to improve the quality of life of young patients, both from a nutritional and oral hygiene point of view, thereby reducing the caries risk factors and DMFT values. A correct intake of fluoride and substitutes, such as biomimetic hydroxyapatite, remains the chief preventive approach thus far.), which were identified through bibliographic research in electronic databases, and by examining the bibliography of articles, on Pubmed (MED-LINE) and Google Scholar. Initially, all the study abstracts were taken into consideration, which evaluated the possible risk factors involved in the development of children's caries.
## Research
We performed the search using the following keywords: "children caries", "early childhood caries", "children caries" AND "risk factors", "early childhood caries" AND Children 2022, 9, 262 5 of 14 "risk factors", "dental caries" AND "primary dentition" AND "risk factors". There is no time limit on the date of publication of the study.
## Screening and selection of articles
The search produced 130 titles matching the search keywords and the information related to the inclusion criteria. The following flowchart shows the selection criteria used to select the final 57 articles that were used for the review analysis, shown in [fig_ref] Figure 1: Flow chart of included studies [/fig_ref].
Pubmed (MEDLINE) and Google Scholar. Initially, all the study abstracts were taken into consideration, which evaluated the possible risk factors involved in the development of children's caries.
## Research
We performed the search using the following keywords: "children caries", "early childhood caries", "children caries" AND "risk factors", "early childhood caries" AND "risk factors", "dental caries" AND "primary dentition" AND "risk factors". There is no time limit on the date of publication of the study.
## Screening and selection of articles
The search produced 130 titles matching the search keywords and the information related to the inclusion criteria. The following flowchart shows the selection criteria used to select the final 57 articles that were used for the review analysis, shown in [fig_ref] Figure 1: Flow chart of included studies [/fig_ref].
# Results
The analysis of the studies included in the review [fig_ref] Table 1: Managing a patient's risk factors [/fig_ref] revealed the main risk factors involved in carious lesions in children.
The objective of this article was to provide guidance to dental professionals, based on the analysis of the risk factors related to the development of caries in children, such as eating habits, oral hygiene practices, and bacterial flora [fig_ref] Table 2: Main risk factors involved in children's caries [/fig_ref]. To this list the socioeconomic conditions and the level of education must surely be added, as well as the social policies-as is evident in the poorest countries. These indications, shown in [fig_ref] Table 3: Determination of the caries risk [/fig_ref] , developed on the basis of the knowledge and indications already provided by ICCMS TM , are used to easily frame the risk of caries in children; from the collected data it is then possible to implement management and treatment plans.
# Discussion
On the basis of these considerations and a preliminary review of the literature, a table has been created of the main risk factors involved in the development of carious lesions in children and the associated caries risk in such a way as to facilitate healthcare professionals into the framework of all patients, while also encouraging a multidisciplinary approach with other health professionals, such as the pediatrician and the nutritionist, following the assessment of the child.
The health professionals-namely dentists and dental hygienists-to whom the following table has been submitted [fig_ref] Table 3: Determination of the caries risk [/fig_ref] have evaluated the simple and linear procedure, promptly managing to draw attention to all the risk factors, both local and systemic, on the basis of literature research for DMFT and caries activity, enamel defects and dental erosion, frequency of sugars and carbohydrates intake, saliva quantity and quality, remineralization (fluoride, substituted fluoride ad biomimetic hydroxyapatite and calcium phosphate, eating disorders, oral hygiene, socioeconomic status, and oral family health or other conditions), erosion by gastroesophageal reflux, and eating disorders in children, which are not frequent. In fact, among the 130 studies analyzed, some factors have been highlighted as strongly related to the development of caries in children, such as those related to diet, oral hygiene, changes in microbial flora (primary outcome of the review), breastfeeding, hypolasia, and socioeconomic factors.
Among the risk factors involved in the development of carious lesions in children, poor oral hygiene, diet, and quality of oral bacterial flora have been recognized.
Oral hygiene and its good practices should begin even before the eruption of the first element with the help of lap pads soaked in physiological solution, and then there should be the switch to the use of a toothbrush and fluorinated toothpastes with the help of parents. In this sense, ethnicity, social status, and degree of family education affect the development of injury to the hard tissue of the tooth. In fact, the prevalence of carious lesions is lower in children who brush their teeth more than once a day and especially those who brush before going to sleep with the help of parents. Children of school age do not have a regular brushing schedule in neither quantity nor quality [bib_ref] Comparison of the Tooth Brushing Habits of Primary School Age Children and..., Özbek [/bib_ref]. Oral hygiene in children must be integrated from the first months of life in order to avoid plaque accumulations that can promote the onset of carious lesions over time. For this reason, it is essential to deal with a trusted dental team and pediatrician who can make the right indications based on the age of the child to maintain proper oral hygiene at home. Schools should also play a proactive role in the education of oral hygiene, seeking to provide programmes aimed at the age of the children and, therefore, relying on the proper health figures for reference.
As for the diet, it is also important for maintaining oral health and it is advisable to start to have good habits and good nutrition from when the child is a newborn. Milk is a complete food that is useful for the achievement and maintenance of oral health, given that it is a source of calcium, which is the basic constituent of the inorganic material of bones and the hard tissues of teeth; breast milk is the first food capable of supporting all nutritional needs, except for iron and vitamin C, until about the fourth month of life [bib_ref] The antioxidant components of milk and their role in processing, ripening, and..., Khan [/bib_ref]. Subsequently, during growth it is preferable to eat foods of every type, preferring consistent foods that require good mastication to stimulate the salivary flow. It is desirable that the frequency of intake of fermentable carbohydrates should be reduced and limited to main meals, after which correct oral hygiene is possible [bib_ref] Sugars and dental caries, Touger-Decker [/bib_ref]. Taking this type of food during the course of the day causes the repeated lowering of the pH, which promotes the demineralization of the hard tissues of the tooth, as well as the use of a sweetened pacifier and the non-nutritional use of a bottle containing sugary beverages. In this case, it is preferrable to use snacks based on fruit and vegetables, milk or cheese. Alternatively, low cariogenic sweeteners, which cannot be fermented, are sugar-alcohols (xylitol, sorbitol, and mannitol), synthetic sweeteners (aspartame, saccharin, cyclamates, and acesulfame-k), natural sweeteners (miracolin, taumatin, monellin, phyllodulcin, stevioside, and glycirryzin), hypo-acidogenic sugars, and hydrogenated derivatives of carbohydrates [bib_ref] Low Caloric Substitutes for Food Sugars: Clinical Implications for Addressing the Incidence..., Roberts [/bib_ref].
Another aspect that affects the development and predisposition of dental caries in children is the composition of oral bacterial flora, such as Streptococcus mutans and Steptococcus sobrinus, as well as Lactobacilli spp. (the latter exploits glucose-derived sucrose to destroy tooth tissue), appear to have an association with the development of enamel lesions as they metabolize sugar acids to produce acids, thereby contributing to the demineralization of the structure of dental enamel. The source for the presence of Streptococcus mutans already in newborns seems to be the transmission of the latter from the mother through saliva. In addition to these, other bacteria appear play a role in the beginning and progression of the lesion, such as species of Actinomyces and Bifidobacterium [bib_ref] Frequency, biofilm formation and acid susceptibility of streptococcus mutans and streptococcus sobrinus..., Ghasempour [/bib_ref] [bib_ref] The Oral Microbiome in Dental Caries, Strużycka [/bib_ref]. In addition, a study published in 2020 found other micro-organisms implicated in the progression of carious lesions: Candida albicans would appear abundant at the level of advanced lesions, along with Streptococcus mutans and Scardovia wiggsiae, while Cryptococcus neoformans and Candida sake seem to prevail in primitive lesions [bib_ref] Supragingival mycobiome and inter-kingdom interactions in dental caries, Baraniya [/bib_ref].
On the basis of these considerations, it is easy to see how important the multidisciplinary approach between dental hygienists and dentists, pediatricians, and nutritionists is in ensuring an overall state of health of the child, orally and systemically.
The importance of diet in children is clear given the repercussions it can have on oral health. It is advisable to give nutritional advice to better understand the eating habits of the whole family, with an on the sugar intake and the number of daily intakes of any food that contains simple carbohydrates. It is essential that there is an adequate introduction of energy and nutrients, such as vitamins, minerals, and calcium, which allow for the growth and formation of the organism; it is known, in fact, that there is a relationship between the consumption of fermentable sugars and the development of carious lesions. The danger from the frequency of consumption of these should be emphasized along with the fact that products such as fruit juices or sugary drinks favors childhood obesity. In Italy, according to data reported by the Istituto Superiore di Sanità (ISS), 20.4% of children are overweight and 9.4% suffer from obesity (including severely obese children, which represents 2.4%).
Having clarified which risk factors are mainly involved in the development of caries in children, [fig_ref] Table 1: Managing a patient's risk factors [/fig_ref] reports on the main problems to focus on to carry out a simple and possible analysis of the risk profile for caries in children. From this same table, and from the analysis of the literature, the relationship between the development of carious lesions and the feeding of the child is now clear. Once the risk profile has been analysed, it is then possible to implement targeted management and treatment plans for the child.
## Possible prevention approches
To date, the local administration of fluoride seems to be one of the most effective treatments in the primary prevention of carious lesions, as reported by the Guidelines of the Ministry of Health: "The post-eruptive preventive effect of fluoride, obtained through the topical route of administration, is considered more effective than the pre-eruptive effect obtained through the systemic route of administration"; "The fluorinated toothpaste, therefore, represents a means of administration of primary importance in the prevention of caries". It is necessary, however, to pay attention to the composition of the toothpaste. In fact, not everyone releases enough local fluoride for the prevention of caries. It is necessary to distinguish free ionic fluoride, profluoride compounds, and fluoride compounds; the first type has the ability to interact with tooth enamel and perform anti-carious activity, the second type precipitates during brushing and releases ionic fluoride that can be effective in the prevention of carious lesions over time, and the third type has no efficacy in terms of prevention. Several studies have drawn attention to the effectiveness of fluorinated toothpastes and it has emerged that the use of fluorinated toothpaste, especially toothpaste containing at least 1500 ppm of fluoride, leads to a significant reduction in the development of carious lesions [bib_ref] Focus on Fluorides: Update on the Use of Fluoride for the Prevention..., Carey [/bib_ref] [bib_ref] Fluorides and Other Preventive Strategies for Tooth Decay, Horst [/bib_ref].
There are also other formulations available for the professional topical administration of fluoride, such as gels or varnishes, which are equally effective. Gels (with about 12,300 ppm of fluoride) are usually applied with the help of a mask, which is placed for about 4 min, while the paints (22,600 ppm), which are easier to apply and eliminate the risk of ingestion of fluoride by the child, are applied and brushed on the teeth clean and dry [bib_ref] Topical Fluorides in Caries Prevention and Management: A North American Perspective, Newbrun [/bib_ref].
Other remineralizing systems used for primary prevention are those based on calcium phosphate, such as casein phosphopeptide-amorphous calcium phosphate, calcium amone phosphate, and bioactive glass containing calcium and sodium phosphosilicate [bib_ref] Calcium phosphate-based remineralization systems: Scientific evidence? Aust, Reynolds [/bib_ref] ; studies have shown that the use of CPP-ACP (casein phosphopeptide-amorphous calcium phosphate) reduces the roughness of the enamel, which favors bacterial adhesion, thereby restoring and gradually repairing the central areas of the enamel [bib_ref] Evaluation of the efficacy of casein phosphopeptide-amorphous calcium phosphate on remineralization of..., Ma [/bib_ref].
Biomimetic mineralization from P 11 -4 is also known, in combination with the application of fluoride, as a non-invasive treatment for initial carious lesions thar is capable of regenerating enamel tissue and preventing the progression of lesions. This self-acting oligopeptide assembler is able to form a biological matrix, causing the regrowth of the crystals of dental minerals. This activity is demonstrated by studies in the literature, which show that primary carious lesions treated with P 11 -4 are significantly reduced [bib_ref] Randomised clinical trial investigating self-assembling peptide P11-4 in the treatment of early..., Bröseler [/bib_ref] [bib_ref] Self-assembling Peptide P11-4 and Fluoride for Regenerating Enamel, Alkilzy [/bib_ref] [bib_ref] Efficacy of P11-4 for the treatment of initial buccal caries: A randomized..., Kondelova [/bib_ref].
Last but not least, hydroxyapatite products that are biomimetic zinc-substituted can be used for the remineralization of the enamel surface, such as toothpastes or mousse that is able to interact with biological tissues. The effectiveness has been demonstrated after one month of treatment with toothpaste, thanks to the deposition of calcium, phosphorus, and silicon ions, thereby forming a real coating layer on the surface of the enamel [bib_ref] EDS Evaluation of the Mineral Deposition on a Polymeric Composite Resin of..., Butera [/bib_ref] ; the mousse, instead, contains a high percentage of microrepair (30%) and is effective if applied daily for 10 days a month for about 10 min-already from the second cycle of application [bib_ref] Biomimetic Effect of Nano-Hydroxyapatite in Demineralized Enamel before Orthodontic Bonding of Brackets..., Scribante [/bib_ref]. The use of these products in primary prevention would seem valid in the reduction of the risk for development of carious lesions in children, having antibacterial properties against Steptococcus mutans and being effective in the prevention of biofilm, based on the biomimetic coating they create on the surface of the enamel-even in children [bib_ref] Enamel remineralization and repair results of Biomimetic Hydroxyapatite toothpaste on deciduous teeth:..., Bossù [/bib_ref]. Remineralizing systems, if there is a low risk, and substances based on biomimetic zincsubstituted hydroxyapatite can be used to take proactive action to avoid the incidence of caries and, above all, an overdose of fluorinated substances, which in the long term can induce irreversible lesions. [fig_ref] Table 4: Indications for primary prevention in accordance with caries risk [/fig_ref] summarizes the remineralizing products and their use for the risk management of caries in children. Finally, to date, non-pharmacological alternatives are available for those children who need chlorhexidine to counter possible gum inflammations, given by the accumulation of bacterial biofilms, such as probiotics-which are equally effective. However, it may be used at 0.05% with the additional fluoride-based composition in the form of mouthwash for an extended period (not more than one month of treatment).
# Conclusions
Caries is a degenerative disease that affects the hard tissues of the tooth and is caused by the acidic action of bacteria that are normally present within the oral cavity. These bacteria proliferate in excess and take advantage of the presence of sugars introduced from the children's diet. As can be seen from this report, it is a multifactorial etiology pathology, mainly linked to poor oral hygiene, bad eating habits, and an alteration of oral bacterial flora. There are, in fact, foods considered strongly cariogenic, such as sugars, sugary drinks, industrial snacks, caramel, candies, sweets, chocolate, industrial fruit juices, breakfast cereals, tomato preserves, sauces or balsamic vinegar, cured meats, canned bread, and many other foods that apparently do not contain sugar, but, in reality, have significant quantities. On the basis of the dietary habits of children and parents, caries risk should be established by taking into account all the relevant factors in order to provide prevention and/or maintenance protocols that are aimed at restoring the optimal conditions in the oral cavity by means of close professional hygiene sessions and remineralizing the areas of greatest risk by using products with fluoride or containing other remineralizing agents, such as biomimetic zinc-substituted hydroxyapatite.
Supplementary Materials: The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/children9020262/s1, [fig_ref] Table 1: Managing a patient's risk factors [/fig_ref] : Studies included. Institutional Review Board Statement: Not applicable for studies not involving humans or animals.
# Informed consent statement:
Not applicable for studies not involving humans.
# Data availability statement:
The data presented in this study are available on request from the corresponding author.
[fig] Figure 1: Flow chart of included studies. [/fig]
[fig] Author: Contributions: Conceptualization, A.B., C.M., A.M. and A.S.; methodology, A.B. and A.M.; software, A.S., C.M. and A.B.; validation, A.B., A.S., M.S., S.M. and A.S.; formal analysis, A.B.; investigation, A.B., M.S., C.M., J.T., S.M. and A.S.; resources, A.M.; data curation, A.B. and J.T.; writing-original draft preparation, C.M.; writing-review and editing, C.M., A.B. and A.S. visualization, A.B. and A.M.; supervision, A.S.; project administration, A.B., M.S. and C.M. All authors have read and agreed to the published version of the manuscript. Funding: This research received no external funding. [/fig]
[table] Table 1: Managing a patient's risk factors (from ICCMS TM ).Motivational interviewing One-to-one dietary intake interventions Altering medication-induced hyposalivationReducing the use of recreational drugsIncrease fluoride varnish to 4 times/year [/table]
[table] Table 2: Main risk factors involved in children's caries. [/table]
[table] Table 3: Determination of the caries risk. [/table]
[table] Table 4: Indications for primary prevention in accordance with caries risk. [/table]
|
Subsets and Splits