text
stringlengths 66
377k
|
|---|
Illuminating novel biological aspects and potential new therapeutic approaches for chronic myeloproliferative malignancies
This review reflects the presentations and discussion at the 14th post-American Society of Hematology (ASH) International Workshop on Chronic Myeloproliferative Malignancies, which took place on the December 10 and 11, 2019, immediately after the 61st ASH Annual Meeting in Orlando, Florida. Rather than present a resume of the proceedings, we address some of the topical translational science research and clinically relevant topics in detail. We consider howThis is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. recent studies using single-cell genomics and other molecular methods reveal novel aspects of hematopoiesis which in turn raise the possibility of new therapeutic approaches for patients with myeloproliferative neoplasms (MPNs). We discuss how alternative therapies could benefit patients with chronic myeloid leukemia who develop BCR-ABL1 mutant subclones following ABL1-tyrosine kinase inhibitor therapy. In MPNs, we focus on efforts beyond JAK-STAT and the merits of integrating activin receptor ligand traps, interferon-α, and allografting in the current treatment algorithm for patients with myelofibrosis.
# | introduction
Despite the development of significant genomic insights and the availability of successful treatment options, chronic myeloproliferative malignancies, with the exception of chronic myeloid leukemia (CML), remain enigmatic. [bib_ref] Chronic myeloid leukemia: reminiscences and dreams, Mughal [/bib_ref] For patients with CML, we now have five licensed ABL1 tyrosine kinase inhibitors (TKIs) that can durably reduce the number of CML cells, and the life expectancy of most patients now approaches that of the normal population; for patients with higher-risk myelofibrosis (MF), the most serious and life-threatening of the myeloproliferative neoplasms (MPNs), there are two licensed JAK inhibitors (JAKi) that accord rapid and substantial clinical benefits in terms of constitutional symptoms and splenomegaly but few molecular complete remissions and no significant impact on the risk of transformation to acute myeloid leukemia. [bib_ref] Emerging translational science discoveries, clonal approaches, and treatment trends in chronic myeloproliferative..., Mughal [/bib_ref] Indeed, there is a paucity of disease-modifying drugs for MF and the disease remains incurable with a median survival of 4 to 5 years. This is similar to the historical scenario that existed when cytotoxic chemotherapy was used to treat newly diagnosed patients with CML who benefitted from a transient control of their leukocytosis and splenomegaly but awaited the inevitable leukemic transformation with equanimity.
Compared with CML, the Ph-negative MPNs and particularly MF demonstrate substantial genetic and bone marrow microenvironment complexity, which impacts upon disease biology and treatment. In this regard, the significant insights into the biochemical and biological consequences of the well-characterized driver mutations found in about 90% of MF patients and the corresponding clinical implications is raising the possibility of new therapeutic approaches. This review, based on the presentations at the 14th post-ASH International Workshop on Chronic Myeloproliferative Malignancies, which took place on the December 10 and 11, 2019, immediately after the 61st American Society of Hematology Annual Meeting in Orlando, Florida presents the important aspects of the biological and therapeutic advances as well as the recent progress made in treating disease-related complications. To illustrate, we address how symptomatic anemia related to MF can be alleviated by targeting transforming growth factor beta (TGF-β) by an activin receptor ligand, luspatercept. We also discuss the emerging role of single cell technologies in MPNs and tools to personalize the management of MF.
## | using single cell technologies to identify novel targets
## In mpns
Since the first studies performing whole transcriptome analysis in individual cells, termed single cell RNA-sequencing (scRNAseq) a decade ago, there has been an explosion of technological advances in single cell methods, resulting in advanced platforms for high-throughput and/or multiomic approaches to study molecular biology at single cell resolution. [bib_ref] mRNA-Seq whole-transcriptome analysis of a single cell, Tang [/bib_ref] Indeed, single cell sequencing was selected as "Method of the Year" by Nature Methods in 2013, and single cell multimodal omics as "Method of the Year" in 2019.In hematology, these approaches have been applied to gain insight into normal and aberrant hematopoiesis, and more recently the bone marrow niche. [bib_ref] Haematopoiesis in the era of advanced single-cell technologies, Jacobsen [/bib_ref] [bib_ref] Single-cell approaches reveal novel cellular pathways for megakaryocyte and erythroid differentiation, Psaila [/bib_ref] [bib_ref] Single-cell sequencing in hematology, Brierley [/bib_ref] [bib_ref] Unravelling intratumoral heterogeneity through high-sensitivity single-cell mutational analysis and parallel RNA sequencing, Rodriguez-Meira [/bib_ref] [bib_ref] A cellular taxonomy of the bone marrow stroma in homeostasis and leukemia, Baryawno [/bib_ref] [bib_ref] Single-cell transcriptomes of murine bone marrow stromal cells reveal niche-associated heterogeneity, Addo [/bib_ref] [bib_ref] The bone marrow microenvironment at single-cell resolution, Tikhonova [/bib_ref] [bib_ref] A comprehensive single cell transcriptional landscape of human hematopoietic progenitors, Pellin [/bib_ref] Technological and computational challenges and high cost have prohibited widespread application in clinical diagnostics or monitoring pipelines. So what is the future potential for clinical application of single cell omics in MPNs? First, it provides the necessary resolution to detect and characterize rare cellular populations such as leukemic stem cells. In addition to understanding disease biology and mechanisms of relapse or treatment resistance, this technology may identify novel leukemia stem cell-specific drug targets. Second, dissection of tumor heterogeneity will enable step changes in personalized medicine, identifying tumors most likely to respond to a given treatment approach, informing the selection of treatment combinations that are most likely to be effective in any given cancer at a particular stage in the disease course. Thirdly, unlike genomic (or proteomic) analyses of cells studied in "bulk," single cell studies provide information on specific combinations or genes or proteins expressed by cellular subtypes. For example, identification of unique combinations of cell surface proteins may enable selective targeting of cancers were specific individual markers have not been found. Finally, comprehensive characterization of bone marrow cells in MPNs may be applied to simultaneously study the MPN clone together with associated immune or nonclonal cells as well as the tissue stroma, leading to key insights into the cell-cell interactions that are important for disease evolution.
Key challenges for the field include reducing the high cost and practical challenges to enable single cell omics to be delivered in the context of routine clinical practice and clinical trials. In addition, further advances in spatial single cell omics are required to enable study of myeloid neoplasms in their native context, without losing key information contained within the tissue architecture as occurs when cells are studied "in suspension." Moreover, just as hematology has led the way in other areas of precision oncology, the application of single cell technologies in MPNs is also highly relevant for other cancer types.
Recently, single cell omics has been applied to study abnormal hematopoiesis in MF. A dramatic bias toward megakaryocyte (Mk) differentiation was identified from early stem cells, present across clinical and molecular MF subgroups. [bib_ref] Single-cell analyses reveal megakaryocyte-biased hematopoiesis in myelofibrosis and identify mutant clone-specific targets, Psaila [/bib_ref] Mk progenitors were heterogeneous, with distinct expression of inflammatory mediators and unique populations present in MF. Aberrant cell surface expression of G6B expression on MF stem and progenitors was identified, which could allow selective immunotherapeutic targeting of the MF clone.
## | targeting specific bcr-abl1 mutant subclones by
## Alternative treatment approaches
Although the introduction of ABL1-TKIs has revolutionized the treatment of BCR-ABL1positive leukemias, including CML and Philadelphia (Ph)-positive acute lymphoblastic leukemia (ALL), they are not panacea. Outgrowth of leukemic clones harboring resistant point mutations in the BCR-ABL1 tyrosine kinase domain (TKD) poses a serious clinical problem. These mutant subclones are believed to expand due to disruption of the TKI/TKD interaction. However, due to the positioning of resistant mutations at functional key positions of the BCR-ABL1 TKD, Lion and colleagues hypothesized that these mutations have a biological effect on BCR-ABL1 downstream signaling and oncogenic gene expression. They generated over 50 Ba/F3 cell lines expressing BCR-ABL1 with various single and compound mutations (CMs) at levels comparable to patient-derived cells. 14 These cell lines were tested against a panel of TKIs, with a particular focus on CMs and their in vitro responsiveness to ponatinib. Their findings identified three categories of CMs displaying either high, intermediate, or low sensitivity to ponatinib. While the intermediate category of CMs was shown to respond to higher, yet clinically achievable concentrations of ponatinib, the inhibitory concentrations required for growth suppression of the least sensitive category of CMs were too high for clinical applicability. [bib_ref] BCR-ABL1 compound mutants display differential and dose-dependent responses to ponatinib, Byrgazov [/bib_ref] The latter category comprised particularly CMs including the gatekeeper mutation T315I in combination with a variety of other TKD mutations, and these CMs were shown to be unresponsive even to the most potent TKI available at present, thus presenting a great clinical challenge. Investigation of the effect of individual CMs on intracellular downstream signaling identified altered patterns in comparison to cells carrying wild-type BCR-ABL1. These findings provided the basis for identifying potentially druggable vulnerabilities, implicating various compounds not directly targeting the kinase activity of BCR-ABL1, such as cell cycle progression inhibitors, proteasome inhibitors, and others. Interestingly, in contrast to native BCR-ABL1 and several mutant subclones, CMs including the T315I mutation was very sensitive in vitro to treatment with hydroxyurea (HU), and molecular analyses provided a biological basis for the intriguing effect of this compound. The observation that, unlike cells with native BCR-ABL1, CMs including the T315I mutation display highly upregulated cyclin-dependent kinase 6 (CDK6), thus raising the possibility that specific inhibitors of the enzyme might be effective in these instances. Indeed, in vitro testing revealed great efficacy of palbociclib and ribociclib at concentrations readily achievable in the clinical setting. Moreover, synergistic effects of ponatinib in combination with HU or ciclibs against cells carrying highly resistant CMs including the T315I mutation have been demonstrated.It appears therefore that the upregulation of CDK6 documented in BCR-ABL1-positive cells displaying the T315I mutation alone or as part of a CM, provides a therapeutically exploitable vulnerability of the mutant cells. If these observations can be confirmed in the clinical setting, the therapeutic armamentarium would facilitate control of mutant BCR-ABL1 subclones resistant to all presently approved ABL1 TKIs. The observations so far demonstrate that the impact of acquiring TKI-resistant mutations may go far beyond hindering the TKI/TKD interaction. Studying the differential downstream signaling in mutant BCR-ABL1 subclones should facilitate the understanding of their biological behavior and help establish their druggable vulnerabilities, as a basis for the development of novel effective treatment options in patients resistant to currently available TKI-based therapies.
## | activin receptor ligand traps for the treatment of anemia associated with chronic myeloproliferative
## Malignancies
Anemia is not only common in patients with MF (and is a WHO minor diagnostic criterion), but contributes adversely to the quality of life, morbidity and mortality, and increased symptom burden. The activin receptor ligand traps sotatercept (ACE-011) and luspatercept belong to a novel class of fusion proteins that are capable of alleviating anemia resulting from a range of benign and malignant conditions. [bib_ref] An activin receptor IIA ligand trap corrects ineffective erythropoiesis in beta-thalassemia, Dussiot [/bib_ref] [bib_ref] Luspatercept for the treatment of anemia in myelodysplastic syndromes and primary myelofibrosis, Fenaux [/bib_ref] These drugs consist of the extracellular domain of the activin receptor (A in the case of sotatercept; B in the case of luspatercept) fused to the Fc domain of human IgG1 and improve anemia by sequestering ligands belonging to the TGF-β superfamily that suppress terminal erythropoiesis. [bib_ref] An activin receptor IIA ligand trap promotes erythropoiesis resulting in a rapid..., Carrancio [/bib_ref] Mechanistically, growth and differentiation factor 11 (GDF11) has been considered to be an important TGF-β superfamily ligand trapped by such agents that ameliorate anemia resulting from diverse causes. [bib_ref] Lack of Gdf11 does not improve anemia or prevent the activity of..., Guerra [/bib_ref] Indeed, luspatercept was approved in November 2019 for the treatment of anemia in patients with β-thalassemia and patients with very low-to intermediate-risk myelodysplastic syndromes with ring sideroblasts (MDS-RS) or MDS/MPN with RS and thrombocytosis (MDS/MPN-RS-T) in April 2020. [bib_ref] The medalist trial: results of a phase 3, randomized, double-blind, placebo-controlled study..., Fenaux [/bib_ref] An important distinguishing feature of this class of drugs that sets them apart from erythropoietin and its analogs is that they act on late-stage erythroid precursors . Based on preclinical studies suggesting anemia benefit secondary to its ability to sequester ligands secreted by the MF bone marrow stroma that inhibit erythropoiesis via Smad signaling, sotatercept was studied in an ongoing clinical trial. [bib_ref] Stromal cell-mediated inhibition of erythropoiesis can be attenuated by sotatercept (ACE-011), an..., Iancu-Rubin [/bib_ref] Eligible patients had to be anemic or transfusion-dependent (TD) per International Working Group for Myelofibrosis Research and Treatment (IWG-MRT) criteria. [bib_ref] Revised response criteria for myelofibrosis: International Working Group-Myeloproliferative Neoplasms Research and Treatment..., Tefferi [/bib_ref] Sotatercept was studied both as monotherapy and in patients on a stable dose of ruxolitinib. [bib_ref] ACE-011) in subjects with MPN-associated myelofibrosis and anemia, Bose [/bib_ref] The response rate, a combination of transfusion independent (TI) by IWG-MRT criteria and anemia response by Gale criteria, was 35% in patients treated with sotatercept alone and 12.5% in patients treated with the combination of sotatercept and ruxoloitinib. [bib_ref] What are RBC-transfusion-dependence and -independence?, Gale [/bib_ref] Grade 3 treatment-emergent adverse events (TEAEs) were limited to hypertension and limb pain. These findings initiated a phase 2 study of luspatercept alone or in combination with ruxolitinib in 76 anemic patients with MF, either alone (cohorts 1 and 2) or in combination with ruxolitinib (cohorts 3A and 3B). [bib_ref] A phase 2 study of luspatercept in patients with myelofibrosis-associated anemia, Gerds [/bib_ref] Patients in cohorts 1 (n = 22) and 3A (n = 14) could not have been receiving any transfusions, while patients in cohorts 2 (n = 21) and 3B (n = 19) had to be TD per IWG-MRT criteria. By intention-to-treat analysis, the anemia response rates (by Gale criteria) in cohorts 1 and 3A were 14% and 21%, respectively, while the RBC TI rates in cohorts 2 and 3B were 10% and 32%, respectively. The proportions of patients who achieved a mean increase in hemoglobin of ≥1.5 g/dL and who achieved a ≥50% reduction in the number of units transfused also favored the ruxolitinib cohorts, a pattern of response not well understood at this time. The most frequent TEAEs included hypertension (12%), bone pain (9%), and diarrhea (5%). Noteworthy, the activin receptor ligand traps discussed above do not target TGF-β per se, a candidate therapeutic target in MF, and discussed in the next section. This cytokine is secreted at high levels, primarily by Mks in the MF bone marrow milieu, and is implicated in both the fibrotic process and in promoting clonal dominance in the classic MPNs. [bib_ref] Bone marrow fibrosis in primary myelofibrosis: pathogenic mechanisms and the role of..., Agarwal [/bib_ref]
## | current and investigational therapies for mf
The introduction into clinical practice in 2011 of the original type I JAKi ruxolitinib was an important therapeutic advance, as with this agent most MF patients achieve substantial spleen volume reduction (SVR) and improvement in total symptom score (TSS), and some may expect prolongation of survival compared with earlier methods of treatment. [bib_ref] A double-blind, placebo-controlled trial of ruxolitinib for myelofibrosis, Verstovsek [/bib_ref] [bib_ref] JAK inhibition with ruxolitinib versus best available therapy for myelofibrosis, Harrison [/bib_ref] Ruxolitinib, an oral JAK1 and JAK2 inhibitor, targets the ATP-binding site of the JAKs under the active conformation of the kinase domain. The median duration of response to ruxolitinib is about 2 to 3 years, and the prognosis of patients who relapse, or are refractory or resistant, or have clonal evolution, have a median survival of 14 months. [bib_ref] Clonal evolution and outcomes in myelofibrosis after ruxolitinib discontinuation, Newberry [/bib_ref] [bib_ref] Between a rux and a hard place: evaluating salvage treatment and outcomes..., Kuykendall [/bib_ref] Patients with monocytosis, accelerated phase, or high-risk genetic features also have a poor prognosis. [bib_ref] Monocytosis is a powerful and independent predictor of inferior survival in primary..., Tefferi [/bib_ref] [bib_ref] Significant impact of the molecular profile on the prognosis of patients with..., Debureaux [/bib_ref] Second-line therapies comprise of an alternative type I JAKi, or agents that target mechanisms beyond JAK/STAT signaling, either alone or in combination with JAKi. Many of these JAKis have had their clinical development discontinued due to the emergence of serious toxicity. Fedratinib, a JAK2/FLT3 inhibitor, demonstrated efficacy in both first and second-line, but a clinical hold was placed in 2013 by the FDA due to concerns over increased risk for Wernicke encephalopathy. [bib_ref] Safety and efficacy of fedratinib in patients with primary or secondary myelofibrosis:..., Pardanani [/bib_ref] [bib_ref] Janus kinase-2 inhibitor fedratinib in patients with myelofibrosis previously treated with ruxolitinib..., Harrison [/bib_ref] This was subsequently revised and the drug licensed. [bib_ref] Case series of potential Wernicke's encephalopathy in patients treated with fedratinib, Harrison [/bib_ref] [bib_ref] Forging ahead or moving back: dilemmas and disappointments of novel agents for..., Mclornan [/bib_ref] Pacritinib, a JAK2/FLT3/IRAK1/CFS1R inhibitor, and momelotinib, a JAK1/2 inhibitor, remain investigational. Pacritinib is being tested in MF patients with thrombocytopenia at a revised dose in a randomized phase III trial (PACIFICA). [bib_ref] Myelofibrosis biology and contemporary management, Gangat [/bib_ref] The drug was previously found to be effective in phase III trials in reducing splenomegaly and symptom control, but there were concerns about cardiotoxicity and hemorrhage. [bib_ref] Pacritinib versus best available therapy for the treatment of myelofibrosis irrespective of..., Mesa [/bib_ref] [bib_ref] Pacritinib vs best available therapy, including ruxolitinib, in patients with myelofibrosis: a..., Mascarenhas [/bib_ref] Momelotinib can improve anemia by improving functional iron availability by decreasing hepcidin production through the ACVR1 pathway. The drug was noninferior to ruxolitinib for SVR but not TSS response in first-line and in second-line setting, there were improvements in TI and TSS, but inadequate SVR. [bib_ref] SIMPLIFY-1: a phase III randomized trial of momelotinib versus ruxolitinib in Janus..., Mesa [/bib_ref] [bib_ref] Momelotinib versus best available therapy in patients with myelofibrosis previously treated with..., Harrison [/bib_ref] There was concern for emergent peripheral neuropathy and further studies are underway to assess the drug's candidacy for second-line anemic/TD patients (MOMENTUM).
At present, a major focus of drug development in MF is identifying new therapeutic targets beyond JAK-STAT inhibition . As illustration, to mention a few, encouraging preliminary results have been noted in trials investigating AURKA inhibitors, HDAC inhibitors, PARP inhibitors, PRMT5 inhibitors, HDM2 inhibitors, hedgehog pathway inhibitors, TGF-β inhibitors, telomerase inhibitors, and immunomodulatory drugs. [bib_ref] Targeted therapies for myeloproliferative neoplasms, Li [/bib_ref] [bib_ref] A pilot study of the telomerase inhibitor imetelstat for myelofibrosis, Tefferi [/bib_ref] Novel immunological therapeutic approaches are also being tested, and include CD123 targeted antibody, mutant CALR blocking antibodies, and peptide vaccination. Preclinical models have also studied type II JAKis, such as NVP-BBT594 and NVPCHZ868, and found them to be effective. [bib_ref] CHZ868, a type II JAK2 inhibitor, reverses type JAK inhibitor persistence and..., Meyer [/bib_ref] depicts some of the agents that have shown safety and efficacy in second line use either as monotherapy or in combination with ruxolitinib ("add-on"/"add-back"). [bib_ref] Novel therapies in myeloproliferative neoplasms (MPN): beyond JAK inhibitors, Economides [/bib_ref] A focus on apoptosis is illustrated by an ongoing phase II study of the BCL-XL inhibitor, navitoclax, combined with ruxolitinib. [bib_ref] Results from a phase 2 study of navitoclax in combination with ruxolitinib..., Harrison [/bib_ref] In this small study, 29% of the patients achieved an SVR ≥35% and a 20% improvement in TSS; 42% patients had SVR ≥35% at any time on study; 25% patients had bone marrow fibrosis improvement; and 42% patients were observed to have major molecular responses. Grade 3/4 TEAE included thrombocytopenia (44%) and anemia (24%). A three-arm phase II trial (MANIFEST), assessing epigenetic modification by a bromodomain and extraterminal inhibitor, CPI-0610, observed 80% of patients to achieve SVR ≥35 at week 12, 50% of patients to achieve TSS response, 43% TD patients became TI and some improvements in bone marrow fibrosis were noted in first-line setting with CPI-060 plus ruxolitinib. [bib_ref] MANIFEST, a phase 2 study of CPI-0610, a bromodomain and extraterminal domain..., Mascarenhas [/bib_ref] The most common TEAEs were thrombocytopenia (23%). Updated results of a phase II trial of ruxolitinib with either thalidomide or pomalidomide, confirmed earlier positive results demonstrating improvements in cytopenias, in particular thrombocytopenia and other clinical benefits. [bib_ref] Updated results from the German MPNSG-0212 combination trial: ruxolitinib plus pomalidomide in..., Stegelmann [/bib_ref] [bib_ref] Safety and efficacy of combined ruxolitinib and thalidomide in patients with myelofibrosis:..., Rampal [/bib_ref] Other studies of note included a phase 1 study of an Mk-targeting therapy, alisertib, a specific inhibitor of aurora kinase (AURKA) demonstrated clinical benefits in a third of patients by reducing splenomegaly and symptom burden. [bib_ref] Aurora kinase A inhibition provides clinical benefit, normalizes megakaryocytes and reduces bone..., Gangat [/bib_ref] Importantly, alisertib reduces bone marrow fibrosis, mutant allele burden and restores normal morphology in atypical MF Mks in some cases. The most common grade 3/4 TEAEs were neutropenia (42%), thrombocytopenia (29%), and anemia (21%). Another novel approach is tagraxofusp, a CD123-directed antibody, composed of human IL-3 and truncated diptheria toxin, licensed in 2018 for children and adults with blastic plasmacytoid neoplasm. [bib_ref] Tagraxofusp in blastic plasmacytoid dendritic-cell neoplasm, Pemmaraju [/bib_ref] Tagraxofusp has been tested in a preclinical model either alone or in combination with ruxolitinib and noted to have activity in primary MF (PMF). [bib_ref] Evaluation of tagraxofusp (SL-401) alone and in combination with ruxolitinib for the..., Krishnan [/bib_ref] Interim clinical results from a phase 2 study of tagraxofusp in patients with relapsed/refractory MF with thrombocytopenia observed spleen responses in 20% and symptom reduction in half of the patients. [bib_ref] Results from a phase 1/2 clinical trial of tagraxofusp (SL-401) in patients..., Pemmaraju [/bib_ref] The most common TEAE was abnormal liver function tests and the most serious was a single patient experiencing grade 4 capillary leak syndrome. The final results of a phase 2 study assessing apoptosis induced by a SMAC mimetic (IAP antagonists) agent in high-risk thrombocytopenic MF patients observed an overall response rate of 32%, a modest anemia response and a median overall survival that has not been reached with median follow-up of about 2 years. [bib_ref] Final results of phase 2 clinical trial of LCVL161, a novel oral..., Pemmaraju [/bib_ref] The most common TEAEs were nausea/vomiting (60%); fatigue syndrome (49%); grade 3/4 thrombocytopenia occurred in 3 pts (6%) and correlative studies confirmed target-inhibition (cIAP1). Results from a small phase I/II study of epigenetic modification by means of a lysine specific demethylase1 inhibitor, bomedemstat, showed 75% to have a modest decrease in SVR and symptom burden, and bone marrow fibrosis improvement. [bib_ref] A phase 2a study of the LSD1 inhibitor Img-7289 (bomedemstat) for the..., Pettit [/bib_ref] A small repurposing effort with a biguanide, metformin, to target bone marrow fibrosis revealed a progressive reduction in bone marrow collagen fibers from 26.9% at baseline to 3.8% at 3 months and 0.84% at 6 months posttherapy. [bib_ref] Analysis of metformin effects on bone marrow fibrosis and disease progression in..., De Melo-Campos [/bib_ref] Finally, AVID200, a TGF-β ligand trap, has shown promise in preclinical studies and is now in a clinical trial in patients with MF. [bib_ref] AVID200, a potent trap for TGF-β ligands inhibits TGF-β1 signaling in human..., Varricchio [/bib_ref]
## | genomic risk scores for allogeneic stem cell transplantation in mf
At present, allografting remains the only potentially curative treatment for patients with MF but is confounded by therapy-related morbidity and nonrelapse mortality, due at least in part, to the older age and the general poor condition of many patients at diagnosis. [bib_ref] Long-term outcome after allogeneic hematopoietic cell transplantation for myelofibrosis, Robin [/bib_ref] Transplant outcomes have improved further with advances in transplant technology which has increased donor availability, reduced transplant toxicity and the introduction of reducedintensity conditioning. [bib_ref] Comparison of reduced intensity conditioning regimens used in patients undergoing hematopoietic stem..., Jain [/bib_ref] Furthermore, the potential use of JAKi therapy, pretransplant to reduce splenomegaly and improve constitutional symptoms, peritransplant to reduce graft vs host disease and posttransplant to improve long-term outcomes appears attractive but requires confirmation. [bib_ref] Allo-SCT for myelofibrosis: reversing the chronic phase in the JAK inhibitor era?, Tamari [/bib_ref] Until recently, there were no suitable risk stratification methodology to optimize the decision-making process for treating MF patients with transplantation. The best-studied method was the Lille score, first introduced in 1996, for nontransplant treatment of patients with PMF. [bib_ref] Prognostic factors in agnogenic myeloid metaplasia: a report on 195 cases with..., Dupriez [/bib_ref] Since then, multiple new risk assessment methods have been introduced for PMF and extrapolated into the treatment-decision-making process for all MF patients, including post-ET/PV MF who comprise about 60% of MF, and adapted with modest success for transplant selection criteria . These have included the International Prognostic Scoring System (IPSS), based on age >65 years, hemoglobin <10 g/L, leucocyte count >25 × 10 9 /L, the presence of circulating peripheral blasts and constitutional symptoms; the Dynamic IPSS (DIPSS), which assesses prognosis during the course of the disease and is based on the prognostic variables as IPSS but ranks anemia as a higher risk factor; and the DIPSS-plus, which added transfusion-dependence, platelet count <100 × 10 9 /L and unfavorable karyotype to the DIPSS model. [bib_ref] New prognostic scoring system for primary myelofibrosis based on a study of..., Cervantes [/bib_ref] [bib_ref] A dynamic prognostic model to predict survival in primary myelofibrosis: a study..., Passamonti [/bib_ref] [bib_ref] DIPSS plus: a refined dynamic international prognostic scoring system for primary myelofibrosis..., Gangat [/bib_ref] Current risk scores have improved prognostic ability compared with IPSS and have incorporated genomic information to optimize risk stratification. These include the updated Mutation-Enhanced Prognostic Score (MIPSS-70 version 2.0), which includes CALR type I mutation, the presence of ASXL1, EZH2, SRSF2, U2AF1, or IDH1/2 mutations and a three-tiered cytogenetic risk classification, myelofibrosis secondary to PV and ET (MYSEC-PM) score, which includes the presence of constitutional symptoms, platelets <150 × 10 9 /L, hemoglobin <11 g/dL, circulating blasts >3% and a CALR-unmutated genotype, and a simpler genetic risk score, the genetically inspired scoring system. [bib_ref] MIPSS70+ version 2.0: mutation and karyotypeenhanced international prognostic scoring system for primary..., Tefferi [/bib_ref] [bib_ref] A clinical-molecular prognostic model to predict survival in patients with post polycythemia..., Passamonti [/bib_ref] [bib_ref] GIPSS: genetically inspired prognostic scoring system for primary myelofibrosis, Tefferi [/bib_ref] These risk models are clearly useful in the nontransplant setting but are not optimal for transplant purposes, possibly due to the omission of transplant-and patient-specific factors that influence clinical outcomes following a transplant. [bib_ref] Early post-transplantation factors predict survival outcomes in patients undergoing allogeneic hematopoietic cell..., Jain [/bib_ref] [bib_ref] MIPSS70+ v2.0 predicts long-term survival in myelofibrosis after allogeneic HCT with the..., Ali [/bib_ref] In 2019, a risk score, the myelofibrosis transplant scoring system (MTSS), was developed specifically for all MF patients to predict outcomes after transplantation. [bib_ref] Comprehensive clinical-molecular transplant scoring system for myelofibrosis undergoing stem cell transplantation, Gagelmann [/bib_ref] MTSS was formulated at diagnosis in a cohort of 361 patients (PMF = 260; post-ET MF = 101; post-PV MF = 46) and excluded patients who had progressed to acute leukemia. It is based on multivariable analysis which identified age (≥57 years), Karnofsky performance status (<90%), pretransplant thrombocytopenia, leucocyte count (>25 × 10 9 /L), HLA-mismatched unrelated donor, ASXL1 mutation and triple-negative, and non-CALR/MPL or JAK2 driver mutation as independent prognostic factors for posttransplant survival. A comparison of the MTSS with some of the nontransplant risk-scoring systems, including those (such as MPISS-70 or MYSEC-PM) incorporating genetic information that now considered an integral facet, suggests an improvement in the patient selection for transplantation for all MF patients with respect to posttransplant outcomes. [bib_ref] Impact of high-molecular-risk mutations on transplantation outcomes in patients with myelofibrosis, Tamari [/bib_ref] depicts a potential treatment algorithm incorporating transplantation for patients with MF. It is of interest that neither constitutional symptoms, nor clinical variables such as transfusion dependence, appear to influence outcome after transplantation. Clearly, the MTSS will require additional validation prior to wider clinical use, but represents conceivably an important step forward in precision cancer medicine for patients with MF.
## | a five decade odyssey of interferon-α in chronic
## Myeloproliferative leukemias
Interferon-α (IFNα) is a member of a large family of glycoproteins of biologic origin with antiviral and antiproliferative properties. Studies in the early 1980s showed that IFNα was active in CML and able to induce major cytogenetic remissions in 5% to 15% of patients with restoration of Ph-negative (putatively normal) hematopoiesis and a survival advantage. IFNα, in particular the long-acting form, has since been tested in MPNs in phase 2 studies, including the MPD-RC 111 trial in which recombinant IFNα−2a was observed to reverse all hematological features of PV in 22% of the 50 HU refractory/intolerant patients. [bib_ref] Pegylated interferon alfa-2a in patients with essential thrombocythemia or polycythemia vera: a..., Masarova [/bib_ref] [bib_ref] Pegylated interferon alfa-2a for polycythemia vera or essential thrombocythemia resistant or intolerant..., Yacoub [/bib_ref] [bib_ref] Final results of prospective treatment with pegylated interferon Alfa-2a for patients with..., Yacoub [/bib_ref] More recently, ropeginterferon-α−2b was compared to HU in a prospective randomized controlled phase 3 trial (PROUD-CONTINUATION) in patients with PV. [bib_ref] Ropeginterferon alfa-2b versus standard therapy for polycythaemia vera (PROUD-PV and CONTINUATION-PV): a..., Gisslinger [/bib_ref] The 4-year follow-up of this study demonstrated sustained complete hematological responses in 61% of patients treated with IFNα vs 43% in the HU group (P = .02), with a similar efficacy for reducing thromboembolic events to very low rates (0.0%, 0.0%, and 1.1% of patients treated with IFNα in the second, third, and fourth years). [bib_ref] Thromboembolic risk reduction and high rate of complete molecular response with long-term..., Kiladjian [/bib_ref] Of particular interest was the observation that 67% of IFNα-treated patients achieved reduction in the JAK2V617F allele burden compared with 26% of the HU-treated cohort (RR: 2.5 [95% CI: 1.7-3.7; P <
.0001]). Serial measurements of the allele burden in the IFNα-treated patients confirmed 14% complete molecular remissions and 36% major molecular response at month 48. This raises the important question of whether these "molecular responders" would have their life prolonged by treatment with IFNα compared with HU. Furthermore, the majority of the "molecular responders" also achieved complete hematological remissions, suggesting that treatment discontinuation could be possible. Toxicity after 4 years was generally mild and reversible, and was similar between the two groups of patients: 28% (IFNα) vs 22.9% (HU); four cases of secondary malignancies have been reported in the HU patients so far. Collectively, these observations signal the need to consider the merits of IFNα replacing HU as the primary treatment for PV and randomized controlled studies to test it against ruxolitinib, the latter of which was approved as a second-line therapy for PV in 2014. [bib_ref] Ruxolitinib versus standard therapy for treatment of polycythemia vera, Vannucchi [/bib_ref]
# | conclusion
MPNs are clonal stem cell neoplasms with heterogenous clinical phenotypes and complex genetic architecture, which currently remain incurable. Treatment with a choice of two first-line orally administered JAKis, ruxolitinib, and fedratinib, for patients with MF has led to substantial symptomatic improvement, improvement in quality of life, but no sustained disease modification or long-term remissions. It seems crucial, therefore, to improve our understanding of the various JAKi resistance mechanisms and develop new treatment approaches beyond JAK-STAT inhibition. Moreover, new MPN biological and clinical entities, such as those associated with the novel JAK2ex13InDel mutation and linked with eosinophilia and erythrocytosis, add to the heterogeneity and underline the need for reconciliation within the treatment algorithm.The improved molecular technologies, such as single-cell transcriptome approaches tracking somatic mutations and characterizing cellular and molecular features are informing on clonal dynamics and architecture in MF and allowing insights into disease progression and treatment. scRNAseq demonstrates a Mk-biased differentiation in MF stem and progenitors that are associated with specific molecular signatures and suitable for Mk-targeted therapies. In contrast to MF, most patients with CML have an overall survival similar to that of the general population, following treatment with any of the ABL1 TKIs licensed for first-line therapy. Resistance, however, from mutant BCR-ABL1 subclones arising from the likely disruption of the TKI/TKD interaction, is increasingly being recognized as a serious problem due to the biological effect on BCR-ABL1 downstream signaling and oncogenic gene expression. These mutant clones display highly upregulated CDK6 and raise the possibility of new therapeutic approaches. Last but not least, the long-term data from the randomized PROUD-CONTINUATION study in patients with PV raises the prospect of IFNα replacing HU as the primary treatment for this MPN.
## Author manuscript
## Author manuscript
Mughal et al.
Page 20
## Table 1
Myelofibrosis clinical drug development programs
## Drug (class) comments reference
Momelotinib (JAK1/2 inhibitor) Ongoing Phase 3 (NCT04173494) [bib_ref] Novel therapies in myeloproliferative neoplasms (MPN): beyond JAK inhibitors, Economides [/bib_ref] [bib_ref] Results from a phase 2 study of navitoclax in combination with ruxolitinib..., Harrison [/bib_ref] Pacritinib (JAK2/ FLT 3 inhibitor) Ongoing Phase 3 (NCT03165734) [bib_ref] A pilot study of the telomerase inhibitor imetelstat for myelofibrosis, Tefferi [/bib_ref] [bib_ref] CHZ868, a type II JAK2 inhibitor, reverses type JAK inhibitor persistence and..., Meyer [/bib_ref] Luspatercept ± ruxolitinib (receptor type IIb and IgG1Fc domain) Ongoing phase 2 (NCT03194542) [bib_ref] A phase 2 study of luspatercept in patients with myelofibrosis-associated anemia, Gerds [/bib_ref] CPI-0610 ± ruxolitinib (BET inhibitor) Ongoing phase 2 (NCT02158858) [bib_ref] MANIFEST, a phase 2 study of CPI-0610, a bromodomain and extraterminal domain..., Mascarenhas [/bib_ref] Navitoclax + ruxolitinib (Bcl-2 inhibitor) Ongoing phase 2 (NCT03222609) [bib_ref] Results from a phase 2 study of navitoclax in combination with ruxolitinib..., Harrison [/bib_ref] Pomalidomide + ruxolitinib (immunomodulatory) Ongoing phase 2 (NCT01644110) [bib_ref] Updated results from the German MPNSG-0212 combination trial: ruxolitinib plus pomalidomide in..., Stegelmann [/bib_ref] Thalidomide + ruxolitinib (immunomodulatory) Ongoing phase 2 (NCT03069326) [bib_ref] Safety and efficacy of combined ruxolitinib and thalidomide in patients with myelofibrosis:..., Rampal [/bib_ref] Tagraxofusp (CD123 targeted antibody) Ongoing phase 2 (NCT02268253) [bib_ref] Results from a phase 1/2 clinical trial of tagraxofusp (SL-401) in patients..., Pemmaraju [/bib_ref] Bomedemstat (LSD1 inhibitor) Ongoing phase 2 (NCT03136185) [bib_ref] A phase 2a study of the LSD1 inhibitor Img-7289 (bomedemstat) for the..., Pettit [/bib_ref] Alisertib (Aurora kinase inhibitor) Ongoing phase 1/2 (NCT02530619) [bib_ref] Aurora kinase A inhibition provides clinical benefit, normalizes megakaryocytes and reduces bone..., Gangat [/bib_ref] Imetelstat (Telomerase inhibitor) Completed phase 1/2 (NCT01731951) 44 AVID200 (TGF-β ligand trap) Ongoing phase 1 (NCT03895112) [bib_ref] AVID200, a potent trap for TGF-β ligands inhibits TGF-β1 signaling in human..., Varricchio [/bib_ref] Abbreviations: BET, bromodomain and extraterminal; LSD1, lysine specific demethylase1; TGF-β, transforming growth factor beta.
Hematol Oncol. Author manuscript; available in PMC 2022 March 04.
ACKNOWLEDGEMENTS |
Utility of Transcranial Magnetic Simulation in Studying Upper Motor Neuron Dysfunction in Amyotrophic Lateral Sclerosis
Citation: Geevasinga, N.; Van den Bos, M.; Menon, P.; Vucic, S. Utility of Transcranial Magnetic Simulation in Studying Upper Motor Neuron Dysfunction in Amyotrophic Lateral Sclerosis. Brain Sci. 2021, 11, 906.
# Introduction
Amyotrophic lateral sclerosis is a progressive neurodegenerative disorder first described by Charcot. The incidence of ALS has been estimated at two per hundred thousand, with a median survival of 3-5 years. The disease can start focally and spread insidiously to other regions over time with respiratory muscle weakness heralding the eventual demise of patients. The majority of ALS patients (65-75%) present with asymmetrical weakness, leading to wasting of the limb muscles, typically spreading along the neuroaxis and affecting contiguous motor neurons. In 20% of ALS patients the disease can involve the bulbar muscles, presenting with progressive dysphagia and dysarthria. Neuronal degeneration in ALS encompasses both the central and peripheral nervous systems, leading to the pathognomonic clinical features of ALS with a combination of upper (UMN) and lower motor (LMN) neuron clinical examination findings..
Cortical hyperexcitability appears to be an early feature of ALS, highlighting the importance of the UMN dysfunction in the disease process. Transcranial magnetic stimulation is a noninvasive tool that allows studying UMN dysfunction and thereby cortical hyperexcitability utilising novel neurophysiological parameters. Cortical hyperexcitability is a unique and specific feature of ALS, differentiating the pathophysiological processes in ALS from other mimic disorders such as Kennedy's disease, oculopharyngeal muscular dystrophy, multifocal motor neuropathy and spinal muscular atrophy.
Whilst the underlying pathogenic process in ALS is still unclear, three different hypotheses have been proposed. After the initial description by Charcot, Eisen and colleagues proposed the "dying forward" hypothesis, where hyperexcitability of descending corticomotoneuronal tracts resulted in motor neuron degeneration via an anterograde transsynaptic glutamate excitotoxic process. Clinical observations have supported this dying forward mechanism including: (i) Sparing of muscles with a sparse corticomotoneuronal supply (extraocular and sphincter muscles) and clinical appearances of the split-hand pattern; (ii) ALS phenotypes with pure LMN findings were infrequent; and (iii) lack of ALS in the animal world, postulated due to developmental differences in corticomotoneuronal projections onto the spinal motor neurons across species. In contrast, the primacy of LMNs in ALS pathogenesis have been suggested in the dying-back hypothesis, albeit only supported by animal models. Additionally, the independent degeneration hypothesis was also advanced which suggested that UMN and LMN degeneration occurs independently.
Supporting the notion for the importance of UMN dysfunction in ALS, recent clinical, neurophysiological, radiological and genetic advances, have suggested the importance of cortical hyperexcitability in ALS pathogenesis. Given the diagnostic challenges seen in ALS, cortical hyperexcitability may be an early diagnostic biomarker in ALS, allowing earlier diagnosis of ALS patients. In this review, we highlight the utility of TMS in identifying cortical hyperexcitability in ALS, furthering the understanding of the pathophysiological processes driving ALS, aiding in the diagnosis of this condition and the utility cortical hyperexcitability as a diagnostic biomarker in this progressive and invariably fatal disease.
## Cortical hyperexcitability
To explore and assess the function of the motor pathways in ALS and thereby gaining better understanding of ALS pathophysiology, TMS can be utilised to assess cortical excitability. The utility of single, paired and triple-pulse transcranial magnetic stimulation (TMS) techniques have significantly enhanced the understanding of cortical hyperexcitability in ALS pathophysiology, thereby resulting in novel diagnostic approaches. TMS is a neurophysiological technique which is performed non-invasively by stimulating the human motor cortex; this technique was first described by Barker and colleagues. TMS activates the motor cortex at a depth of approximately 1.5 to 2.1 cm, resulting in the generation of direct (D) and/or indirect (I) waves, reflecting the function of cortical output cells (Betz cells) and intracortical neuronal networks within the primary motor cortex (M1). In a clinical setting, to assess cortical excitability and evaluate the integrity of corticospinal pathways, the following TMS parameters are most useful: (i) short interval intracortical inhibition; (ii) motor threshold; (iii) motor evoked potential (MEP) amplitude; (iv) cortical silent period (CSP) duration and (v) short interval intracortical facilitation (SICF),.
## Tms neurophysiological parameters suspected physiological mechanisms
## Motor threshold
Density of corticomotoneuronal projections on spinal motor neurons as well as cortical excitability being influenced by glutamatergic neurotransmission and Na + channel conductance.
Motor evoked potentials (MEP) Summation of complex corticospinal volleys consisting of D and/or I waves onto the spinal motor neuron.
## Central motor conduction time (cmct)
Prolongation of the CMCT probably reflects degeneration of the fastest conducting corticomotoneuronal axons in ALS patients, resulting in increased desynchronization of corticomotoneuronal volleys due to axonal loss.
## Cortical silent period (csp)
Probably multifactorial in nature, with the early segment mediated by spinal processes, whilst the later segment being mediated by long-lasting inhibitory post-synaptic potentials generated via GABAergic B receptors.
## Short interval intracortical inhibition (sici)
Mediated by inhibitory GABAergic intracortical circuits located within the primary motor cortex, acting via GABA A receptors.
Intracortical facilitation (ICF) Generated at the motor cortex, associated with increases in the I-wave amplitude, being reduced by GABA A receptor agonists.
## Short interval intracortical facilitation (sicf)
A cortical origin has been proposed reflecting activity of facilitatory cortical circuits, potentially from axonal-axonal reinforcement of interneuronal activations resulting in synergistic increases in corticospinal output.
Short interval intracortical inhibition (SICI) is the most robust TMS parameter of cortical excitability, and is generated by the paired-pulse TMS paradigm where a subthreshold conditioning stimulus precedes a suprathreshold test stimulus at a preset interstimulus interval (ISI). Using this paradigm, when TMS pulses are delivered between an ISI of 1-5 ms, the test response is inhibited, thereby generating SICI. As the ISI is increased to intervals of 7 and 30 ms, a facilitation of the test responses is noted resulting in intracortical facilitation (ICF). In the original constant stimulus technique, the condition and test stimulus intensity were kept constant, but this technique was affected by MEP amplitude variability. The threshold tracking TMS technique overcame this limitation by maintaining a constant target MEP response (0.2 mV) and varying the suprathreshold stimulus test intensity. Physiological studies exploring the generation of SICI have shown it is mediated by inhibitory GABAergic intracortical circuits acting via GABA A intracortical circuits, located within the primary motor cortex. In addition, a weak effect of glutaminergic neurotransmission has also been reported.
A reduction, or absence of SICI in sporadic ALS patients has been reported as indicative of cortical hyperexcitability. The reduction in SICI is an early feature of ALS; it precedes the onset of LMN dysfunction and correlates with neurophysiological biomarkers of peripheral neurodegeneration. Abnormalities of SICI have also been associated with patterns of disease spreadand development of clinical features seen in ALS such as the split-hand phenomenon. Reduction of SICI was also evident in atypical sporadic ALS phenotypes, including patients with pure LMN syndromes, thereby indicating subclinical UMN dysfunction.
The CSP duration is a distinct neurophysiological biomarker of cortical excitability. It is mediated by long-lasting inhibitory post-synaptic potentials generated via GABA B receptors. The CSP duration can also be modulated by the density of the corticomotoneuronal projections onto motor neurons, motor attention, the extent of voluntary motor drive and neurotransmitters such as dopamine. Reduction of CSP duration has been noted in sporadic and familial ALS cohorts, being most prominent in early disease stages. Abnormalities of ipsilateral CSP have also been reported as an early feature in ALS, signifying loss of transcallosal function. Disinhibition at a cortical level, due to degeneration or dysfunction of cortical inhibitory interneurons, appears to underlie the reduction of CSP duration in ALS.
Marked changes in motor thresholds have been also described in ALS. The resting motor threshold reflects the density of corticomotoneuronal projections onto the spinal motor neurons as well as cortical excitability, influenced by glutamatergic neurotransmission and Na + channel conductance. When motor thresholds have been studied longitudinally in TMS studies, ALS patients had reduced motor thresholds in the early stages of the disease, progressing to an increase in motor threshold and eventual cortical inexcitability. This early reduction in motor thresholds, likely reflects cortical hyperexcitability, and appears to be most prominent in patients with profuse fasciculations, exaggerated deep-tendon reflexes and preserved muscle bulk, thereby supporting the notion that fasciculations are derived from cortical dysfunction in the early stages of ALSand that cortical hyperexcitability may underlie the development of neurodegeneration in ALS. Similarly, increased MEP amplitudes have been noted in ALS. Similar to motor thresholds, the MEP amplitude reflects the density of corticomotoneuronal projections onto motor neurons, and it appears to be influenced by neurotransmitters such as glutamate. Increases in MEP amplitude have been reported as an early feature in ALS, correlating with surrogate neurophysiological biomarkers of axonal degeneration.
Another novel neurophysiological parameter, SICF, is also generated by the pairedpulse TMS technique. This neurophysiological parameter is obtained when a conditioning stimulus is set to peri-and suprathreshold levels followed by a test stimulus set at threshold intensity. The threshold tracking TMS technique has been utilised to generate SICF, revealing two distinct peaks at ISI of 1.5 and 3 ms. Whilst the exact physiological processes underlying SICF development still remain undetermined, a cortical origin has been proposed with SICF likely reflecting activity of facilitatory cortical circuits. Increases in SICF were reported in sporadic ALS patients, noticed in combination with a reduction in SICI. A novel index of excitation, a neurophysiological biomarker of cortical excitability expressing SICF as a function of SICI, was recently described, revealing an increased index of excitation in ALS patients, suggesting overactivity of facilitatory circuits contributing to cortical hyperexcitability.
Other more complex TMS stimulation paradigms include the triple stimulation technique, which has also shown utility in detecting sub-clinical UMN dysfunction in the early stages of ALS.
Some have argued that cortical hyperexcitability is a compensatory mechanism secondary to motor neuron degeneration. However, given that cortical hyperexcitability was not evident in ALS mimicking disorders, despite a comparable degree of LMN burden, this would argue against a compensatory mechanism. In addition, the partial normalization of cortical hyperexcitability with riluzole, an anti-glutaminergic agent which has modest therapeutic benefits in ALS, provides further support for the pathogenic role of cortical hyperexcitability in ALS. The presence of cortical hyperexcitability as measured as reductions in SICI has utility as a prognostic marker, predicting a worse prognosis in an ALS cohort.
## The split-hand pattern
A role for corticomotoneuronal dysfunction in ALS pathogenesis is supported by the clinical observation that the split-hand pattern of muscle wasting is a specific feature in ALS. The split-hand phenomenon refers to preferential wasting of the intrinsic hand muscles the abductor pollicis brevis (APB) and first dorsal interosseous (FDI) muscles, when compared to the hypothenar muscles. This dissociated pattern of intrinsic hand muscle atrophy is specific for ALS and is not noted in ALS mimic disorders. Transcranial magnetic studies have shown that cortical hyperexcitability appears to underlie the development of the split-hand phenomenon, suggesting a greater cortical representation of the thenar muscles, utilised for fine precision tasks, thereby reaffirming the importance of UMN dysfunction in ALS. Additionally, cortical hyperexcitability was reported to be a specific feature of the split-hand plus phenomenon, where preferential weakness of the thenar muscle is observed when compared to flexor pollicis brevis. More recently, a split-leg (preferential weakness of posterior calf muscles) and a split-elbow (preferential weakness of the biceps brachii compared to the triceps muscle) phenomenon were reported as specific clinical features of ALS and also attributed to cortical dysfunction.
## Insights gathered from studying cortical excitability in familial als
The recent advances in the genetic understanding of ALS have provided further insights into the role of the corticomotor neurons. In particular, the identification of an increased hexanucleotide repeat expansion (GGGGCC) in the c9orf72 gene on chromosome 9p21.1which accounts for over 40% of familial and approximately 8% of apparently "sporadic" ALS caseshas radically altered the understanding of ALS pathogenesis. The c9orf gene expansion is also the most common cause of familial frontotemporal dementia, thereby placing ALS in the continuum of an underlying cortical neurodegenerative disorder. At one end of the spectrum of this gene mutation lies frontotemporal dementia, on the other end ALS, with some patients presenting with a mixed picture of cognitive and motor deficits, FTD-ALS. The overwhelming majority of ALS cases and approximately 50% of FTD cases are characterized by inclusions consisting of the RNA-binding protein TDP-43 (TAR DNA-binding protein 43) in neurons and glia, suggesting a shared pathogenic process. The role of cortical dysfunction in ALS is further evident with studies showing that up to 30-50% of ALS patients present with cognitive dysfunction, suggesting extra motor involvement.
Specifically SICI reduction appears to be an early and prominent feature in familial ALS, including patients expressing mutations in the superoxide dismutase-1, fused in sarcoma (FUS)and c9orf72 expansion, and correlating with peripheral neurodegeneration. The presence of cortical hyperexcitability was noted in a cohort of c9orf72 FALS patients, but not in asymptomatic c9orf72 expansion carriers. These findings confirm that cortical hyperexcitability is a feature of familial forms of ALS and suggests that it develops prior to clinical onset of FALS.
As such, irrespective of the underlying genetic mutation in familial ALS patients, cortical hyperexcitability appears to represent a uniform pathophysiological process in ALS. Of relevance, recent studies looking at complex mathematical modeling of ALS pathophysiology have inferred a six-step process in ALS, with a prolonged prodromal period, potentially extending to the perinatal period, where patients may develop an increased risk of developing ALS in utero. This multistep model of disease progression in ALS was also recently suggested in another study looking at Australia, Japanese and Korean patients. Cortical hyperexcitability probably represent an important step in ALS pathogenesis, developing just prior to or at onset of neuronal degeneration. This notion is supported by findings of significant correlations between neurophysiological features of cortical hyperexcitability and motor amplitude and clinically measured muscle strength scores. Recent animal studies further support this notion, identifying neuronal hyperexcitability at a pre-clinical stage, upstream of the spinal motor neurons.
## The role of tms in aiding diagnosis
The current diagnosis of ALS relies on the clinical and neurophysiological identification of a combination of UMN and LMN signs in multiple body regions, along with evidence of ongoing disease progression noted over time. Importantly, in confirming ALS, any mimic disorders need to be excluded prior to establishing a definitive diagnosis of ALS. The clinically based ALS diagnostic criteria (El-escorial and the Awaji criteria) can be insensitive, particularly in the early stages of the disease or in the setting of atypical ALS phenotypes. This can then unfortunately result in significant diagnostic delays. Consequently, delayed diagnosis may slow the implementation of adequate management strategies, including commencement of therapies such as riluzole, and recruitment into therapeutic trials, perhaps beyond the therapeutic window period. The recently reported Gold Coast criteria aim to simplify diagnostic criteria by requiring UMN and LMN dysfunction in at least one body region or two regions with LMN findings. The clinical identification of UMN signs in ALS may be limited by complex physiological factors, including marked muscle wasting along with dysfunction of descending motor pathways and local spinal circuits that usually facilitate UMN signs. Separately, adaptive changes within the neuromuscular system may further complicate assessment.
Consequently, direct assessment of UMN function by TMS techniques could overcome the limitations of the disease process, providing an objective diagnostic biomarker of UMN dysfunction. Studies have shown that utilising TMS may allow identifying a greater proportion of patients with ALS at an earlier stage. The most robust marker in the aid of diagnosing ALS was a reduced SICI or an inexcitable motor cortex. Other changes documented included reductions in resting motor threshold and CSP duration, along with an increases in MEP amplitude and ICF.
## Utilising tms to differentiate between clinical disorders
Whilst the most common phenotype of ALS includes involvement of both the UMN and LMNs, there are other forms of the disease that can present with atypical phenotypes. One particular variant includes predominant involvement of the UMNs, termed primary lateral sclerosis (PLS), which is characterized by a slowly progressive UMN syndrome with absence of LMN features for up to four years. Utilising TMS, a greater frequency of motor cortex inexcitability was noted in the PLS phenotype than in ALS patients. This finding is probably related to a greater degree of neurodegeneration within the motor cortex and the corticospinal tracts, as the resting motor threshold reflects the density of corticomotoneuronal projections onto spinal motor neurons as well as excitability of large motor cortical neurons (Betz cells). Furthermore, TMS is useful in differentiating ALS from a mimic disorder such as hereditary spastic paraparesis, which does not have any evidence of cortical dysfunction as measured by TMS.
Another atypical variant of ALS is the flail leg (FL) syndrome also referred to as leg amyotrophic diplegia; this phenotype was first described in the early 20th century. Clinically, the FL syndrome is characterised by a predominant LMN phenotype with absent or subtle upper motor neuron features. The disease remains limited to the lower limbs for a prolonged period. The median survival for the FL phenotype is more favourable that the typical ALS phenotype, estimated to be 90 months. There appears to be a heterogeneity of the FL phenotype, and some studies have reported a greater degree of UMN dysfunction and shorter survival, approximating the disease course of the more classical ALS. Cortical hyperexcitability measured as a significant reduction of SICI, was only evident in the FL patients exhibiting UMN signs, with the degree of cortical hyperexcitability similar to other ALS phenotypes. In contrast, in FL patients with absent UMN signs, cortical excitability was preserved, despite a prolonged central motor conduction time in the entire cohort. These findings suggest heterogeneity in the clinical phenotype and underlying pathophysiological processes of FL patients. Similarly, in patients with the flail arm variant, cortical excitability was evident similar to patients with classical ALS with reductions in SICI and resting motor threshold documented.
# Conclusions
Cortical hyperexcitability is an important pathophysiological and diagnostic biomarker in sporadic and familial ALS. Cortical hyperexcitability is heralded by development of cortical disinhibition and increases in activity of cortical facilitatory circuits with evolution of the disease course. It precedes the development of LMN dysfunction and correlates with neuronal degeneration. Future studies looking at transcranial magnetic stimulationelectroencephalography, molecular/proteomics, genetic studies evaluating familial forms of ALS and advanced neuroimaging techniques may help provide further insights into the pathophysiological mechanisms underlying ALS. A better understanding of the underlying pathophysiological processes could lead to identification of novel pathogenic biomarkers that serve as therapeutic targets and be utilised as biomarkers in measuring disease activity or disease progression in future ALS therapeutic trials. |
Influence of polymer size, liposomal composition, surface charge, and temperature on the permeability of pH-sensitive liposomes containing lipid-anchored poly(2-ethylacrylic acid)
Background: Liposomes containing pH-sensitive polymers are promising candidates for the treatment of tumors and localized infection. This study aimed to identify parameters influencing the extent of contents release from poly(ethylacrylic acid) (PEAA) vesicles, focusing on the effects of polymer size, lipid composition, vesicle surface charge, and temperature. Methods: Anchored lipid pH-sensitive PEAA was synthesized using PEAA with a molecular weight of 8.4 kDa. PEAA vesicles were prepared by insertion of the lipid-anchored PEAA into preformed large unilamellar vesicles. The preformed liposomes were manipulated by varying the phosphocholine and cholesterol content, and by adding negative or positive charges to the liposomes. A calcein release assay was used to evaluate the effects of polymer size, liposome composition, surface charge, and temperature on liposomal permeability. Results: The release efficiency of the calcein-entrapped vesicles was found to be dependent on the PEAA polymer size. PEAA vesicles containing a phosphatidylcholine to cholesterol ratio of 60:40 (mol/mol) released more than 80% of their calcein content when the molecular weight of PEAA was larger than 8.4 kDa. Therefore, the same-sized polymer of 8.4 kDa was used for the rest of study. The calcein release potential was found to decrease as the percentage of cholesterol increased and with an increase in the phosphocholine acyl chain length (DMPC . DPPC . DSPC). Negatively charged and neutral vesicles released similar amounts of calcein, whereas positively charged liposomes released a significant amount of their contents. pH-sensitive release was dependent on temperature. Dramatic content release was observed at higher temperatures. Conclusion: The observed synergistic effect of pH and temperature on release of the contents of PEAA vesicles suggests that this pH-sensitive liposome might be a good candidate for intracellular drug delivery in the treatment of tumors or localized infection.
# Introduction
The design of nanoliposomal drug delivery systems able to undergo controlled fusion and release has been the subject of intense research. [bib_ref] Microencapsulated liposomes in controlled drug delivery: strategies to modulate drug release and..., Dhoot [/bib_ref] [bib_ref] Fusogenic liposome delivers encapsulated nanoparticles for cytosolic controlled gene release, Kunisawa [/bib_ref] [bib_ref] Photoactivated enhancement of liposome fusion, Bennett [/bib_ref] [bib_ref] Fusion of liposomes with mitochondrial inner membranes, Schneider [/bib_ref] [bib_ref] Control of the interaction between membranes or vesicles: adhesion, fusion and release..., Volodkin [/bib_ref] Considering that the primary route of liposome uptake in cells is through relatively low pH endosomes, pH-sensitive liposomes may be a favorable approach. [bib_ref] Current status of pH-sensitive liposomes in drug delivery, Drummond [/bib_ref] [bib_ref] Cytosolic drug delivery using pH-and light-sensitive liposomes, Gerasimov [/bib_ref] [bib_ref] Trigger lipids inducing pH-dependent liposome fusion, Ogawa [/bib_ref] It has been reported that pH-sensitive vesicles with polymers immobilized on their surfaces can effectively mediate this type of release and fusion. [bib_ref] On the formulation of pH-sensitive liposomes with long circulation times, Simo -Es [/bib_ref] In previous studies, we constructed pHsensitive liposomes by either conjugating thiolated poly(ethylacrylic acid) (PEAA) to vesicles containing maleimide 10 or inserting lipid-anchored polymers into preformed liposomes. [bib_ref] Alkylated derivatives of poly(ethylacrylic acid) can be inserted into preformed liposomes and..., Chen [/bib_ref] The resulting PEAA liposomes are stable under physiological conditions and can fuse with adjacent membranes and release their contents upon acidification. Importantly, it has been demonstrated that the association of lipid-anchored PEAA with vesicles promotes intracellular delivery of vesicular contents in vitro in a cultured cell line. [bib_ref] Cationic poly(ethyleneglycol) lipids incorporated into pre-formed vesicles enhance binding and uptake to..., Fenske [/bib_ref] This study attempted to identify parameters influencing the extent of release of contents from PEAA vesicles, focusing on the effects of polymer size, lipid composition, vesicle surface charge, and temperature. Such information is important for optimizing liposomal PEAA formulations as intracellular drug delivery carriers that are able to release their contents efficiently in a pH-dependent manner.
## Materials and methods chemicals
All chemical reagents were commercially available products of high purity and were used without further purification. In order to quantify the polymer, PEAA was fluorescentlabeled with small amounts of pyrene, and synthesized as described previously. Different phospholipids, including egg phosphatidylcholine, 1,2 dimyristoyl-sn-glycero-3-phosphocholine (DMPC), 1,2 dipalmitoyl-sn-glycero-3-phosphocholine (DPPC), 1,2 distearoyl-sn-glycero-3-phosphocholine (DSPC), phosphatidylinositol (PI), dioleoyl-1,2-diacyl-3dimethylammonium-propane (DODAP), and 1,2-distearoylsn-glycero-3-phosphoethanolamine (DSPE) were obtained from Northern Lipids (Vancouver, Canada). N,N-dioleyl-N,N-dimethylammonium chloride (DODAC) was a gift from Helican Biotechnology Inc (Vancouver, Canada).
## Synthesis of lipo-peaa
Anchored PEAA (lipo-PEAA) was synthesized using 8.4 kDa PEAA as described previously. [bib_ref] Fluorescently labeled poly(ethylene glycol) lipid conjugates with distal cationic headgroups, Chen [/bib_ref] Briefly, 100 mg (100 mmol unit, unit molecular weight 100) of PEAA and 3 mmol of 1-decylamine were dissolved in water (pH 7.0). A 20 mg/mL 1,3-dicyclohexylcarbodiimide solution was then slowly added to the reaction mixture until the amine disappeared (monitored by thin layer chromatography, CHCl 3 :MeOH:triethylamine, 8:2:0.2, visualized by ninhydrin). The resulting derivate was precipitated by adjusting the pH of the solution to pH 2-3. The supernatants were then removed and the pellets were redissolved in 2 M NaOH solution, stirred for 30 minutes, and readjusted to pH 2-3. The suspension was centrifuged and the resulting pellets were washed 3-5 times in water and lyophilized. A typical yield was 60%-70%. In our paper, PEAA-C 10 (lipo-PEAA with a decyl chain introduced) was used for post-insertion processing and forming PEAA liposomes.
## Preparation of peaa liposomes
Large unilamellar vesicles (LUVs) were prepared by extrusion as described by Hope et al. [bib_ref] Production of large unilamellar vesicles by a rapid extrusion procedure. Characterization of..., Hope [/bib_ref] Appropriate amounts of lipid mixtures with trace amounts of 3 H-CHE (1.33 µCi/4 µmol) in chloroform were dried under a stream of nitrogen gas to form a homogeneous lipid film. Any trace solvent was then removed under vacuum overnight. The lipid film was hydrated in HEPES-calcein buffered saline (20 mM HEPES, 100 mM calcein, pH 7.5) by vortex mixing. The resulting multilamellar vesicles were frozen/thawed (liquid nitrogen/55°C) five times and extruded 10 times at 55°C through two stacked 100 nm polycarbonate filters (Nuclepore™, Whatman, Clifton, NJ) using an extrusion device (Vancouver, Canada). Untrapped free calcein was removed by chromatography using a 1.1 × 20 cm Sepharose CL-6B column (Sigma Chemical Corporation) equilibrated with HEPES-buffered saline (20 mM HEPES, 150 mM NaCl, pH 7.5).
PEAA-LUVs were then prepared by incubation of preformed LUVs with lipo-PEAA solution (PEAA-C 10 , pH 7.4) in an appropriate ratio at room temperature overnight as previously described. [bib_ref] Fluorescently labeled poly(ethylene glycol) lipid conjugates with distal cationic headgroups, Chen [/bib_ref] The mixture was eluted on a Sepharose CL-6B column (1.5 × 20 cm) and the fractions were assessed for lipid and polymer content. Noninserted free polymer was separated from PEAA-LUVs on a Sepharose CL-6B column as described in our previous work. [bib_ref] Fluorescently labeled poly(ethylene glycol) lipid conjugates with distal cationic headgroups, Chen [/bib_ref] The percentage of polymer insertion varied according to the nature of the preformed vesicles, especially for the charged ones. Hydrophobic PEAA segments are unstable in an aqueous environment and are therefore expected to penetrate into the hydrophobic interior of a lipid bilayer through Van der Waals forces. [bib_ref] Membrane solubilization by hydrophobic polyelectrolyte: surface activity and membrane binding, Thomas [/bib_ref] [bib_ref] Kinetics of membrane micellization by hydrophobic polyelectrolyte poly(2-ethylacrylic acid), Thomas [/bib_ref] A high PEAA concentration would lead to solubilization of the lipid membrane, while a low PEAA concentration would probably cause defective lipid packing, resulting in release of the vesicle contents. Therefore, a 10 mol% PEAA composition is used in our research. 17 submit your manuscript | www.dovepress.com
## Dovepress
## Dovepress
Determination of liposome size Liposome size was determined by quasi-elastic light scattering using a Nicomp 370 submicron particle sizer (Santa Barbara, CA).
## Calcein release
A calcein release assay was used to evaluate liposomal permeability. [bib_ref] Methods to monitor liposome fusion, permeability, and interaction with cells, Düzgüneş [/bib_ref] Calcein, a nonpermeable aqueous fluorescent dye, was used as a marker for determining the stability of the vesicles and the permeability of the different vesicles on acidification and changes in temperature. Calcein was trapped in the liposomes at a self-quenching concentration (100 mM). When released from the vesicles, it was diluted and the fluorescent intensity of the solution increased. Two release assays were used in this study. One assessed the potency of different sized, nonanchored polymers for triggering calcein release. The other measured the calcein release from PEAA-LUVs.
## First release assay
For quantification of the vesicle permeability induced by polymers of different size (without a lipid anchor), fluorescent aqueous calcein (100 mM) was encapsulated into LUV (egg phosphatidylcholine to cholesterol ratio 60:40, mol/mol) and free calcein was removed using a gel column as already described for the preparation of PEAA liposomes. An assay solution of 1.0 mg/mL LUV containing calcein was prepared by diluting the stock formulation with HEPES-buffered saline (20 mM HEPES, 150 mM NaCl, pH 7.5). Next, 10 µL of polymer (1 mg/mL, pH 7.5) was added to 1.0 mL of the diluted LUV containing calcein. A 100 µL aliquot of the test solution was diluted to 1.2 mL with HEPES-buffered saline (pH 7.5). The fluorescence intensity of calcein (F 0 ) as the initial background was determined using an Aminco Bowman luminescence spectrofluorometer (SLM-Aminco, Urbana, IL) at 520 nm (emission) and 495 nm (excitation). The test solution was then adjusted to pH 4.5 with 1% HCl and allowed to equilibrate for 20 minutes at room temperature. A 100 µL aliquot was diluted to 1.2 mL with HEPES-buffered saline (pH 7.5) and the fluorescence intensity was measured (F t ). Maximal fluorescence intensity (F max ), representing complete release of encapsulated calcein, was measured after solubilization of the vesicles (a 100 µL aliquot diluted with 1.2 mL of HEPES-buffered saline) with 10 µL of 10% Triton X-100. The potency of polymer in inducing vesicle permeability was calculated as the percentage of calcein released as shown in equation 1.
[formula] % % max release F F F F t o o = − − × 100 (1) [/formula]
## Second release assay
The permeability of the polymer-associated PEAA-LUV containing calcein liposomes and of the control samples were acidified to pH 4.5 with 1% HCl. After 20 minutes of equilibration, an 100 µL aliquot was withdrawn and diluted with 1.2 mL of HEPES-buffered saline and the fluorescence intensity of the diluted solution was measured using an Aminco Bowman Series 2 luminescence spectrofluorometer at 530 nm (slit width 4 nm) under steady-state excitation at 495 nm (slit width 4 nm). The initial and final fluorescence intensities were determined using the method described above and the percentage release was calculated as in equation 1.
To determine the kinetics of calcein release at different temperatures, the test sample was incubated in a temperaturecontrolled water bath at a predetermined temperature. The initial pH of 7.5 was maintained for a period of 10 minutes before acidifying the solution to pH 4.5. At different time points, a 100 µL aliquot was withdrawn, diluted with 1.2 mL of HEPES-buffered saline, and the fluorescence intensity of the solution was measured (F t is the fluorescence intensity at time t). The maximum fluorescence intensity (F max ), representing complete release of the encapsulated calcein, was determined following solubilization of the vesicles with Triton X-100 (10% of lipid concentration). The percentage of calcein released was calculated using equation 1.
# Results
## Effect of polymer size on vesicle permeability
It was expected that the ability of PEAA to trigger release of vesicular contents would be dependent on polymer size, but the polymer size cutoff for inducing efficient release of commonly used liposomes was unknown. Therefore, using an in vitro calcein release assay, we quantitatively measured content release from vesicles composed of egg phosphatidylcholine and cholesterol (ratio 60:40) with different-sized PEAA. [fig_ref] Figure 1: Effect of polymer size on pH-sensitive calcein release from large unilamellar vesicles [/fig_ref] shows that the tendency of calcein to be released under acidic conditions increased as the polymer size increased. The extent of calcein release reached a plateau when PEAA was larger than 8.4 kDa. In order to evaluate the effects of lipid composition, charge, and temperature on the permeability of pH-sensitive vesicles containing lipo-PEAA, we used the same-sized polymer of 8.4 kDa for the rest of study.
## Formulation and characterization of ph-sensitive vesicles containing lipo-peaa
Because the pH-sensitive PEAA vesicles were intended for pharmaceutical use, it was important to identify parameters influencing formulation of characteristic PEAA vesicles. These parameters included the efficiency of polymer association, formulation stability, size change during polymer insertion, and efficient content release under acidic conditions. Different vesicles with either a positive, neutral, or negative net surface charge were used for our post-insertion study. As indicated in [fig_ref] Figure 2: Characteristics of pH-sensitive PEAA vesicles with differing lipid composition by the post-insertion... [/fig_ref] , positively charged liposomes with either a permanent charge (DODAC) or protonable amine (DODAP) resulted in a much higher amount of associated polymers, while neutrally (DPPC) and negatively (PI or DSPE) charged vesicles contained smaller amounts of inserted polymers. The different phosphocholine acyl chain lengths, eg, DPPC, PI, or DSPE, did not affect the efficiency of polymer insertion. Upon mixing DODAC-containing liposomes and lipo-PEAA, extensive aggregation took place [fig_ref] Figure 2: Characteristics of pH-sensitive PEAA vesicles with differing lipid composition by the post-insertion... [/fig_ref] , anticipated because of the strong electron interaction between the multiple oppositely charged DODAC-containing liposomes and lipo-PEAA. Surprisingly, the size increase of DODAPcontaining liposomes after polymer insertion was much smaller than that for DODAC, although the value was still higher than that of both the neutrally and negatively charged submit your manuscript | www.dovepress.com Dovepress Dovepress liposomes. Because the amine of DODAP is protonable and its charged form is environmentally dependent, it may exist in a partially charged form in the bilayer. A much lower pKa for a similar amino lipid, AL1, in the membrane bilayer has been reported by This may also be the case for DODAP. The net amount of charge would be lower for DODAP near the membrane surface and the interaction with the negatively charged PEAA would be subsequently reduced. Other factors, such as the different orientation of DODAC and DODAP in the lipid bilayer, could also contribute to the different interactions between these two amino lipids and negatively charged PEAA but, at present, we do not have evidence to support or rule this out.
In all cases, except for the DODAC-containing vesicles, only small amounts of calcein leaked out after insertion of the polymer, indicating that polymer insertion did not affect the stability of most of the liposomes. As demonstrated in [fig_ref] Figure 2: Characteristics of pH-sensitive PEAA vesicles with differing lipid composition by the post-insertion... [/fig_ref] , the encapsulated calcein was completed released from DODAC vesicles upon incubation with lipo-PEAA. The interaction between PEAA and DODAC was so strong that, even without a pH change, the lipid bilayer was already greatly disrupted, causing calcein release and vesicular aggregation. Due to failure to prepare a satisfactory formulation, we did not study DODAC-containing liposomes in a further experiment. The formulations used in the rest of the study were all purified using the gel filtration column to remove the noninserted polymers and trace amounts of released calcein. In summary, PEAA liposomes with different lipid compositions were successfully prepared, with good polymer association, a narrow particle size distribution, and their contents encapsulated using a post-insertion method.
## Effect of cholesterol content on vesicle permeability
It has previously been shown that cholesterol increases the stability of liposomes when added as a component. [bib_ref] A study of calcein as an index for evaluation of the pH-sensitivity..., Chen [/bib_ref] [bib_ref] Mechanical properties and stability measurement of cholesterol-containing liposome on mica by atomic..., Liang [/bib_ref] This prompted us to insert lipo-PEAA without concern about loss of vesicular content. An increase in cholesterol content may also reduce or prevent release of contents when PEAA-LUV is acidified, because the rigidity and fluidity of the liposomal membrane is changed. Mills et al found that cholesterol reduced the potency of PEAA with regard to pH-induced liposomal permeability in an experiment using free PEAA and liposomes with encapsulated fluorescent dye. [bib_ref] Effect of liposomal composition on the critical fusion temperature of photoactivated liposome..., Miller [/bib_ref] [bib_ref] Effect of bilayer cholesterol and surface grafted poly(ethylene glycol) on pH-induced release..., Mills [/bib_ref] The critical pH, ie, the point at which the polymer is hydrophobic enough upon partial acidification to interact with the liposomal bilayer and trigger release of the contents, was shifted from pH 6.7 for the pure SOPC formulation to pH 6.1 for liposomes containing 60% cholesterol. Therefore, in the present study, the effect of cholesterol on the permeability of PEAA vesicles comprising DSPC and cholesterol, as shown in the release kinetics study in [fig_ref] Figure 3: Effect of cholesterol content on pH-induced calcein release of PEAA-LUVs [/fig_ref] , was examined using a calcein release assay. The higher the cholesterol content, the lower the permeability of the PEAA liposomes; 18%, 38.5%, and 48% of calcein leaked out from DSPC vesicles containing 20, 40, and 45 mol%, respectively, of cholesterol 15 minutes after acidification (pH 4.5). Vesicular stability resulting from addition of cholesterol was also observed for liposomes containing DMPC and DPPC. Formulations with cholesterol were stable at neutral conditions (pH 7.4), while the calcein was almost completely released upon acidification. Although a very stable liposome (DSPC and cholesterol) was used in the experiment [fig_ref] Figure 3: Effect of cholesterol content on pH-induced calcein release of PEAA-LUVs [/fig_ref] , a significant amount of calcein was still released under acidic conditions, indicating that PEAA has a very efficient pHdependent membrane permeability potential.
## Effect of different lipid chain lengths on vesicle permeability
pH-sensitive calcein release from PEAA-LUV was also dependent on the acyl chain length of phosphatidylcholine. shows that the amount of calcein released was proportional to the chain length of phosphocholine, regardless of whether 40 mol% or 45 mol% cholesterol was included in the liposomes. It is known that membrane permeability is higher in a liquid phase state (temperature higher than the gelsolid transition temperature). The calcein release observed was induced by pH-sensitive PEAA, not by temperaturecontrolled slow release. This was confirmed by a control experiment in which liposomal vesicles with no polymer content did not release significant amounts of calcein when the test solution was acidified to pH 4.5.
## Effect of charge on vesicle permeability
pH-induced calcein release from PEAA-LUV was strongly influenced by a positive surface charge, but not a negative charge . The effect of charge on lipid behavior is governed by the overall surface charge density of the liposome, the lipid head group, [bib_ref] Production of large unilamellar vesicles by a rapid extrusion procedure. Characterization of..., Hope [/bib_ref] and the polymer-liposome interaction before and after acidification. We selected negatively and positively charged lipids to compare the effect of different charges on PEAA vesicle formulation and their pH-induced permeability. It seemed that there were more problems with the positively charged preformed vesicles. As mentioned earlier, we were unable to prepare DODACcontaining PEAA-LUV due to aggregation and content release, and vesicles containing DODAP released much lower amounts of calcein than either neutral or negatively charged vesicles. It was considered that negatively charged PEAA could form a stable structure by partial neutralization of the positive charge of DODAP, reducing its sensitivity to protonation, and losing polymer fusion activity and permeability.
## Effect of temperature on vesicle permeability
Given that release of vesicular contents from PEAA liposomes has already been demonstrated in our previous work and the potency of release has also been shown to be regulated by polymer size, lipid composition, and the surface charge at room temperature, it would be interesting to assess the permeability of PEAA vesicles at body temperature and under hyperthermic conditions. Hayashi et al constructed a thermosensitive liposome by using poly(N-isopropylacrylamide), a structurally related polymer. [bib_ref] Temperature-controlled release property of phospholipid vesicles bearing a thermo-sensitive polymer, Hayashi [/bib_ref] Temperature-sensitive liposomes for use with mild hyperthermia have also been reported by fine-tone of lipid composition to obtain good permeability for marching desired temperature. [bib_ref] Hyperthermia and thermosensitive liposomes for improved delivery of chemotherapeutic drugs to solid..., Koning [/bib_ref] It was anticipated that PEAA in its acid form would be more hydrophobic and could also be responsible for the effects of temperature change, such that the efficiency of release could be further enhanced. In our study, three different temperatures, ie, 25°C, 37°C, and 47°C, were used to test the effect of temperature on pH-induced calcein release from PEAA-LUVs. DSPC has the high transition temperature (55°C), so formulations containing DSPC were used to test the effect of temperature on vesicle permeability. As shown in , calcein release increased as temperature increased. Enhancement of release by temperature was more efficient for DSPC formulations containing a lower amount of cholesterol , 40 mol% cholesterol), with 50% more calcein being released at 37°C than at room temperature. When tested at 47°C, complete release of contents was observed. For liposomes with saturated cholesterol concentrations , 45 mol% cholesterol), enhanced permeability was also seen, although the amount was much less than that shown in . Vesicles containing cholesterol 40 mol% showed more calcein release than those containing cholesterol 45 mol% at the same temperature and with the same incubation time, indicating high vesicle permeability with lower cholesterol content. Cholesterol acts as a fluid buffer in liposomes and reduces the freedom of motion of the liposome membrane. Cholesterol causes changes in the molecular order of the membrane and dynamics in the fluid and gel phases, as well as introducing a new thermodynamic phase when the cholesterol content is high. [bib_ref] The effects of liposome composition and the temperature on the stability of..., Atrouse [/bib_ref] Thus, the higher transition temperature in with a high cholesterol content resulted in low vesicle permeability. summarizes the effects of temperature on the permeability of PEAA liposomes comprised of different phosphocholines and 45 mol% cholesterol after 20 minutes of incubation in acid conditions (pH 4.5). Because DMPC has a lower transition temperature (23°C), calcein release was near maximal at room temperature, and the effect of increased temperature was not significant. Compared with calcein release at 25°C, there was a roughly 25% increase of calcein obtained at 37°C for DPPC with a transition temperature of 41°C. Calcein release was 22%, 25%, and 63% for DSPC at temperatures of 25°C, 37°C, and 47°C, respectively. The limited calcein release from the DSPC vesicle might be due to its high transition temperature (55°C).
# Discussion
The development of drug carrier systems with an ability to deliver and release their therapeutic contents intracellularly is an important strategy and an interesting research project. [bib_ref] A new liposome-based gene delivery system targeting lung epithelial cells using endothelin..., Allon [/bib_ref] [bib_ref] Polymer based pH-sensitive carriers as a means to improve the cytoplasmic delivery..., Roux [/bib_ref] As a continuation of our investigation of pH-sensitive liposomes containing the PEAA polymer, in the present study we have evaluated parameters influencing release of vesicular contents. We demonstrated that pH-sensitive vesicles comprised of different phosphatidylcholines, cholesterol, and negatively charged and protonable amino lipids can be constructed by insertion of lipid-anchored PEAA without loss of vesicular contents. The resulting formulations show a small and uniform particle size, a constant but controllable amount of membrane-inserted polymer, good stability under neutral conditions, and efficient content release in a pH-dependent manner. Considering that a more complicated lipid composition (such as phosphocholine) with a different phase transition temperature, cholesterol, and charged lipids may be required for actual drug entrapment and application, the information acquired here may have an impact on further development of pH-sensitive vesicles with surface-bound PEAA.
The pH-induced release of liposomal contents by PEAA is dependent on the molecular weight of the polymer. Under acidic conditions, the PEAA conformation collapsed from an expanded hydrophilic form to a globular hydrophobic coil as a result of protonization of the carboxylic groups of PEAA. [bib_ref] Tir rell DA. pH-induced fusion and lysis of phosphatidylcholine vesicles by the..., Linhardt [/bib_ref] The hydrophobic portion of PEAA penetrates and disturbs the liposomal membrane bilayer, resulting in release of the liposomal contents. Previously, Chung et al found that the critical pH at which PEAA starts to induce release of liposomal contents decreases with decreasing molecular weight of the polymer. [bib_ref] Opsahl-Ong LR. pH-sensitive, cation-selective channels formed by a simple synthetic polyelectrolyte in..., Chung [/bib_ref] In the present study, we also observed dependence of liposomal content release on the molecular weight of PEAA. As shown in [fig_ref] Figure 1: Effect of polymer size on pH-sensitive calcein release from large unilamellar vesicles [/fig_ref] , the potency of PEAA required to trigger release of the contents of the liposome, ie, membrane permeability, increases as the molecular weight of the PEAA increases. At molecular weights exceeding 8.4 kDa, the percentage release reaches a plateau. We selected a relative lower molecular weight of PEAA, which still induces efficient content release under acidic conditions. In practice, using reasonably sized polymers in a liposomal PEAA preparation reduces the risk of aggregation caused by larger molecular polymers and assures even distribution of the polymer on the surface of the liposome. A later experiment confirmed that the 8.4 kDa molecular size of PEAA is well suited to producing a well characterized liposomal PEAA formulation resulting in efficient release of its contents in a pH-dependent manner.
Because bilayer fluidity and rigidity can be an important determinant of the release of liposome-encapsulated compounds, 31,32 the effects of fluidity (using phosphocholine with different phase transition temperatures) and rigidity (varying cholesterol content) on the characteristics of pH-sensitive PEAA liposomes were also studied. In a previous study, we found that neither natural nor synthetic phosphocholine is an ideal choice for construction of pHsensitive liposomes with surface-bound PEAA because the PEAA polymers will still have some degree of interaction with the liposomal membrane, resulting in leakage of contents, even in neutral solution. [bib_ref] Cationic poly(ethyleneglycol) lipids incorporated into pre-formed vesicles enhance binding and uptake to..., Fenske [/bib_ref] In contrast, the stability of liposomes incorporating cholesterol enabled us to prepare PEAA vesicles which maintain their integrity but can still release their contents under acidic conditions. As in other liposomal release studies, pH-induced membrane permeability is also dependent on fluidity and rigidity. A trend of calcein release from PEAA vesicles under acidic conditions was observed in the order of DMPC . DPPC . DSPC, which reflects the fluidity of liposomes composed of DMPC, DPPC, and DSPC with a phase transition temperature of 23°C, 41°C, and 55°C, respectively. A similar conclusion can be drawn with regard to rigidity, ie, a higher cholesterol content results in greater membrane rigidity and lower permeability. More importantly, it indicates that the permeability of PEAA liposomes can be controlled by modifying the cholesterol content using different types of phosphocholine according to their intended applications.
When investigating pH-induced release from PEAA liposomes with different lipid compositions, the bilayer surface charge has also been found to be important for vesicle permeability. [bib_ref] Antibacterial activity of liposome-encapsulated ampicillin in vitro and in vivo in relation..., Bakker-Woudenberg [/bib_ref] [bib_ref] Effects of PEG-lipids on permeability of phosphatidylcholine/cholesterol liposomes in buffer and in..., Silvander [/bib_ref] In general, it is not possible or at least very difficult to prepare PEAA vesicles with prominent positive charged preformed LUVs. Multiple positive-negative charge interactions cause premature release of liposomal contents and vesicle-vesicle fusion during the polymer insertion step. However, PEAA liposomes with preformed LUVs incorporating amine-containing lipids were successively prepared with desirable characteristics, ie, small size and retention of the encapsulated drug. Permeability of the liposomes caused by protonation of PEAA was similar for both neutral and negatively charged liposomes, but reduced for LUVs containing DODAP. Whether the reduction in permeability for liposomes containing DODAP was caused by partial ionpairing of the amines in DODAP with the carboxylic groups in PEAA is not clear. However, we still obtained adequate release of the liposomal contents from PEAA-LUVs containing DODAP under acidic conditions. Furthermore, pH-sensitive vesicles with the out-leaflet incorporated PEAA also show temperature-dependent release. [bib_ref] Characteristics, phase behavior and control release for copolymer-liposome with both pH and..., Zhou [/bib_ref] The use of thermal-sensitive liposomal carriers for enhanced local release of drugs has been studied by several groups. [bib_ref] Enhancement of the phase transition permeability of DPPC liposomes by incorporation of..., Anyarambhatla [/bib_ref] [bib_ref] Effect of poly(ethylene glycol) grafts on temperature-sensitivity of thermosensitive polymer-modified liposomes, Kono [/bib_ref] [bib_ref] Thermo-sensitive polymer-modifed liposomes that release contents around physiological temperature, Kono [/bib_ref] [bib_ref] Design of liposomes for enhanced local release of drugs by hyperthermia, Yatvin [/bib_ref] Enhancement of therapeutic efficacy has been submit your manuscript | www.dovepress.com Dovepress Dovepress reported for many antitumor drugs. [bib_ref] Thermosensitive liposomes: Extravasation and release of contents in tumor microvascular networks, Gaber [/bib_ref] [bib_ref] Thermosensitive polymer-modified liposomes, Kono [/bib_ref] [bib_ref] A new temperature-sensitive liposome for use with mild hyperthermia: characterization and testing..., Needham [/bib_ref] Temperature-sensitive vesicles containing a synthetic polyacrylic polymer have also been developed. Considering that there is a relatively higher temperature and lower local pH at most disease sites, the combination of pH and temperature sensitivity may be advantageous, and illustrates a new development in liposome research. Although far from optimized, the liposomal carrier developed in the current study has demonstrated an ability to respond to both pH and temperature changes by releasing its vesicular contents, and there may be a benefit to be derived from differences in pH and temperature between healthy tissues and sites of disease. In the future, more sensitive pH-responsive and temperature-responsive vesicles may be constructed by partially modifying the carboxylic groups of PEAA with diisopropylamine. It is well known that poly(diisopropylacrylamide) is a good thermal-sensitive polymer, and liposomes containing poly(diisopropylacrylamide) have shown temperaturedependent release of contents. [bib_ref] Por ter TM. Thermosensitive liposomes modified with poly(n-isopropylacrylamideco-propylacrylic acid) copolymers for triggered..., Ta [/bib_ref] PEAA has a chemical structure similar to that of poly(acrylic acid), with extra ethyl groups at the 2-position. Therefore, by carefully controlling the amount of diisopropylamide in the modified PEAA, the new PEAA liposomes would still be pH-sensitive and also thermal-sensitive.
## Conclusion and further perspectives
In summary, this study has demonstrated that polymer size, liposome composition, surface charge, and temperature all influence the permeability of pH-sensitive liposomes containing lipid-anchored PEAA. Because the polymer size was kept at 8.4 kDa, the permeability of the PEAA liposome was dependent on its fluidity and rigidity, which could be modified by the cholesterol content and surface charge on the vesicles, and was varied using different types of phosphocholine. The PEAA liposome was also shown to be temperature-sensitive by careful control of the amount of diisopropylamide introduced into the modified PEAA. These new PEAA liposomes are both pH-sensitive and thermalsensitive, suggesting possible applications in the treatment of tumor or localized infection. Moreover, the information obtained in this study should be useful for the design of liposomes that are highly sensitive to both ambient temperature and an acidic environment. We plan to explore these issues in our future in vivo experiments.
[fig] Figure 1: Effect of polymer size on pH-sensitive calcein release from large unilamellar vesicles. Notes: Calcein release was measured by a calcein release assay, with calcein entrapped in large unilamellar vesicles and free polymers as described in the text. Liposomes were composed of egg phosphatidylcholine and cholesterol in a ratio of 60:40 (mol/mol) and calcein 100 mM. Free PEAA with a designed size was mixed with large unilamellar vesicles containing calcein at a polymer-to-lipid ratio of 10 µg/mg. The fluorescence intensity of the vesicles was measured after 20 minutes of incubation after acidification to pH 4.5. Abbreviation: PEAA, poly(ethylacrylic acid). [/fig]
[fig] Figure 2: Characteristics of pH-sensitive PEAA vesicles with differing lipid composition by the post-insertion method. LUVs with encapsulated calcein 100 mM were prepared using a hydration extrusion method. PEAA-LUVs were then prepared by incubating lipo-PEAA with preformed LUVs at a polymer-to-lipid ratio of 20 µg/mg. (A) Percentage polymer insertion when different phospholipids were used to prepare the PEAA-liposomes. The noninserted polymer was removed, the liposomes were determined by radioactivity, and the polymer was quantified by fluorescence intensity. (B) Particle size of PEAA-LUVs. (C) Calcein released from PEAA-LUVs overnight at room temperature with incubation at neutral conditions (pH 7.4). Calcein release was measured as described in the text. Abbreviations: DODAP, dioleoyl-1,2-diacyl-3-dimethylammonium-propane; DODAC, N,N-dioleyl-N,N-dimethylammonium chloride; DPPC, 1,2 dipalmitoyl-snglycero-3-phosphocholine; PI, phosphatidylinositol; DSPE, 1,2-distearoyl-sn-glycero-3phosphoethanolamine; LUVs, large unilamellar vesicles; PEAA, poly(ethylacrylic acid). [/fig]
[fig] Figure 3: Effect of cholesterol content on pH-induced calcein release of PEAA-LUVs. Notes: Preformed LUVs with DSPC and differing amounts of cholesterol were prepared using an extrusion method with calcein 100 mM. PEAA-LUVs containing calcein were then obtained by post-insertion of PEAA-C 10 at a final polymer concentration of 7 mol in PEAA-LUVs containing calcein at a polymer-to-lipid ratio of 20 µg/mg. Calcein released from PEAA-LUVs was measured at room temperature (25°C) as described in the text. Abbreviations: Chol, cholesterol; DSPC, 1,2 distearoyl-sn-glycero-3-phosphocholine; LUVs, large unilamellar vesicles; PEAA, poly(ethylacrylic acid). submit your manuscript | www.dovepress.com Dovepress Dovepress [/fig]
[fig] Figure 4, Figure 5: Effect of lipid chain length on pH-induced calcein release from PEAA-LUVs. Notes: Preformed LUVs with differing cholesterol content (45 mol%, 40 mol%) were prepared using an extrusion method with calcein 100 mM. PEAA-LUVs containing calcein were then obtained by post-insertion of PEAA-C 10 at a final polymer concentration of 7 mol in PEAA-LUVs containing calcein at a polymer-tolipid ratio of 20 µg/mg. Liposomal vesicles without PEAA inserted served as controls. The amount of calcein released from PEAA-LUVs was measured after 20 minutes of incubation at room temperature (25°C) as described in the text. Abbreviations: Chol, cholesterol; DMPC, 1,2 dimyristoyl-sn-glycero-3phosphocholine; DSPC, 1,2 distearoyl-sn-glycero-3-phosphocholine; DPPC, 1,2 dipalmitoyl-sn-glycero-3-phosphocholine; LUVs, large unilamellar vesicles; PEAA, poly(ethylacrylic acid). Effect of charged lipids on pH-induced release of calcein from PEAA-LUVs. Notes: Preformed neutral LUVs (40 mol% cholesterol) or charged LUVs (40 mol% cholesterol, and 10 mol% charged lipid introduced) were prepared using an extrusion method with calcein 100 mM. PEAA-LUVs containing calcein were then obtained by post-insertion of PEAA-C 10 at a final polymer concentration of 7 mol in PEAA-LUVs containing calcein at a polymer-to-lipid ratio of 20 µg/mg. The calcein released from the PEAA-LUVs was measured after 20 minutes of incubation at room temperature (25°C) as described in the text. Abbreviations: DODAP, dioleoyl-1,2-diacyl-3-dimethylammonium-propane; DPPC, 1,2 dipalmitoyl-sn-glycero-3-phosphocholine; PI, phosphatidylinositol; DSPE, 1,2-distearoyl-sn-glycero-3-phosphoethanolamine; LUVs, large unilamellar vesicles; PEAA, poly(ethylacrylic acid). submit your manuscript | www.dovepress.com Dovepress Dovepress [/fig]
[fig] Figure 6, Figure 7: Effect of temperature on pH-induced release of calcein from PEAA-LUVs. Preformed LUVs with (A) a DSPC to cholesterol ratio of 60:40 (mol/mol) and (B) a DSPC to cholesterol ratio of 55:45 (mol/mol) was prepared using an extrusion method with calcein 100 mM. PEAA-LUVs containing calcein were then obtained by post-insertion of PEAA-C 10 at a final polymer concentration of 7 mol in PEAA-LUVs containing calcein at a polymer-to-lipid ratio of 20 µg/mg. Calcein released from PEAA-LUVs was measured at different temperatures as described in the text.Abbreviations: DSPC, 1,2 distearoyl-sn-glycero-3-phosphocholine; LUVs, large unilamellar vesicles; PEAA, poly(ethylacrylic acid). Effect of temperature on pH-induced calcein release of PEAA-LUVs containing different chain lengths of phosphocholine. Notes: The preformed LUVs (45 mol% cholesterol) were prepared using an extrusion method with calcein 100 mM. PEAA-LUVs containing calcein were then obtained by post-insertion of PEAA-C 10 at a final polymer concentration of 7 mol in PEAA-LUVs containing calcein at a polymer-to-lipid ratio of 20 µg/mg. The calcein released from PEAA-LUVs was measured after 20 minutes of incubation at differing temperatures as described in the text. Abbreviations: DMPC, 1,2 dimyristoyl-sn-glycero-3-phosphocholine; DPPC, 1,2 dipalmitoyl-sn-glycero-3-phosphocholine; DSPC, 1,2 distearoyl-sn-glycero-3phosphocholine; LUVs, large unilamellar vesicles; PEAA, poly(ethylacrylic acid). submit your manuscript | www.dovepress.com Dovepress Dovepress [/fig]
|
Models of turfgrass seed germination related to water content
Turfgrasses have been widely utilized by humans to enhance the environment for more than several centuries. Seed germination is of great importance in the development of the turfgrass industry. In this study, the seed germination models of the responses of three types of turfgrasses to water were studied. The results indicated that irrigation water was mainly related to the variation of seed volume rather than a fixed value in this experiment. The maximum k value of tall fescue, Kentucky bluegrass and perennial ryegrass is 1.1548, 1.6946 and 4.0940, respectively. The optimal value of k is 1.0617 for tall fescue, 1.4610 for Kentucky bluegrass, and 1.6614 for perennial ryegrass. Perennial ryegrass seeds are more sensitive to water than those of tall fescue and Kentucky bluegrass, and this turfgrass is the fastest to reach the maximum value of the germination rate. Based on the results from the present experiment, the seed germination function can describe the response of turfgrass seed germination to external water content variation and their sensitivities. The function obtained could be used to perform quantitative studies on the dynamic changes of seed germination under different water conditions that will contribute to improved predictions of the optimal fitting curves of the germination over a range of water.
# Introduction
Turfgrasses are widely used throughout the world to protect our environment, and they have been an issue of major national and international importance to modern societies for many centuries [bib_ref] The role of turfgrasses in environmental protection and their benefits to humans, Beard [/bib_ref]. Seed germination is an important step in the plant life cycle, and seeds can only germinate normally in an appropriate ecological environment [bib_ref] Water content and GA3-induced embryonic cell expansion explain Euterpeedulis seed germination, rather..., Roberto [/bib_ref]. Water is a prerequisite for seed germination since seeds must absorb a certain amount of water to germinate [bib_ref] Seed germination response to high temperature and water stress in three invasive..., Yuan [/bib_ref]. A seed germination bed, as a place to obtain water, definitively affects seed germination by controlling the water content; if there is too little water, the seeds will not germinate without reaching the minimum water requirement; however, when there is too much water, factors such as hypoxia will reduce the germination rate [bib_ref] Fullerol improves seed germination, biomass accumulation, photosynthesis and antioxidant system in Brassica..., Xiong [/bib_ref].
Within a certain range of water content, the seed germination rate is significantly positively correlated with water supply [bib_ref] Effect of temperature and water potential on Carthamustinctorius L. seed germination: quantification..., Bidgoly [/bib_ref]. If the supply is inadequate, the seed germination will be restricted by the water content [bib_ref] Seed germination, seedling establishment and growth patterns of wrinklegrass (IschaemumrugosumSalisb), Bakar [/bib_ref]. Previous studies suggested that seed germination can occur a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 after the completion of imbibition and absorption, that is, the germinating seeds need time to prime, and this only occurs when the seeds absorbed water for a period of time [bib_ref] Water availability shifts the optimal temperatures for seed germination: a modelling approach, Mesgaran [/bib_ref] [bib_ref] Environmental signals for seed germination reflect habitat adaptations in four temperate Caryophyllaceae, Vandelook [/bib_ref]. The chemical composition of the seeds as well as the membrane permeability, seed size, 1000-grain weight, seed dormancy, temperature, oxygen, carbon dioxide, light, salinity, and water depth will affect the rate of water absorption and the water use efficiency of the seeds as well as their interactions with water during seed germination [bib_ref] Seed dormancy and the control of germination, William [/bib_ref]. Research has indicated the minimum water requirement for seed germination in various plants using different modeling methods limited to different launch conditions. Research has also provided several models and empirical formulas, and the hydrothermal threshold model has been widely used to describe the temperature and water potential effect on seed germination [bib_ref] A water relations analysis of seed germination rates, Bradford [/bib_ref]. In these models, researchers studied crop seeds based on the hydrothermal time (HTT) and virtual osmotic potential (VOP) models. These models assume that the variable conditions of the water potential depend on the current water environment. Therefore, when the difference value between the water and osmotic potentials exceeds a fixed threshold, the seed radicle will grow. Assessing the effect of the promotion of seed growth to establish the precise model would help to predict the variables of temperature and water potential on the influence of germination, which will advance the predicted simulation of seed priming time, which is conducive to seed germination, seedling emergence and seedling growth, and increase crop yields [bib_ref] Development of combined imbibition and hydrothermal threshold models to simulate maize (Zea..., Finch-Savage [/bib_ref]. However, previous studies largely focused on prediction and improved the methods of seed priming time, ignoring the role of the external water environment during the changes of water content, such as excessive evapotranspiration due to a shortage of water or insufficient irrigation [bib_ref] Development of combined imbibition and hydrothermal threshold models to simulate maize (Zea..., Finch-Savage [/bib_ref] [bib_ref] Hydrothermal threshold models can describe the germination response of carrot (Daucuscarota) and..., Rowse [/bib_ref].
By simulating water stress conditions, it is possible to study the biochemical and physiological processes involved in the drought resistance of plants, which are similar to the effect of external water content variation. As a common herbaceous ground cover plant, turfgrass is a plant community that is often used in mixed sowing (different species or varieties). The relative germination time and germination capacity of each turfgrass plant population will play a key role in the dynamic competition of the community structure [bib_ref] The effect of saline stress on the germination and embryo growth of..., Li [/bib_ref] [bib_ref] Sample size and water dynamics on germinating diaspores: the first step for..., Ribeirooliveira [/bib_ref]. The extremely sensitive reaction of turfgrass to water suggests that their adaptability is closely related to the response of the water content variation during turfgrass seed germination and emergence. Considering the characteristics of water parallel diffusion in layered soil under subsurface drip irrigation, it can maintain relatively stable and uniform soil water contents in high-density planted turfgrass.
The objectives of this study were to study the seed germination rate under different water conditions, establish models between seed germination and water content, provide valuable information to evaluate the relationships between turfgrass seed germination and soil water content, and, finally, provide a theoretical basis for the application of subsurface drip irrigation to turfgrass.
# Materials and methods
## Theoretical considerations
Under suitable conditions, seeds absorb water and begin to germinate [bib_ref] Combined effect of water potential and temperature on seed germination and seedling..., Gurvich [/bib_ref]. Seeds require external water for this process. At the same time, the external water may change due to environmental reasons. When the water is saturated, the seeds will quickly germinate. Subsequently the instantaneous germination rate will decrease as it reaches the optimal level, while the cumulative seed germination rate will gradually attain its maximum value. Considering that the seeds are not infinitely sprouting in the limited space and resources, when the germination rate reaches a certain point, the seeds will inhibit their growth. Therefore, it can be assumed that the environment can support the maximum of cumulative seed germination rate designated a, while y represents the current cumulative germination rate. Therefore, a − y is the germination rate left which the environment can support, and the changes in germination rate positively correlate with these two items. Thus:
[formula] dy dt ¼ rða À yÞyð1Þ [/formula]
where t is the germination time, and r is a positive constant. Separation of the variables and the integral for the Eq (1) results in:
[formula] y ¼ a 1 þ e À raðtÀ cÞð2Þ [/formula]
where c is a constant term arising from integration. The deformation to Eq (2) leads to:
[formula] y ¼ a 1 þ e À kðtÀ t c Þð3Þ [/formula]
where k is a positive constant, and k expresses the sensitivity of the reaction of the seeds to environmental conditions. While the other conditions remain constant, k relates to the environment water, and t c is the time when the instantaneous germination rate gains the maximum value. k = ar, Thus, Eq (1) becomes:
[formula] dy dt ¼ k a ða À yÞyð4Þ [/formula]
Seeking its second derivative:
[formula] d 2 y dt 2 ¼ k a dy dt ða À 2yÞð5Þ [/formula]
as dy dt ! 0, when d 2 y dt 2 ¼ 0, then dy dt 6 ¼ 0, y ¼ a 2 ; this is the maximum point of the instantaneous germination rate.
The changes in the seed germination function and the characteristics of the curves are shown in [fig_ref] Fig 1: Seed germination function [/fig_ref] When the seeds begin to germinate, the instantaneous germination rate (slope of the curve) will increase, and the cumulative germination rate also increases [fig_ref] Fig 1: Seed germination function [/fig_ref]. The slope increases with k. The time it takes to achieve the maximum cumulative germination becomes shorter as the k increases. From [fig_ref] Fig 1: Seed germination function [/fig_ref] it can be concluded that when y ¼ a 2 , results in t = t c , the instantaneous germination rate will reach its maximum and start to decrease. The cumulative seed germination rate will gradually reach its maximum.
# Materials
The seeds of the three turfgrasses used in this study were obtained from the Beijing Clover Technical Development Center. The detailed information of these seeds is listed in [fig_ref] Table 1: Turfgrass seeds used in the study [/fig_ref].
## Analytical procedures
Determination of the seed volumes. To test the average volume of the seed of each species, the seed specific gravity of each species was first determined. A total of 100 seeds of each species was randomly collected and weighed in three replicates and placed in a flask containing 50% ethyl alcohol. The flasks were shocked to exhaust the air (at 20˚C). The specific gravities of each species were calculated using the following formula:
[formula] S ¼ W 1 W 2 þ W 1 À W 3 Â Gð6Þ [/formula]
where W 1 is the weight of 100 seeds of each species (g); W 2 is the weight of flask with 50% ethyl alcohol (g); W 3 is the weight of flask with ethyl alcohol and seeds (g), and G is the proportion of 50% ethyl alcohol. The volume of the seeds (V) was calculated using the following formula:
[formula] V ¼ W Sð7Þ [/formula]
where W is the average weight of seed, and S is the specific gravities of the seed. Determination of the seed equilibrium water content. One hundred seeds of each species were weighed and wrapped with gauze (three replicates). The seeds were placed in a thermostatic container at 25˚C with the appropriate amount of water to imbibe it and weighed after centrifugation at a 12 h-interval until their weight would not change. The seed equilibrium water content (f e )was calculated using the following equation:
[formula] f e ¼ EW À DW DWð8Þ [/formula]
where EW is the equilibrium water weight (g), and DW is the dried weight of seeds.
Response of the seed germination to irrigation. One hundred seeds of each species were randomly chosen (three replicates). The volume of a 90 mm (diameter) Petri dish was first determined, and the weight of the seeds was calculated under the same volume. Finally, we weighed the dish which was covered with a layer of the seeds. The water content of a soaked layer of the seeds (Y s ) was calculated as: where W 4 is the seed weight (g) of volume size of the Petri dish; W 5 is the weight (g) of a layer seed covered Petri dish, and V 2 is the volume of the Petri dish (mL). The seeds were placed in the Petri dishes whose volume has been determined. Irrigation was conducted at ten levels (y s × 10%, y s × 20%, y s × 30%, y s × 40%, y s × 50%, y s × 60%, y s × 70%, y s × 80%, y s × 90% and y s × 100%). The Petri dishes were placed in a light and constant temperature incubator. The temperature was controlled at 25±1˚C, and the light was controlled at 12 h light / 12 h dark. These conditions were close to the optimal growth condition of the three turfgrasses studied. The radicle breaking through the seed coat was selected to be the germination standard, and the germination status was monitored and recorded at approximately 24-hour intervals. At the same time, each Petri dish was weighed to determine the evapotranspiration, and the loss of water caused by measurements was added. The observations continued until the germination rate remained the same for a week. The germination was assessed based on the seed germination rate and germination priming time, which was the time from sowing to the first seed germinated (days).
[formula] Y s ¼ W 5 W 4 Â V 2ð9Þ [/formula]
# Results
## Imbibition and equilibrium water of seed germination
The change in the external water minimum value (variation rate) of the three turfgrasses gradually decreased under the different irrigation standards [fig_ref] Fig 2: The change of external water of three turfgrasses [/fig_ref]. The external water did not change along with the increase in the irrigation water content. However, the variation rate decreased. This result helped to maintain the stability of the external environment water. It should be noted that the irrigation water in this experiment was primarily relative due to the variation of the seed volume rather than a fixed value. Thus, during practical applications, the seed germination is primarily determined by the water use efficiency rather than the water quantity. For example, if there is excessive evapotranspiration, the final storage capacity of the water will be low. The water content helped the seeds to germinate and improved the germination rate only when it remained stable to supply the water for a long time [bib_ref] Evapotranspiration and water use of full and deficit irrigated cotton in the..., Oweis [/bib_ref]. Therefore, when choosing different irrigation methods, it should be initially considered that the external water changes are smaller over a long period (irrigation cycle) and maintain high water use efficiency. As shown in [fig_ref] Fig 1: Seed germination function [/fig_ref] there was no significant discrepancy among the three turfgrasses studied, and they share a similar variation tendency.
The water balance of the plant communities could be expressed by the water equilibrium equation [bib_ref] On concepts of ecological water demand, Zheng [/bib_ref]. There is only a small difference in the form of specific performance. In this experiment, irrigation is assumed to be the only water input for seed germination; thus, the water input (irrigation water) is equal to the loss of evapotranspiration water. However, when the quantity of the irrigation water is more than evapotranspiration or is too low to meet the growth needs of seeds, water storage would increase or decrease to maintain the water balance. Therefore, the irrigation water quantity equilibrium equation is:
[formula] W I ¼ W S ðtÞ þ W E ðtÞð10Þ [/formula]
where W I is the irrigation water quantity, W S is the variation of water storage, and W E is the evapotranspiration water quantity. Under stable external environmental conditions, such as when the temperature and humidity are relatively constant, the variation of evapotranspiration water is considered a constant, namely:
[formula] W E ðtÞ ¼ vtð11Þ [/formula]
where v is a constant to indicate the evapotranspiration rate, and t is time. Therefore, when the irrigation water remains stable, the seed germination mainly depends on the water storage capacity, and the W S consists of the seeds imbibing water (f) and the external environment water (φ)which is not absorbed, namely:
[formula] W S ¼ f ðtÞ þ φðtÞ ð12Þ [/formula]
f can be expressed as a percentage of the dry weight of seeds. At the time of seed germination, there is an equilibrium between the seed imbibing water and the external environmental water, which is the maximum capacity of the seeds to imbibe water based on the seed itself and the environmental conditions, which is the seed equilibrium water content. Assuming that the water imbibition rate only depends on the current seed water content and equilibrium water content (f e ), thus:
[formula] df dt ¼ Lðf e À f Þð13Þ [/formula]
where L is the imbibition exponential coefficient of the seed; thus, we integrate Eq (13):
[formula] f ¼ f e ð1 À e À Lt Þð14Þ [/formula]
When the seed imbibes water, the seed water content does not start from zero; however, from a fixed value f p (initial water content), when t = 0, f = f p , as t infinity, f = f e , then the Eq [bib_ref] Sample size and water dynamics on germinating diaspores: the first step for..., Ribeirooliveira [/bib_ref] can be changed to:
[formula] f ¼ f p þ ðf e À f p Þð1 À e À Lt Þð15Þ [/formula]
where L is a constant related to the seed size and the external environment water. Since the same seeds are in each Petri dish, the impact of volume could be ignored, and the L is affected by the external environmental water. Then:
[formula] φðtÞ ¼ W I À W E ðtÞ À f ðtÞð16Þ [/formula]
The seed germination occurred after the completion of imbibition and absorption. Therefore, when t = 0, φ(t) = W I , this is the maximum value of the water in the external environment. When f = f e , the minimum value appeared, φ(t) = W I − v t − f e . The same species in similar environments, φ(t) is mainly affected by the irrigation water W I , which indicates that the irrigation water content decides the rate of seed germination.
## Seed germination rate and equilibrium water
Population growth is restricted in a limited space and resources. When the population is excessive, it would restrain the growth. Thus, when the water varies, the germination rate also differs. In practice, Eq (3) often does not accurately determine the instantaneous time t c when the germination rate reaches the maximum value, which is due to the germination rate increasing (Δy) maintaining a certain level of growth over a certain time interval. Therefore, the Δy of the two observation points may be similar or equal to the condition. If using the time (t m ) to observe, and the maximum value of Δy simply corresponds to t c , it would lead to a large error for k. However, in the observations, it was determined that t c must be near t m , and as shown in [fig_ref] Fig 1: Seed germination function [/fig_ref] where dy dt is the slope; when the increment Δt is close to 0, then the slope is approximately equal to Dy Dt ; with equal interval sampling, t c can be evaluated based on Δy; in the model simplification, because the increasing value of the two sampling points were similar, it can determine that the maximum value is between the two points and the practice of t c , whose accuracy has been able to meet the demand of this test.
The variations of the cumulative germination rate of the three turfgrasses under different irrigation levels are shown in [fig_ref] Fig 3: Mathematical model of the germination rate for three turfgrasses under study [/fig_ref] The cumulative germination rate of all the three turfgrasses shows a "slow-fast-slow" change rule along with the water level, and the germination rate of different kinds of turfgrass seeds with varying water content all meet the same function. At the same time, when the amount of water increases to a certain degree, the cumulative germination rate gradually tends to slow and exhibits a downward trend. Thus, the existence of the optimal amount of water can make the seed reach the maximum germination rate; however, when it exceeds the boundary, it will inhibit the germination of the seeds.
## Factor of water sensitivity
The parameters and original data for the simulation in the seed germination function are listed in [fig_ref] Table 2: Model parameters of different turfgrasses [/fig_ref].
The results indicated that k reflected the difference of the sensitivities of the different seed species to external water content. If k increases, the cumulative germination rate changes more quickly with time. Thus, the instantaneous germination rate (the slope of the curve) changes more. Considering the relationship between k and the water, there is an assumption that when the seed imbibition and absorption of water are completed but the seeds have not started to germinate, k = 0; before the change rate of φ(t) reaches its optimal value, k increases as the irrigation water content increases; and when the change rate of φ(t) fluctuates around the optimal value (decrease or increase), k will decrease. The regression equations based on the data measured or the three turfgrasses are (Tall fescue, Eq 17; Perennial ryegrass, Eq 18; and Kentucky bluegrass, Eq 19):
k ¼ À 0:0242φ 2 þ 0:3437φ À 0:1293 ðR 2 ¼ 0:8091Þ ð17Þ
[formula] k ¼ 0:0079φ 2 þ 0:2284φ þ 0:9051 ðR 2 ¼ 0:9554Þð18Þ [/formula]
k ¼ À 0:0329φ 2 þ 0:4602φ À 0:1158 ðR 2 ¼ 0:6411Þ ð19Þ
As described above, Eq (3) becomes:
[formula] y ¼ a 1 þ e À kðφÞðtÀ t c Þð20Þ [/formula]
The maximum k value of tall fescue, Kentucky bluegrass and perennial ryegrass is 1.1548, 1.6946 and 4.0940, respectively. This result indicates that the germination rate of perennial ryegrass is the quickest, and its cumulative seed germination rate first reaches the steady state. In this experiment, perennial ryegrass germinates too fast, while there are longer intervals between the observation points. Therefore, the test may change the value of k from high to low, and the k of perennial ryegrass is much higher than those of tall fescue and Kentucky bluegrass. The k of perennial ryegrass is higher than the others in the beginning of germination, which indicates that the response speed of perennial ryegrass is faster. In practice, perennial ryegrass can be used as a pioneer of mixed sowing, and the k of tall fescue is low, thus explaining why its reaction is less sensitive to water. This outcome could indicate that it possesses stronger drought resistance. Based on Eq (20), a relationship can be obtained for the germination rate and irrigation water content; assuming that y optimal = βy, β can be obtained, and the original data are modified to determine the optimal regression equations for the three turfgrasses (Tall fescue, Eq 21; Perennial ryegrass, Eq 22; and Kentucky bluegrass, Eq 23): y optimal ¼ 73:3556 1 þ e À 1:0617ðtÀ 1:5Þ ðR 2 ¼ 0:8314Þ ð21Þ y optimal ¼ 89:1278 1 þ e À 1:6614ðtÀ 1:5Þ ðR 2 ¼ 0:7142Þ ð22Þ y optimal ¼ 70:0540 1 þ e À 1:4610ðtÀ 1:
[formula] 5Þ ðR 2 ¼ 0:8659Þð23Þ [/formula]
These three equations show the cumulative germination rate in response to the different sensitivities of water for the three turfgrasses. The optimal value of k is 1.0617 for tall fescue, 1.4610 for Kentucky bluegrass, and 1.6614 for perennial ryegrass. This result indicates that perennial ryegrass seeds are more sensitive to water than tall fescue and Kentucky bluegrass. Therefore, perennial ryegrass will be the fastest to reach the maximum value of the germination rate. Using the experimental k of the optimal water on its fitting curve that represents k, t c is the time for optimal water from this experiment. In addition, the error of a may come from the fitting error of k and the estimation error of t c . [fig_ref] Fig 4: The optimal fitting curves of the cumulative germination in a range of... [/fig_ref] the optimal fitting curves of the cumulative germination for the three turfgrasses, and it indicated that the slope increases with k. When k is growing, it would take less time to achieve the maximum value of the germination rate.
# Conclusions
Water content significantly affects the germination of turfgrass. Turfgrass has a high germination rate when the water conditions are appropriate. The results from the present experiment indicate that the seed germination function can describe the response of turfgrass seed germination to external water content variation and their sensitivities. This function can be used to quantitatively study the dynamic changes in seed germination under different water conditions. In addition, this function can confirm the characteristic parameters to express the sensitivities of seeds with water. The seed germination function has been tested with the data obtained. Both the function and the measured data showed that the variation of the cumulative seed germination rate is "slow-fast-slow". At the beginning of germination, the cumulative germination rate increased quickly and gradually tended to remain steady. The cumulative germination rate increased rapidly as the water content increased. When the water reached a certain value, it reached its maximum value and tended to slightly and gradually decrease. The results indicated that there was an optimal water content that led to the seed germination rate reaching its maximum.
The germination of the turfgrass seeds was mainly affected by the rate of change of the water rather than the amounts of total water. Therefore, choosing appropriate irrigation methods is useful to maintain stable external water contents and high water use efficiency to meet the needs of seed germination. Based on this study, the optimal value of k was 1.6614 for perennial ryegrass, which implies that perennial ryegrass seeds are more sensitive to water content than those of tall fescue and Kentucky bluegrass. The k value of tall fescue was lowest, which suggested that tall fescue is less sensitive to water. However, tall fescue exhibits more drought resistance than the other two types of turfgrasses. Based on the seed germination function, further research should be conducted using experiments under different cultivation media conditions to improve turfgrass seed germination models through field experiments.
# Author contributions
Conceptualization: Derong Su.
Formal analysis: Ying Jiang.
[fig] Fig 1: Seed germination function (A) and the characteristics of curves (B).https://doi.org/10.1371/journal.pone.0204983.g001 [/fig]
[fig] Fig 2: The change of external water of three turfgrasses. https://doi.org/10.1371/journal.pone.0204983.g002 [/fig]
[fig] Fig 3: Mathematical model of the germination rate for three turfgrasses under study. (A) Tall fescue (TF), (B) Perennial ryegrass (PR) and (C) Kentucky bluegrass (KB). [/fig]
[fig] Fig 4: The optimal fitting curves of the cumulative germination in a range of Irrigation Water Contents (WC Indicates Water Content) for three turfgrasses. (A) Tall fescue (TF), (B) Perennial ryegrass (PR) and (C) Kentucky bluegrass (KB). [/fig]
[table] Table 1: Turfgrass seeds used in the study. [/table]
[table] Table 2: Model parameters of different turfgrasses. [/table]
|
LncRNA MALAT1 functions as a biomarker of no-reflow phenomenon in ST-segment elevation myocardial infarction patients receiving primary percutaneous coronary intervention
MALAT1 was reported to sponge miR-30e, miR-126 and miR-155 in the pathogenesis of many diseases. Plasma miR-30e can indicate the risk of no-reflow during primary percutaneous coronary intervention (pPCI), while miR-126 can be used as a predictor of coronary slow flow phenomenon. In this study, we compared the diagnostic value of above genes in the prediction of no-reflow phenomenon in ST-segment elevation myocardial infarction (STEMI) subjects receiving pPCI. Quantitative real-time PCR, ELISA, Western blot and luciferase assays were performed to explore the regulatory relationship of MALAT1/miR-30e, MALAT1/miR-126, MALAT1/miR-155, miR-126/ HPSE, and miR-155/EDN1. ROC analysis was carried out to evaluate the potential value of MALAT1, miRNAs and target genes in differentiating normal reflow and no-reflow in STEMI patients receiving pPCI. Elevated MALAT1, CRP, HPSE, and EDN1 expression and suppressed miR-30e, miR-155 and miR-126 expression was found in the plasma of STEMI patients receiving pPCI who were diagnosed with no-reflow phenomenon. ROC analysis showed that the expression of MALAT1, miR-30e, miR-126 and CRP could be used as predictive biomarkers to differentiate normal reflow and no-reflow in STEMI patients receiving pPCI. MALAT1 was found to suppress the expression of miR-30e, miR-126 and miR-155, and HPSE and EDN1 were respectively targeted by miR-126 and miR-155. This study demonstrated that MALAT1 could respectively sponge the expression of miR-30e, miR-126 and miR-155. And miR-30e, miR-126 and miR-155 respectively targeted CRP, HPSE and EDN1 negatively. Moreover, MALAT1 could function as an effective biomarker of no-reflow phenomenon in STEMI patients receiving pPCI.AbbreviationsHUVECs Human umbilical vein endothelial cells HAECs Human aortic endothelial cells STEMI ST-segment elevated myocardial infarction PCI Percutaneous coronary intervention HPSE Heparanase EDN-1 Endothelin No-reflow is defined by insufficient myocardial perfusion despite mechanical opening of the corresponding lesion via percutaneous coronary intervention (PCI) 1 . No-reflow occurs in about 10% of primary PCI patients 2 . Also, OPEN Scientific Reports | (2022) 12:3294 | https://doi.org/10.1038/s41598-022-06923-zwww.nature.com/scientificreports/ microvascular obstructions as a result of distal embolization, thrombosis, as well as microvascular contraction have been suggested as mechanisms underlying no-reflow, which are associated with many adverse results such as stroke, heart failure, as well as cardiac mortality 3,4 . Moreover, parameters such as CHA2DS2-VASc score, which estimates the risk of thromboembolism in patients with atrial fibrillation, can also predict no-reflow in patients who underwent pPCI 5 . Although many factors have been suggested as risk factors of no-reflow, no widely approved procedure is available to stratify the risk of no-reflow. Long non-coding RNAs (lncRNAs) are transcripts containing ≥ 200 nucleotides. In recent studies, lncRNAs have attracted great recognition due to their important role in many biological processes such as the differentiation, growth, homeostasis, as well as embryonic development of cells 6,7 . Metastasis-associated lung adenocarcinoma transcript 1 (MALAT1) is a conserved lncRNA implicated in lung metastasis and poor prognosis of lung cancer patients8,9. Latest studies revealed that MALAT1 exerted an essential impact on the regulation of many pathophysiological processes such as neurologic disorders, vascular diseases, and cancers by affecting the proliferation, migration, apoptosis and invasion of cancer cells 10 .Just recently, the elevated pre-operative value of C-reactive protein (CRP) was shown to predict the prognosis of survival in pNEN patients 11 . It was also shown that CRP caused upregulated COX-1 and COX-2 expression to reduce inflammation 12 . Some studies disclosed that COX-2 exerted an essential effect on the angiogenesis induced by IL-1b 13,14 . Heparanase (HPSE) was an enzyme recently found to play an important role to degrade the heparan sulfate (HS) chains in glycosaminoglycans. As a biologically active proteoglycan, HS is implicated in many important processes of adhesion reactions by interacting with cytokines, adhesion molecules, and signaling molecules in cells, thus impacting cell differentiation, proliferation and migration. HPSE can also promote the invasion, metastasis, chemotaxis and division of tumor cells15,16. The levels of HPA were additionally shown to be higher in PCI patients with the no-reflow phenomenon 17 . Also, elevated levels of HPA in STEMI patients were linked to a higher incidence of TB. On top of that, raised HPA levels may be linked to thrombotic complications including no-reflow in STEMI patients. Endothelin-1 (ET-1) is an effective vasoconstrictor, mitogen, and pro-inflammatory mediator generated upon vessel wall stress and hypoxia. ET-1 also plays an important role in endothelial inflammation and the formation of atherosclerotic plaques 18 . Additionally, the network structure of stents can aggravate microcirculation dysfunction to induce chemotaxis and trigger monocyte activation to secrete various vasoactive factors like ET-1 and VCAM-1, which in turn induce the onset of reperfusion injury 19,20 .MALAT1 has been reported to regulate the miR-155 expression and promote the proliferation and migration of cardiac stem cells under hypoxia 21 . Apart from miR-155, many miRNAs including miR-30e and miR-126 were proved to be 'sponged' by MALAT1 in the pathogenesis of many diseases 22,23 . Moreover, these miRNAs have been reported to function as biomarkers in STEMI patients. Plasma miR-30e could indicate the risk of no-reflow during pPCI, while miR-126 could be used as a predictor of coronary slow flow phenomenon 24 . Furthermore, several proteins were also implicated in cardiac diseases. For example, increased HPSE expression was shown to be associated with no-reflow phenomenon in STEMI patients, and elevated EDN-1 indicated increased coronary instability17,25. In this study, we compared the diagnostic value of the above genes in the prediction of no-reflow phenomenon in STEMI subjects receiving primary PCI.
# Materials and methods
Human subjects and sample collection. In this study, we recruited 198 ST-segment elevation myocardial infarction (STEMI) patients receiving primary percutaneous coronary intervention (pPCI) during August 2018 to May 2021 in the department of cardiology at the Affiliated Hospital of Guilin Medical University. The patients were continuously enrolled and divided into two groups according to their reflow status as the normal reflow group (n = 152) and the no-reflow group (N = [bib_ref] Advances in the vascular pathophysiology of ischemic stroke, Del Zoppo [/bib_ref]. No-reflow is diagnosed as the persistence of forward blood flow disorder [TIMI flow grade ≤ 2] after coronary artery without mechanical obstruction and no significant residual stenosis or dissection. The age, gender, diabetes history, hypertension history, hyperlipidemia history, familial history, smoking history, chest pain to hospitalization time, and door to balloon time, the angiography and procedural characteristics including infarct-related artery, multi-vessel disease, stent implantation and total stent length were collected and summarized for subsequent comparison. And serum samples were collected from all participants before their pPCI treatment. The ethical committee of Affiliated Hospital of Guilin Medical University has approved the protocol of this study (Approval ID: 2018-08-012). All procedures were performed in strict accordance with the last vision of the Declaration of Helsinki. Written informed consent was obtained from all patients or their first-degree relatives before the study.
## Rna isolation and real-time pcr.
In this study, real-time PCR was done to compare the expression of MALAT1, miR-30e, miR-126, miR-155, CRP mRNA, HPSE mRNA, and EDN1 mRNA in each sample. In brief, total RNA in each sample was separated by utilizing a Trizol ® reagent (Invitrogen, Carlsbad, CA) following the standard assay procedure provided on the instruction manual of the supplier. In the next step, the quality and content of separated RNA was quantified at 260 nm and 280 nm absorbance by using a Nanodrop ND-3000 device (Thermo Fisher Scientific, Waltham, MA) following the standard assay procedure provided on the instruction manual of the instrument supplier. Then, 300 ng of separated total RNA from each sample were used to synthesis cDNA templates using a cDNA reverse transcription assay kit (Thermo Fisher Scientific, Waltham, MA) following the standard assay procedure provided on the instruction manual of the assay kit supplier. Finally, a Taqman Gene Expression Assay kit (Invitrogen, Carlsbad, CA) was used following the standard assay procedure provided on the instruction manual of the assay kit supplier to determine the relative expression of MALAT1, miR-30e, miR-126, miR-155, CRP mRNA, HPSE mRNA, and EDN1 mRNA in each sample via real time qPCR, which was carried out on a StepOnePlus (Invitrogen, Carlsbad, CA) real time qPCR machine following the standard assay procedure provided on the instruction manual of the instrument manufacturer. The Cell culture and transfection. HUVEC and HAEC cells were bought from Lonza and maintained following the standard incubation procedure provided by the supplier. In brief, the cells were maintained at 37 °C and 5% CO 2 in DMEM media (Gibco, Thermo Fisher Scientific, Waltham, MA) added with 10% FBS and penstrep. When HUVEC and HAEC cells reached confluency, they were divided into two groups, i.e., 1. NC siRNA group (HUVEC and HAEC cells transfected with negative control siRNA); and 2. MALAT1 siRNA group (HUVEC and HAEC cells transfected with MALAT1 siRNA). The transfection was carried out using Lipofectamine 3000 (Invitrogen, Carlsbad, CA) following the standard transfection procedure provided on the instruction manual of the transfection reagent supplier, and transfected cells were collected 48 h later to assay the expression of target genes.
Vector construction, mutagenesis, and luciferase assay. Our binding site screening results showed that miR-30e could potentially bind to MALAT1, so the luciferase vectors containing wild type and mutant promoter of MALAT1 carrying the miR-30e binding site were established and transfected into HUVEC and HAEC cells with miR-30e. In brief, the wild-type promoter sequence of MALAT1 carrying the miR-30e binding site was inserted into a pcDNA vector (Promega, Madison, WI) downstream the luciferase reporter gene to generate the wild type plasmid for the MALAT1. Then, a Quick Change III mutagenesis kit (Stratagene, San Diego, CA) was used following the standard assay procedure provided on the instruction manual of the assay kit supplier to induce a site-directed mutation in the miR-30e binding site of MALAT1 promoter, and the mutant sequence was also inserted into a pcDNA vector downstream the luciferase reporter gene to generate the mutant type plasmid for the MALAT1 promoter. In the next step, both luciferase vectors containing wild type and mutant MALAT1 were co-transfected into HUVEC and HAEC cells along with miR-30e, and the luciferase activity of transfected HUVEC and HAEC cells was assayed 48 h later using a Dual-Luciferase Report gene assay kit (Promega, Madison, WI) following the standard assay procedure provided on the instruction manual of the assay kit. Similarly, our binding site screening results showed that miR-126 could potentially bind to MALAT1, so the luciferase vectors containing wild type and mutant promoter of MALAT1 carrying the miR-126 binding site were established and transfected into HUVEC and HAEC cells with miR-126, according to the method similar to the one described above. In addition, to study the regulatory relationship of MALAT1/miR-155, miR-126/HPSE, and miR-155/EDN1, luciferase vectors containing wild type and mutant MALAT1, HPSE or EDN1 were established and co-transfected into HUVEC and HAEC cells along with miR-155, miR-126 and miR-155, respectively, and the luciferase activity of transfected HUVEC and HAEC cells was assayed 48 h later using the Dual-Luciferase Report gene assay kit. The reading of luciferase activity of transfected HUVEC and HAEC cells was done in a Turner luminometer (Turner Biosystems Luminometer, Promega, Madison, WI) following the standard assay procedure provided on the instruction manual of the luminometer.
Western blot analysis. The total protein from each sample was isolated first by using a RIPA lysis buffer (Sangon Biotech, Shanghai, China) following the standard assay procedure provided on the instruction manual of the lysis buffer supplier. In the next step, the concentration of isolated protein was quantified by utilizing a bicinchoninic acid (BCA) protein assay kit (Thermo Fisher Scientific, Waltham, MA) following the standard assay procedure provided on the instruction manual of the assay kit supplier. Then, the total protein isolated from each sample was resolved on a 10% SDS-PAGE gel (Thermo Fisher Scientific, Waltham, MA) and blotted onto a PVDF membrane, which was then blotted in 5% skim milk and then incubated in sequence with primary anti-CRP, anti-HPSE, and anti-EDN1 antibodies as well as HRP-conjugated secondary antibodies (all antibodies were purchased from Santa Cruz biotechnology, Dallas, TX) following the standard incubation procedures provided on the instruction manual of the antibody supplier. Finally, after the PVDF membrane was developed in an enhanced chemiluminescence reagent (Thermo Fisher Scientific, Waltham, MA) following the standard assay procedure provided on the instruction manual of the reagent supplier, the protein bands were imaged and analyzed by using a chemiluminescence imaging system to calculate the relative protein expression of CRP, HPSE, and EDN1 in each sample.
ELISA. The plasma concentrations of CRP, HPSE, and EDN1 proteins in each patient were determined by utilizing corresponding enzyme-linked immunosorbent assay kits (R&D systems, Minneapolis, MN) following the standard assay procedure provided on the instruction manual of the assay kit supplier.
## Statistical analysis. all statistical analyses were done by utilizing spss version 19 (spss, chicago, il) and
Prism version 7.0 (GraphPad, San Diego, CA). Differences with P < 0.05 were deemed statistically significant differences in this study. The comparison of non-continuous data between two or more groups was evaluated by using Chi square test. The comparison of continuous data between two and more groups was evaluated by using student t-test and one-way ANOVA, respectively. The normal distribution of each group was evaluated by using Kolmogorov-Smirnov analysis. Inter-group comparisons were carried out using Student's t-tests. The diagnostic value including both sensitivity and specificity was evaluated by using ROC analysis. The sample size was evaluated by using https:// www. stat. ubc. ca/ ~rollin/ stats/ ssize/ n2. html.
# Results
Multivariate logistic regression analysis of reflow-related features of STEMI patients. As shown in [fig_ref] Table 1: Basic clinical features of STEMI patients [/fig_ref] , no significant differences were observed in respect to parameters including age, gender, diabetes history, hypertension history, hyperlipidemia history, familial history, smoking history, chest pain to hospitalization time, and door to balloon time between the normal reflow and no-reflow patient groups. Moreover, no remarkable difference was observed in respect to the angiography and procedural characteristics including infarct related artery, multi-vessel disease, stent implantation and total stent length were compared between the two groups . However, the expression of MALAT1, miR-30e, miR-126, CRP, HPSE and EDN1 showed apparent differences between the two groups [fig_ref] Table 3: Logistic regression analysis of reflow-related features of STEMI patients [/fig_ref].
Differential expression of MALAT1, miR-30e, miR-126, miR-155, CRP, HPSE and EDN1 in the plasma of STEMI patients receiving pPCI. The expression of MALAT1 was remarkably enhanced in the plasma of STEMI patients receiving pPCI who were diagnosed with no-reflow phenomenon [fig_ref] Figure 1: Differential expression of MALAT1, miR-30e, miR-126, miR-155, CRP, HPSE and EDN1 in... [/fig_ref]. However, the expression of miR-30e [fig_ref] Figure 1: Differential expression of MALAT1, miR-30e, miR-126, miR-155, CRP, HPSE and EDN1 in... [/fig_ref] , miR-126 [fig_ref] Figure 1: Differential expression of MALAT1, miR-30e, miR-126, miR-155, CRP, HPSE and EDN1 in... [/fig_ref] and miR-155 [fig_ref] Figure 1: Differential expression of MALAT1, miR-30e, miR-126, miR-155, CRP, HPSE and EDN1 in... [/fig_ref] was notably suppressed in the plasma of no-reflow group. Moreover, the abundance of CRP [fig_ref] Figure 1: Differential expression of MALAT1, miR-30e, miR-126, miR-155, CRP, HPSE and EDN1 in... [/fig_ref] , HPSE [fig_ref] Figure 1: Differential expression of MALAT1, miR-30e, miR-126, miR-155, CRP, HPSE and EDN1 in... [/fig_ref] and EDN1 [fig_ref] Figure 1: Differential expression of MALAT1, miR-30e, miR-126, miR-155, CRP, HPSE and EDN1 in... [/fig_ref] was significantly increased in the plasma of STEMI patients receiving pPCI who were diagnosed with no-reflow phenomenon.
The expression of MALAT1, miR-30e, miR-126 and CRP could be used as the predictive biomarker to differentiate normal reflow and no-reflow in STEMI patients receiving pPCI. Furthermore, the AUC for MALAT1 was as high as 0.95, indicating that MALAT1 was an ideal biomarker for the prediction of reflow phenomenon [fig_ref] Figure 2: ROC analysis showed considerable efficiency of MALAT1, miR-30e, miR-126 and CRP to... [/fig_ref]. Besides, the AUCs of miR-30e [fig_ref] Figure 2: ROC analysis showed considerable efficiency of MALAT1, miR-30e, miR-126 and CRP to... [/fig_ref] , miR-126 [fig_ref] Figure 2: ROC analysis showed considerable efficiency of MALAT1, miR-30e, miR-126 and CRP to... [/fig_ref] and miR-155 [fig_ref] Figure 2: ROC analysis showed considerable efficiency of MALAT1, miR-30e, miR-126 and CRP to... [/fig_ref] were 0.85, 0.80 and 0.55, respectively, while the AUCs of CRP [fig_ref] Figure 2: ROC analysis showed considerable efficiency of MALAT1, miR-30e, miR-126 and CRP to... [/fig_ref] , HPSE [fig_ref] Figure 2: ROC analysis showed considerable efficiency of MALAT1, miR-30e, miR-126 and CRP to... [/fig_ref] and EDN1 [fig_ref] Figure 2: ROC analysis showed considerable efficiency of MALAT1, miR-30e, miR-126 and CRP to... [/fig_ref] were 0.75, 0.60 and 0.65, respectively. These results indicated that the expression of MALAT1, miR-30e, miR-126 and CRP could be used as the biomarker to differentiate normal reflow and no-reflow in STEMI patients receiving pPCI.
The luciferase activities of MALAT1 were suppressed by miR-30e, miR-126 and miR-155. The luciferase activity of HPSE was inhibited by miR-126 and the luciferase activity of EDN1 was repressed by miR-155. Binding site screening respectively identified potential binding site of miR-30e [fig_ref] Figure 3: Luciferase assays were used to explore the relationship of MALAT1/miR-30e, MALAT1/miR-126, MALAT1/miR-155,... [/fig_ref] , miR-126 [fig_ref] Figure 3: Luciferase assays were used to explore the relationship of MALAT1/miR-30e, MALAT1/miR-126, MALAT1/miR-155,... [/fig_ref] and miR-155 [fig_ref] Figure 3: Luciferase assays were used to explore the relationship of MALAT1/miR-30e, MALAT1/miR-126, MALAT1/miR-155,... [/fig_ref] in MALAT1, and the luciferase activity of wild type MALAT1 was remarkably suppressed by the respective transfection of miR-30e [fig_ref] Figure 3: Luciferase assays were used to explore the relationship of MALAT1/miR-30e, MALAT1/miR-126, MALAT1/miR-155,... [/fig_ref] , miR-126 [fig_ref] Figure 3: Luciferase assays were used to explore the relationship of MALAT1/miR-30e, MALAT1/miR-126, MALAT1/miR-155,... [/fig_ref] and miR-155 [fig_ref] Figure 3: Luciferase assays were used to explore the relationship of MALAT1/miR-30e, MALAT1/miR-126, MALAT1/miR-155,... [/fig_ref] in HUVEC and HAEC cells. Meanwhile, potential binding site of miR-126 was found in the 3′ UTR of HPSE, and luciferase activity of wild type HPSE was remarkably suppressed by miR-126 in HUVEC and HAEC cells [fig_ref] Figure 3: Luciferase assays were used to explore the relationship of MALAT1/miR-30e, MALAT1/miR-126, MALAT1/miR-155,... [/fig_ref]. Also, miR-155 was found to potentially bind to EDN1, and the luciferase activity of wild type EDN1 was remarkably suppressed by miR-155 in HUVEC and HAEC cells [fig_ref] Figure 3: Luciferase assays were used to explore the relationship of MALAT1/miR-30e, MALAT1/miR-126, MALAT1/miR-155,... [/fig_ref].
MALAT1 siRNA activated the expression of miR-30e, miR-126 and miR-155, but suppressed the expression of CRP, HPSE and EDN1. The expression of MALAT1 was dramatically suppressed by MALAT1 siRNA in HUVEC and HAEC cells [fig_ref] Figure 4: MALAT1 siRNA altered the expression of miR-30e, miR-126, miR-155, CRP, HPSE and... [/fig_ref] , but the expression of miR-30e [fig_ref] Figure 4: MALAT1 siRNA altered the expression of miR-30e, miR-126, miR-155, CRP, HPSE and... [/fig_ref] , miR-126 [fig_ref] Figure 4: MALAT1 siRNA altered the expression of miR-30e, miR-126, miR-155, CRP, HPSE and... [/fig_ref] and miR-155 [fig_ref] Figure 4: MALAT1 siRNA altered the expression of miR-30e, miR-126, miR-155, CRP, HPSE and... [/fig_ref] was remarkably activated by MALAT1 siRNA in HUVEC and HAEC cells. On the contrary, the expression of CRP [fig_ref] Figure 4: MALAT1 siRNA altered the expression of miR-30e, miR-126, miR-155, CRP, HPSE and... [/fig_ref] , HPSE [fig_ref] Figure 4: MALAT1 siRNA altered the expression of miR-30e, miR-126, miR-155, CRP, HPSE and... [/fig_ref] and EDN1 [fig_ref] Figure 4: MALAT1 siRNA altered the expression of miR-30e, miR-126, miR-155, CRP, HPSE and... [/fig_ref] mRNA was notably inhibited by MALAT1 siRNA in HUVEC and HAEC cells. Furthermore, the protein expression of CRP, HPSE and EDN1 was notably inhibited by MALAT1 siRNA in HUVEC and HAEC cells [fig_ref] Figure 4: MALAT1 siRNA altered the expression of miR-30e, miR-126, miR-155, CRP, HPSE and... [/fig_ref]. The original blots of CRP, HPSE and EDN1 are present in the Supplementary Information.
# Discussion
In this study, we recruited 198 ST-segment elevation myocardial infarction (STEMI) patients receiving primary percutaneous coronary intervention (pPCI) and divided them into two groups according to their reflow status. QPCR was performed to evaluate the expression of MALAT1, miR-30e, miR-126, and miR-155 in the plasma of patients with normal reflow and no-reflow. The expression of MALAT1 was remarkably enhanced, while the expression of miR-30e, miR-126 and miR-155 was significantly suppressed in the plasma of STEMI patients receiving pPCI who were diagnosed with no-reflow phenomenon. In addition, ELISA was performed to analyze the expression of CRP, HPSE and EDN1 in the plasma of patients with normal reflow and no-reflow. The expression of CRP, HPSE and EDN1 was notably upregulated in the plasma of STEMI patients receiving pPCI who were diagnosed with no-reflow phenomenon. MALAT1 was confirmed to be associated with numerous forms of tumors. For instance, MALAT1 is overexpressed in many cancer cells to control their proliferation, invasion, migration, and apoptosis [bib_ref] MALAT1: A long non-coding RNA highly associated with human cancers, Zhao [/bib_ref]. Latest research also showed that lncRNA MALAT1 enhanced the invasion and proliferation of pancreatic, ovarian and glioma cancer cells by activating autophagy [bib_ref] Long noncoding RNA MALAT1 promotes aggressive pancreatic cancer proliferation and metastasis via..., Li [/bib_ref] [bib_ref] Malat1 activates autophagy and promotes cell proliferation by sponging miR-101 and upregulating..., Fu [/bib_ref] [bib_ref] Interaction of E3 ubiquitin ligase MARCH7 with long noncoding RNA MALAT1 and..., Hu [/bib_ref]. A previous research presented that MALAT1 regulated ATG5 expression www.nature.com/scientificreports/ and autophagy through miR-30e [bib_ref] 3,3'-Diindolylmethane induces anti-human gastric cancer cells by the miR-30e-ATG5 modulating autophagy, Ye [/bib_ref]. Likewise, it was additionally discovered that miR-30e may be inhibited by MALAT1 to up-regulate ATG5 as well as activate autophagy in GC. Furthermore, MALAT1 contains two miR-126-5p miRNA response elements (MREs). Dual luciferase, biotin-RNA pull-down, as well as RNA-IP assays all presented that MALAT1 can sponge miR-126-5p 23 . MiR-155 was first identified to play an immune regulatory role in adaptive and innate immune responses [bib_ref] microRNA-155 regulates the generation of immunoglobulin class-switched plasma cells, Vigorito [/bib_ref] [bib_ref] Regulation of the germinal center response by microRNA-155, Thai [/bib_ref]. It was also suggested that miR-155 plays an essential role in the onset of fibrosis [bib_ref] Mir-155 is overexpressed in systemic sclerosis fibroblasts and is required for NLRP3..., Artlett [/bib_ref] [bib_ref] Identification of keratinocyte growth factor as a target of microRNA-155 in lung..., Pottier [/bib_ref]. Over-expressed in fibrotic tissues, miR-155 is essential to collagen metabolism by mediating TGF-1 signaling [bib_ref] Smad2 and Smad3 have differential sensitivity in relaying TGFbeta signaling and inversely..., Liu [/bib_ref] [bib_ref] Competition between Ski and CREB-binding protein for binding to Smad proteins in..., Chen [/bib_ref] [bib_ref] The mechanism of TGF-beta/miR-155/c-Ski regulates endothelial-mesenchymal transition in human coronary artery endothelial..., Wang [/bib_ref]. In this study, we used luciferase assay to explore the regulatory relationship of MALAT1/ miR-30e, MALAT1/miR-126, MALAT1/miR-155, miR-126/HPSE, and miR-155/EDN1. The luciferase activity of MALAT1 was effectively suppressed by miR-30e, miR-126 and miR-155. The luciferase activity of HPSE was inhibited by miR-126. The luciferase activity of EDN1 was repressed by miR-155. Moreover, we suppressed the expression of MALAT1 using MALAT1 siRNA. MALAT1 siRNA up-regulated the expression of miR-30e, miR-126 and miR-155, but suppressed the expression of CRP, HPSE and EDN1. It was show that pre-PCI plasma miRNA-30e expression in STEMI patients could be used to identify the risk of no-reflow. Furthermore, it was shown that the plasma levels of miRNA-30e were positively and significantly correlated with the level of LVEF upon admission, while negatively and considerably correlated with the expression level of hs-CRP. As a result of the rapid rise of CRP level in inflammatory conditions, CRP is considered as an important protein involved in the acute phase of inflammation by playing a significant role in mental stress, anxiety, neoplastic disease and myocardial infarction [bib_ref] Novel insights into an "old" phenomenon: The no reflow, Durante [/bib_ref] [bib_ref] Effect of no-reflow during primary percutaneous coronary intervention for acute myocardial infarction..., Brosh [/bib_ref] [bib_ref] The relation between no-reflow phenomenon and complete blood count parameters, Celik [/bib_ref].
Indicators like endothelin as well as CRP play an important regulatory role during the progression of noreflow [bib_ref] Association of endothelial microparticle with NO, eNOS, ET-1, and fractional flow reserve..., Song [/bib_ref]. In a recent study, the hs-CRP and ET-1 levels in the post-PCI peripheral blood were significantly elevated in the no-reflow group (P < 0.05). Moreover, the serum levels of hs-CRP showed the largest reduction within 3 h in the reflow group after PCI 41 . A number of mechanisms were proposed to support the hypothesis that high ET-1 levels may be used as a potential biomarker for the prediction of no-reflow phenomenon. As one of strongest vasoconstrictors in the body, ET-1 is produced and secreted from the vascular endothelia upon injury [bib_ref] The interaction between endothelin-1 and nitric oxide in the vasculature: New perspectives, Bourque [/bib_ref]. ET-1 also targets coronary arteries with low resistance. It has been revealed that the level of ET-1 was elevated in patients with ischemia ailments [bib_ref] Increased plasma endothelin-1 in acute ischemic stroke, Ziv [/bib_ref]. In patients with renal ischemia/reperfusion injury, the deletion of ET-1 from endothelial tissues protects the normal function of kidneys [bib_ref] ET-1 deletion from endothelial cells protects the kidney during the extension phase..., Arfian [/bib_ref]. Moreover, ET-1 can additionally interact with polymorphonuclear leukocytes to induce no-reflow phenomenon by boosting the adherence of PMN leukocytes to endothelial tissues, therefore promoting the plugging of PMN [bib_ref] Endothelin-1 changes polymorphonuclear leukocytes' deformability and CD11b expression and promotes their retention..., Sato [/bib_ref]. Finally, ET-1 may enhance microvascular compression by elevating the permeability of microvasculature and by inducing edema [bib_ref] Advances in the vascular pathophysiology of ischemic stroke, Del Zoppo [/bib_ref]. Consequently, it is likely that the secretion of ET-1 in endothelia injured by ischemia may trigger sustained and rigorous microvascular constriction to potentiate no-reflow.
Heparanase is ubiquitously expressed to regulate angiogenesis, tumorigenesis, autophagy, fibrosis, and inflammation by cleaving HS groups from proteoglycans. The HS removal then lowers the stability of endothelial cell barriers while regulating the communication with cytokines, growth factors and chemokines [bib_ref] Potential roles in multiple sclerosis, Changyaleket [/bib_ref]. The logistic regression revealed that miR-126 and also hs-CRP are independent risk factors of coronary slow flow. In addition, the plasma expression of miR-126 was significantly associated with the level of CSF, thus miR-126 may be utilized as a valuable predictor of coronary slow flow. Elevated level of HPA in STEMI patients was associated with a high level of TB. Moreover, elevated level of HPA may predict thrombotic issues like no-reflow in STEMI patients [bib_ref] Heparanase is a predictive marker for high thrombus burden in patients with..., Gurbuz [/bib_ref]. In this study, we performed ROC analysis to evaluate the diagnostic potential of MALAT1, miR-30e, miR-126, miR-155, CRP, HPSE and EDN1 in differentiating normal reflow and no-reflow in STEMI patients receiving pPCI. MALAT1, miR-30e, miR-126 and CRP showed considerable efficiency in differentiating normal reflow and no-reflow in STEMI patients receiving pPCI.
The computational analysis indicated the MALAT1 as a regulator of miR-30e/CRP, miR-126/HPSE, and miR-155/END1. Furthermore, luciferase reporter assay results showed that there are interactions between MALAT1 and those three miRNAs (miR-30e, miR-126, and miR-155) and the CRP, HPSE, and END1 are the direct genes of miR-30e, miR-126, and miR-155, respectively. Those results were also confirmed by the results of cellular transfection of miRNA mimics. All those results that MALAT1 is a upstream regulator, which might be the basis for its role as a better biomarker for no-reflow phenomenon.
# Conclusion
In this study, we demonstrated that MALAT1 could respectively sponge the expression of miR-30e, miR-126 and miR-155. The subsequent assays validated that miR-30e, miR-126 and miR-155 targeted CRP, HPSE and EDN1 in a negative manner. Therefore, since MALAT1 was found to present the highest AUC value among all studied genes, MALAT1 was shown to function as a biomarker of no-reflow phenomenon in STEMI patients who were subjected to pPCI.
## Data availability
The data of this study are available from the corresponding author upon reasonable request.
[fig] Figure 1: Differential expression of MALAT1, miR-30e, miR-126, miR-155, CRP, HPSE and EDN1 in the plasma of STEMI patients receiving pPCI (*P value < 0.05 vs. normal reflow group). (A) The expression of MALAT1 was enhanced in the plasma of STEMI patients receiving pPCI who were diagnosed with no-reflow phenomenon when compared with patients with normal reflow. (B) The expression of miR-30e was suppressed in the plasma of STEMI patients receiving pPCI who were diagnosed with no-reflow phenomenon when compared with patients with normal reflow. (C) The expression of miR-126 was suppressed in the plasma of STEMI patients receiving pPCI who were diagnosed with no-reflow phenomenon when compared with patients with normal reflow. (D) The expression of miR-155 was suppressed in the plasma of STEMI patients receiving pPCI who were diagnosed with no-reflow phenomenon when compared with patients with normal reflow. (E) The expression of CRP was increased in the plasma of STEMI patients receiving pPCI who were diagnosed with no-reflow phenomenon when compared with patients with normal reflow. (F) The expression of HPSE was increased in the plasma of STEMI patients receiving pPCI who were diagnosed with no-reflow phenomenon when compared with patients with normal reflow. (G) The expression of EDN1 was increased in the plasma of STEMI patients receiving pPCI who were diagnosed with no-reflow phenomenon when compared with patients with normal reflow. [/fig]
[fig] Figure 2: ROC analysis showed considerable efficiency of MALAT1, miR-30e, miR-126 and CRP to differentiate normal reflow and no-reflow in STEMI patients receiving pPCI. (A) ROC analysis of MALAT1 expression in differentiating normal reflow and no-reflow in STEMI patients receiving pPCI (95% confidence interval: 0.94 to 0.98; P value < 0.0001). (B) ROC analysis of miR-30e expression in differentiating normal reflow and no-reflow in STEMI patients receiving pPCI (95% confidence interval: 0.81 to 0.88; P value < 0.0001). (C) ROC analysis of miR-126 expression in differentiating normal reflow and no-reflow in STEMI patients receiving PCI (95% confidence interval: 0.77 to 0.85; P value < 0.0001). (D) ROC analysis of miR-155 expression in differentiating normal reflow and no-reflow in STEMI patients receiving pPCI (95% confidence interval: 0.49 to 0.60; P value = 0.1257). (E) ROC analysis of CRP expression in differentiating normal reflow and no-reflow in STEMI patients receiving pPCI (95% confidence interval: 0.81 to 0.80; P value < 0.0001). (F) ROC analysis of HPSE expression in differentiating normal reflow and no-reflow in STEMI patients receiving pPCI (95% confidence interval: 0.54 to 0.65; P value = 001,525). (G) ROC analysis of EDN1 expression in differentiating normal reflow and no-reflow in STEMI patients receiving pPCI (95% confidence interval: 0.60 to 0. [/fig]
[fig] Figure 3: Luciferase assays were used to explore the relationship of MALAT1/miR-30e, MALAT1/miR-126, MALAT1/miR-155, miR-126/HPSE, and miR-155/EDN1 (* P value < 0.05 vs. Control group). (A) The luciferase activity of MALAT1 was suppressed by miR-30e in HUVEC and HAEC cells. (B) The luciferase activity of MALAT1 was suppressed by miR-126 in HUVEC and HAEC cells. (C) The luciferase activity of MALAT1 was suppressed by miR-155 in HUVEC and HAEC cells. (D) The luciferase activity of HPSE was suppressed by miR-126 in HUVEC and HAEC cells. (E) The luciferase activity of EDN1 was suppressed by miR-155 in HUVEC and HAEC cells. [/fig]
[fig] Figure 4: MALAT1 siRNA altered the expression of miR-30e, miR-126, miR-155, CRP, HPSE and EDN1 in HUVEC and HAEC cells (*P value < 0.05 vs. NC siRNA). (A) MALAT1 siRNA dramatically suppressed the expression of MALAT1 in HUVEC and HAEC cells. (B) MALAT1 siRNA notably enhanced the expression of miR-30e in HUVEC and HAEC cells. (C) MALAT1 siRNA notably enhanced the expression of miR-126 in HUVEC and HAEC cells. (D) MALAT1 siRNA notably enhanced the expression of miR-155 in HUVEC and HAEC cells. (E) MALAT1 siRNA remarkably suppressed the expression of CRP mRNA in HUVEC and HAEC cells. (F) MALAT1 siRNA remarkably suppressed the expression of HPSE mRNA in HUVEC and HAEC cells. (G) MALAT1 siRNA remarkably suppressed the expression of EDN1 mRNA in HUVEC and HAEC cells. (H) Western blot analysis showed that the protein expression of CRP, HPSE and EDN1 was suppressed by MALAT1 siRNA in HUVEC and HAEC cells. (I) Quantitative analysis showed that the protein expression of CRP was suppressed by MALAT1 siRNA in HUVEC and HAEC cells. (J) Quantitative analysis showed that the protein expression of HPSE was suppressed by MALAT1 siRNA in HUVEC and HAEC cells. (K) Quantitative analysis showed that the protein expression of EDN1 was suppressed by MALAT1 siRNA in HUVEC and HAEC cells. Scientific Reports | (2022) 12:3294 | https://doi.org/10.1038/s41598-022-06923-z [/fig]
[table] Table 1: Basic clinical features of STEMI patients.Table 2. Angiography and procedural features of STEMI patients. [/table]
[table] Table 3: Logistic regression analysis of reflow-related features of STEMI patients. [/table]
|
A pulse of mid-Pleistocene rift volcanism in Ethiopia at the dawn of modern humans
LNE and 31-01-LE are 3shown. Ages shown are the weighted mean ± standard error of the mean (1σ), and in brackets 4 the arithmetic mean ± one standard deviation. While both methods of error reporting are 5 equally valid, for consistency throughout the article, we take precedent from previous 6 hominin literature 1 and report ages as arithmetic mean ± one standard deviation. 7 8 9 [2] 10 Supplementary Figure 2. Age probability spectrum for USGS 40 Ar/ 39 Ar sanidine 11 analyses from Corbetti. a. Sample COI2E and b. Sample Chabbi-7. Ages shown are the 12 weighted mean ± standard error of the mean (1σ), and in brackets the arithmetic mean ± one 13 standard deviation.14 15 16 Supplementary Figure 3. Calibration of the radiocarbon age for the sample 31-01-08. 17 The charcoal sample was collected beneath young pumiceous pyroclastic deposits on the east 18 of the Aluto edifice. Radiocarbon dates were calibrated with IntCal13 2 and OxCal v.4.2. 19 20 [3] 21 22 23 24 25 26 Supplementary Figure 4. Topographic profiles constructed away from the caldera walls 27 of Gedemsa volcano. Profile are used to estimate the slope of the pre-collapse volcano, and 28 combined with thickness estimates to calculate the volume of the pre-collapse edifice, see 29 Methods section for details. For each profile the slope measurement (i.e., the darker coloured 30 segment of line) was made between the caldera rim and the point at which flat lying rift floor 31 sediments were intersected (i.e., light coloured segment of line). Profile lines are shown in 32 map view in Supplementary Figure 5. Slope = 4.3°D istance from caldera rim (m) Height (m) Line 1 Line 2 Line 3 Line 4 [4] 34 Supplementary Figure 5. DEM differencing method used to mask out post-caldera 35 deposits at Gedemsa and calculate their volume. a. represents the present-day DEM, and 36 lines 1-4 correspond to the topographic profiles given in Supplementary Figure 4. b. shows 37 the masked area (grey). c. shows the present-day DEM with interpolated masked surface and 38 d. shows the difference between a and c, shown as residual height. The area of the masked 39 region is given in the bottom left, while the volume of masked region (i.e., the volume of 40 post-caldera deposits) is given in the bottom right. Note that volume shown is not corrected to 41 a dense rock equivalent (DRE).42 43 [5] 44 Supplementary Figure 6. DEM differencing method used to mask out and calculate the 45 volume of the main edifice of Aluto. We approximate this as the volume of all post-caldera 46 deposits (see Supplementary Table 3 for further details). a. represents the present-day DEM. 47 101 volcano and the surrounding region. Major elements were determined by XRF and are 102 shown in wt. % while trace elements were determined by both XRF and ICP-MS methods 103 and are shown in ppm (data after refs. 3,4). The upper left plot shows a total alkalis versus 104 silica (TAS) diagram 5 . The grey dashed line shows the alkaline-sub-alkaline divide of ref. 6. 105 The right hand plots show incompatible elements (Nb, Rb and La) plotted against Zr. While 106 the Munesa ignimbrite (grey inverted triangle, analysed by ref. 3) is broadly comparable to 107 the Aluto rhyolites in major element chemistry, incompatible element plots demonstrate that 108 the Munesa sample falls off the main linear array and evidences a different fractionation trend 109 from the Aluto samples. We conclude that the available geochemical evidence does not 110 require a genetic link between the Munesa (Pliocene) and pre-, syn-and post-caldera 111 (Pleistocene) magmas from Aluto.112 113
The ring structure is hypothesized to be 9 km × 6 km and with a minimum height of ~100 m (ref. 12). Deep well stratigraphy records 13,14 reveal grey-green welded ignimbrites, variably consolidated tuffs and lithic breccias overlying lacustrine sediments and pre-Aluto mafic lava sequences. The deep well ignimbrites also underlie the obsidian lava flows that characterize the post-caldera deposits. The deep well ignimbrites have a fine, devitrified groundmass, and contain sanidine phenocrysts as well as rhyolitic and trachytic lithic fragments. We correlate these units with the surface deposits of welded grey and green ignimbrite sequences that yielded overlapping [bib_ref] Ar*/ 40 K for the Fish Canyon sanidine standard, and improved accuracy..., Pr Renne [/bib_ref] Ar/ 39 Ar ages of 316 ± 13 ka and 306 ± 12 ka . Post-caldera volcanism has generated a set of small coalescing edifices (200-250 m high) in the centre of the caldera . The post-caldera eruptive units are composed of alternating rhyolitic lavas and pumice-fallout deposits .
## Aluto
## 27
Several phases of post-caldera silicic volcanism have occurred at Aluto over the last ~60 ka , and deep wells sited in the centre of the complex reveal that post-caldera units (peralkaline rhyolite lavas and pyroclastic units) are ~250 m in thickness 3,17 . Estimating the volume of these units using the DEM masking approach is complicated because the post-caldera deposits do not have a clearly defined morphology (i.e., they mantle the pre-existing volcanic collapse feature, rather than fill the caldera floor). The total volume of the present edifice is 32 km 3 (Supplementary , this represents a relatively high estimate of the total volume of post-caldera products because the collapse topography, buried beneath the surface, may be included as part of the volume. A simplified approach to calculate the total volume of the post-caldera units at Aluto, is to assume they have a constant thickness (250 m, after deep well observations 3,17 ) proximal to the volcanic centre, equivalent to a cylinder of 6 km radius, and this yields a total volume of 28 km 3 . This value is almost identical to the total volume of the current edifice (Supplementary , and while the good agreement is probably partly fortuitous, it indicates that an estimate of ~30 km 3 is the correct order of magnitude and therefore, following conversion to DRE, we take the volume of post-caldera collapse products at Aluto as 27 km 3 .
## Shala
## 56
3 Post-caldera volcanism has been limited to the Tulu Fike pyroclastic cone on the northern shore of Shala lake 9 .
## Corbetti
## 15
Post-caldera activity at Corbetti is represented by two volcanic cones (Urji and Chabbi, formed of pumice flows, pumice fall and obsidian coulees 18 .
[14]
Supplementary
[fig] Supplementary, Figure 7 Supplementary, Figure 9: shows the masked area (grey). c. shows the present-day DEM with interpolated masked48 surface and d. shows the difference between a and c, shown as residual height. The area of 49 the masked region is given in the bottom left, while the volume of masked region (i.e., the 50 volume of edifice) is given in the bottom right. Note that volume shown is not corrected to a 51 dense rock equivalent (DRE). Topographic profiles constructed away from the caldera walls 59 of Shala volcano. Profile are used to estimate the slope of the pre-collapse volcano, and 60 combined with thickness estimates to calculate the volume of the pre-collapse edifice, see 61 Methods section for details). For each profile the slope measurement (i.e., the darker coloured 62 segment of line) was made between the caldera rim and the point at which flat lying rift floor 63 sediments were intersected (i.e., light coloured segment of line). Profile lines are shown in Shala and calculate their volume. a. represents the present-day DEM, and lines 69 1-3 correspond to the topographic profiles given in Supplementary Figure 7. b. shows the 70 masked area (grey). c. shows the present-day DEM with interpolated masked surface and d. 71 shows the difference between a and c, shown as residual height. The area of the masked 72 region is given in the bottom left, while the volume of masked region (i.e., the volume of 73 post-caldera deposits) is given in the bottom right. Note that volume shown is not corrected to 74 a dense rock equivalent (DRE). Topographic profiles constructed away from the caldera walls 83 of Corbetti volcano. Profiles are used to estimate the slope of the pre-collapse volcano, and 84 combined with thickness estimates to calculate the volume of the pre-collapse edifice, see 85 Methods section for details. For each profile the slope measurement (i.e., the darker coloured 86 segment of line) was made between the caldera rim and the point at which flat lying rift floor 87 sediments were intersected (i.e., light coloured segment of line). Profile lines are shown in 88 map view in Supplementary Figure 10. [/fig]
[fig] Figure 4, Figure 7, Figure 9 *: pre-caldera volcanic sequences (Fig. 2d) are dominated by rhyolitic lavas, although pumice fall and ignimbrite deposits are also present 7,8 . Individual rhyolite lava flows are thick (20-30 m), 2-5 km long and show a clear flow folding 8 . Based on the geological map of ref. 7 the minimum thickness of the pre-caldera deposits, identified in the northern caldera wall of Gedemsa, is 250 m. Slope measurements, made on the northern and southern back walls of the caldera (i.e., the non-faulted sections) range between 4-7.5° (Supplementary for a pre-collapse edifice is observed along faults on the margins of the complex (Figs. 2a, 3) where a minimum thickness (~100 m) of trachyte lavas and tuffs ~8 km from the centre of the present edifice are found. Our interpretation is that trachyte lava flows and tuffs built up a low relief silicic complex (or lava shield) upon the faulted rift terrain. The pre-caldera edifice has been largely covered by recent volcanic deposits, and so we assume that the slope of the pre-caldera edifice is in line with that of neighbouring complexes (and colleagues (ref. 9) consider the pre-caldera edifice to have comprised a thick pile of flows and domes of rhyolite. Sites sampled at the caldera wall show well preserved sequences and a minimum thicknesses of trachyte and rhyolite for the pre-collapse edifice is 50 -150 m (ref. 9). Measurements of slope on the southern caldera wall (N.B. other wall sections are deeply eroded and/or faulted) show values of 2-4° (Supplementary lava domes and flows are reported in caldera wall sequences of Corbetti and pre-date the caldera-forming eruption 10,11(Fig. 2b). The pre-caldera lava exposures are well in excess of 50 m, and we consider a range of 100-200 m as a reasonable approximation of their true thickness. The range of slopes for the Corbetti caldera walls are shown in Supplementary At Aluto recent (post-caldera volcanism) has obscured much of the original caldera wall and so we use fault exposures off the main edifice to evaluate pre-caldera thicknesses[12] Supplementary [/fig]
[table] Table 2: Caldera collapse volume estimates. [/table]
[table] Table 3: Post-caldera volume estimates. [/table]
[table] Table 4: Values used in the data regression. Supplementary Table 5. Reactor production ratios and interfering isotope production McDougall, I., Brown, F. H. & Fleagle, J. G. Stratigraphic placement and age of modern humans from Kibish, Ethiopia. Nature 433, 733-736 (2005). Reimer, P. IntCal13 and Marine13 Radiocarbon Age Calibration Curves 0-50,000 [/table]
|
Coexistence of granuloma annulare and chronic spontaneous urticaria*
[bib_ref] Etiology, diagnosis, and therapeutic management of granuloma annulare: an update, Thornsberry [/bib_ref] [bib_ref] Interstitial type granuloma annulare associated with Sjögren's syndrome, Sumikawa [/bib_ref] [fig_ref] Figure 1: Annular erythematous to violaceous plaques on the right forearm and urticarial papules,... [/fig_ref] [bib_ref] Etiology, diagnosis, and therapeutic management of granuloma annulare: an update, Thornsberry [/bib_ref] [bib_ref] Localized granuloma annulare and autoimmune thyroiditis in adult women: a case-control study, Vázquez-López [/bib_ref] [bib_ref] Etiology, diagnosis, and therapeutic management of granuloma annulare: an update, Thornsberry [/bib_ref] [bib_ref] Granuloma annulare: tissue distribution of factor XIIIa+ dermal dendrocytes, thrombomodulin+ dermal cells..., Soub [/bib_ref] [bib_ref] Idiopathic chronic urticaria and thyroid autoimmunity: Experience of a single center, Nuzzo [/bib_ref]
[fig] Figure 1: Annular erythematous to violaceous plaques on the right forearm and urticarial papules, and plaques on the trunk eration of collagen bundles surrounded by multinuclear giant cells, histiocytes and lymphocytes in the dermis (Figure 2). These features were consistent with granuloma annulare. [/fig]
[fig] FIgure 2: Focal degeneration of collagen bundles surrounded by multinuclear giant cells, histiocytes and lymphocytes in the dermis (HEx40) [/fig]
|
Multilayer Dye Aggregation at Dye/TiO2 Interface via π…π Stacking and Hydrogen Bond and Its Impact on Solar Cell Performance: A DFT Analysis
Multilayer dye aggregation at the dye/TiO 2 interface of dye-sensitized solar cells is probed via firstprinciples calculations, using p-methyl red azo dye as an example. Our calculations suggest that the multilayer dye aggregates at the TiO 2 surface can be stabilized by π…π stacking and hydrogen bond interactions. Compared with previous two-dimensional monolayer dye/TiO 2 model, the multilayer dye aggregation model proposed in this study constructs a three-dimensional multilayer dye/TiO 2 interfacial structure, and provides a better agreement between experimental and computational results in dye coverage and dye adsorption energy. In particular, a dimer forms by π…π stacking interactions between two neighboring azo molecules, while one of them chemisorbs on the TiO 2 surface; a trimer may form by introducing one additional azo molecule on the dimer through a hydrogen bond between two carboxylic acid groups. Different forms of multilayer dye aggregates, either stabilized by π…π stacking or hydrogen bond, exhibit varied optical absorption spectra and electronic properties. Such variations could have a critical impact on the performance of dye sensitized solar cells.Dye sensitized solar cells (DSSCs) utilize molecular dye absorbers to convert photo energy into electrical power 1-4 . The dye/TiO 2 interface in the working electrode of a DSSC is responsible for light absorption and charge separation in a DSSC 5,6 . Various researches have been carried out to understand the interfacial dye/TiO 2 structures at nanoscale toward formulating structure-property relationship of the photoelectrode and optimizing DSSC performance 7,8 .Substantial first principles calculations have been performed in order to establish an accurate model to describe the nanoscale structures of dye/TiO 2 interface 8-11 . Monomeric dye adsorption on TiO 2 surface has been extensively evaluated for various dyes to obtain adsorption geometry and dye/TiO 2 binding mode 12-14 . Yet, the monomeric dye adsorption model is often inadequate to describe the dye/TiO 2 interface; this is due to the fact that many dye molecules are readily adsorbed on the TiO 2 surface and form closed packed dye aggregates via intermolecular interactions 15,16 . It has been identified that molecular dye aggregates are responsible for novel optoelectronic properties for functional molecular dyes in DSSCs and related devices, and numerous researches have been performed to manipulate molecular aggregates and fine-tune their materials properties[17][18][19][20][21][22]. To this end, dimeric dye aggregates at different anchoring positions on the TiO 2 surface have been analyzed by Filippo et al.23, offering nanoscopic insights on experimental phenomena that is inconsistent with the monomeric adsorption model. To more closely reassemble experimental scenarios, higher degrees of dye aggregates, such as trimer, tetramer and pentamer, have been constructed via first principles calculations 24 . These aggregation models qualitatively explain the "H-aggregation" phenomena of dyes at the dye/TiO 2 interface, where blue-shifted light
Scientific RepoRts | 6:35893 | DOI: [bib_ref] What Is the Best Anchoring Group for a Dye in a Dye-Sensitized..., Ambrosio [/bib_ref].1038/srep35893 absorption spectra are observed. However, the existing aggregation models are limited in the lateral direction relative to the TiO 2 surface, while dye aggregate structures built outward from the TiO 2 surface are neglected. In other words, a monolayer model is assumed. The monolayer model sets an intrinsic limitation on the amount of dye adsorption on TiO 2 surface, leadings to theoretical predications that are two orders of magnitude smaller than the experimental results [bib_ref] Variation in Optoelectronic Properties of Azo Dye-Sensitized TiO 2 Semiconductor Interfaces with..., Zhang [/bib_ref] [bib_ref] Adsorption Properties of p-Methyl Red Monomeric-to-Pentameric Dye Aggregates on Anatase (101) Titania..., Zhang [/bib_ref]. An improved multilayer model to describe the interfacial dye aggregates is thus required to bridge the gap between experimental and computational results.
Azo dyes are frequently used in textile and coating industries due to its stability, vivid color, high molar extinction, and low cost [bib_ref] Enhancement of photogenerated mechanical force in azobenzene-functionalized polyimides, Lee [/bib_ref] [bib_ref] Torsionally responsive C3-Symmetric azo dyes: azo-hydrazone tautomerism, conformational switching, and application for..., Lee [/bib_ref] [bib_ref] Photo-responsive carbon nanomaterials functionalized by azobenzene moieties: structures, properties and application, Feng [/bib_ref]. Recently, it has been successfully incorporated into DSSCs and convert light into electricity [bib_ref] Low band gap dyes based on 2-styryl-5-phenylazo-pyrrole: synthesis and application for efficient..., Mikroyannidis [/bib_ref] [bib_ref] Relating Electron Donor and Carboxylic Acid Anchoring Substitution Effects in Azo Dyes..., Zhang [/bib_ref]. The aggregation of p-methyl red azo dyes has been evaluated in previous research in the lateral direction relative to the TiO 2 surface [bib_ref] Adsorption Properties of p-Methyl Red Monomeric-to-Pentameric Dye Aggregates on Anatase (101) Titania..., Zhang [/bib_ref] , and the study on the p-methyl red azo dyes serves as a good platform to further analyze their intermolecular interactions and thereby construct the multilayer aggregation model.
In this manuscript, p-methyl red phenyl]diazenyl]-benzoic acid] is used to probe the dye aggregation on TiO 2 surface [fig_ref] Figure 1: The molecular structure of 4-[2-[4- [/fig_ref]. The nanoscale structures of different dye aggregates stabilized via intermolecular forces will be presented, followed by a detailed discussion on their electronic and optical properties that impact solar cell performance.
# Results and discussion
Structure of dye adsorbates on a (101) TiO 2 surface. In the optimized structure of the azo dye/TiO 2 [fig_ref] Figure 2: Optimized geometries of monomer 1, dimer 2 and trimer 3 of methyl... [/fig_ref] , the molecular axis of the adsorbates lies perpendicular to the TiO 2 surface. Two molecules in dimer 2 are stabilized by π … π interactions. The top azo molecule in dimer 2 has it carboxylic acid group facing outward away from the TiO 2 surface. The separations between two oxygen atoms in the carboxylic acid group and TiO 2 surface are similar among 1-3 (2.08 Å, 2.07 Å and 2.05 Å for 1-3, respectively), consistent with previously reported values in the literature [bib_ref] Adsorption Properties of p-Methyl Red Monomeric-to-Pentameric Dye Aggregates on Anatase (101) Titania..., Zhang [/bib_ref]. The bond lengths of the azo group (-N= N-bond), an indicator of intramolecular charge Both front view and side view are depicted for clarify. The π … π stacking and hydrogen bond interactions are highlighted in red rectangles. The red arrow points to a potential adsorption site for further aggregate expansion via π … π interaction.
Scientific RepoRts | 6:35893 | DOI: 10.1038/srep35893 transfer 30 , are also similar (1.27-1.28 Å) in 1-3. The π … π distance between the two azo molecules are 4.1 Å for both aggregates 2 and 3, while the H… O distance of the hydrogen bond is 1.53 Å for 3. Interestingly, the π … π stacking mode and the hydrogen bond interactions between neighboring carboxylic acid groups have been observed in the crystal structures of azo dyes determined from single crystal X-ray diffraction experiment [bib_ref] Relating Electron Donor and Carboxylic Acid Anchoring Substitution Effects in Azo Dyes..., Zhang [/bib_ref]. The monomeric form 1 demonstrates more perpendicular adsorption geometry compared with 2 and 3. The angle between its molecule axis and TiO 2 surface is 3°-4° larger in 1 than that in 2 and 3. The molecule chemisorbed onto TiO 2 surface is more distorted in the case of 2 and 3, due to steric effects introduced by additional azo molecules.
Dye adsorption energies and the amount of dye loading on TiO 2 surface. The adsorption energy of the three aggregates follows the ranking: 1 (0.60 eV) > 3 (0.56 eV) > 2 (0.43 eV; . This is because 1 possesses chemisorption which is more stable in nature than physisorption. 2 is the most unstable system because the intermolecular π … π interaction is based on relatively weak van der waals forces. Molecules in 3 are connected by hydrogen bond which has interaction strength between chemical bonding and van der waals forces.
In terms of dye coverage, 3 has the highest degree of aggregation (0.43 nmol/cm 2 ) among 1-3 because it involves three molecules in the unit cell. In contrast, 1 demonstrates the lowest dye coverage (0.14 nmol/cm 2 ). The experimental dye coverage is 147 nmol/cm 2 for p-methyl red on TiO 2 nanoparticles in a DSSC photoelectrode [bib_ref] Variation in Optoelectronic Properties of Azo Dye-Sensitized TiO 2 Semiconductor Interfaces with..., Zhang [/bib_ref]. Although a large discrepancy exists between experimental and calculated values, it is of note that 3 provides adsorption sites for further vertical aggregate expansion [fig_ref] Figure 2: Optimized geometries of monomer 1, dimer 2 and trimer 3 of methyl... [/fig_ref]. The outermost molecule in 3 with respect to TiO 2 surface can interact with a forth molecule via π … π interactions (indicated by a red arrow in [fig_ref] Figure 2: Optimized geometries of monomer 1, dimer 2 and trimer 3 of methyl... [/fig_ref] ; so on and so forth. In this way, the π … π stacking and hydrogen bond interactions offers a route to construct periodic structures in the c-axis, leading to much higher dye loading. As a result, the multilayer aggregation model proposed in this manuscript agrees better with the experimental results [bib_ref] Variation in Optoelectronic Properties of Azo Dye-Sensitized TiO 2 Semiconductor Interfaces with..., Zhang [/bib_ref]. In contrast, the monolayer model in the literature produces a maximum dye coverage of 0.42 nmol/cm 2 for p-methyl red on TiO 2 2D slab 8 , three orders of magnitude smaller than experimental values.
We also noticed that after rinsing dye/TiO 2 photoelectrodes during DSSC fabrications (using washing solvent, such as ethanol, to remove weakly adsorbed dye molecules during the sensitization process), the washing solvent displays vivid color due to dye dissolution; this phenomenon indicates that abundant azo dyes are weakly adsorbed on the TiO 2 surface such that they can be easily removed from TiO 2 substrate. Previous aggregation models, that involve only stable chemisorption, are not sufficient to explain these experimental observations. In contrast, the multilayer aggregation model which incorporates physisorption (such as π … π interaction and hydrogen bond) provides a plausible explanation to the large amount of weakly adsorbed azo dyes on TiO 2 substrates.
Optical Properties. The calculated UV/vis absorption spectra of 1-3 exhibit two major bands: one in the UV region near 400 nm and the other in the visible region from 500 nm to 900 nm . The intense band in the UV region corresponds to the light absorption in TiO 2 substrate, while the visible band corresponds to the absorption of the dye adsorbates. Dimer 2 has the most red-shifted visible band (with peak wavelength at 693 nm). This peak wavelength is red shifted by 22 nm compared with that of monomer 1, demonstrating that the π … π stacking interaction is effective in tuning optical properties of DSSC (a red shifted UV/vis absorption spectra is desirable for DSSC device performance for a close match with solar spectrum). Trimer 3 stabilized by hydrogen bond between molecules exhibits the most blue-shifted light absorption spectra among 1-3, indicating that hydrogen bond between neighboring carboxylic acid anchoring groups in azo dyes is detrimental for improving the light absorption of a DSSC device. The experimental light absorption spectra of the p-methyl red/TiO 2 film exhibit severe H-aggregation and its peak wavelength approaches the UV region at 400 nm [bib_ref] Variation in Optoelectronic Properties of Azo Dye-Sensitized TiO 2 Semiconductor Interfaces with..., Zhang [/bib_ref]. As a result, the blue-shifted absorption spectrum of 3 implies that the aggregate 3 represents a possible interfacial structure. Further aggregation expansion, which is computationally expensive and unreachable at this stage, is expected to further narrow the gap between experimental and theoretical results. Note that the aggregate 3 does not conform to the typical H-aggregate model, which usually exhibit a head-to-head and tail-to-tail conformation. Instead, the direct "face-to-face" interactions via the hydrogen bonds in aggregate 3 essentially lead to partial dipole-dipole cancellation, because both carboxylic acid groups carry partial negative charge as a result of intramolecular charge transfer in azo dyes [bib_ref] Relating Electron Donor and Carboxylic Acid Anchoring Substitution Effects in Azo Dyes..., Zhang [/bib_ref] [bib_ref] Graph-set analysis of hydrogen-bond patterns in organic crystals, Etter [/bib_ref]. The net result of such interactions affords a blue-shift [bib_ref] Tuning Solvatochromism of Azo Dyes with Intramolecular Hydrogen Bonding in Solution and..., Zhang [/bib_ref] [bib_ref] Photoinduced Intramolecular Charge Transfer in 4-(Dimethyl)aminobenzonitrile − A Theoretical Perspective, Rappoport [/bib_ref] in the light absorption spectra of aggregate 3.
We also notice that the calculated UV/vis spectra of 1-3 (~600 nm) demonstrate a large red shift against experimental data (~400 nm), which has also been experienced in the simulation based on the monolayer model [bib_ref] Variation in Optoelectronic Properties of Azo Dye-Sensitized TiO 2 Semiconductor Interfaces with..., Zhang [/bib_ref]. This is not surprising, because the pure functional PBE (used in our study) neglects Hartree-Fock like exact exchange and tends to severely underestimate the band gap and the excitation energies of charge-transfer dyes (such as azo dyes) [bib_ref] The calculations of excited-state properties with Time-Dependent Density Functional Theory, Adamo [/bib_ref] [bib_ref] Assessment of Functionals for Optical 0-0 Transitions in Solvated Dyes, Jacquemin [/bib_ref] [bib_ref] Density functional theory and the band gap problem, Perdew [/bib_ref]. Nevertheless, the relative spectral difference among 1-3 is more reliable and provides a strong support to our multi-layer aggregation model.
## Band structure and density of states. the band structures of 1-3 exhibit a series of straight band lines
below Fermi level [fig_ref] Figure 4: Band Structures, and projected density of states [/fig_ref]. These straight band lines are discrete and are contributed by organic molecules [bib_ref] Electronic and optical properties of the triphenylamine-based organic dye sensitized TiO 2..., Liang [/bib_ref]. Nevertheless, these band lines are only defect states and do not affect the electronic band gap of the material. 2 has dense band lines near valence bands contributed by organic dyes. The band lines in 2 are denser than those in 1, due to more atoms available in 2. As a result, the optical absorption spectra, which are related to the difference between conduction band and valence band [bib_ref] Adsorption Properties of p-Methyl Red Monomeric-to-Pentameric Dye Aggregates on Anatase (101) Titania..., Zhang [/bib_ref] , are expected to demonstrate a red-shift in 2 with respect to that in 1. In the case of 3, most bands contributed by the chromophore are localized in the deep region between − 1.5 and − 2.0 eV, which are further away from the conduction band. This could explain the blue shifted absorption spectrum of 3 compared with that of 2. The band lines with large curvature in the upper and lower regions of the band structure plots are contributed by periodic TiO 2 2D slab.
In the PDOS spectra [fig_ref] Figure 4: Band Structures, and projected density of states [/fig_ref] , the HOMO (highest occupied molecular orbital) is contributed by carbon, oxygen and nitrogen atoms in the organic molecules. The deep valence bands between − 6 eV and − 2 eV are mainly contributed by TiO 2 oxygen atoms. The conduction band is mainly contributed by Ti atoms in the TiO 2 substrate. These features are common in a dye/TiO 2 system. Orbital Distributions. The orbital distributions for molecular orbitals HOMO-2, HOMO-1, HOMO, LUMO, and LUMO+ 1 are depicted to provide an intuitive visualization on the electronic properties of the dye/ TiO 2 system [fig_ref] Figure 5: Orbital distributions of the dye/TiO 2 systems [/fig_ref]. Electron distributions in the unoccupied orbitals are almost the same in 1-3: the LUMO and the LUMO+ 1 are distributed in the TiO 2 region, displaying Ti 3d characters. The occupied molecular orbitals vary among the three structures: for 1, HOMO is localized in the azo group of the molecule; HOMO-1 is delocalized across the entire azo molecule; HOMO-2 is in the aromatic ring next to the TiO 2 surface. In the case of 2, the HOMO-2, HOMO-1 and HOMO are localized in the upper molecule interacting with the rest of the system via π … π stacking forces. As a result, this additional molecule in 2 plays an important role in the electronic and optical properties when azo dyes are incorporated in a DSSC device. In the case of 3, HOMO and HOMO-1 are distributed in the uppermost molecule interacting with the rest system via hydrogen bond, while HOMO-2 resides in the middle molecule. For all the three structures, the outermost molecule is the most active in electronic properties due to HOMO localization in this region, thus having an important impact on the electronic and optical properties of the dye/TiO 2 system.
# Conclusions
In this study, we propose a multilayer dye aggregation model to describe the aggregation phenomena in the dye/ TiO 2 interface. We show that multilayer aggregates may form along the c-axis of the dye/TiO 2 unit cell via intermolecular interactions, such as π … π stacking and hydrogen bond interactions. Our calculations suggest that further aggregation growth may take place in the c-axis, forming a large 3D dye/TiO 2 interfacial structure. This multilayer dye aggregation model could thus bridge the gap between theoretical and experimental dye loading on TiO 2 surface, in contrast to the monolayer dye adsorption model. Moreover, the weak adsorptions of azo dyes on TiO 2 substrate, reported in previous experiments, can be rationalized in the new multilayer model in terms of weak physisorption (in contrast to strong chemisorption in the monolayer model). We also investigated the spectral and electronic properties of the multilayer molecular aggregates as well as their impacts on the DSSC device performance. We show that the π … π stacking induce a red shift, while hydrogen bond leads to a blue shift of the light absorption spectra in the dye/TiO 2 system. It is expected that this multilayer model will pave a new way for a deeper understanding of large molecular aggregation systems in DSSC and other relevant applications.
# Methods
A unit cell of (TiO 2 ) 36 exposing anatase 101 surface is constructed to analyze the multilayer dye aggregation. A very long vacuum layer and c-axis (69.4 Å) of the unit cell is built in order to accommodate the multiple dye molecules in the c-axis extending outward from the TiO 2 surface. A bidentate bridging anchoring mode for the carboxylic acid anchoring group on anatase 101, suggested by previous research [bib_ref] Density functional theory study of formic acid adsorption on anatase TiO 2..., Gong [/bib_ref] [bib_ref] Influence of the sensitizer adsorption mode on the opencircuit potential of dye-sensitized..., De Angelis [/bib_ref] , is used in this project. Three types of aggregates 1-3 are investigated in this study, where 1 represents the monomeric adsorption form; 2 is a dimer joined via π … π stacking effects; 3 is a trimer where a third molecule is placed on top of dimer 2 via hydrogen bond. Upon adsorption, the vacuum layer thickness values are ca. 44 Å for 1, 39 Å for 2, and 23 Å for 3, which are sufficient to minimize interactions between neighbouring layers [bib_ref] Atomically Thin Arsenene and Antimonene: Semimetal-Semiconductor and Indirect-Direct Band-Gap Transitions, Zhang [/bib_ref] [bib_ref] Semiconducting Group 15 Monolayers: A Broad Range of Band Gaps and High..., Zhang [/bib_ref] [bib_ref] Semiconductor-topological insulator transition of two-dimensional SbAs induced by biaxial tensile strain, Zhang [/bib_ref] [bib_ref] Hydrogenated arsenenes as planar magnet and Dirac material, Zhang [/bib_ref] [bib_ref] Lateral black phosphorene P-N junctions formed via chemical doping for high performance..., Yu [/bib_ref] [bib_ref] A promising two-dimensional solar cell donor: Black arsenic-phosphorus monolayer with 1.54 eV..., Xie [/bib_ref]. A 1 × 3 × 1 supercell was constructed to understand the multilayer aggregation, corresponding to a surface area of 10.2 × 11.3 Å 2 . The three interfacial structures 1, 2 and 3 are geometrically optimized in VASP 45 , using PBE functional (energy cutoff: 500 eV; ionic displacement convergence criterion: 0.02 eV/Å). During geometric optimization, oxygen atoms at the bottom of the unit cell are frozen while all other atoms are freely relaxed to their equilibrium positions. The band structures and density of states of 1-3 are calculated via an 8 × 8 × 1 monkhorst-pack grid. The optical absorption spectra are obtained from dielectric constant in Castep 46 , using 430 eV as the energy cutoff. Graeme D2 dispersion correction [bib_ref] GGA-type density functional constructed with a long-range dispersion correction, Grimme [/bib_ref] is included in all the calculations. The dye adsorption energy on TiO 2 surface is calculated using: The first term in the bracket corresponds to the total energy of molecular dyes; the second is the total energy of bare TiO 2 substrate, and the third term is the total energy of the dye/TiO 2 system; and n is the number of dye adsorbates.
[formula] = × + − E n E E E n ( ) / [/formula]
[fig] Figure 1: The molecular structure of 4-[2-[4-(dimethylamino)phenyl]diazenyl]-benzoic acid (p-methyl red). [/fig]
[fig] Figure 2: Optimized geometries of monomer 1, dimer 2 and trimer 3 of methyl red molecules. [/fig]
[fig] Figure 4: Band Structures, and projected density of states (PDOS) of C (carbon), H (hydrogen), O (oxygen), N (nitrogen) and Ti (titanium). The vertical dotted lines correspond to the Fermi Energy, which is set to 0 eV. Scientific RepoRts | 6:35893 | DOI: 10.1038/srep35893 [/fig]
[fig] Figure 5: Orbital distributions of the dye/TiO 2 systems. Five orbitals are presented from left to right of each structure: HOMO-2, HOMO-1, HOMO, LUMO and LUMO+ 1. Scientific RepoRts | 6:35893 | DOI: 10.1038/srep35893 [/fig]
|
Firearm bullet settling into the lumbar spinal canal without causing neurological deficit: A report of two cases
Background: Uncertainty still exists regarding the treatment of the patients presenting with gunshot wounds to the spine. Neurological insults, cerebrospinal fluid fistula, infection, lead or copper toxicity, migration of bullets, and spinal instability are included among the common challenging issues. Case Description: An 18-year-old woman was admitted with low back pain following a gunshot injury five days ago. She was neurologically intact. Radiological examinations showed that a bullet was settled in L4-5 disc space. The bullet was removed with a unilateral L4-5 partial hemilaminectomy and discectomy from the left side. The second case was of a 29-year-old man admitted with radiating leg pain on the right side following a gunshot injury from his left side of lower back four months ago. He had only positive straight leg raising test. Radiological studies showed two bullets, one was in the psoas muscle on the left side and the other was in spinal canal that had caused a burst fracture of the L5 vertebra. Following L5 laminectomy and bilateral L5-S1 facetectomy, the bullet was removed from the spinal canal and L5-S1 transpedicular posterior stabilization was performed. The postoperative period of both patients was unremarkable. Conclusion: Bullet settling into the lumbar spinal canal without causing neurological deficit may require surgical intervention. Removal of bullets provided not only pain relief in both the cases but also prevented future complications such as migration of the bullets, plumbism, and neuropathic pain and instability.
# Introduction
Ambrose Pare described a penetrating spinal cord injury caused by a gunshot for the first time in 1557. [bib_ref] Recovery from spinal cord injury, Ducker [/bib_ref] Since then, many combat-related or civilian gunshot injuries to the spine have been reported; however, uncertainty still exists for the treatment of such patients. [bib_ref] Can surgery improve neurological function in penetrating spinal injury? A review of..., Klimo [/bib_ref] The level of the spine primarily determines the severity of the neurological deficit and outcome. [bib_ref] Current concepts in penetrating and blast injury to the central nervous system, Rosenfeld [/bib_ref] Neurological insults, cerebrospinal fluid (CSF) fistula, infection, lead or copper toxicity, migration of bullets, and spinal instability How to cite this article: Hakan T, Çerçi A, Gürcan S, Akçay S. Firearm bullet settling into the lumbar spinal canal without causing neurological deficit: A report of two cases. Surg Neurol Int 2016;7:S251-4. http://surgicalneurologyint.com/Firearm-bullet-settling-into-the-lumbar-spinal-canalwithout-causing-neurological-deficit:-A-report-of-two-cases/ This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-ShareAlike 3.0 License, which allows others to remix, tweak, and build upon the work non-commercially, as long as the author is credited and the new creations are licensed under the identical terms.
## For reprints contact: [email protected]
## Firearm bullet settling into the lumbar spinal canal without causing neurological deficit: a report of two cases
Tayfun Hakan 1,2 , Ajlan Çerçi 2 , Serkan Gürcan 3 , Serkan Akçay 3 [bib_ref] Delayed neurological symptoms from the spontaneous migration of a bullet in the..., Avci [/bib_ref] Okan University, The Vocational School of Health Services, Tuzla, 2 Neurosurgery Clinic, [bib_ref] Gunshot wounds to the spine, Bono [/bib_ref] Orthopedic Clinic, Kurtkoy Ersoy Hospital, Istanbul, Turkey E-mail: *Tayfun Hakan [email protected]; Ajlan Çerçi [email protected]; Serkan Gürcan [email protected]; Serkan Akçay [email protected] *Corresponding author are included among the common challenges. Presence of CSF fistula or progressive neurological worsening correlated with radiological compression of neural elements on radiological studies are regarded as absolute indicators for surgery. [bib_ref] Can surgery improve neurological function in penetrating spinal injury? A review of..., Klimo [/bib_ref] Treatment of the spine injuries with missile should be chosen based on the nature of the injury. [bib_ref] Gunshot injuries of the spine, Jakoi [/bib_ref] Bullet in the intervertebral disc space and/ or existing spinal instability require reconstruction and stabilization using instrumentation and fusion.
In this study, two cases of a gunshot wound in which the bullets lodged in the lumbar spine without any neurological injury are reported.
## Case reports
## Case report 1
An 18-year-old woman was admitted with low back pain following a gunshot injury to the back in the Libyan combat five days ago. During physical examination, an entry wound in her posterior left flank was detected. She had pain around the bullet entry zone and in the low back. Her examination was completely normal. Radiological studies revealed a bullet in the left side of the L4-5 disc space [ [fig_ref] Figure 1: The bullet lodged at L4-5 disk space in coronal [/fig_ref] ]. During operation, a unilateral L4-5 partial hemilaminectomy and discectomy from the left side and removal of the bullet was performed. The bullet was completely embedded into the disc space and partly destroyed the posterior-inferior end plate of the L4 vertebra; otherwise, the other structures of the spine were anatomically intact; no dural tear and/or neural injury was identified. The bullet entry wound in the back of the patient was also repaired during the same session by a surgeon without any complication. Her postoperative period was unremarkable, and she was free from pain originating from the low back.
## Case report 2
A 29-year-old man was admitted with radiating leg pain on his right side following a gunshot injury from his back in the Libyan combat four months ago. His neurological examination was completely normal except a positive straight leg raising test on the right side. Radiological studies showed two bullets, one was in the psoas muscle on the left side and the other was in the spinal canal that had caused a burst fracture of the L5 vertebra [ ]. A congenital L5 lamina defect in the midline was also seen in plain anteroposterior radiograph. He underwent a spinal operation. Following L5 laminectomy and bilateral L5-S1 facetectomy, the bullet was removed from the spinal canal and L5-S1 transpedicular posterior stabilization was performed. There was neither neural nor dural injury. His right leg pain disappeared immediately following operation.
# Discussion
We present two referred cases of combat-related gunshot injury of the spine in the Libyan conflict. In civilian practice, incidence of gunshot wounds to spine accounts 13-17% of all gunshot injuries. [bib_ref] Current concepts in penetrating and blast injury to the central nervous system, Rosenfeld [/bib_ref] Similarly, the incidence of spinal injuries from gunshot were found to be 10 and 17% for the Afghanistan and Iraq war, respectively. [bib_ref] Spinal injuries in United States military personnel deployed to Iraq and Afghanistan:..., Schoenfeld [/bib_ref] Penetrating gunshot injuries to the spine with high caliber bullets have mostly permanent neurological defects. The degree of neural and mechanical damage depends upon the physical properties and the direct impact of the bullet, the pressure shock waves, and the temporary cavitation. [bib_ref] Gunshot injuries in the spine, De Barros Filho [/bib_ref] The degree of transection and contusion of the spinal cord, compression of neural elements by hematoma, bone and/or foreign body fragments, disruption of the spinal cord vasculature, and spinal instability are possible from extension of the injury. [bib_ref] Current concepts in penetrating and blast injury to the central nervous system, Rosenfeld [/bib_ref] Missile settling in the disc space may provoke pain and neurological deficit by causing injury in nucleus pulposes and annulus. [bib_ref] Migration of a bullet in the lumbar intervertebral disc space causing back..., Ceylan [/bib_ref] The first case reported here presented with only low back pain apart from the bullet Thoracic region is the most commonly affected area [bib_ref] Civilian gunshot injuries of the spinal cord: A systematic review of the..., Sidhu [/bib_ref] with most risk of complete injuryfor civilian gunshot wounds. In an epidemiological investigation involving combat casualties and including explosions and gunshot injury, Schoenfeld et al. [bib_ref] Spinal injuries in United States military personnel deployed to Iraq and Afghanistan:..., Schoenfeld [/bib_ref] reported that the most involved area was the lumbar spine, followed by thoracic and cervical spine. The presence of intracanalicular bullet fragments was accepted as an important predictor of neurological injury. [bib_ref] How bullet trajectory affects outcomes of civilian gunshot injury to the spine, Chittiboina [/bib_ref] Both of the presented cases suffered from injury to the lumbar area and had bullet in the spinal canal without any neurological deficit.
Plain films can be used for determining the existence and/or level of the bullet and/or damage of the spine. [bib_ref] Gunshot wounds to the spine, Bono [/bib_ref] For determining the integrity of spine in means of vertebral body, spinal canal, facets and alignment, and better localization computed tomography in axial, coronal, and sagittal planes should be sought. [bib_ref] Can surgery improve neurological function in penetrating spinal injury? A review of..., Klimo [/bib_ref] There is still controversy for optimal treatment for bullets lodged in the spinal canal. The treatment modalities for such patients need to be managed individually. The treatment strategies-bullet removal from lumbar disc space and bullet removal from spinal canal and stabilization-were decided after detailed evaluation of each patient in view of clinical and radiological findings.
Klimo et al. [bib_ref] Can surgery improve neurological function in penetrating spinal injury? A review of..., Klimo [/bib_ref] reported that there is no need to perform exploratory surgery for removal of bullet or fragments in order to reduce infection and prevent lead or copper toxicity. The authors suggested laminectomy in patients with cervical injury who demonstrated compressive pathology as well as in patients with incomplete neurological deficit, and instrumentation if the injury deemed unstable. In a recent series involving civilians, Bumpass et al.concluded the surgical indication as post-gunshot spinal infections and persistent CSF leaks. They pointed that injury caused by a small-caliber bullet to the spine infrequently destabilizes the spine, even if all the three columns are impacted. Bono and Heary [bib_ref] Gunshot wounds to the spine, Bono [/bib_ref] recommended the removal of the bullet that was located in the spinal canal between T12-L5, and instrumentation for unstable cases in all levels. Some authors suggested surgery for the cases showing migration [bib_ref] Delayed neurological symptoms from the spontaneous migration of a bullet in the..., Avci [/bib_ref] and toxicity [bib_ref] Lead toxicity and management of gunshot wounds in the lumbar spine, Rentfrow [/bib_ref] or the cases with bullets in the disc space. [bib_ref] Migration of a bullet in the lumbar intervertebral disc space causing back..., Ceylan [/bib_ref] The bullet may show cranial migration in the spinal canal depending on the position of the victim; it may cause a change in neurological status and pain due to mechanical reactions. [bib_ref] Delayed neurological symptoms from the spontaneous migration of a bullet in the..., Avci [/bib_ref] Even though it is rare, the bullet fragment primarily located within an intervertebral disc space may cause plumbism [bib_ref] Lead toxicity and management of gunshot wounds in the lumbar spine, Bellabarba [/bib_ref] and back pain. [bib_ref] Migration of a bullet in the lumbar intervertebral disc space causing back..., Ceylan [/bib_ref] Reactive tissue formation and compression due to the retained bullet can cause progressive neuropathic pain. [bib_ref] Spinal cord stimulation for radicular pain following retained bullet in the spinal..., Keel [/bib_ref] Such studies help in the accumulation of scientific data. Studies reported regarding wartime spinal gunshot injuries are limited, and it seems there is no ethical and scientific way of conducting prospective, double blind clinical studies in the near future. Indeed, it should not to be; humanity must find a way of living without this kind of conflict.
# Conclusion
Bullet settling into the lumbar spinal canal without causing neurological deficit may require surgical intervention. In the presented study, removal of bullets provided not only pain relief in both the cases but also prevented future possible complications such as migration of the bullets, plumbism, and neuropathic pain and instability problems. The problem reported here is caused by one of the most important issues of our world, the war economy. There is a direct relationship between war economy and people with gunshot spine injuries. If we would spend our energy for construction of a better world to live in peace, this paper would not exist.
## Financial support and sponsorship
Nil.
[fig] Figure 1: The bullet lodged at L4-5 disk space in coronal (a) and sagittal (b) CT scan, and postoperative a unilateral L4-5 partial hemilaminactomy defect (black arrow) on the left side (c) after removal of bullet a b c Figure 2: The bullet (arrow head) in spinal canal with L5 corpus fracture (arrow) in axial CT scan (a), the bullets lodged in spinal canal and the congenital lamina defect just medial to the bullet at S1 level and the bullet lodged in psoas muscle on left side in plain anteroposterior radiography (b), L4-S1 posterior stabilization seen in postoperative lateral plain anteroposterior radiography (c) [/fig]
|
Participatory Development and Analysis of a Fuzzy Cognitive Map of the Establishment of a Bio-Based Economy in the Humber Region
Fuzzy Cognitive Mapping (FCM) is a widely used participatory modelling methodology in which stakeholders collaboratively develop a 'cognitive map' (a weighted, directed graph), representing the perceived causal structure of their system. This can be directly transformed by a workshop facilitator into simple mathematical models to be interrogated by participants by the end of the session. Such simple models provide thinking tools which can be used for discussion and exploration of complex issues, as well as sense checking the implications of suggested causal links. They increase stakeholder motivation and understanding of whole systems approaches, but cannot be separated from an intersubjective participatory context. Standard FCM methodologies make simplifying assumptions, which may strongly influence results, presenting particular challenges and opportunities. We report on a participatory process, involving local companies and organisations, focussing on the development of a bio-based economy in the Humber region. The initial cognitive map generated consisted of factors considered key for the development of the regional bio-based economy and their directional, weighted, causal interconnections. A verification and scenario generation procedure, to check the structure of the map and suggest modifications, was carried out with a second session. Participants agreed on updates to the original map and described two alternate potential causal structures. In a novel analysis all map structures were tested using two standard methodologies usually used independently: linear and sigmoidal FCMs, demonstrating some significantly different results alongside some broad similarities. We suggest a development of FCM methodology involving a sensitivity analysis with different mappings and discuss the use of this technique in the context of our case study. Using the results and analysis of our process, we discuss the limitations and benefits of the FCM methodology in this case and in general. We conclude by proposing an extended FCM methodology, including multiple functional mappings within one participant-constructed graph.
# Introduction
## Bio-based economy in the humber region
The Humber region surrounds the tidal estuary of the UK's largest river system. It is a large active industrial area comprising a diverse set of industries ranging from the UK's highest concentration of food processing industries to oil refining and chemical and bio-chemical production facilities. The port of Immingham is the UK's largest by tonnage and, along with the other Humber ports of Grimsby, Goole and Hull, forms one of the largest and busiest port complexes in Europe. The estuary provides infrastructure for 20% of national gas landing and 27% of UK oil refining capacity. The wider region is a net energy exporter and, due to the large number of coal-fired power stations and heavy industrial facilities such as steel making and cement production, the source of 27% of total UK CO 2 emissions emanating from industries subject to Integrated Pollution, Prevention and Control regulations(This figure is based on a recalculation of 2008 Environment Agency IPPC data available from the reference and also available, on request, from the UK Environment Agency. Total commercial and industrial CO 2 emissions in Yorkshire and the Humber, including IPPC and non-IPPC registered companies, was approximately 27 million tonnes (in 2007) (11% of total UK industrial CO 2 emissions). These figures are derived from DECC 2009 data also presented in the reference).
The estuary is of national and international biodiversity and conservation importance and due to climate change presents increasing flood risk management issues, both of which issues can cause friction over proposed development. Neighbouring communities face significant socio-economic problems including unemployment and fuel poverty. Development of the region is affected by, and affects, linked biophysical, industrial, economic, social and governance systems, populated by many diverse actors. The region faces significant new challenges and opportunities with transition to a low carbon economy and national energy security as current key and potentially controversial policy issues. It is one of the UK's most important energy hubs, with strategic energy generation facilities and infrastructure, significant potential for carbon capture and storage and new investment in large-scale renewable energy technologies from offshore-wind to biofuels. The development of a bio-based economy has been recognised as a key opportunity for regional economic growth by regional industrial fora. This is due to both the presence of required infrastructure and support industries and also availability of feedstock from the substantial agricultural hinterland and bulk imports via the port. Numerous biodiesel and bioethanol facilities already exist or are under construction and the region expects to become the centre of an emerging UK biofuel industry responsible for 50% of UK production within the next five years. Significant investment is also underway in energy from biomass and biowaste alongside developments in biorefinery for high value chemicals.
As this sector emerges, managing interactions between policy, society, technology and economics within the system will be central in addressing the balance between economic development, efficient use of resources, reduction in environmental impacts and job creation on a regional and national scale. Hence decision making about the region and its possible future scenarios will have impacts on sustainability goals locally, nationally and globally. In this context we are using the development of a bio-based economy in the region as a case study to address the factors characterizing current key and potentially controversial, policy issues. By understanding the inter-relations between these factors and their consequent development we aim to provide decision support tools for the region to facilitate effective management of this transition. Although data obtained from academic and public sources will be invaluable in developing such an understanding, in this rapidly changing and highly regionally-specific context the input of expert stakeholders is vital. One particularly effective way to solicit such input is via participatory modelling; a process in which stakeholders collaborate in model framing and production.
## Participatory modelling
Participatory modelling refers to any number of techniques by which stakeholders in a system of study are actively involved in some aspect of the creation or evaluation of models of that system. It is widely accepted that stakeholders can bring valuable firsthand knowledge (lay perceptions, expertise etc.) to a research process [bib_ref] The Design of Participatory Agent-Based Social Simulations, Ramanath [/bib_ref] [bib_ref] International Rice Research Institute: Los Banos International Rice Research Institute, chapter Companion..., Bousquet [/bib_ref] [bib_ref] Fostering Industrial Symbiosis With Agent-Based Simulation and Participatory Modeling, Batten [/bib_ref] [bib_ref] The joint use of role-playing games and models regarding negotiation processes: characterization..., Barreteau [/bib_ref]. They can have meaningful ideas for selecting and developing a model, can help in collecting and integrating data, and can be involved in the development of scenarios, interpretation of results, and formulation of collective strategies or policy alternatives. On the other hand, engaging stakeholders is timeconsuming, may bring plural perceptions to the research process rather than unambiguous data, may be difficult to manage and might be perceived to be difficult to carry out in research teams that are not interdisciplinary. Despite these potential pitfalls 'participatory modelling, with its various types and clones, has emerged as a powerful tool that can (a) enhance the stakeholders' knowledge and understanding of a system and its dynamics under various conditions, as in collaborative learning, and (b) identify and clarify the impacts of solutions to a given problem, usually related to supporting decision making, policy, regulation or management' [bib_ref] Modelling with stakeholders, Voinov [/bib_ref].
In many social domains, including our case study, data needed to construct a model may commonly be sparse, commercially sensitive or not centrally collected. In such situations engagement with stakeholders can increase the value of a research project by improving access to data and hence the reliability of the simulation emerging from it. Moreover, it may also improve the chances for implementation of a model's results as stakeholders become more personally connected to and interested in the goals of the research. Our aims in running a participatory modelling exercise were thus twofold: to gather information about what a variety of local stakeholders considered to be key in understanding how a particular local industrial system will develop; and to enhance their understanding of, and engagement with, modelling and complexity approaches to their region.
Most participatory modelling techniques require extensive and ongoing engagement with stakeholder groups in order to iteratively frame, produce and refine a model of the system in question [bib_ref] The design of participatory agent-based social simulations, Ramanath [/bib_ref]. For example story and simulation [bib_ref] Scenarios as tools for international environmental assessments, Alcamo [/bib_ref] , or companion modelling and participatory multi agent modelling approaches [bib_ref] International Rice Research Institute: Los Banos International Rice Research Institute, chapter Companion..., Bousquet [/bib_ref] [bib_ref] Role-playing games for opening the black box of multi-agent systems: Method and..., Barreteau [/bib_ref] [bib_ref] Our companion modelling approach, Barreteau [/bib_ref] [bib_ref] Synthesizing experiences: Lessons to be learned from internetmediated simulation games, Asakawa [/bib_ref]. For the most part stakeholders participate in framing and repeatedly evaluating detailed models of particular types produced by expert modellers, rather than being involved in producing models themselves. Other approaches, such as Bayesian belief networks for example [bib_ref] Coupling real time control and socioeconomic issues in participatory river basin planning, Castelletti [/bib_ref] [bib_ref] Bayesian networks and participatory modelling in water resource management, Castelletti [/bib_ref] [bib_ref] Bayesian networks in environmental modelling, Aguilera [/bib_ref] [bib_ref] A review of bayesian networks as a participatory modeling approach in support..., Düspohl [/bib_ref] , allow stakeholders themselves to be fully involved in model construction, but still require an extensive participatory process and good data availability or in depth empirical knowledge for determining conditional probabilities on system variables.
Due to the complex nature of our case study coupled with the scarcity of system data, we required a methodology which could capture qualitative knowledge from a variety of domains, social, economic, political, environmental and engineering. Additionally, given the very limited time that our stakeholders had available and our goals of increasing stakeholder engagement in and awareness of 'whole systems' or complexity approaches we chose to use a methodology in which the stakeholders themselves would be able to construct the model and which could produce preliminary results within the course of a one day workshop. After consideration of all these factors, the specific participatory modelling methodology which we chose to use for our initial approach to the case study was fuzzy cognitive mapping, or FCM.
## Fuzzy cognitive mapping
Fuzzy cognitive mapping was originally developed by Kosko [bib_ref] Fuzzy Cognitive Maps, Kosko [/bib_ref] as an extension of Axelrod's cognitive maps, which were designed to represent social scientific knowledge. FCM has since been widely used for problem solving in situations in which numerous interdependencies are thought to exist between the important components or variables of a system, but quantitative, empirically-tested information about the forms of these interdependencies is unavailable [bib_ref] Knowledge processing with fuzzy cognitive maps, Taber [/bib_ref] [bib_ref] Modeling organizational behaviour with fuzzy cognitive maps, Craiger [/bib_ref] [bib_ref] Automatic construction of FCMs, Schneider [/bib_ref] [bib_ref] Fuzzy cognitive mapping as a tool to define management objectives for complex..., Hobbs [/bib_ref] [bib_ref] A fuzzy cognitive mapping analysis of the impacts of an ecoindustrial park, Fons [/bib_ref] [bib_ref] Participatory modeling and analysis for sustainable forest management: Overview of soft system..., Mendoza [/bib_ref] [bib_ref] Using fuzzy cognitive maps to describe current system dynamics and develop land..., Soler [/bib_ref]. The method aims to encapsulate the qualitative knowledge of expert participants or system stakeholders in order to rapidly construct a simple systems dynamics model of a specified issue. In the context of environmental management, it has been suggested that the technique is particularly useful in four types of situation [bib_ref] Ecological models based on peole's knowledge: a multi-step fuzzy cognitive mapping approach, Ö Zesmi [/bib_ref] : firstly, when behaviour and decisions of stakeholders play an important role in determining the outcome of a system's development; secondly when detailed local knowledge, but not scientific data, is available; thirdly in 'wicked' environmental problems, which are complex and have no 'right' answers; and finally, in problems in which public or stakeholder participation is desirable or required. All of these situations could be said to be true of our problem domain. The model produced via an FCM process is said to be semi-quantitative because the values of factors and the links between them can only be interpreted in relative terms [bib_ref] The potential of Fuzzy Cognitive Maps for semi-quantitative scenario development, with an..., Kok [/bib_ref]. Such a model can be used for projection or scenario testing purposes and to facilitate further discussion and interaction within or with a stakeholder group. The process of model construction consists of several stages. Firstly the generation and selection by stakeholders of key concepts or factors which are important influences on, or parts of, the system of interest. Importantly, factors can be from any domain (social, economic, physical etc.) and may be qualitative or quantifiable. Secondly, discussion of, and decisions on, what the causal influences, or links, between factors are and whether they are positive or negative (that is, does an increase in one factor cause an increase or a decrease in a second factor to which it is causally connected). This allows the construction of a directed graph. Finally participants rank and verbally describe the strengths of these influences between factors, ultimately producing a directed graph with weighted links which we refer to as the cognitive map or FCM. This graph is then used as the basis for a simple model which is iterated forward to infer the possible, logical outcome of the system interconnections that participants have described, as well as the outcomes if links or their strengths are modified to represent alternative scenarios. FCMs may be generated collaboratively by a group of stakeholders at a workshop [bib_ref] The potential of Fuzzy Cognitive Maps for semi-quantitative scenario development, with an..., Kok [/bib_ref] [bib_ref] Submitted) Fuzzy cognitive maps for futures studies -a methodological assessment of concepts..., Jetter [/bib_ref] , or elicited from individuals via questionnaires or interviews [bib_ref] Ecological models based on peole's knowledge: a multi-step fuzzy cognitive mapping approach, Ö Zesmi [/bib_ref] [bib_ref] Mapping public participation in the water framework directive: A case study of..., Mouratiadou [/bib_ref]. Disparate maps of the same system from different sources can be combined and normalized [bib_ref] Mapping public participation in the water framework directive: A case study of..., Mouratiadou [/bib_ref] [bib_ref] Application of fuzzy cognitive maps to factors affecting slurry rheology, Banini [/bib_ref]. Alternatively, conflicting structures resulting from different expert opinions, or different suggested policy interventions in the system, can be investigated as alternative scenarios [bib_ref] Submitted) Fuzzy cognitive maps for futures studies -a methodological assessment of concepts..., Jetter [/bib_ref].
It is clear that any graph that stakeholders produce, collaboratively or singly, will be a representation of their own opinions and expertise about their system. The cognitive maps produced must therefore be explicitly understood as representing stakeholders' subjective opinions on the area in question, with consequent potential differences between stakeholders from different domains. Maps may not represent reality, for example stakeholders may be sensitized to current controversial factors or infrequent, but high impact, factors which have recently occurred, and hence overestimate the number and weight of their connections. The nature of this technique then, produces a potential weakness for quantitative modelling if the goal is to obtain a 'definitive' model via stakeholder interaction. It has significant strengths however, in its ability to engage stakeholders, promote learning and discussion amongst disparate groups, enhance understanding of whole systems approaches and extract a starting point for systems modelling where data on system structure is not available and where important variables are qualitative or hard to quantify [bib_ref] Ecological models based on peole's knowledge: a multi-step fuzzy cognitive mapping approach, Ö Zesmi [/bib_ref] [bib_ref] Mapping public participation in the water framework directive: A case study of..., Mouratiadou [/bib_ref]. Additionally, structural biases in the map or disagreements between experts give important information on stakeholders' opinions, which can give a guide to points of intervention important for more socially effective policy or decision making in areas in which stakeholder involvement is crucial [bib_ref] Submitted) Fuzzy cognitive maps for futures studies -a methodological assessment of concepts..., Jetter [/bib_ref]. Discussions of the causal 'stories' associated with the maps may also provide more subtle information on perceptions of how the system operates which can aid with future model development and engagement. It must also be emphasized that what stakeholders in a system believe about its causal structure, and the effects of that structure, is in fact crucial to the decisions that they make, and hence to the actual structure and function of that system. This is true for any social system, but is particularly important in cases such as these when a stakeholder group includes key decision makers or when stakeholder participation is vital for successful decision or policy implementation. Despite their intersubjective nature therefore, FCMs and other participatory models have the potential to provide thinking tools that may change stakeholder behaviour and have a powerful impact on the system. FCMs can be understood or used in different ways either as models of a system which can be used for decision support or for promoting organisational learning and discussion amongst stakeholders. In the context of much participatory work, the FCM is primarily an organisational learning tool and an aid to engagement. It is as valuable (or more) in its role in making explicit, then clarifying, mental models and provoking discussion amongst stakeholders as it is at providing a 'definitive' model of a given human system. The rapid construction of a simple mathematical model from such a cognitive map still serves an important function however, in making explicit to stakeholders what the consequences of their beliefs about lower level causal structure actually entail for the whole system. The benefit of using a mathematical analysis is to check the internal consistency of stakeholders' cognitive maps of the system. If these maps are incorrect or incomplete, then an exploration using simple mathematical techniques can quickly expose potential inconsistencies with respect to the stakeholders' own system knowledge and allow discussion, learning and clarification and redrawing of the map to more effectively represent their thinking. These models thus constitute an important part of the verification process. Standard methodology for producing a mathematical model from a fuzzy cognitive map is described below.
## Mathematical model of the fcm
From the cognitive map produced at a participatory workshop, one would like to investigate the interaction of all the key concepts and links on a system in a systematic manner. Koskosuggested using models drawn from neural networks as a means to mathematically explore the interactions of the concepts produced by experts. These models successively update each value of a concept using the previous value of the concept plus a sum of all incoming concept values and application of a thresholding function i.e., a step function centred at a half. The output of the model is a steady state from which a ranking of the most important factors in the system can be derived. Various different scenarios can then be tested and investigated on the system quickly and simply allowing participants of a workshop to develop a systems-level understanding of the implications of their mental models of the system in question. Subsequent research (see Hobbs et al. [bib_ref] Fuzzy cognitive mapping as a tool to define management objectives for complex..., Hobbs [/bib_ref] and Mendoza et al. [bib_ref] Participatory modeling and analysis for sustainable forest management: Overview of soft system..., Mendoza [/bib_ref] for a review of FCMs) has focused on choosing different thresholding functions (ramp, sigmoid, etc.), introducing a weighted sum of the incoming concepts, and on learning algorithms for improving the weights used.
In this study we follow [bib_ref] Fuzzy cognitive mapping as a tool to define management objectives for complex..., Hobbs [/bib_ref] [bib_ref] Participatory modeling and analysis for sustainable forest management: Overview of soft system..., Mendoza [/bib_ref] [bib_ref] Using fuzzy cognitive maps to describe current system dynamics and develop land..., Soler [/bib_ref] [bib_ref] A new hybrid method using evolutionary algorithms to train Fuzzy Cognitive Maps, Papageorgiou [/bib_ref] and turn the FCM produced by our workshop participants into a dynamical model
[formula] x nz1~f (Ax n ), x 0 given ,ð1Þ [/formula]
where x n [R m , A[R m|m is the connectivity matrix created by participants, f : R m ?V5R m is the thresholding function, and n is the discrete time step. The state vector x n contains real values for all the key factors identified by participants. The weighted connectivity matrix A is formed by placing a value a in A ij for every link from state j to state i. The value of a depends on the strength of the link and conventionally lies between 21 and 1. Although in principle there is no specific necessity for this restriction we choose to follow convention. In this paper we initially use a linear function; f (Ax)~Ax following the methodology described by [bib_ref] Using fuzzy cognitive maps to describe current system dynamics and develop land..., Soler [/bib_ref] [bib_ref] The potential of Fuzzy Cognitive Maps for semi-quantitative scenario development, with an..., Kok [/bib_ref] [bib_ref] Using fuzzy cognitive maps as a system model for failure modes and..., Peláez [/bib_ref]. The values of the states are usually interpreted in three ways; active/inactive, good/bad or important/not important [bib_ref] Participatory modeling and analysis for sustainable forest management: Overview of soft system..., Mendoza [/bib_ref] , and when interpreting the results of (1), all three ways are used interchangeably given the 'Fuzzy' nature of the modelling. Hobbs et al. [bib_ref] Fuzzy cognitive mapping as a tool to define management objectives for complex..., Hobbs [/bib_ref] suggest that for long-term policy decisions, the initial transient temporal dynamics are not of interest. This means that in most cases, one is interested in stable fixed points of (1) i.e., x ?~A x ? and x n ?x ? as n??. These stable fixed points allow one to rank the importance of the factors and establish dependences. This is useful information in evaluating and feeding back the model to the participants.
In the course of our participatory work in the Humber region, we ran both an FCM construction workshop and a follow-up verification exercise. For ease of explanation of the process on the day of the first workshop we made use of a linear mapping following [bib_ref] Using fuzzy cognitive maps to describe current system dynamics and develop land..., Soler [/bib_ref] [bib_ref] The potential of Fuzzy Cognitive Maps for semi-quantitative scenario development, with an..., Kok [/bib_ref] [bib_ref] Using fuzzy cognitive maps as a system model for failure modes and..., Peláez [/bib_ref]. For model production from the modified cognitive maps generated from our verification exercise we compared the results using both linear and sigmoidal FCMs, due to certain problematic properties of linear maps (See the Section on Comparison of Results of Linear and Sigmoidal FCMs). Using both of these functions gave us the opportunity to consider the sensitivity of our results to the functions used and to use this information to strengthen the verification process.
# Methods
## Humber region fcm workshop
During a one year period of engagement with regional industrial and political stakeholders, we identified potential study participants through a process of snowballing [bib_ref] Rational Analysis: The Study of Social Organizations with Survey Methods, Coleman [/bib_ref]. After in depth interviews with eighteen of these stakeholders, we decided to focus the workshop on the drivers and barriers surrounding the replacement of fossil fuel products by bio-based alternatives in the domains of energy, chemicals and food. The study involved data collection from primary sources (stakeholder interviews), but was considered to be exempt from the need for ethical clearance by the rules of the University of Surrey Ethics Committee for the following reasons: No deception was used in the research design; participants were not considered to be vulnerable; questions could not be deemed as sensitive or potentially offensive; there was no risk to volunteers' health or wellbeing; no payments or benefits in kind were given to participants and issues of confidentiality and anonymity were guaranteed. Participants were invited to interview with an email setting out the scope and aims of the exercise. Interviews were then carried out and recorded with the explicit, recorded, verbal consent of participants for the purposes of scoping the participatory research project and providing general information for model building. This is standard practice in this kind of participatory work, in which no ethical issues have been identified and which is carried out as a collaborative activity with stakeholders. An open, written, invitation was made to our contact lists and regional industrial fora and environmental managers groups. Eleven participants attended representing industry, local authorities and nongovernmental organisations. During a day-long facilitated workshop participants produced a 'cognitive map' of interrelations between important factors in the development of a bio-based economy in the Humber region.
The workshop followed a standard form as follows:
1. Identifying Important Factors: Participants were first asked to make a list of factors (physical, political, social, and/or economic) that they considered important in the development of a bio-based economy in the Humber region. 2. Grouping the Factors: The factors identified were then grouped in relation to themes and system levels to consolidate the group's ideas. The stakeholders then discussed the factors that arose, agreeing on 16 'dominant' factors to focus on. 3. Linking the Factors: Participants discussed and decided on connections between the factors (relationships or edges) and directions of those connections (positive or negative influences). 4. Weighting the Links: Participants then ranked and defined the relative strengths of these interrelations according to a fixed scale (weak, medium or strong). 5. Creating the Model: The weighted graph produced was represented as an adjacency matrix which was used to update a vector of factor 'values', thus allowing a simple linear model to be rapidly produced and demonstrated on the day. Participants were then able to view and interrogate the model dynamically to evaluate different scenarios (see Section on Development of a Linear Model).
# Results
## Cognitive map of the humber bio-based energy system
The Map developed during the workshop consisted of 16 factors considered key for the development of the regional bio-based economy (See [fig_ref] Table 1: Key factors generated by participants [/fig_ref]. The proportion of energy produced from bio-based as opposed to fossil sources was selected as a focal issue around which to construct the map and the directional, weighted, causal interconnections between factors were added starting from this point.
The cognitive map produced is illustrated in [fig_ref] Figure 1: Humber region bio-based economy FCM from first workshop [/fig_ref]. Several notable features are visible on a first examination. International instability (vs. UK stability) and fossil fuel price were identified as key external drivers of the regional system, a driver being defined as a factor with outgoing links only (These are denoted by selfreinforcing links). The map constructed consists of 3 relatively separate parts connected through bio-based energy production:
1. International instability and associated national funding and regulations that drives bio-based energy production 2. Competitiveness, oil price, and technology that drives biobased energy production 3. Ecological sustainability, community acceptance, and infrastructure that reacts to changes in bio-based energy production but does not drive the system.
## Development of a linear model
Although the cognitive map is a useful starting point for discussion, we can interrogate the structure produced and the interaction of the key factors more effectively by constructing a dynamical model of the system. Following [bib_ref] Fuzzy cognitive mapping as a tool to define management objectives for complex..., Hobbs [/bib_ref] [bib_ref] Participatory modeling and analysis for sustainable forest management: Overview of soft system..., Mendoza [/bib_ref] [bib_ref] Using fuzzy cognitive maps to describe current system dynamics and develop land..., Soler [/bib_ref] [bib_ref] A new hybrid method using evolutionary algorithms to train Fuzzy Cognitive Maps, Papageorgiou [/bib_ref] we turn the FCM into a dynamical linear model as described above in the Section on Mathematical Model of the FCM. The particular modifications used in our process are described below.
Within the weighted connectivity matrix A we have chosen a to be a slightly modified version of that found in [bib_ref] Using fuzzy cognitive maps to describe current system dynamics and develop land..., Soler [/bib_ref] a~À 0:7 À0:5 À0:2
[formula] 0 z0:2 z0:5 if the link is strong fnegative, if the link is medium fnegative, if the link is weak fnegative, if there is no link, if the link is weak fpositive, [/formula]
if the link is medium fpositive, z0:7 if the link is strong fpositive:
[formula] 8 > > > > > > > > > > > < > > > > > > > > > > > :ð2Þ [/formula]
The modification that we make compensates for an ambiguity in the modelling process. Where a (positive) link is not simply strong or medium but is instead described as strong/medium we assign a~z0:6; for instance the link between 'Number of Jobs' and 'Community Acceptance'. Similarly a weak/medium link is given a weight of a~z0:3 and a medium/weak link a strength of a~z0:4.
One amendment also needs to be made to the weighted adjacency matrix before we use it in the linear FCM: x nz1~A x n . This change concerns the drivers of the system. As drivers (by definition) have no edges going into them, they will immediately be killed off (have value zero) under the first iteration of the linear FCM. To prevent this happening we provide the drivers with a self-reinforcing edge, a loop of strength one. So, for instance, concept 16 (fossil fuel prices) is a driver, so we set A 16,16~1 . Having made this change, all that is required to run the linear FCM is an initial condition, the iterative map can then be run to a fixed point. If we stipulate that the drivers initially have the same value, then the choice of initial condition does not effect the ordering of the concepts at the fixed point. So we initially set the value of all drivers to one and the value of all other concepts to zero.
The time series output of the linear model produced from the original FCM graph was simulated and shown to the participants at the end of the workshop in order to provide a preliminary visual result. Participants were also shown output from models produced from the original graph with the effect of the two drivers reversed (that is, the signs of their outgoing links reversed within the adjacency matrix) both singly and simultaneously. All graphs were found to produce an output with a stable fixed point. [fig_ref] Figure 2: Output of the linear model of the FCM from the first workshop [/fig_ref] shows the output of each of these cases. As discussed above, although the absolute values of factors are not meaningful, we can gain an understanding of the consequences of the factors' interrelations by considering their ranking at the fixed point.
In the original model as produced by the participants, bio-based energy production is seen to be maintained at a high level along with those factors on which it has positive causal influence, jobs, by-products and feedstock availability. Competitiveness is high with a consequent positive impact on bio-based energy production. This high production seems to be at the expense of sustainability, ranked lowest of the factors, existing symbiotic industries and land availability for development. When the effects of fossil fuel price were reversed bio-based energy production still maintained a relatively high, although diminished, ranking, as consequently do jobs, by-products and feedstock availability. Sustainability increases within the system whilst existing symbiotic industry remains low. The lowest ranked factor in this scenario becomes competitiveness, explaining the relatively decreased biobased energy production. Reversing the effects of international instability produces a quite different result. Bio-based energy production, by-products, jobs and feedstock availability decrease much further in ranking as legislation and funding supporting biobased energy production decrease. Sustainability increases in ranking whilst existing symbiotic industry and competitiveness become higher still. If the effects of both drivers are reversed then, in very marked contrast to our original scenario, bio-based energy production is driven down to become the lowest ranked factor. Excepting the drivers, which are maintained at 1, the factor rankings are reversed. By-products, jobs, competitiveness and feedstock availability are thus also low, with sustainability consequently becoming high.
According to this interpretation of the FCM graph, both drivers are required to maintain a high level of bio-based energy production, competitiveness and jobs simultaneously, although at the expense of sustainability. The influence of international instability and its associated group of factors is most crucial however as it contains two strong reinforcers of bio-energy production, favourable legislation and funding as opposed to the single, strong reinforcing factor of competitiveness associated with the fossil fuel-driven group.
In the context of the workshop discussion of these results, participants generally agreed that these would be expected outcomes and that the cognitive map effectively represented their thinking on the bio-based economy in a useful fashion. Due to time constraints on the day no further analysis or scenario exploration was possible at that time. For this reason a verification exercise was carried out as described below.
## Verification and scenario generation
An FCM is at its heart a representation of the opinions of a particular group of stakeholders on the causal structure of their system and as such cannot be separated from its intersubjective context. In many real situations it is impossible to define what the 'right' structure is and different stakeholders may hold different views on this. However, the limited time, specific participants and group dynamics of a workshop may bias the map produced in particular ways. In order to attempt to mitigate these biases and validate the map's structure with a different group of experts, a feedback and verification exercise on the FCM was carried out at the local Environmental Managers Group. The group consists of environmental and technical managers from local heavy industries, as well as representatives of local authorities, network organisations and interest groups. There was a modest degree of overlap between participants in this second group and the original FCM workshop. The aim of this workshop was twofold. Firstly, to confirm that the structure of the map seemed reasonable to other local stakeholders with similar expertise and to determine whether any links had been overlooked in the first workshop. Secondly, to gather information on different potential scenarios for the region's biobased economy which imply distinct, different causal structures and hence distinct differences in subsets of factors and links within the cognitive map.
Feedback, after a presentation of the methodology and results of the original workshop, was obtained from both an unstructured and a structured exercise. Firstly on a diagram of the original map, participants were asked to add additional links that they felt should be present and to delete or alter the weights of links which they felt should not be present or were incorrectly weighted. They were also invited to add additional factors or future factors and their links to the map and to comment on the rationale for changes that they had made. Secondly, they were given a structured questionnaire asking them to comment on the absence of particular links which we felt to be noteworthy based on our understanding of the system and its context. We present results from this first unstructured exercise.
In general the basic structure of the map was approved by participants with no suggestions to remove links, although different participants considered that a wide diversity of further causal connections should be represented in the map. The majority of responses agreed however that a connection should exist between international instability and fossil fuel price, meaning that fossil fuel price could no longer be considered a driver. We thus considered this to be a valid update to the map. Additionally, a large number of participants described and commented extensively on two potential scenarios for land use dynamics. Both included the addition of negative influences on the availability of land for development from policy (via habitat regulations) and from potential flood risk (a new factor). The first scenario also considered the possibility of locally-grown feedstock and consequent competition for land between industrial and agriculture use (Scenario 1), whilst the second scenario considered that feedstock would be imported from outside the system (Scenario 2). These scenarios were explicitly drawn out by stakeholders as possible alternative causal maps for the region under different possible futures and so are worthwhile to compare. The amended graphs illustrating each of these scenarios are illustrated in figures 3, 4 and 5.
## Analysis with linear and sigmoidal fcms
In order to compare the results from our original workshop and aid with the verification process, we again construct dynamical models using the graphs produced by participants to form adjacency matrices. For this second round of model construction however, we decided to address problems that had become apparent in the use of a simple linear map. For our first workshop we followed the methodology and model production procedure described by , which included the use of a linear mapping. This can easily be explained to non-expert participants and the update rule for each factor is just the weighted sum of all its inputs. It does however have certain problematic properties. In particular, it is possible for the value of the factors to become negative. The product of a negative factor and a negative link then needs to be carefully rationalised as it will evidently result in a net positive influence on the factor to which it is connected. Furthermore, the factor values may become large in magnitude and the weights taken in the connectivity matrix (2) may no longer distinguish between strong and weak links.
As mentioned briefly in the Section on Mathematical Model of the FCM there are several different functions that are commonly used in the construction of a mathematical model from the cognitive map (1). One such is a sigmoidal function, which may overcome some of the limitations of a linear mapping.
A sigmoidal mapping is given by using
[formula] f (x)~1 1ze {k(x{h) , [/formula]
for constants k and h in (1). We take k~0:05 and h~0:5 which guarantees that the map (1) has a unique (stable) fixed point (seefor a justification). Two slightly different procedures for implementing a sigmoidal FCM are described in the literature. The first is the same as the linear FCM, in that a 1 is put in the diagonal entry of the adjacency matrix for each driver. The second sets all the diagonal elements of the adjacency matrix to unity (see [bib_ref] A new hybrid method using evolutionary algorithms to train Fuzzy Cognitive Maps, Papageorgiou [/bib_ref] for another example of the same procedure) and hence all the factors become drivers in the model. We chose to use this second methodology since we had not asked the participants in the workshop which factors they consider to be drivers (rather inferred them from their lack of incoming links) and all factors could potentially be maintained by influences external to the cognitive map.
The main advantage that the sigmoidal map possesses over the linear map is that the values of concepts are bounded (between 0 and 1). This means that the values of concepts cannot become negative and also that the effects of strong and weak links can always be distinguished. However the sigmoidal map also has some disadvantages; the main one being that the update rule is not as intuitive and harder to explain to participants (possibly leading to unexpected conclusions). Given the advantages and disadvantages of these two choices for f , we compared the results of the original workshop, the updated map and the two scenarios under both a linear and sigmoidal mapping. This has the added benefit of allowing us to determine the sensitivity of our results to the form of the function f , important given the previously stated aim of using these models to check the internal consistency of stakeholders' cognitive maps.
For the updated map, with international instability and fossil fuel price linked, we used the original network as a basis and add a strong positive connection into the adjacency matrix from international instability to the factor representing fossil fuel price. For the linear FCM, this requires that we remove fossil fuel prices as a driver (removing its self-reinforcing link) as it now has incoming connections. Consequently we then use the initial condition of fossil fuel price zero, rather than the previous fossil fuel price of 1. Scenarios 1 and 2, locally and non-locally produced feedstock with habitat regulations and flood risk (which we will refer to as local and non-local), were combined one at a time with the updated network including a link from international instability to fossil fuel prices. In adding the scenarios we kept the base network the same and simply added the extra edges and concepts. In neither scenario did this cause international instability to cease to be a driver. However the additions created two new drivers in each scenario, Flood Risk and Policy -Habitat Regulations. As a result the initial conditions used were one for the concepts
## Comparison of results of linear and sigmoidal fcms
When comparing the output of the linear and sigmoidal mappings applied to the same graph, it quickly becomes apparent that the functional form of the mapping may make a large difference to the results. [fig_ref] Figure 6: Results of adding in a link from international instability to fossil fuel... [/fig_ref] shows the fixed points of the dynamical models (linear or sigmoidal) created from the original map from the first workshop compared with those produced from the updated map [fig_ref] Figure 3: Modified FCM from the Humber Environmental Managers' Meeting showing the addition of... [/fig_ref] , in which international instability influences fossil fuel price. Using a linear mapping we can see that the ranking of the factors at the fixed point is changed only minimally, with principal factors of interest such as the level of biobased energy production and ecological sustainability remaining unchanged in ranking, and jobs and competitiveness reversed in order with each other. Neither is significant change on adding the additional link apparent using a sigmoidal FCM. Adding the link from international instability to fossil fuel price leaves the ranking of the majority of factors, including bio-based energy production, competitiveness, ecological sustainability and jobs, unchanged. This suggests that the link between the two drivers has a minimal impact on the outcome for the system as a whole. However, the overall ranking of factors is changed by using a sigmoidal rather than a linear mapping. Although bio-based energy production remains highest ranked and ecological sustainability remains low (although not at the same rank) in both mappings, suggesting that these results do not depend on the form of f , other factors undergo significant changes in rank. For example, finance rises from 13th in the linear mapping to 3rd under a sigmoidal mapping and land availability rises from 14th to 10th. These similarities and differences suggest the possibility of using comparison between the mappings as a form of sensitivity analysis. We shall expand on this below after presentation of our initial findings. [fig_ref] Figure 7: Results of phasing in the extra links and factors associated with the... [/fig_ref] compares the local and non-local land use scenarios (as shown in [fig_ref] Figure 4: Modified FCM from the Humber Environmental Managers' Meeting showing the locally grown... [/fig_ref] analysed with both a linear and sigmoidal mapping. Rankings of factors are shown at the fixed point as before, but the specific sets of links and weights suggested by stakeholders for each scenario are gradually phased in and new fixed points calculated. 'Confidence' refers to the value of a multiplier on those new links from zero to one, thus the factor ranks at confidence zero are simply the ranks at the fixed point of the updated map. Ranks at confidence 1 are the relative values of factors at the fixed points of Scenarios 1 and 2, with weights as described by stakeholders.
Phasing in Scenario 1, local land use for feedstock, with a linear FCM changes little from the output of the base map. Land availability for development decreases in rank from 16th to 19th and ecological sustainability increases slightly, but other factors remain largely unchanged. Using a sigmoidal FCM, however, Scenario 1 gives rise to an increase in rank of feedstock availability, and significant decreases in land availability for both development and feedstock with consequent small increases in rank for ecological sustainability and existing symbiotic industry (defined in the workshop as existing industry supporting or connecting to bio-based industry). Land availability for feedstock maintains a lower rank than land availability for development as the confidence is increased.
Phasing in Scenario 2, non-local land use for feedstock, affects the ranking of more factors in the linear case than does Scenario 1. With a linear mapping we again see a decrease in land availability for development to the lowest ranking and a small increase in ecological sustainability. Presumably in both scenarios this change is caused by the new influence of habitat regulations. We also see a decrease in the ranking of feedstock availability and a slight decrease in ranking of jobs. It is notable that the absolute values at the fixed point of numerous factors are decreased by phasing in the new scenario, but without changing their ranking relative to others. Conversely, using a sigmoidal mapping, we see an increase in the availability of feedstock and again a decrease in land availability for development and feedstock. This is a similar result to the local scenario, however land availability for feedstock decreases at a lower rate, meaning that it is equally ranked with land availability for development when confidence is one.
It is noteworthy that Scenario 1 shows signs of the possible unexpected effects of a linear mapping discussed in the Section on Analysis with Linear and Sigmoidal FCMs. The change from decreasing to increasing value of some factors with increasing confidence can be traced back to the existence of a negative link between land availability for development and land availability for feedstock. As the value of land for development is driven negative by its strong negative link from habitat regulations, it will begin to have a positive influence on land availability for development and its connected factors. In Scenario 2, the two land availability factors, as well as existing symbiotic industry and sustainability, also have negative values at the fixed points. However in the map associated with this scenario, none of these negative factors have outgoing negative links. Such reversals in direction of influence as factors become negative, although potentially possible, certainly requires careful justification as it may radically change model output.
When we compare all four cases, two scenarios under two different functional mappings, it is clear that the functional form of the mapping may make an equal or even greater difference to the results than the scenarios themselves. For example feedstock availability increases under both scenarios with a sigmoidal mapping, but decreases or remains unchanged under a linear mapping. Consequently, if we consider possible interpretations of these results considering the mapping types one at a time we might draw quite different conclusions. With a linear mapping we might conclude that additional pressures on land availability from habitat regulations, flood risk and competition between land availability for development and feedstock make little difference to factors which concern us in the system as a whole (with the exception of land availability for development, which is significantly decreased by habitat regulations for only a small gain in ecological sustainability). Feedstock availability decreases only slightly in this scenario with no change in ranking for bio-based energy production, competitiveness or jobs. In the non-local scenario, land availability for development is again decreased by the impact of habitat regulations with a consequent increase in sustainability, but feedstock availability is significantly decreased. This is caused by the compound effects of new weak, negative links between the driver international instability and both land available for growing feedstock and feedstock availability, in a situation in which feedstock is mostly imported. Overall this does not lead to a decrease in rank for bio-based energy production however, although stronger links between feedstock availability and international instability might do so if they were present. The factor representing jobs decreases slightly to exchange its ranking with competitiveness, but both remain high. In the linear case then, we might conclude that bio-based energy production, jobs and competitiveness remain high whether feedstock is imported or locally-sourced. And that land availability for development is decreased whilst ecological sustainability is increased by the imposition of habitat regulations.
If we compare the two scenarios using the sigmoidal mapping we might conclude that again, key system indicators such as biobased energy production, competitiveness and jobs are unaffected by whether feedstock is sourced locally or imported. We would conclude that, as might be expected, land availability for feedstock and development both decrease as the local land use scenario is phased in, as the land use types are now in direct competition with each other. With a sigmoidal mapping however we would conclude that both the scenarios of local and non-local feedstock production would lead to an overall increase in feedstock availability. This notable difference in results emphasises the value of comparing these two different functional forms. We can clearly see that the conclusion of whether feedstock availability increases or decreases is sensitive to the function f , meaning that we need to look at these conclusions in more detail.
Such points of disagreement between the conclusions drawn from different functions may be used as a basis for further discussion and investigation. However, despite some large differences, there are certain similarities which are preserved in the results from individual maps treated with the two different functions. If we consider the application of the two functions as a form of sensitivity analysis we can have greater confidence in the results indicated by these similarities and hence draw preliminary conclusions. For example, in the analysis of the original map from the first workshop, there are five concepts which are in the top seven in both of the rankings (that from the linear FCM, and that from the sigmoidal FCM) and four which are consistently in the bottom seven. We say that these are the five most important concepts, and the four least important concepts respectively, to the development of a bio-based economy. The five factors in the top seven are bio-based energy production, by-products, feedstock availability, competitiveness and jobs. Similarly from our analysis, four of the least important of the stakeholders' key concepts are ecological sustainability, existing symbiotic industries, land availability and knowledge. Community acceptance retains a moderate importance under both mappings. The same analysis performed on the updated map with the link between the two drivers would again suggest that bio-based energy production, feedstock availability, by products, competitiveness and jobs are the five most important factors and that knowledge, land availability, existing symbiotic industry and ecological sustainability are the least important factors. Community acceptance also remains relatively unchanged in position as a factor of moderate importance. Repeating this process for our two feedstock supply scenarios gives similar results: in the locally supplied feedstock scenario comparing the analysis of linear and sigmoidal FCMs suggests that bio-based energy production, competitiveness, byproducts, feedstock availability and jobs are the five most important factors and that land availabilities for development and feedstock production, existing symbiotic industry, sustainability and knowledge are the five least important factors with community acceptance again relatively stable in a moderate position. In the non-local feedstock supply scenario, we have fewer certainties regarding important factors with only bio-based energy production, competitiveness and feedstock availability ranked as in the top eight under both mappings. However, the least important factors are more certain and remain unchanged as knowledge, ecological sustainability, existing symbiotic industry and land availabilities for feedstock production and development. None of the factors of particular importance to the bio-based economy seem particularly surprising. However, it is extremely interesting to note that stakeholders' own models show the various forms of land availability to be relatively unimportant as this is a highly controversial issue in the region. Similarly promoting knowledge and training are thought to be particularly important to successful development of the area. Future participatory interrogation and possible re-interpretation of these results in the context of stakeholders' own mental models of the system may either indicate a deficiency in the model or in stakeholders' own perceptions of the system.
# Discussion
The models presented in this paper represent a first attempt in ongoing efforts to understand the bio-based economy in the Humber region and in this respect, despite their subjective nature, the cognitive maps produced by the group of expert stakeholders are in fact a highly useful output, both for us and the stakeholders (as evidenced by personal communication). The selection and verification of key factors, and the structure of their interactions, by a diversity of stakeholders provides a solid basis for further modelling work. Stakeholders identified a large number of factors supporting the development of a bio-based economy, with strong influences from policy, funding and fossil fuel price (via competitiveness). Connections which were perhaps more unexpected were also emphasised, for example from international instability to positive political and financial support for bio-based industries. Also perhaps unexpectedly, issues concerning ecological sustainability were considered to have only weak interactions with the rest of the system. Considerable variation of opinion amongst stakeholders, with regards to issues around feedstock availability and land use, was also revealed. Although in the initial workshop, discussion was eventually resolved with bio-based energy production driving feedstock availability without any negative influences, this was strongly questioned in the verification exercise and indeed in personal communication from stakeholders after the first workshop. The two different land use and feedstock source scenarios which were produced from the verification exercise demonstrate that a wide variety of opinion exists on the subject. Time limitations within the workshop meant that other key components of the bio-based economy, such as chemical and food production, were not considered. The potential for competition between these sectors for feedstock was thus not taken into consideration in model construction. It was however mentioned during our verification exercise as having an uncertain, but potentially important, impact on the evolution of the system.
The construction of the FCM and production of a dynamic output in the context of the initial participatory modelling workshop and subsequent verification exercise has additionally provided a means for our stakeholders to experience 'systems modelling' concepts and has increased project engagement. Feedback from participants has confirmed the usefulness of the process as a thinking tool for those involved. Despite this however, large discrepancies in results between the different mappings and different implementation procedures commonly used to create model output make interpretation of these simple models difficult. We compared the results from two different mappings, both of which have advantages and disadvantages (and neither of which can be considered as 'correct' due to the subjective and incomplete information used to create the model). The linear FCM is easy to explain to non-expert participants and rapid production of results and analysis can be carried out in a workshop context. However, the effect of factors becoming negative can profoundly change the model output and could be difficult to justify in many cases. It is likely that participants assumed that factors such as jobs or availability of land would be non-negative and would not have considered the implications of this possibility when constructing the interaction structure of the FCM graph. More qualitative factors such as, for example, community acceptance or ecological sustainability, or indeed factors such as price, could plausibly be modelled as either negative or positive. The effects of such factors could also justifiably be symmetric around zero. For example, if community acceptance of bio-based industry were to have a negative impact on habitat regulations, then we could justifiably expect community dis-acceptance (a negative factor value) to have a positive influence on the amount of habitat regulation. In the linear case, concepts are also unconstrained in magnitude. This could certainly be considered plausible when considering factors such as ecological sustainability or price, for which it might be difficult to assign particular upper or lower limits. Evidently the use of a linear mapping needs careful justification on a case by case basis.
The sigmoidal FCM does not have the same drawbacks as a linear mapping, as factor values are constrained to the unit interval and may correspond more closely to a functional response that participants might describe in some circumstances. However, since the sigmoidal FCM is nonlinear the analysis is significantly more difficult. It also requires additional parameters k and h which arguably should be fixed by the participants. Only a limited amount of work has been done on comparing the use of different functional mappings in an FCM context. Tsadiras [bib_ref] Comparing the inference capabilities of binary, trivalent and sigmoid fuzzy cognitive maps, Tsadiras [/bib_ref] discussed the appropriateness of binary, trivalent (in which factors can only take values of 0 or z1 and {1, 0 or z1 respectively) and sigmoidal FCMs for different situations. He concluded that binary and trivalent functions were useful in highly qualitative situations, whereas sigmoidal FCMs could be useful in both qualitative and quantitative problems and for strategic planning. No comparison was made between a sigmoidal FCM and any other continuous mapping. McNeildiscusses a wide range of 'squashing functions' (that is functions which constrain the factor values to between 0 and 1) for use in FCMs and suggests different verbal labels which might be used to describe their effect. However he makes no mathematical comparison between either the functions or their use in model construction. A more extensive comparison of the implications of different functions on model output could certainly aid in the choice of function for participatory FCM construction and interpretation.
We have discussed the primary usefulness of FCM's as representations of stakeholders' beliefs and knowledge about a given system which have the additional strength of allowing a testing of the internal consistency of these beliefs. The production of a model and generation of output from a cognitive map makes explicit the consequences, both direct and indirect, of a given system structure and thus allows discussion, learning and reevaluation of how this system may actually function. As such, it is vital that the conversion into a mathematical model represents the causal connections that stakeholders propose well enough to be able to perform these activities usefully. Yet, as we have shown, different commonly used FCM methodologies may have a larger impact on model output than changes in the structure of the cognitive maps themselves. As a first step towards overcoming this issue, a comparison of the output of the model generated using differing functional mappings could provide a form of sensitivity analysis as demonstrated above. Factors of interest which retain their approximate relative positions under different mappings could be considered as robust model output. Such an analysis is certainly useful in the 'offline' analysis of models outside a workshop context. It could also potentially be performed in a workshop setting with appropriate design of the feedback, although it might render understanding the output more difficult and hence discussion less productive.
When producing an FCM we must walk a fine line between keeping the map and model construction simple and understandable, yet producing output robust enough that participants in a workshop can usefully interrogate it and compare it to their own ideas about system function. The analysis of our case study and interaction with the stakeholders involved has suggested a possible methodological improvement which could meet both criteria. Given the differing nature of the factors within an FCM, it seems likely that for different factors and links within one network, different functional responses might more accurately represent the particular interactions. The process of constructing a linear FCM essentially forces participants to fit their system knowledge to a linearised version of reality. However, both during the course of the FCM workshop and via personal communications afterwards, participants suggested the possibility of non-linear mappings such as threshold functions for particular factors. This aspect of their expertise could represent a significant resource to be tapped in the construction of more useful models. Both this opportunity and the issues with standard functional mappings suggest a need to develop new methods which uncover and capture different functional relationships between factors beyond just strength and sign.
We propose an extension of standard FCM methodology, in which participants not only produce the factors and their interconnections, but choose from a set of possible functional mappings between factors for each link. A possible set of useful functions could include not only linear and sigmoidal functions, but tanh-like, step and Gaussian functions. Determining these functional responses in a participatory context would of course be challenging for participants and a significant part of developing this methodology would involve presenting the functions in intuitively understandable ways, as well as creating a tool kit to allow them to be added easily to the map (for example by the use of flashcards). Some of these functions would require additional information from the participants in order to specify them, for example the mid-point of a Gaussian, adding another level of detail. Additionally, to retain the facility for rapid generation of results, the production of dynamical output from the map during the course of the workshop would require the development of an easily-usable interface for the more complex modelling required. Participatory workshops such as these are often performed under time constraints and a two stage process of constructing a cognitive map and then addressing the issue of functional mappings might ensure better output and be more understandable to participants. Both mathematical and facilitative issues must be explored and resolved however, before such an extended methodology could be deployed.
# Conclusions
In summary, the fuzzy cognitive mapping exercise has produced major steps forward in our understanding of the potential development of a bio-based economy in the Humber region and in participants' engagement with modelling and systems, and proved useful in promoting discussion of the issues involved. The work has given us large amounts of data on the significant factors and interrelations which we need to consider in constructing models of the Humber system. It has also revealed potential differing scenarios of land use and feedstock production which should be explored with further work. It has successfully provided what is perhaps the most useful aspect of the methodology, engagement and discussion within a group of disparate stakeholders and their co-construction of a systems-representation of their reality. These benefits should certainly not be underestimated and provide a solid platform for further work in the region, as well as benefits for the stakeholders themselves. Despite this, however, it has also revealed significant issues with the standard methodology used to create dynamic models of the FCM. Our analysis highlighted that different functional mappings commonly used to construct FCM output may give rise to large differences in the output and thus change the interpretation of different scenarios. Linear mappings in particular may give rise to results which affect the system output in ways that require careful justification and may be misleading. Sigmoidal mappings however, may not be appropriate for the interaction of all factors. This limits the usefulness of the approach when attempting to gain stakeholder feedback on model output in the context of a workshop. We suggest that a comparison between the output produced using different functions can act as a useful part of the validation process by highlighting which model outputs are more or less robust to the mapping used. In this case, bio-based energy production, competitiveness, by-products, feedstock availability and jobs were found to be the most important factors in the original and updated maps and in the locally-produced feedstock scenario under both mappings, whereas only bio-based energy production, competitiveness and feedstock availability were reliably important in the non-local feedstock scenario. Land availability (for feedstock production or development), knowledge, existing symbiotic industry and environmental sustainability were robustly found to be the least important factors in all maps, whereas community acceptance consistently retained a moderate ranking. Much work remains to be done to improve the methodology of using FCMs in a participatory context to produce more reliable mappings from stakeholders mental models to system-wide consequences of the interacting effects of the factors and interconnections that they describe. In pursuit of this goal, an expansion of the standard methodology has been suggested in which multiple functional relationships between different factors could be determined by participants and incorporated into a model. There will inevitably be numerous technical issues, both mathematical and in terms of participatory methods, which must be solved in order to develop this new approach. This work is under development in the course of our ongoing engagement with the Humber region and its stakeholder groups.
[fig] Figure 1: Humber region bio-based economy FCM from first workshop. Thickness of the links denotes the strength of the influence. doi:10.1371/journal.pone.0078319.g001 [/fig]
[fig] Figure 2: Output of the linear model of the FCM from the first workshop. a) Output from graph as drawn by participants, b)output showing the effects of reversing the effects of fossil fuel price alone, c) political instability alone and d) both fossil fuel price and political instability. doi:10.1371/journal.pone.0078319.g002 international instability, flood risk and policy -habitat regulations, and zero for all other concepts. Due to the addition of extra factors the maps for Scenarios 1 and 2 now contain 19 rather than 16 factors. [/fig]
[fig] Figure 3: Modified FCM from the Humber Environmental Managers' Meeting showing the addition of a link from international instability to fossil fuel price. Thickness of the links denotes the strength of the influence. doi:10.1371/journal.pone. [/fig]
[fig] Figure 4: Modified FCM from the Humber Environmental Managers' Meeting showing the locally grown feedstock scenario. Thickness of the links denotes the strength of the influence. doi:10.1371/journal.pone.0078319.g004 [/fig]
[fig] Figure 5: Modified FCM from the Humber Environmental Managers' Meeting showing the non-local feedstock scenario. Thickness of the links denotes the strength of the influence. doi:10.1371/journal.pone.0078319.g005 [/fig]
[fig] Figure 6: Results of adding in a link from international instability to fossil fuel price. Figures show ranking of factors at a stable fixed point using a linear map without (a) or with (b) the additional link, or using a sigmoidal map without (c) or with (d) the additional link. doi:10.1371/journal.pone.0078319.g006 [/fig]
[fig] Figure 7: Results of phasing in the extra links and factors associated with the local and non-local land use scenarios. Figures show ranking of factors at a stable fixed point using either a linear or a sigmoidal map as a function of confidence (see text). doi:10.1371/journal.pone.0078319.g007 [/fig]
[table] Table 1: Key factors generated by participants. [/table]
|
Chronotherapy for morning blood pressure surge in hypertensive patients: a systematic review and meta-analysis
Background:The morning blood pressure surge (MBPS) is related to an exaggerated risk of cardiovascular diseases and mortality. With increasing attention on circadian change in blood pressure and extensive use of ambulatory blood pressure monitoring (ABPM), chronotherapy that administration of medication according to biological rhythm, is reported to improve cardiovascular outcomes. The aim of this study is to evaluate the influence of chronotherapy of antihypertensive drugs upon MBPS in hypertensive patients.Methods:A search strategy was applied in Ovid MEDLINE, EMBASE, Cochrane (Wiley) CENTRAL Register of Controlled Trials, Cochrane Database of Systematic Reviews, and the Chinese Biomedical literature database. No language and date restrictions. Randomized controlled trials (RCT) assessing the efficacy of evening and morning administration of the same medications in adult patients with primary hypertension were included.Results: A total of ten trials, comprising 1724 participants with a mean age of 61 and 51% female, were included in this study. Combined analysis observed significant reduction of MBPS (− 5.30 mmHg, 95% CI − 8.80 to − 1.80), night-time SBP (− 2.29 mmHg, 95% CI − 4.43 to − 0.15), night-time DBP (− 1.63 mmHg, 95 %CI − 3.23 to − 0.04) and increase in night blood pressure dipping (3.23%, 95% CI 5.37 to 1.10) in evening dosage compared with traditional morning dosage of blood pressure-lowering drugs. No significant difference was found in the incidence of overall adverse effects (RR 0.65, 95% CI 0.30 to 1.41) and withdrawal due to adverse effects (RR 0.95, 95% CI 0.53 to 1.71).Conclusions:Our study suggested that evening administration of antihypertensive medications exerted better blood pressure-lowering effect on MBPS compared with conventional morning dosage. Safety assessment also indicated that the evening regimen did not increase the risk of adverse events. However, endpoint studies need to be carried out to confirm the significance and feasibility of this treatment regimen in clinical practice.
# Background
The circadian rhythm plays a critical role in multiple neurohormonal processes, thus modulating the cardiovascular system. The circadian changes in blood pressure have received increasing interest. Additionally, with the introduction of ambulatory blood pressure monitoring (ABPM) into regular hypertension management, daynight patterns of blood pressure can be observed. Early morning is the time of the highest incidence of cardiovascular events during the day. Accumulating evidence demonstrated that an exaggerated morning blood pressure surge (MBPS) is closely related to the increased risk of cardiovascular diseases and all-cause mortality. Besides, systematic review also indicated that when using a continuous variable to test correlations, a 10 mm Hg increase in MBPS was related to high risk of stroke. Thus, the control of MBPS is clinically relevant and important. In addition, night blood pressure dipping, an index of blood pressure decline, has also been reported as an important prognostic marker for cardiovascular events and morbidity in hypertensives.
Chronotherapy is the adaption of medication to biological rhythm, to achieve maximal effectiveness by altering the time of drug administration, which could influence the pharmacokinetic properties of antihypertensive medications. Multiple works have evaluated the administration-time effects in hypertension therapy. Hermida et al. revealed that administration of at least one prescribed antihypertensive medications in hypertensives at bedtime, compared to upon wakening, significantly improved blood pressure control (especially bedtime blood pressure) and remarkably diminished the occurrence of major cardiovascular events, such as stroke, myocardial infarction, death, etc.. A Cochrane systemic review showed that bedtime ingestion of antihypertensive drugs was more effective to decrease 24-h blood pressure without additional adverse effects than morning regimen.
Some studies documented that bedtime ingestion of blood pressure-lowering agents had a more efficient antihypertensive effect during night-time and at the early morning period. Most studies working on the administration time of antihypertensive medications were small-scaled and single-centered. And the bedtime administration has not been recommended in guidelines. Whether the bedtime administration of antihypertensive medications has a more significant effect on lowering MBPS than conventional morning administration has not been reported in any systematic review and metaanalysis. Therefore, this systematic review was conducted to investigate the efficacy of chronotherapy of once-daily antihypertensive drugs on reducing MBPS, systolic and diastolic blood pressure, and adverse effects in hypertensive patients, thus providing more evidence to the bedtime regimen.
# Methods
The protocol for this systematic review and meta-analysis was registered in PROSPERO (CRD42020180166).
## Search strategy
The searching databases included: Ovid MEDLINE, EMBASE, Cochrane (Wiley) CENTRAL Register of Controlled Trials and Cochrane Database of Systematic Reviews, the Chinese Biomedical literature database up to April 2020. Besides, we also tracked completed clinical trials that met our inclusion criteria on ClinicalTrials.gov.
The major search terms included: "hypertension", ''morning surge", "morning blood pressure surge", "MBPS", "night* decline", "Chronotherapy", "morning OR day OR am OR diurnal OR daytime OR awake", "evening OR bedtime OR night OR nocturnal OR pm" (shown in Additional file 1:. In addition to databases, studies in references lists of relevant articles that met our criteria were also hand-searched and screened as a significant supplement. No language and date restrictions were applied to avoid missing any related investigations.
## Inclusion and exclusion criteria
Study selection was performed independently by two reviewers (Z.X. and J.Z.) by viewing the titles and the abstracts of the search strategies. The disagreements were judged by a third reviewer who was blinded to the first two reviewers' decisions. The selection process was repeated twice by each reviewer. Studies meeting the inclusion criteria were checked in detail (full text).
Only randomized control trials (RCT) were included to assess the effects of chronotherapy. Randomized crossover trials that only had two treatment periods (two interventions) were also considered. Adults (more than 18) with primary hypertension, which is defined as systolic and/or diastolic blood pressure levels more than 140/90 mmHg were included. Secondary hypertension, alternating shift workers, and severe cardiac insufficiency (NYHA III-IV) were the exclusion criteria. Studies reporting one-drug therapy or combined therapy (two or more drugs) with the antihypertensive drug(s) administered once daily at bedtime when the control group was the same drug(s) at the same dose(s) once a day upon awakening were included. Antihypertensive drugs comprised diuretics, adrenergic beta-antagonists (β-blockers), alpha-antagonists, calcium channel blockers (CCB),vasodilator agents and renin-angiotensin system inhibitors (RASI, including angiotensin II receptor blockers and angiotensin-converting enzyme inhibitors). Administration time in the evening was from 18:00 to 24:00, while in the morning was from 6:00 to 12:00.
Included studies must clearly define and measure MBPS. Both continuous and categorical variables were included. Definitions of MBPS must belong to any one of the following: (1) the sleep-trough surge, calculated by the mean value of morning blood pressure within 2 h after waking minus the lowest night-time blood pressure;
(2) the prewaking surge, the mean blood pressure within 2 h after waking to subtract the mean blood pressure within 2 h before waking; (3) the rising blood pressure surge, morning blood pressure reading upon rising minus blood pressure reading in the lying position 30 min before rising. The measurement of night dipping is the percentage of reduction in mean nighttime blood pressure relative to mean daytime blood pressure.
## Outcomes
The primary outcome was the change in MBPS from baseline to the end of treatment (or the value of MBPS at the end of treatment if the baseline is comparable) when it was calculated as a continuous variable; or the ratio of patients whose MBPS exceeded the settled threshold after treatment when MBPS was defined as a categorical variable.
The secondary outcomes contained night blood pressure dipping, 24-h mean systolic blood pressure (SBP) and diastolic blood pressure (DBP), daytime mean SBP and DBP (measured by ABPM from the time patients wake up in the morning to the time they fall asleep in the evening or from 6-8:00 to 22-24:00), night-time SBP and DBP (measured by ABMP from the time patients fall asleep in the evening to the time they wake up in the morning or from 22-24:00 to 6-8:00). Safety outcomes include overall adverse effects and withdrawals due to adverse effects during treatment. Adverse effects can be any unplanned and unfavorable symptom, or disease temporally associated with the use of medicine. Withdrawals due to adverse effects are reported as events leading to permanent trial discontinuation.
## Data extraction and management
Data were extracted by two independent reviewers (Z.X. and J.Z.) based on a standard form. Necessary information of studies was extracted, such as patients' demographics (age, sex, region, race, hypertensive status, medication history), study methods (patients' recruitment, randomization, crossover, treatment duration), interventions (drugs, dose, administration time), outcomes (definition and measurement of MBPS, night blood pressure dipping, 24-h mean SBP/DBP, daytime mean SBP/DBP, night-time SBP/DBP) and safety endpoints (headache, nasopharyngitis, edema, bronchitis, pain and withdrawals due to adverse effects). The numeric data (e.g., blood pressure) were collected from text and tables. Data in graphs were not extracted due to a possible measurement error. All the studies were double coded by the two reviewers. If there were missing data, we attempted to contact the authors to provide the missing information. Regarding missing data for the standard deviation of the change in MBPS, imputation was conducted based on other similar trials.
## Risk of bias assessment
Evaluation of the risk of bias in all included trials was executed by two independent reviewers (Z.X. and J.Z.) based on the guide of the Cochrane risk of bias tool.
Characteristics of assessment included random sequence generation, allocation concealment, blinding, incomplete outcome data, selective reporting, and other bias (e.g., conflict of interest). The assessment was done at the study level.
# Data analysis
All data analysis and synthesis in the meta-analysis was conducted by RevMan 5.4 (RevMan 5.4; Cochrane Collaboration, Oxford, UK). Comparisons of continuous MBPS changes, 24-h, daytime and night-time SBP and DBP between groups were presented as mean ± standard deviation (SD) with corresponding 95% confidence intervals. These data were entered using a generic inverse variance. Categorical MBPS, overall adverse effects, and withdrawals were presented as relative risk ratios (RR). The outcomes were combined using a fixed-effect model (I 2 ≤ 25% ) and a random-effect model (I 2 > 25%).
# Subgroup analysis
Subgroup analysis were conducted by classifying the trials into those using different MBPS definitions, and those using antihypertensive drugs from different classes. Subgroup analysis can help locate the sources of heterogeneity and assess the effectiveness of different kinds of medicines.
# Results
# Search results
636 records were detected after database searching, and 4 additional articles were hand-searched in reference lists of published papers. After the removal of duplicates, 569 articles were viewed by titles and abstracts by two independent reviewers (Z.X. and J.Z.), 401 of which were excluded, and the remaining 168 references entered full-text screening for intensive evaluation. Eventually, 10 articles (10 trials) were considered for analysis and involved in our study. The reasons for excluding 158 references are shown in.
## Description of included studies
Included studies were 8 parallel-designed RCTs and 2 crossover-designed RCTwith 1724 participants in 8 different countries. The key characteristics of the trials are listed in. The study size ranged from 31 to 639. The mean age of patients was 61 years, and females accounted for 43-61%. The duration of followup periods differed from studies (ranged from 2 to 48 weeks). Eight studies described ethnicity. One trial was conducted in the USA, with 38 participants (2%), half of which were African American. One trial (37%) was a multi-center study conducted in 94 centers in five countries with nearly 99% white. Six studies. The remaining two trials without reporting race were organized in Europe (14%) and Japan (2%). The definitions and measurements of MBPS were listed in Additional file 1:. Among all the trials, seven studies presented secondary outcomes .
## Risk of bias in the included studies
The overall assessment of the risk of bias in the included studies are shown in Additional 1: . One study, as a crossover RCT, only the first period data were available. One study had the risk of baseline imbalance; six studies provided insufficient information about other bias. Good random sequence generation, concealed allocation and blinding were ascertained in all the studies. And none of the trials had attrition bias.
## Effects of interventions
## Changes in mbps
The most common definition of MBPS (sleep-trough surge) were compared in included studies. The overall analysis showed that the evening administration significantly decreased MBPS by 5.30 mmHg (95% CI − 8.80 to − 1.80). But a significant heterogeneity (I 2 = 98%) was detected. The subgroup analysis observed significant differences in the evening dosage regimen of CCB and RASI compared with morning administration. The evening administration of CCB lowered MBPS by 4.21 mmHg (95% CI − 4.66 to − 3.76), and RASI reduced MBPS by 6.90 mmHg (95% CI − 11.49 to − 2.31). No heterogeneity (I 2 = 0%) was observed in the CCB subgroup, while high heterogeneity (I 2 = 97%) was detected in the RASI subgroup. No differences in blood pressure reduction were observed with β-blockers (P = 0.67) and combined CCB and ARB (P = 0.52) in the evening versus morning administration.
The effects of intervention were also analyzed when MBPS was measured as categorical variable (Additional file 1:. No significance was detected in overall analysis. Due to the low number of trials analyzing MBPS as a categorical variable and different thresholds of MBPS, it might be insufficient to carry out a combined analysis.
## Changes in sbp
The data and results of SBP are presented in. The overall effects of evening administration showed no statistically significant reduction in 24-h SBP (P = 0.27) and daytime SBP (P = 0.78). However, evening regimen significantly decreased night-time SBP by 2.29 mmHg (95% CI − 4.43 to − 0.15), compared with morning administration. The subgroup analysis (Additional file 1: showed that CCB evening regimen significantly reduced 24-h SBP, daytime SBP and night-time SBP by 4.1 mmHg (95% CI − 5.28 to − 2.92, I 2 = 0%), 3.72 mmHg (95% CI − 5.04 to − 2.39, I 2 = 0%) and by 5.37 mmHg (95% CI − 6.92 to − 3.82, I 2 = 0%), respectively. There were no significant differences of SBP reduction were detected in RASI and CCB + ARB combined therapy.
## Changes in dbp
The details of efficacy in DBP are shown in. The analysis of overall effects demonstrated that evening dosage regimen significantly reduced night-time DBP by 1.63 mmHg (P = 0.04) and no significant differences were observed in 24-h DBP (P = 0.31), daytime DBP (P = 0.35). The subgroup analysis (Additional file 1: showed that the evening CCB was associated with significant 24-h DBP and night-time DBP reduction by 3.32 mmHg (95% CI − 5.85 to − 0.78) and 3.81 mmHg (95% CI − 5.45 to − 2.18), respectively; and no significant differences of DBP reduction were found in RASI or CCB + ARB combined therapy subgroups.
## Changes in night bp dipping
6 studies with 3674 hypertensive subjects were included and analyzed. The result indicated that bedtime dosing remarkably enhanced nocturnal blood pressure decline by 3.23% (95% CI 5.37 to 1.10), compared with morning dosage regimen (Additional file 1:. And a high heterogeneity (I 2 = 89%) was detected.
## Adverse effects
The meta-analysis showed that the incidence of overall adverse effects (RR 0.65, 95% CI 0.30 to 1.41, I 2 = 69%) and discontinuations due to adverse effects (RR 0.95, 95% CI 0.53 to 1.71, I 2 = 0%) had no significant differences between evening and morning regimens (Additional file 1:and . Among all the kinds of adverse effects reported, headache was the most frequently reported symptom, and a slightly significant increase of nausea was observed in the morning regimen (P = 0.05).
# Discussion
This systematic review and meta-analysis investigated the effects of evening and morning dosage of antihypertensive medications on MBPS and blood pressure control in patients with essential hypertension.
Morning administration of medication is commonly used in antihypertensive treatment in clinical practice. However, some recent clinical trials reported evening administration of the antihypertensive drug to be efficient in lowering blood pressure, improving organ functions, and preventing cardiovascular events. This study indicated that the evening regimen of antihypertensive drugs exerted better effects on decreasing night-time blood pressure than morning therapy. Since studies showed that night-time blood pressure was significantly better than daytime blood pressure in predicting all-cause, cardiovascular and non-cardiovascular mortality, stroke and cardiovascular events. Besides, this analysis also demonstrated that bedtime administration significantly better enhanced night blood pressure dipping, relative to morning regimen, which was consistent with several previous studies. Increased nocturnal decline drives blood pressure towards more of a dipper pattern, thus improving blood pressure variability and lowering cardiovascular risk. Therefore, chronotherapy targeting night-time blood pressure control may favorably affect cardiovascular morbidity and mortality in hypertensive patients.
This study found that evening administration of blood pressure-lowering medication significantly improved MBPS, which has not been systematically reviewed before. The increase of MBPS was first associated with. A clinical trial with 519 hypertensive participants found that a 10 mm Hg raised in MBPS caused a 25% increased stroke incidence. Several but not all subsequent studies indicated that exaggerated MBPS had been related to increased risk for cardiovascular events and all-cause mortality. However, in recent years, the prognostic role of MBPS is controversial.
One possible reason is the distinct definitions and cut-off points of MBPS. There are four definitions commonly used (the sleep-trough surge, the prewaking surge, the rising blood pressure surge, and the morning-evening difference), and the thresholds are various. The thresholds of MBPS could be defined as top decile, quartiles of MBPS in participants, or determined according to hypertensive guidelines, which have not reached a consensus. Among included studies, two articles analyzed MBPS as categorical variable with different predetermined threshold, 23.58 mmHg and 35 mmHg, respectively. Due to the limited number and different thresholds of studies, the heterogeneity between studies of categorical MBPS is very high. And a previous metaanalysis found that there was no significant association between MBPS and all-cause mortality or cardiovascular events when MBPS was analyzed as categorical variable. Therefore, a single threshold dichotomizing MBPS as normal and exaggerated may not be powerful to define the significance of MBPS. Besides, in patients with nondipper hypertension and nocturnal hypertension, which are described as a sleep-to-awake SBP ratio of less than 10%, and night-time SBP more than 120 mmHg and DBP more than 70 mmHg respectively, the MBPS may be low and not suitable for the evaluation of cardiovascular prognosis. This was also supported by the Jackson Heart Study, that no significant association between MBPS and the incidence of cardiovascular events was found in the black population with non-dipper and nocturnal hypertension. Consequently, more research is needed to analyze MBPS's definitions and reach a consensus on the target value of MBPS control. In patients with nocturnal dipping hypertension, a prognostic indicator needs to be developed to predict cardiovascular events and serve as a target of hypertension management.
On the other hand, ethnicity is a critical factor affecting morning blood pressure surge. The Ohasama Studyand JMS-ABPM Studyfound significant relationship between exaggerated MBPS and stoke incidence in Japanese patients. Conversely, Bombelli et al.reported that high MBPS was not associated with increased mortality and cardiovascular events in a white population. Furthermore, systematic review showed that the degree of sleep-trough surge was higher in Japanese than in European patients with hypertension. The Jackson Heart Study also revealed that there was no clear evidence for the associations of sleep-trough MBPS, prewaking MBPS, and rising blood pressure surge with the incidence of cardiovascular events and all-cause mortality in black adults. Discrepancies in the pathogenesis and manifestations of hypertension and related cardiovascular outcomes have been consistently reported among different races. Asians have more active sympathetic nerve activity during the morning period and higher incidence of stroke, which may account for the high degree and the prognostic role of MBPS. Therefore, therapeutic strategies targeting MBPS control are of considerable significance in Asian population.
This meta-analysis discovered a prominent role of evening CCB in lowering blood pressure. The CCBs in included studies were nifedipine gastrointestinal therapeutic system (GITS) and amlodipine, which both belongs to the dihydropyridine calcium blocker family with long duration of action. Amlodipine has been shown to be more effective for the MBPS control compared with valsartan in the VALUE trial. Besides, previous evidence observed that the pharmacokinetic pharmacodynamics relationship of CCB is circadian rhythm-dependent that a night dosing of amlodipine displayed a longer half-life and higher peak plasma concentration compared to the morning dosage. This may also explain the remarkable effects of CCB in lowering night-time and MBPS when given at bedtime.
# Limitations
This study has some potential limitations. First, the searching results indicated that only ten studies were eligible for combined analysis, and six of them were conducted in China, which reflects the racial differences in the attention to MBPS. The lack of evidence also posed a limitation to the analysis of distinct definitions and thresholds of MBPS and sensitivity analysis according to methodological quality. Besides, the blood pressure-lowering efficacy alone is not adequate to judge the clinical significance of antihypertensive of chronotherapy. Thus long-term, larger scale and multi-racial clinical trials are needed to explore the correlation between MBPS reduction and the incidence of cardiovascular endpoint.
# Conclusions
MBPS has received considerable attention from clinicians. Our meta-analysis provided the evidence that evening administration of antihypertensive medication improves blood pressure variability by significantly reducing MBPS and increasing nocturnal blood pressure decline without increasing adverse effects. |
Elements of Psychological Medicine; an Introduction to the Practical Study of Insanity, Adapted for Students and Junior Practitioners
Medical Officer to the Clifton Hall Retreat, &c
M.DMedical Officer to the Clifton Hall Retreat, &c
181 Review XII.
Medicine, in Manchester, and which he has published at the request of his ^edical brethren, who constituted a large proportion of his auditory. We happen to know that these lectures were listened to with great interest by several of the most discinguislied members of the profession in that city \ aQd that the request for their publication was not an idle compliment, but a genuine expression of their sense of the value of what they had heard, ai*d of their desire that others should benefit by it. To the valuable rec?rmnendation of the book which this fact conveys, we are glad to find our-Selves able, after a careful examination of its pages, to add our own very favourable testimony. Dr. Noble does not put it forth as a systematic tfeatise on the subject, but designs it merely as an introduction, arranged upon a plan that shall take the student up, as it were, at the point where the completion of his lectures and hospital attendance leaves him. He has connected the pathology of mental maladies, as much as possible, with tlie present state of our knowledge of cerebral physiology ; adopting in Nearly every particular, the views enunciated by Dr. Carpenter, in his reviews of "Noble on the Brain," in our twenty-second volume; and ahandoning, with a candour most creditable to him, the greater part of the phrenological doctrines which he formerly upheld, but which his more Matured judgment and larger experience (especially in regard to insanity), r^ake him regard as 110 longer tenable. And he has explained the principles of treatment by constant reference to those more general ones, which are applicable in the ordinary practice ol medicine, and with which, there- This little volume contains Eight Lectures on Insanity, which were delivered by Dr. Noble during last session, at the Chatham-street School of Medicine, in Manchester, and which he has published at the request of his ^edical brethren, who constituted a large proportion of his auditory. We happen to know that these lectures were listened to with great interest by several of the most discinguislied members of the profession in that city \ aQd that the request for their publication was not an idle compliment, but a genuine expression of their sense of the value of what they had heard, ai*d of their desire that others should benefit by it. To the valuable rec?rmnendation of the book which this fact conveys, we are glad to find our-Selves able, after a careful examination of its pages, to add our own very favourable testimony. Dr. Noble does not put it forth as a systematic tfeatise on the subject, but designs it merely as an introduction, arranged upon a plan that shall take the student up, as it were, at the point where the completion of his lectures and hospital attendance leaves him. He has connected the pathology of mental maladies, as much as possible, with tlie present state of our knowledge of cerebral physiology ; adopting in Nearly every particular, the views enunciated by Dr. Carpenter, in his reviews of "Noble on the Brain," in our twenty-second volume; and ahandoning, with a candour most creditable to him, the greater part of the phrenological doctrines which he formerly upheld, but which his more Reviews.
[Jan.
consisting in " chronic disorder of the brain, inducing perversion of ideas prejudicial to, or destructive of, the freedom of the will;" each point ot this definition being more fully discussed in subsequent lectures. Thus, after giving in the second lecture such a sketch of the present state of our knowledge of the functions of the brain and nervous system, as may serve to make his readers comprehend their probable connexion with the healthful activity of the mind, he proceeds in the third lecture to discuss the General Pathology of Insanity ; and under this head he makes good his position as to the essential connexion between this disease and cerebral disorders. His arguments on this point are extremely acute; and we can o:ily regret that he did not fortify his position a little more, by showing how cerebral disorders, that may leave nothing for the morbid anatomist to discover, may be excited either by perverted blood, or by morbid sympathies with other parts of the system. These points, it is true, are not altogether passed over ; but they scarcely receive the prominence to which they are entitled. The fourth and fifth lectures embrace an account ot the Varieties and Particular Characteristics of Insanity: the primary forms of the disease recognised by Dr. Noble being the notional, the intelUgentiol, and the emotional.
In the first there is a primary perversion of ideas, giving rise to settled illusions, of which the mind cannot get rid by any effort of its own. In the second there is a primary perversion of the reasoning faculty, quite irrespective of any fixed or characteristic illusion. The term emotional insanity is not used by Dr. Noble in precisely the sense in which it is employed by many recent writers ; since he refuses to admit into the category of insanity those forms of mental disorder which do not involve some amount of perversion of ideas,?a limitation in which we cannot accord with him,?and applies the term emotional insanity to those cases in which, while the prominent derangement obviously consists m disorder of the emotions, there is at the same time some perversion of ideas.
The sixth lecture treats of the Diagnosis, Prognosis, and Etiology > the seventh lecture gives a general view of the Exciting Causes and Physical Treatment; and in the eighth lecture the principles of the Moral Management of the insane are discussed,?the whole being summed up in the following admirable formula : " Deal with the physical characteristics which may accompany insanity, as y?u would deal with them under other circumstances, and act in correspondence with sound principles of medical practice, always remembering that, with high nervous susceptibility, depletion will be but indifferently tolerated, more especially when the ailment refers itself to causes exclusively psychical. Eor the relief of insanity itself, properly considered, trust almost entirely to hygienic and moral treatment; withdrawing circumstances likely to aggravate the special features of individual cases, and supplying to the mind such objects of attention and excitants to activity [and, we should add, such motives to self-control], as may be best calculated to arouse and sustain a new and more healthful mode of operation."
Holding it, as we do, to be the imperative duty of every practitioner to qualify himself to give a sound diagnosis, and to direct the treatment in any case which may present itself to him, and feeling sure that without some special guidance even those most generally well informed may be led into grievous error, we have great satisfaction in being able warmly to recommend Dr. Noble's little treatise, as extremely well adapted to supply the desideratum which we have long felt and lamented.
Matured judgment and larger experience (especially in regard to insanity), r^ake him regard as 110 longer tenable. And he has explained the principles of treatment by constant reference to those more general ones, which are applicable in the ordinary practice ol medicine, and with which, therefore, the advanced student, as well as the practitioner, ought to be acquainted.In his preface Dr. Noble makes some very pertinent remarks regarding the ignorance which prevails through a large part of the profession, in regard to the diagnosis and treatment of insanity. "Well educated Persons out of the profession," he truly remarks, " will frequently judge as soundly concerning the condition of a patient, as many who are within 1t-" Our hospitals, colleges, and universities make no provision for training the future physician or surgeon in a practical knowledge of insanity, so that men who possess adequate information upon this branch of their duties, are necessarily self-educated to a great extent. And yet any two practitioners, having no special knowledge of the disorder, are empowered hy law to take away the personal freedom of any individual, by certifying to the fact of his insanity ; a fact about which those who have no iamiliarity with the disease are peculiarly liable to be deceived, the most accurate diagnostic powers being often required even by the most experienced.The first lecture contains a general introduction to the study ot insanity, aud embodies Dr. Noble's view of its essential nature. He defines it as |
Physical activity and the prevention, reduction, and treatment of alcohol and/or substance use across the lifespan (The PHASE review): protocol for a systematic review
Background: Alcohol and substance use results in significant human and economic cost globally and is associated with economic costs of £21 billion and £15billion within the UK, respectively, and trends for use are not improving. Pharmacological interventions are well researched, but relapse rates across interventions for substance and alcohol use disorders are as high as 60-90%. Physical activity may offer an alternative or adjunct approach to reducing rates of alcohol and substance use that is associated with few adverse side effects, is easily accessible, and is potentially cost-effective. Through psychological, behavioural, and physiological mechanisms, physical activity may offer benefits in the prevention, reduction, and treatment of alcohol and substance use across the lifespan. Whilst physical activity is widely advocated as offering benefit, no systematic review exists of physical activity (in all forms) and its effects on all levels of alcohol and substance use across all ages to help inform policymakers, service providers, and commissioners. Methods: The objectives of this mixed methods systematic review are to describe and evaluate the quantitative and qualitative research obtained by a diverse search strategy on the impact of physical activity and its potential to:1. Reduce the risk of progression to alcohol and/or substance use (PREVENTION) 2. Support individuals to reduce alcohol and/or substance use for harm reduction (REDUCTION), and 3. Promote abstinence and relapse prevention during and after treatment for an alcohol and/or substance use disorder (TREATMENT).With the input of key stakeholders, we aim to assess how what we know can be translated into policy and practice. Quantitative, qualitative, service evaluations, and economic analyses will be brought together in a final narrative synthesis that will describe the potential benefits of physical activity for whom, in what conditions, and in what form.Discussion: This review will provide details of what is known about physical activity and the prevention, reduction, and treatment of alcohol and/or substance use. The synthesised findings will be disseminated to policymakers, service providers, and commissioners in the UK.
(Continued from previous page) Systematic review registration: PROSPERO number: CRD42017079322.
Keywords: Physical activity, Exercise, Alcohol use, Substance use, Addiction, Prevention, Harm reduction, Treatment, Mixed method Background Rationale Alcohol and substance use is common: globally, 5.9% and 1% of deaths are attributable to alcohol and illicit drug use, respectively. In the UK, alcohol use is attributed to more than one in five deaths of men aged 16-54 years old , and alcohol harms are associated with an economic annual cost of around £21 billion (£3.5billion in healthcare. Illicit drug use in the UK has an economic cost of around £15 billion(£488 million through healthcare, with nearly one in ten adults aged in England and Wales having used illicit drugs in the past year. Worldwide, alcoholattributable deaths increased from 3.8% in 2004to 5.9% in 2012, and illicit drug use levels have failed to decline between 2005 and 2010, with a slight increase in the UK in recent years.
## Scope for identifying new interventions
Pharmacological interventions for alcohol and substance use disorders have been well researched and reported on for the management of withdrawal, dependence, and relapse prevention. The Cochrane Drug and Alcohol Group has published 11 and 30 reviews of pharmacological interventions for alcohol and substance use, respectively, whilst psychosocial interventions (e.g. brief interventions and motivational interviewing) are less well reported, with six and eight published reviews, respectively. Preventive interventions only have five reviews for alcohol use, and three reviews for substance use. Due to the heterogeneity of the types of drugs used and style of intervention, it is hard to summarise meaningfully the available data of existing interventions. However, with relapse rates as high as 60% 1 year after treatment for substance use disorders (SUD) [bib_ref] Drug dependence, a chronic medical illness: implications for treatment, insurance, and outcomes..., Mclellan [/bib_ref] [bib_ref] Classes of substance abuse relapse situations: a comparison of adolescents and adults, Ramo [/bib_ref] [bib_ref] Conceptualizations of relapse: a summary of psychological and psychobiological models, Connors [/bib_ref] and 60-90% for alcohol use disorders (AUD) [bib_ref] Understanding and preventing relapse, Brownell [/bib_ref] [bib_ref] Alcohol treatment, changes in coping skills, self-efficacy, and levels of alcohol use..., Maisto [/bib_ref] [bib_ref] How effective is alcoholism treatment in the United States?, Miller [/bib_ref] [bib_ref] Substance abuse relapse in a ten-year prospective follow-up of clients with mental..., Xie [/bib_ref] and drug substitution therapies being associated with innate complications [bib_ref] Benefits of retention in methadone maintenance and chronic medical conditions as risk..., Fareed [/bib_ref] [bib_ref] The barriers to smoking cessation in Swiss methadone and buprenorphine-maintained patients, Wapf [/bib_ref] [bib_ref] Eyes wide shut?-a conceptual and empirical critique of methadone maintenance treatment, Fischer [/bib_ref] [bib_ref] Hypertension, chronic obstructive pulmonary disease, diabetes and depression among older methadone maintenance..., Maruyama [/bib_ref] , there is a need for evidence for new treatments and preventive interventions to help address the growing burden of alcohol and/or substance use.
Physical activity (PA; defined as any bodily movement produced by skeletal muscles that requires energy expenditure, inclusive of organised sportand healthoriented exercise interventions could impact on the prevention, reduction, and treatment of alcohol and/or substance use and have the potential to be cost-effective, flexible, accessible, acceptable across the range of levels of use and have a lower risk of adverse events compared to pharmacological treatment [bib_ref] Comparison of aerobic exercise, clomipramine, and placebo in the treatment of panic..., Broocks [/bib_ref]. In 2001 (with updates in 2005 and 2008), AT (with co-researchers) reviewed and reported the effects of exercise on smoking from eight randomised controlled trials (RCTs) as part of a Cochrane Review [bib_ref] Exercise interventions for smoking cessation, Ussher [/bib_ref]. This evidence contributed to a 2008 report to the US Surgeon General on smoking cessation which highlighted the value of exercise as an option to support smoking cessation [bib_ref] Treating tobacco use and dependence: 2008 update. Rockville: Department of Health and..., Fiore [/bib_ref] , and in the UK, many NHS Stop Smoking Services now advocate exercise [bib_ref] Determinants of physical activity promotion by smoking cessation advisors as an aid..., Everson [/bib_ref]. In 2014, an update to the Cochrane Review revealed there were 20 RCTs of exercise and smoking cessation, suggesting a rapid growth of interest in the topic [bib_ref] Exercise interventions for smoking cessation, Ussher [/bib_ref]. A first systematic review is now needed of physical activity interventions for the prevention, reduction, and treatment of alcohol and/or substance use that also includes a comprehensive search of grey literature and service evaluations to generate practical implications for practice and policy.
Evidence for the role of PA for preventing alcohol and/or substance use
Prospective studies indicate that sports participation in adolescents and young people is associated with an increase in alcohol use but decrease in illicit drug use [bib_ref] Sport participation and alcohol and illicit drug use in adolescents and young..., Kwan [/bib_ref]. However, such studies may fail to eliminate confounding factors (e.g., specific sports may attract those more predisposed to engaging in 'risky' behaviours).
In contrast, a rigorous study in Finland tracked 1870 twin pairs from 16 to 27 years of age and concluded that low levels of physical activity increased the risk of both alcohol and illicit drug use [bib_ref] Physical activity in adolescence as a predictor of alcohol and illicit drug..., Korhonen [/bib_ref]. This further demonstrates the need for a robust, systematic review assessing the role of physical activity (not just participation in sports) on progression to alcohol and/or substance use disorders.
Evidence for the effects of PA interventions for harm reduction and treatment of alcohol and/or substance use There is increasing interest in the role of physical activity as a treatment and reduction strategy for alcohol and/or substance use. In 2011, the US National Institute on Drug Abuse (NIDA) invested $4.3 millionon a programme of work including high-quality RCTs such as STRIDE [bib_ref] Stimulant reduction intervention using dosed exercise (STRIDE) -CTN 0037: study protocol for..., Trivedi [/bib_ref] which is investigating stimulant use reduction using exercise. A recent systematic review by Wang et al. [bib_ref] Impact of physical exercise on substance use disorders: a meta-analysis, Wang [/bib_ref] was limited by an incomplete search strategy (e.g., not CINAHL) and thus omitted key papers. They identified three studies with a focus on alcohol, five with a focus on illicit drug use, four on multiple drug use, and 11 on smoking. The data from each of these studies were pooled using meta-analysis, despite considerable apparent heterogeneity across interventions and outcomes. Harm-reduction studies were not considered. In another systematic review, Zschucke and colleagues [bib_ref] Exercise and physical activity in the therapy of substance use disorders, Zschucke [/bib_ref] found nine studies reporting the effects of PA on AUD and eight on SUD, but again, that review did not include key search engines and did not consider grey literature that may be most informative for the UK context. Both these reviews focussed on AUD and SUD and did not consider the broader spectrum of use that may not meet the classification of a disorder, e.g. recreational use that is still associated with risk of harm. A rigorous review of the evidence encompassing all aspects of alcohol and substance use is still needed.
Plausible mechanisms for the effects of exercise on the use of any addictive substance Physical activity may affect alcohol and/or substance use through various psychological mechanisms, such as an acute reduction in cravings and urges, an increase in positive affect, and a chronic improvement in co-morbid depression and anxiety which may moderate outcomes related to alcohol and substance use [bib_ref] Exercise-based treatments for substance use disorders: evidence, theory, and practicality, Linke [/bib_ref]. From the behavioural perspective, exercise involvement may help avoidance of cues which trigger cravings and relapse, and provide exposure to new environments, which provide diversionary safe and immediately rewarding experiences [bib_ref] Exercise-based treatments for substance use disorders: evidence, theory, and practicality, Linke [/bib_ref]. Participation in meaningful structured activities are a key part of overcoming AUD and SUD, and some physical activities may offer the chance for identity transformation through exposure to meaningful routine activities, informal social controls, and promoted personal agency. From the physiological perspective, there is evidence from animal studies to suggest that neurobiological changes associated with exercise [bib_ref] Exercise-based treatments for substance use disorders: evidence, theory, and practicality, Linke [/bib_ref] [bib_ref] Exercise as a novel treatment for drug addiction: a neurobiological and stage-dependent..., Lynch [/bib_ref] [bib_ref] The neurobiology of exercise and drug-seeking behaviour, Smith [/bib_ref] help to explain the consistent evidence that exercise acutely reduces consumption of cocaine, morphine, nicotine, and alcohol [bib_ref] Exercise as a novel treatment for drug addiction: a neurobiological and stage-dependent..., Lynch [/bib_ref] [bib_ref] Effects of different exercise protocols on ethanol-induced spatial memory impairment in adult..., Hashemi Nosrat Abadi [/bib_ref] [bib_ref] Chronic wheel running reduces maladaptive patterns of methamphetamine intake: regulation by attenuation..., Engelmann [/bib_ref] [bib_ref] Daily treadmill exercise attenuates cocaine cue-induced reinstatement and cocaine induced locomotor response..., Thanos [/bib_ref] [bib_ref] Effect of wheel-running during abstinence on subsequent nicotine-seeking in rats, Sanchez [/bib_ref].
Finally, recent studies indicate that physical activity interventions can be acceptable for those with AUD and SUD [bib_ref] Exercise preferences of patients in substance abuse treatment, Abrantes [/bib_ref] [bib_ref] Exercise attitudes and behaviors among persons in treatment for alcohol use disorders, Read [/bib_ref] [bib_ref] Heroin users' views and experiences of physical activity, sport and exercise, Neale [/bib_ref] , but no review exists of this published and grey literature to help inform the design of the most feasible and acceptable interventions across the spectrum of levels of use.
PA may influence alcohol and/or substance use in similar ways and through common mechanisms and therefore form the focus of this review. However, due to the different way in which alcohol and substance use are viewed, approached, and treated within the UK, they will be considered separately within this review and not combined in any analyses in order to ensure the most pertinent findings for policy and practice.
## Impact of stakeholder engagement
Stakeholder engagement has many benefits and can contribute towards the development of systematic reviews. Stakeholders can be funders, service users, healthcare professionals, or charities, i.e. anyone who will implement interventions based on the findings of the reviews. A research white paper looking at the benefits of stakeholder engagement in systematic reviews was published by Cottrell and colleagues in 2014. They reviewed papers and suggested that the benefits of stakeholder engagement included identifying and prioritising potential research topics, helping to recruit participants, and providing useful feedback on the systematic review protocol. Other benefits included helping the researchers to understand the perspective of the service users/participants and ensuring the accessibility of the results with wider dissemination. Most studies reviewed were in the UK, and they suggested that the scope for the reviews was refined due to stakeholder engagement and that generally the overall quality of the review was improved. Given the broad focus of this review across several sectors and service providers in what is an underresearched area, involving stakeholders will maximise the applicability and impact of the findings.
As part of this process, our original questions were based on concerns highlighted by the Plymouth City Council Public Health team and the Plymouth NHS Hospital Trust due to the high national prevalence of alcohol and/or substance use and resulting hospital admissions in the area. We worked with a local third sector organisation which provides day support for persons in the community affected by the use of drugs and/or alcohol, as well as an education service as an alternative to pupil referral units in Plymouth. Stakeholder groups within this service (three service providers and eight service users) supported the focus on our three key research questions about prevention, harm reduction, treatment and relapse prevention, and highlighted the importance of PA through their own narratives, independent of, and in addition to, standard treatment. Further engagement with co-applicants Gary Wallace (Senior Specialist Drugs and Alcohol Team Manager in the Plymouth Public Health), Julia Sinclair (Honorary Consultant in Alcohol Liaison and Wessex Alcohol (AHSN) Lead), Joanne Neale (lead for PPI addiction research group at KCL), and local third sector leaders refined the scope and methods for the review. In addition, coapplicants Adrian Taylor and Joanne Neale have previously conducted and published qualitative research involving people with SUD which highlighted the need to further develop appropriate interventions [bib_ref] Determinants of physical activity promotion by smoking cessation advisors as an aid..., Everson [/bib_ref] [bib_ref] Heroin users' views and experiences of physical activity, sport and exercise, Neale [/bib_ref]. The scope and methods for this review have been strongly influenced by both service provider and user perspectives throughout the development of the application and this protocol.
## Aim and objectives
Our overarching aim is to describe and evaluate the quantitative and qualitative research on the impact of physical activity on the prevention, reduction, and treatment of alcohol and/or substance use across the lifespan.
Physical activity interventions (including those involving sport, exercise, or general lifestyle physical activity) may have the potential to impact three domains of alcohol and/or substance use:
1. Reduce the risk of progression to alcohol and/or substance use (PREVENTION); 2. Support individuals to reduce alcohol and/or substance use for harm reduction (REDUCTION), and 3. Promote abstinence and relapse prevention during and/or after treatment of AUD and SUD (TREATMENT).
We aim to describe and evaluate the available quantitative and qualitative research for each of these scenarios and seek to assess how what we know can best be translated into policy and practice with the input of key stakeholders, presenting where possible any cost-effectiveness data. This will be achieved by the following objectives: a) To quantify and describe quantitative data relating to the impact of physical activity on alcohol and/or substance use outcomes (completing meta-analyses where possible); b) To analyse and describe qualitative data relating to the acceptability, feasibility, mechanisms, mediators, and moderators of physical activity in relation to alcohol and/or substance use (completing metasyntheses where possible); c) To describe and analyse service evaluations which may not meet peer-reviewed quantitative or qualitative inclusion criteria relating to the implementation and impact of physical activity interventions relating to alcohol and/or substance use; d) To quantify and describe potential cost-effectiveness data relating to physical activity and its impact on alcohol and/or substance use; e) To produce practical recommendations about what is known about what works for who, when, where, and how through a narrative synthesis informed by stakeholder input.
Each of these objectives will address the three domains of prevention, reduction, and treatment, separately, but some crossover will be expected. Outcomes tabled by population, intervention, control, and outcome against the three domains of physical activity and its possible impact on alcohol and substance use
## 1) prevention
(2) Reduction Treatment (and relapse prevention) Population Adolescents, at risk groups General population who use alcohol or substances but not receiving acute or long-term care for a diagnosed AUD and/ or SUD, at risk groups Levels of subsequent use of alcohol and/ or substances, prevalence rates % reduction in alcohol and/or substance use, prevalence Abstinence rates, % days abstinent, % reduction in alcohol and/or substance use, relapse rates Costs, cost-effectiveness, net benefit, resource use and cost, economic measures of benefit (e.g. Quality-adjusted-life-years (QALYs), and willingness to pay (WTP)), summary economic measures (e.g. incremental cost-effectiveness ratios ICERs)
Costs, cost-effectiveness, net benefit, resource use and cost, economic measures of benefit (e.g. Quality-adjusted-life-years (QALYs), and willingness to pay (WTP)), summary economic measures (e.g. incremental cost-effectiveness ratios ICERs)
Costs, cost-effectiveness, net benefit, resource use and cost, economic measures of benefit (e.g. Quality-adjusted-life-years (QALYs), and willingness to pay (WTP)), summary economic measures (e.g. incremental cost-effectiveness ratios ICERs) a Outcomes listed are indicative and not exhaustive
# Methods
This protocol has been prepared using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses protocols (PRISMA-P) guidelines [bib_ref] Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015 statement, Moher [/bib_ref] (see Additional file 1).
## Eligibility criteria
We will not limit our searches by country; however, we will only include papers published in English. Whilst we recognise there is a potential for bias to be introduced because of limiting the searches to English, the direction and degree of such bias are unknown. As outlined in the Cochrane Handbook for Systematic Reviews of Interventions, there is conflicting evidence about the potential bias introduced by an English language limit: Juni [bib_ref] Direction and impact of language bias in meta-analyses of controlled trials: empirical..., Juni [/bib_ref] reported that non-English trials were more likely to report significant results, whilst Moher [bib_ref] What contributions do languages other than English make on the results of..., Moher [/bib_ref] reported no significant difference in meta-analyses which excluded trials in languages other than English. Studies will be restricted from 1978 to the present: 1978 was chosen as a cutoff point based on a frequency analysis on a subsample of relevant literature.
## Types of studies
We will include (a) quantitative studies (RCTs, quasi-RCTs, non-randomised controlled trials, controlled before and after studies, prospective or retrospective cohort studies that include a control group, historically controlled trials, nested case-control studies, casecontrol studies, and before-and-after comparisons); (b) qualitative investigations (of any recognised qualitative methodology); (c) local service evaluations; and (d) economic evaluations (full and partial).
## Type of setting
We will not limit the setting or country in which interventions are delivered (although this will be impacted considerably by the English language restriction), and this variation will be considered in the narrative synthesis. Studies may include inpatient and outpatient programmes, public health interventions, and communitybased interventions. We will not place any limitations on who delivers the intervention and in what format.
## Participants/population
No limit on participants will be applied; all adults and children will be considered. We expect most studies to include adolescents at risk of alcohol and substance use (prevention), adults in acute rehabilitation for SUD/ AUD and post-acute rehabilitation for SUD/AUD (relapse prevention and supporting abstinence), and any other adults receiving support or intervention for reducing alcohol and substance use (reduction). We will record and consider these diversities in the synthesis of results. We expect certain populations to be of particular significance in the research (e.g. people who are homeless, have mental health problems, or belong to groups experiencing complex needs or disadvantages), where alcohol and substance use may not be the primary outcome and physical activity may be part of a more complex intervention. Where this type of study is identified, it will be assessed for relevance on a case-by-case basis, discussed within the research team, and included if it contains viable data that can be included within the review's defined primary outcomes.
## Intervention(s) and comparator(s)
We will include any studies evaluating and comparing interventions that include a physical activity promotion element either explicitly targeting a reduction in alcohol and substance use or implicitly resulting in a reduction in alcohol and substance use. This could be within one of the three domains of prevention, reduction, or treatment. The comparator could be no intervention, treatment as usual (e.g. pharmacotherapy and psychological therapies), or alternative physical activity interventions (e.g. running vs walking).
The scope of this review is to include research on alcohol and substance use in its broadest sense. We plan to include data on alcohol and substance use which may not be considered a 'disorder' which reflects levels and prevalence of use, as well as including research on AUD and SUD as classified in the diagnostic and statistical manual of mental health disorders, fifth edition (DSM-V).
## Outcomes
The primary outcomes are mapped against the four planned analyses by the three domains of PA and its possible impact on alcohol and/or substance use in .
## Secondary outcomes
Secondary outcomes will be collected during the data extraction phase in addition to the primary outcomes above where present. These include:
Physical activity levels/fitness; Biomedical outcomes (e.g. liver function, hepatitis C status); Mental health and wellbeing; Adverse events.
We will also extract data referring to the identification of the underlying psychological theory informing interventions; intervention structure and content; information relating to how an intervention may work including challenges, barriers, and facilitators of behaviour change (process evaluations); the mechanisms of change (mediators and moderators), acceptability, and feasibility data; and any evidence of a dose-response relationship.
## Information sources
We will develop and test a highly sensitive search strategy of published and grey literature using background scoping searches, previously identified relevant research, and in consultation with subject experts and public and patient involvement. The strategy will include searches of the following sources:
## Search strategy
An Information Specialist (AW) will design and conduct the search strategy with expert consultation. The strategy will be translated for use in each database stated above, and a modified keyword-only strategy will be used for grey literature searching. The search strategy will be designed to encompass the three aims of the review (i.e. prevention, reduction, and treatment).
See Appendix 1 for a sample search strategy.
## Searching other resources
Extensive grey literature searching will be conducted to ensure maximum coverage of the subject area. The grey literature strategy will encompass focused searches in Google, several specialised databases, and consultation of subject experts for recommendations. This process will generate grey literature publications as well as relevant websites of local and national organisations in the UK, which will be hand-searched for additional citations. We will also conduct backwards and forwards citation chaining of all included studies to identify further relevant articles, as well as directly contact-known experts in the field and the lead authors of key publications for knowledge of any other relevant work. We will include PhD theses, but exclude MSc theses. All grey literature websites and search engines will be searched with targeted keywords and phrases generated from our original search strategy. The first 100 hits of each search will be screened by title and abstract. If a high proportion of the first 100 hits (≥ 10%) can potentially be included, then a further 100 hits will be searched continuing until the next 100 hits contain ≤ 10% of potentially includable hits. If the initial search produces fewer than 100 hits, then all hits will be searched.
See Appendix 2 for sample grey literature search strategy.
## Study records data management
Exported citations from traditional databases will be entered and de-duplicated into EndNote X8 (Clarivate Analytics). Grey literature results will be manually entered or, where available, captured through a browser-based citation management plug-in (such as Zotero [https:// www.zotero.org/]) then imported into EndNote. Using a structured and piloted data extraction form, we will extract relevant outcome data, study characteristics, and participant characteristics from each included paper. Data will be extracted by one reviewer and checked by another.
## Selection process indexed and academic databases
Two waves of study selection will be undertaken. Titles and abstracts will be screened by two reviewers independently and disagreements resolved by discussion or, where necessary, a third reviewer (title and abstract screening will be conducted using Rayyan software (QCRI; Doha, Qatar; https://rayyan.qcri.org/)). Two initial subsets of 500 results will be screened by two reviewers and inclusion and exclusion discrepancies discussed following each in order to ensure good agreement between reviewers. Following this, a set of 1000 will be completed and discussed before the remaining results being screened independently by two reviewers. This will help ensure reliable and consistent screening. Full texts will be obtained for studies appearing to meet the criteria above and screened by two reviewers (each paper reviewed by one member of the team and checked by another). Disagreements are resolved through discussion and a third reviewer (AT). RK will be consulted in relation to uncertainty over economic evaluations arising from the two independent reviewers.
The same process will apply for grey literature searching.
## Appraisal of studies (quality and bias)
We will evaluate risk of bias at the level of outcomes. Randomised controlled trials will be assessed for quality and risk of bias using the Cochrane Risk of Bias Tool, and non-randomised studies will be assessed using the ROBINS-I [bib_ref] ROBINS-I: a tool for assessing risk of bias in non-randomised studies of..., Sterne [/bib_ref]. Any economic evaluations will be assessed for study quality using the Consolidated Health Economics Evaluation Reporting Standards (CHEERS) [bib_ref] Consolidated Health Economic Evaluation Reporting Standards (CHEERS) statement, Husereau [/bib_ref] checklist. Qualitative studies will be assessed for quality using a ten-item checklist for qualitative studies developed and published by the Critical Skills Appraisal Programme (CASP)which focuses on rigour, credibility, and relevance without being overly restrictive. CASP will be used to appraise studies but not to exclude any studies.
## Data synthesis
Data synthesis will be adapted from the multilevel approach as suggested in the Cochrane Handbook of Systematic Reviews: Quantitative and qualitative evidence will be reviewed separately and then combined into an overall narrative synthesis. A narrative synthesis of service evaluations and economic data will be integrated into the main synthesis to aid in contextualising the results in terms of implementation.
## Quantitative studies (analysis a)
Where data allow (e.g. data on the same outcome from at least two studies of similar design, intervention, and population), we will conduct a meta-analysis to estimate the overall effect and consistency of the intervention effect across studies. As the population and setting of studies are likely to be different, we will use a randomeffects model to obtain the summary result as an estimate of the average intervention effect rather than the common effect estimated from a fixed effects model [bib_ref] A basic introduction to fixed-effect and random-effects models for meta-analysis, Borenstein [/bib_ref]. Where possible, we will create and examine funnel plots for the association between study size and estimated effect size, which could be due to publication bias. Where possible, we will explore the extent to which the intervention characteristics, study setting (country, socioeconomic status, healthcare system), and participant characteristics moderate the effect of interventions, through conducting meta-regressions or subgroup analyses.
We will not combine data from non-randomised trials which used different study designs, or data from randomised trials and non-randomised trials, in a metaanalysis, as the estimated intervention effects from different study designs can be influenced by different sources of bias and/or increased heterogeneity. In those cases, where suitable numerical data are not available for pooling, or if pooling is considered inappropriate, we will use other approaches to provide a systematic summary of the studies, including tabulation, transformation of data into common rubric (e.g. days abstinent), groupings and clusters (e.g. different population to assess influence of country, age, socioeconomic status, type/intensity of intervention, setting), and textual descriptions including a detailed narrative synthesis.
## Qualitative studies (analysis b)
The qualitative synthesis aims to describe qualitative data relating to the acceptability, feasibility, mechanisms, mediators, and moderators of physical activity in relation to alcohol and/or substance use. Data on the development, design, methods, and the populations involved will be extracted from qualitative studies using a bespoke data extraction form. The complete "findings" or "results" sections of the qualitative study reports will be exported into NVivo 10 (QSR International Pty Ltd.). Each section will then be read and re-read by two reviewers, in conjunction with the data extraction form, to enable the reviewers to familiarise themselves with the study findings in the context of the study population, setting, and methods. Adopting a thematic analysis approach, reviewers will code and identify emergent themes and concepts independently (extracting associated quotes). The reviewers will come together to consolidate the findings into one summary of overarching themes. Associated quotes will be presented to support the identified themes. The review team will then draw out implications of the themes for policy and practice.
## Service evaluations (analysis c)
Service evaluations will be considered separately from the academic literature, and through a thematic synthesis approach will be summarised to help understand contextual and implementation issues surrounding the delivery of PA for alcohol and/or substance use. It will also be used, where possible, to contextualise data from the academic literature within the UK context to aid with the final narrative synthesis.
## Economic evaluations (analysis d)
The review of resource use, costs, relative effectiveness and cost-effectiveness will include a descriptive summary of the (economic) study questions, methods, and results, culminating in a narrative synthesis. Since the purpose of our review is to provide clear and concise information on the existing economic evidence base, we will also consider partial economic evaluations. These may include cost comparisons, as well as studies with an exclusive focus on relative benefits, i.e. studies that discuss willingness to pay or preference-based outcome measures. Summary tables will not be limited to description of economic outcomes alone and will include all relevant information integral to the economic study. We will extract detail on analytic methods, study perspective, price year, country, currency, and time horizon with further extraction fields informed by section headings within the CHEERS [bib_ref] Consolidated Health Economic Evaluation Reporting Standards (CHEERS) statement, Husereau [/bib_ref] checklist. Since we do not anticipate a substantial amount of economic literature, our methods may focus on translating findings from the review for the purposes of dissemination and stakeholder input. All types of comparative economic study design, including decision-analytic modelling approaches, will be included.
## Narrative synthesis
The analysis of the quantitative (analysis A), qualitative (analysis B), service evaluation (analysis C), and economic (analysis D) data will be integrated to develop a narrative synthesis. This will be summarised for dissemination to PPI groups and key stakeholders and used as a basis for generating critical input to help understand the implications of the findings for different groups.
## Measures of intervention effects (quantitative data) dichotomous data
We will present dichotomous data as risk ratios with their associated 95% confidence intervals (CI).
## Continuous data
For continuous data, we will calculate the mean differences (MD) for outcomes measured by the same scale or the standardised mean differences (SMD) for outcomes measured by different scales and present both with a 95% CI.
## Outcomes at multiple time points
If outcomes were collected at multiple time points, we will attempt to present a summary effect over all time points. If this is not possible, we will choose one time point that is the most appropriate one and report the corresponding summary effect at that time point.
## Unit of analysis issues
Cluster randomised trials are susceptible to unit-ofanalysis errors if the analysis was performed at the level of the individual without accounting for the clustering in the data. If the clustering effect has been accounted for in the analysis, the estimated intervention effect will be obtained from the reported summary data. If the clustering effect has not been accounted for, we will conduct an approximate analysis using the intra-cluster correlation coefficient (ICC), as suggested in the current guidelines. If the ICC is available in the study reports or can be obtained from similar studies, we will use the available ICC to calculate the inflated standard error or effective sample size to account for the clustering effect. If a relevant ICC is not available, we will report the estimated intervention effect as presented but report the issue of unit of analysis error.
## Dealing with missing data
If a study did not provide the summary data of the intervention effects, we will contact the study authors on one occasion to request these data. Where individual-level data are missing due to participant dropout, we will conduct available case analyses and record any issues of missing data in the 'Risk of bias' table. If standard error is available but standard deviations are not reported in a study, we will estimate the standard deviation from the reported standard error and the sample size. We will calculate the effect estimate and its standard deviation if these are not reported, but the 95% CI is reported.
## Assessment of statistical heterogeneity
We will assess statistical heterogeneity by inspection of forest plots, and by formal statistical tests of homogeneity (Chi-squared), measures of inconsistency (I 2 ) [bib_ref] Quantifying heterogeneity in a meta-analysis, Higgins [/bib_ref] , and between-study variance (tau 2 ). We interpret the level of heterogeneity as follows: the heterogeneity is not important if I 2 is lower than 40%; there is moderate heterogeneity if I 2 is between 30 and 50%; and there is substantial or considerable heterogeneity if I 2 is greater than 50%. If substantial heterogeneity is identified among studies, we will explore the potential causes of heterogeneity by conducting subgroup analyses or metaregression where possible.
## Assessment of reporting biases
We will examine the possibility of publication bias and other small study effects using funnel plots of the intervention effect estimates against the inverse of their standard errors and test funnel plot asymmetry using Egger's method [bib_ref] Bias in meta-analysis detected by a simple, graphical, Stuck [/bib_ref] when there is a sufficient number of studies. This is considered at least ten, as a smaller number would leave the power of the test too low to distinguish chance from real symmetry.
# Sensitivity analysis
We will conduct a sensitivity analysis based on quality indicators thought to be significant by the review team. Studies thought to be at high risk of bias due to specific quality indicators (e.g. lack of randomisation) will be removed to ascertain their effect on the estimated overall effect. We also intend to conduct a sensitivity analysis using fixed-effect models.
## Confidence in cumulative evidence
The quality of evidence will be assessed using domains of the Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) [bib_ref] GRADE: an emerging consensus on rating quality of evidence and strength of..., Guyatt [/bib_ref] guidelines. Strength of evidence will be judged as 'high' (further research is very unlikely to change confidence in our findings), 'moderate' (further research is likely to have an important impact on our findings), 'low' (further research is likely to have an important impact and change our findings), 'very low' (further research is needed to draw any conclusions).
## External validity/generalisability
We will explore generalisability at the study level and on the level of the aggregated evidence.
We will collect the following data from each study: details of the intervention, fidelity of the intervention and adherence to it (this includes who delivers the intervention), rationale supporting the choice of outcome measures, and setting in which the study was conducted (country, socioeconomic status of the setting, healthcare system). Depending on the type of available data, we will analyse them using subgroup analyses, meta-regression, or contextualise it through narrative synthesis.
## Advisory groups
We will establish advisory groups: consisting of academics, service providers, and service users. These groups will offer varying perspectives on pertinent issues arising and will provide input into different aspects of the review.
The academic group will be asked to provide a critical voice on the interpretation of findings, to ensure no aspect has been overlooked, and to aid in establishing networks for finding data and disseminating results.
The service provider group will offer insight into the UK context in which alcohol and substance use exists, what the commissioning landscape looks like, current and future reach and structure of services and interventions, contextualise findings to the current UK climate, and to aid in dissemination and implementation of results.
The service user group will offer input into what outcomes we should be searching for, approaches to promote physical activity (i.e. sport, exercise, and daily activity) would be most acceptable to people; and in what other ways could support services (who, where, when) be set up to maximise the reach and effectiveness for promoting physical activity for the prevention, reduction, and treatment of alcohol and substance use.
To maximise the impact of stakeholder, public, and patient involvement, we will develop a user-friendly synthesis of the findings and nature of interventions and their apparent strengths and weaknesses working with the advisory groups. Once a user-friendly synthesis has been generated, it will be disseminated to key stakeholder groups and individuals and their feedback and input will be used to gain further insights into what the evidence suggests and where any gaps may exist. Specifically, people will represent different stages of addiction, including those who are occasional users at risk of progressing to regular users of alcohol and/ or substance use, non-treatment seekers who wish to minimise harm, and those currently receiving treatment or who are in recovery. This synthesis will then be disseminated to key stakeholder groups and individuals and used as the basis for several group and individual meetings to gain further input into what the evidence suggests and where any gaps may exist.
## Dissemination and intended publications
Upon completion of the review, we will develop a summary of key findings of the review of literature, a summary of the PPI assessment of the findings, and issues associated with service development and delivery highlighted by policy makers and service managers. We will present the findings at relevant academic conferences and a website that will be established to summarise the findings and implications, with links to access a final report. We intend also to organise a one-day conference to which key stakeholders, advisory board members, and any interested party will be invited. In addition to a final report, we anticipate submitting articles for publication in peer-reviewed open access journals.
We will also disseminate the findings by phone or Skype to Directors of Public Health (or leads for alcohol and substance misuse) and managers of organisations across the UK who do or could involve physically active interventions to gain a further insight into the issues associated with securing the necessary resources.
# Discussion
The scope and methods for this review have been strongly influenced by both service provider and user perspectives throughout the development of the application and this protocol. The review will generate important and timely information to inform the provision of services for alcohol and substance use. Through the continued engagement with stakeholders, the information produced will have relevance across a variety of settings in addressing the prevention, reduction, and treatment of alcohol and/or substance use throughout the UK. A wide variety of dissemination plans will ensure the information is accessed by the most relevant services, as well as aiding to direct future research efforts.
The size and scope of the review, whilst challenging, will ensure that the information brought together in this review will be as encompassing as possible and will provide all the necessary information about what we know about what works, for who, when, where, and how in an accessible and appropriate way. Information generated from this review will have the potential to directly impact on provision in several domains and address (where possible) issues of acceptability, feasibility, implementation, and cost-effectiveness. Scottish Addiction Studies online library HRB National Drugs Library (use "all fields" search) NIDA International Drug Abuse Research Abstract Database (press enter to search; press enter with blank search to see total contents of the database)
Searches for the above databases: **Note: Each of these searches use "physical activity" related terminology because the databases above are drug and alcohol related. Feel free to search for other relevant terminology as needed and record results in Excel. Search for the above database:
1 Download all items from the database and create smart group in EndNote (n = 216,850)Any field ➔ word begins with ➔ substance*; alcohol*; drug*; opioid*; addict*ANDAny field ➔ word begins with ➔ exercise*; fitness; "physical activity"; fitness; sedentary; sport*= 263 hits
## Strategy 7:
To be conducted by AW.
## Proquest dissertations and theses
Using modified search strategies from the traditional database searches. Availability of data and materials Not applicable.
## Grey lit searching procedures
Authors' contributions TT Principal Investigator (PI) is responsible for overseeing the project at all stages, including the development of the search strategy, screening of identified studies, applying the quality assessment tools, the analysis plan, writing of reports, and managing PPI and stakeholder input. AT provides input and guidance on all aspects of the research, including the search strategy, quality appraisal methods, qualitative and quantitative syntheses, and the production of the final reports. YW oversees the development of the statistical analysis methods of the synthesis, leads the statistical analysis, and input to interpretations and supports report production. SC as a trialist and statistician provides input into the development of the statistical analysis methods, data capture, interpretation, analyses, and input to reports. MN advises on all methodological aspects of the review, the search strategy, quality appraisal and bias assessment, the synthesis methods, and report production. KH works on guiding the search methods and strategies and supports report production. AW as the Information Specialist leads and guides on search strategy. JN advises on the search strategy, qualitative analysis, identification of PPI groups for input and dissemination, with a focus on substance use disorders, synthesis of findings, and report production. JS advises on the search strategy and the development of PPI input, with a focus on alcohol use disorders, the synthesis of findings, and report production. RK oversees and guides the economic review aspects of the research and supports report production. GW provides input to maximise stakeholder input, impact, and relevance, as a local Public Health Service lead, and supports report production. All authors contributed to, reviewed, and approved the final manuscript.
[table] Appendix 1: Sample search strategy Databases to be searchedTo be conducted by AW. [/table]
|
Identification of QTNs and Their Candidate Genes for 100-Seed Weight in Soybean (Glycine max L.) Using Multi-Locus Genome-Wide Association Studies
100-seed weight (100-SW) in soybeans is a yield component trait and controlled by multiple genes with different effects, but limited information is available for its quantitative trait nucleotides (QTNs) and candidate genes. To better understand the genetic architecture underlying the trait and improve the precision of marker-assisted selection, a total of 43,834 single nucleotide polymorphisms (SNPs) in 250 soybean accessions were used to identify significant QTNs for 100-SW in four environments and their BLUP values using six multi-locus and one single-locus genome-wide association study methods. As a result, a total of 218 significant QTNs were detected using multi-locus methods, whereas eight QTNs were identified by a single-locus method. Among 43 QTNs or QTN clusters identified repeatedly across various environments and/or approaches, all of them exhibited significant trait differences between their corresponding alleles, 33 were found in the genomic region of previously reported QTLs, 10 were identified as new QTNs, and three were detected in all the four environments. The number of seed weight (SW) increasing alleles for each accession ranged from 8 (18.6%) to 36 (83.72%), and three accessions (Yixingwuhuangdou, Nannong 95C-5, and Yafanzaodou) had more than 35 SW increasing alleles. Among 36 homologous seed-weight genes in Arabidopsis underlying the above 43 stable QTNs, more importantly, Glyma05g34120, GmCRY1, and GmCPK11 had known seed-size/weight-related genes in soybean, and Glyma07g07850, Glyma10g03440, and Glyma10g36070 were candidate genes identified in this study. These results provide useful information for genetic foundation, marker-assisted selection, genomic prediction, and functional genomics of 100-SW.
# Introduction
Soybean (Glycine max L. Merr.), which provides 69% dietary protein and 30% oil [bib_ref] Resequencing of 31 wild and cultivated soybean genomes identifies patterns of genetic..., Lam [/bib_ref] , is economically imperative food and oilseed crop worldwide. The 100 seed weight (100-SW) is an essential trait in soybean yield component, affected by seed size and shape, and positively correlates with seed yield [bib_ref] Mapping quantitative trait loci for seed size traits in soybean (Glycine max..., Xu [/bib_ref]. There are numerous soybean food items for various seed sizes, for (Glysoja.04g010563) was identified to be associated with SW and highly expressed during seed developmental stages [bib_ref] Genetic dissection of yield-related traits via genome-wide association analysis across multiple environments..., Hu [/bib_ref].
To date, several studies reported QTLs/QTNs regarding soybean SW using linkage and association mapping, but the related genes in soybean are relatively limited. The possible reason for this, firstly, a high degree of linkage disequilibrium (LD) in the soybean genome makes it difficult to detect the QTNs and genes precisely. Secondly, the number of molecular markers used in soybean GWAS is relatively small with low density, which reduces the efficiency of GWAS. Thirdly, single-locus GWAS (SL-GWAS) models were used in soybean for yield traits [bib_ref] Resequencing 302 wild and cultivated accessions identifies genes related to domestication and..., Zhou [/bib_ref] [bib_ref] Identification and validation of candidate genes associated with domesticated and improved traits..., Zhou [/bib_ref] [bib_ref] Genomewide association studies dissect the genetic networks underlying agronomical traits in soybean, Fang [/bib_ref] , and these models are single-locus genome-wide scanning and need multiple tests correction (e.g., Bonferroni correction) that removes many significant small-effect QTNs [bib_ref] Editorial: The applications of new multi-locus gwas methodologies in the genetic dissection..., Zhang [/bib_ref]. To overcome these limitations, Zhang's group developed a series of multi-locus GWAS (ML-GWAS) methods, such as mrMLM [bib_ref] Improving power and accuracy of genome-wide association studies via a multi-locus mixed..., Wang [/bib_ref] , and these ML-GWAS methods were used to dissect the genetic foundations of complex traits in different crops [bib_ref] Editorial: The applications of new multi-locus gwas methodologies in the genetic dissection..., Zhang [/bib_ref] [bib_ref] The application of multi-locus GWAS for the detection of salt-tolerance loci in..., Cui [/bib_ref] [bib_ref] Genetic dissection of maize embryonic callus regenerative capacity using multi-locus genome-wide association..., Ma [/bib_ref] [bib_ref] The applications of new multi-locus GWAS methodologies in the genetic dissection of..., Zhang [/bib_ref]. Currently, most of the studies have used SL-GWAS methods for the detection of 100-SW QTNs. However, almost no ML-GWAS articles have been found to detect QTNs for 100-SW in soybean. Therefore, more efficient studies are required to dissect the genetic basis of 100-SW, and exploring the QTNs/candidate genes associated with SW will be paramount for the genetic improvement and production of this crop.
Therefore, the objectives of this study were: (a) to dissect the genetic basis for 100-SW using the ML-GWAS methods, and compare the QTN results with those in all the previous studies; (b) to identify the seed weight (SW) increasing alleles of these QTNs for MAS, and (c) to find the potential candidate genes regulating 100-SW in the region of stable QTNs. The findings in this study will provide reliable information for MAS in soybean breeding and functional gene validation/cloning.
# Materials and methods
# Plant materials
A total of 250 soybean accessions were selected from different geographic regions of China. These soybean accessions came from 23 provinces and were disseminated in six eco-regions of China [bib_ref] Identification of domestication-related loci associated with flowering time and seed size in..., Zhou [/bib_ref] , and were obtained from the National Center for Soybean Improvement and Linyi Academy of Agricultural Sciences with 139 landraces and 111 cultivars.
All the accessions were planted at the Jiangpu Experimental Station of Nanjing Agricultural University (from June to October) and Experimental Station of Huazhong Agricultural University (from May to October) in 2014 (denoted as E1 and E3) and 2015 (E2 and E4). Plants were grown in 150 cm wide and 200 cm long plots according to the randomized complete block design with three replicates. The flowering time was started after six to eight weeks of emergence. The trait phenotypes were measured from five plants in the middle row of each plot, and 100-SW for each accession was averaged based on three replicates.
## Statistical analysis and heritability estimation
The best linear unbiased prediction (BLUP) of 100-SW for each accession was calculated using the R (http://www.R-project.org/, v3.5.0) package lme4 [bib_ref] Fitting linear mixed-effects models using lme4, Bates [/bib_ref] with the following model:
[formula] Phenotype~(1|Genotype) + (1|Year) [/formula]
The aov function in the R software was used to calculate the variances of 100-SW, and mixed linear model (MLM) was used to estimate polygenic variance components and heritability [bib_ref] Improving power and accuracy of genome-wide association studies via a multi-locus mixed..., Wang [/bib_ref] with the following equation:
[formula] y Xα φ ε = + + [/formula]
where y is the phenotypic vector; X is an incident matrix for fixed (non-genetic) effects, and α is a vector of fixed effects;
[formula] 2 MVN(0, ) g φ σ K [/formula]
is the polygenic effect with a multivariate normal distribution with zero mean, 2 g σ is polygenic variance, and kinship matrix K was calculated from marker information [bib_ref] Mapping quantitative trait loci by controlling polygenic background effects, Xu [/bib_ref] ;
[formula] 2 MVN(0, ) e ε σ I [/formula]
is the vector of residues, and 2 e σ was residual variance. The above two variance components were estimated from restricted maximum likelihood [bib_ref] Underestimation of heritability using a mixed model with a polygenic covariance structure..., Ryoo [/bib_ref]. The broad-sense heritability was calculated using the following equation:
[formula] 2 2 2 2 g B g e h σ σ σ = + [/formula]
## Population structure analysis and genome-wide association studies
RAD-seq was used to obtain high-density SNPs, while RAD-seq, genotyping of soybean accessions, methods of sequencing data calling variations, and the quality control were described in Zhou et al. [bib_ref] Identification of domestication-related loci associated with flowering time and seed size in..., Zhou [/bib_ref]. In this study, a total of 43,834 SNPs with minor allele frequency (MAF) > 0.05 were used to construct a population structure using the STRUCTURE 2.3.4 software [bib_ref] Inference of population structure using multilocus genotype data, Pritchard [/bib_ref]. The hypothetical number of subgroups (k) ranged from 1 to 10. The length of the burn-in period for each run was set to 10,000, and the number of Markov chain Monte Carlo replications after burn-in was set to 100,000. The best k in this population was identified according to Evanno et al. [bib_ref] Detecting the number of clusters of individuals using the software Structure: A..., Evanno [/bib_ref] using STRUCTURE HARVESTER [bib_ref] STRUCTURE HARVESTER: A website and program for visualizing STRUCTURE output and implementing..., Earl [/bib_ref]. Six ML-GWAS approaches with population structure (Q) and kinship (K) were used to detect significant QTNs, including mrMLM [bib_ref] Improving power and accuracy of genome-wide association studies via a multi-locus mixed..., Wang [/bib_ref] , FASTmrEMMA [bib_ref] Methodological implementation of mixed linear models in multi-locus genome-wide association studies, Wen [/bib_ref] , pLARmEB, ISIS EM-BLASSO [bib_ref] Iterative sure independence screening EM-Bayesian LASSO algorithm for multi-locus genome-wide association studies, Tamba [/bib_ref] , FASTmrMLM, and pKWmEB [bib_ref] Integration of Kruskal-Wallis test with empirical Bayes under polygenic background control for..., Ren [/bib_ref]. These methods were included in package mrMLM (https://cran.r-project.org/web/packages/mrMLM/index.html, v4.0). In the above methods, the first step was to select all the potentially associated markers, and kinship matrix K was automatically calculated. In the second step, the effects of all the selected markers were estimated by empirical Bayesian, the significances of the effects apart from zero were obtained by likelihood ratio test, and the threshold level LOD ≥ 3 (p = 0.0002) was used to determine significant QTNs [bib_ref] Methodological implementation of mixed linear models in multi-locus genome-wide association studies, Wen [/bib_ref] [bib_ref] Improving power and accuracy of genome-wide association studies via a multi-locus mixed..., Wang [/bib_ref] [bib_ref] Editorial: The applications of new multi-locus gwas methodologies in the genetic dissection..., Zhang [/bib_ref] [bib_ref] The applications of new multi-locus GWAS methodologies in the genetic dissection of..., Zhang [/bib_ref] [bib_ref] Iterative sure independence screening EM-Bayesian LASSO algorithm for multi-locus genome-wide association studies, Tamba [/bib_ref] [bib_ref] Integration of Kruskal-Wallis test with empirical Bayes under polygenic background control for..., Ren [/bib_ref].
# Elite allele analysis
Based on the QTN effect value and code 1 for genotype, SW increasing alleles of each stable QTN can be determined. If the QTN effect value is positive, the genotype with the code of 1 is regarded as the SW increasing allele; if the QTN effect value is negative, then alternative genotype is viewed as the SW increasing allele [bib_ref] Improving power and accuracy of genome-wide association studies via a multi-locus mixed..., Wang [/bib_ref] [bib_ref] Editorial: The applications of new multi-locus gwas methodologies in the genetic dissection..., Zhang [/bib_ref]. The average seed weight of the accessions with one allele was calculated to verify the QTN [bib_ref] Genome-wide association mapping for seed protein and oil contents using a large..., Li [/bib_ref]. For each QTN, the SW increasing allele percentage in mapping population was measured as the number of accessions having SW increasing allele divided by the total number of accessions. The SW increasing allele percentage for each accession was equal to the number of SW increasing alleles divided by the total number of stable QTNs. Using the stable QTN information, the best cross combinations were predicted for the soybean breeding program. If we want to add seed weight, SW increasing allele is elite allele, while SW decreasing allele is elite allele if we want to decrease seed weight.
## Prediction of candidate genes for 100-seed weight
Prediction of candidate genes for 100-SW was performed in 100 kb downstream and upstream of each stable QTN in SoyBase (http://soybase.org/; Wm82.a1.v1.1). For the screening of genes, the transcriptomic datasets of seven different seed developmental stages such as 4, 12-14, 22-24 (DAF: Days after flowering), seed weight 5-6 mg period (5-6 mgWS), cotyledon weight 100-200 mg period (100-200 mgCOT), cotyledon weight 400-500 mg period (400-500 mgCOT), and full seed maturity period (Dry seed) of soybean Williams 82 [bib_ref] Using RNA-seq to profile soybean seed development from fertilization to maturity, Jones [/bib_ref] were retrieved from the Gene Expression Omnibus database (https://www.ncbi.nlm.nih.gov/geo/; accession no GSE42871). This is because genes with high RPKM at these stages are related to seed size, seed weight, cotyledon, seed coat tissues, embryo, endosperm, seed storage proteins, and seed maturation protein [bib_ref] RNA-seq Atlas of Glycine max: A guide to the soybean transcriptome, Severin [/bib_ref]. Thus, candidate genes were determined as below [bib_ref] Editorial: The applications of new multi-locus gwas methodologies in the genetic dissection..., Zhang [/bib_ref]. Firstly, we removed all the genes with expression level < 1 in all the seven stages and selected those genes with a higher expression levels double their average expression levels in at least one seed developmental stage. Then, homologous genes related to seed weight in Arabidopsis were identified using BLAST analysis with the critical E value 1E-30. Finally, all homologous genes from soybean accompanied seed weight were selected, and considered candidate genes for 100-SW.
## Gene expression level analysis
The freely available RNA-Seq datasets of 14 soybean tissues [bib_ref] RNA-seq Atlas of Glycine max: A guide to the soybean transcriptome, Severin [/bib_ref] , including whole seeds from 11 stages of reproductive tissue development (flower, pod, and seed) and three vegetative tissues (leaves, root, and nodules) were obtained from SoyBase (http://soybase.org/), in order to analyze candidate genes with special higher gene expression levels in soybean seeds. The heat maps were generated by using R software packages "pheatmap".
## Kyoto encyclopedia of genes and genomes (kegg) pathway analysis
The Kyoto Encyclopedia of Gene and Genomes (KEGG) enrichment analysis was conducted for potential candidate genes to identify the functional categories, implemented by KEGG Orthology-Based Annotation System network software (KOBAS v3.0) [bib_ref] KOBAS 2.0: A web server for annotation and identification of enriched pathways..., Xie [/bib_ref] (http://kobas.cbi.pku.edu.cn/kobas3), with adjusted p value < 0.05 as threshold criteria.
# Results
## Phenotype variation of 100-seed weight
The 100-SW phenotype of each accession was the average of three replicates in each environment. The mean phenotypic values of 100-SW across 250 accessions in E1 to E4 environments were , with standard deviations of 5.96, 5.58, 5.07, and 5.83 (g), respectively, and their coefficient of variations ranged from 28.08-29.07 (%) [fig_ref] Table 1: Phenotypic analysis of soybean 100-seed weight in 250 accessions in four environments [/fig_ref]. The highest phenotypic value was observed in E4, whereas the lowest phenotypic value seen in E2 [fig_ref] Figure 1: Phenotypic distributions of 100-seed weight in 250 soybean accessions in four environments... [/fig_ref]. The continuous distribution was found in these environments [fig_ref] Figure 1: Phenotypic distributions of 100-seed weight in 250 soybean accessions in four environments... [/fig_ref]. Two-way ANOVA showed the significant difference of 100-SW across all the accessions (p-value < 0.01), indicating the existence of genetic variation among these accessions [fig_ref] Table 1: Phenotypic analysis of soybean 100-seed weight in 250 accessions in four environments [/fig_ref]. Meanwhile, the estimates of broad-sense heritabilities (h 2 B) for 100-SW in E1 to E4 environments were 93.70, 88.51, 90.15, and 83.20 (%), respectively, using polygenic and residual variances [fig_ref] Table 1: Phenotypic analysis of soybean 100-seed weight in 250 accessions in four environments [/fig_ref] , suggesting that the genetic effects play an essential role in phenotypic variation.
## Population structure analysis and genome-wide association studies
To define the subpopulations within the panel of 250 accessions, as described by Pritchard et al. [bib_ref] Inference of population structure using multilocus genotype data, Pritchard [/bib_ref] , we selected 16,174 of the 43,834 SNPs that were randomly distributed across the 20 soybean chromosomes and had better polymorphisms. STRUCTURE 2.3.4 software was used to calculate delta K (ΔK) [fig_ref] Figure 2: Population structure of 250 soybean accessions using 16,174 SNPs on soybean genome [/fig_ref] ; k = 1-10), revealing the existence of three subpopulations (selected k = 3) based on ΔK values [fig_ref] Figure 2: Population structure of 250 soybean accessions using 16,174 SNPs on soybean genome [/fig_ref]. All of the high-quality 43,834 SNPs in 250 accessions were used to conduct GWAS for 100-SW. As a result, respectively 66, 76, 45, 55, and 70, QTNs were detected to be associated with 100-SW in five situations (E1-E4 and BLUP) (Tables S1-S6). These represented 218 unique QTNs, of which 156 overlapped with previously reported QTNs, and 62 were found newly in this study; were identified by the mrMLM, FASTmrMLM, FASTmrEMMA, pLARmEB, pKWmEB, and ISIS EM-BLASSO, respectively, in all the situations (E1-E4 and BLUP) (Tables 2 and S1-S6). The LOD values ranged from 3.01 to 18.08, and the proportion of phenotypic variance explained (PVE) by each QTN ranged 0.38-7.88 (%). All these QTNs were distributed on 20 chromosomes, and more than 10 QTNs were found to be located on each of eleven chromosomes, including chromosomes 01, 04-07, 10, 11, 13, 14, 17 and 20 (Tables S1-S6).
Using SL-GWAS (MLM), 1, 1, 3, 5, and 1 QTNs were identified in E1 to E4 and BLUP, respectively . Among the eight QTNs, two QTNs (qcHSW-10-4 and qHSW- were overlapped with those from ML-GWAS.
## Stable qtns for 100-sw in soybean
Two types of QTNs were defined as stable QTNs. One is environmentally-stable QTN (esQTN), which is identified by at least three ML-GWAS methods, while another is methods-stable QTN (msQTN), which is detected in at least three environments/BLUP. In the present study, a total of 43 QTNs (37 QTNs and 6 QTN clusters) were identified as stable QTNs and listed in . Among the 43 stable QTNs, 36 were msQTNs, 22 were esQTNs, and 15 were common between msQTNs and esQTNs; [fig_ref] Table 3: Stable QTNs for soybean 100-seed weight identified in multiple environments and the... [/fig_ref]. Moreover, eight QTNs were identified in one environment by at least three ML-GWAS methods, while 3 QTNs were detected by one ML-GWAS method in at least three environments/BLUP, and their LOD scores were 3.58-13.31 and 3.19-15.00, respectively . Interestingly, three QTNs, qHSW-4-1, qcHSW-7-3 and qcHSW-10-4, were identified by six ML-GWAS methods to be associated with 100-SW in all the environments (E1 to E4) and BLUP model, whereas their LOD scores were 3.01-8.64, 4.56-18.08, and 3.26-10.73, respectively, and their PVE values were 1.05-5.34, 2.42-5.91, and 1.37-5.90 (%), respectively [fig_ref] Table 3: Stable QTNs for soybean 100-seed weight identified in multiple environments and the... [/fig_ref]. Furthermore, seven QTNs qcHSW-1-1, qHSW-2-2, qHSW-4-2, qHSW-6-1, qcHSW-6-3, qHSW-8-1, and qHSW-11-3, were detected, respectively, by five, six, three, six, five, three and one ML-GWAS methods to be associated with 100-SW in three environments/BLUP model, whereas their LOD scores were 4.71-6.30, 3.14-5. [bib_ref] Transcriptome-wide functional characterization reveals novel relationships among differentially expressed transcripts in developing..., Aghamirzaie [/bib_ref] [fig_ref] Table 3: Stable QTNs for soybean 100-seed weight identified in multiple environments and the... [/fig_ref]. Approximately 33 (76.74%) QTNs were overlapped or located near the genomic region of previously reported QTLs for 100-SW, while 10 (23.26%) were newly identified. . Stable QTNs of soybean 100-seed weight identified in four environments and BLUP values by using six ML-GWAS methods. The black color dots were the stable QTNs near previously reported genes (Glyma05g34120, GmCRY1, and GmCPK11), while the red color dots were the stable QTNs near predicted candidate genes in this study. The sky blue color dots were other stable QTNs, while the light green color dots were remaining significant QTNs.
## Validation of stable qtns for 100-sw in soybean
The above 43 stable QTNs were used for SW increasing allele analysis in order to validate these stable QTNs. The 100-SW average of accessions having SW increasing alleles was 1.64-20.05 (g) higher than that of the accessions with SW decreasing alleles and 0.3-18.83 (g) higher than that of all the accessions [fig_ref] Table 3: Stable QTNs for soybean 100-seed weight identified in multiple environments and the... [/fig_ref]. These QTNs showed the significant differences of 100-SW between SW increasing and decreasing alleles at the 0.01 level [fig_ref] Figure 1: Phenotypic distributions of 100-seed weight in 250 soybean accessions in four environments... [/fig_ref]. For example, for eleven stable [fig_ref] Figure 1: Phenotypic distributions of 100-seed weight in 250 soybean accessions in four environments... [/fig_ref] , the 100-SW averages of accessions with SW increasing alleles A, C, C, A, C, G, T, T, C, C, and A significantly increased 5. [bib_ref] Development and application of a novel genomewide SNP array reveals domestication history..., Wang [/bib_ref] [fig_ref] Table 3: Stable QTNs for soybean 100-seed weight identified in multiple environments and the... [/fig_ref] ; for seven stable QTNs (qHSW-7-4, qcHSW-10-1, qHSW-14-2, qHSW-15-1, qHSW-16-1, and qHSW- newly identified in this study, the 100-SW averages of accessions with SW increasing alleles A, C, A, C, A, and G were significantly higher than those with SW decreasing alleles G, G, G, A, T, and C, respectively; .
## Prediction of the best parental combinations for 100-seed weight in two directions
The number of SW increasing allele for each stable QTN in 250 accessions ranged from 3 (1.2%) to 225 (90.00%). Among the above 43 stable QTNs, 21 had more than 50% SW increasing alleles in the 250 accessions, while 22 had less than 50% SW increasing alleles in the 250 accessions . The number of SW increasing alleles for each accession ranged from 8 (18.60%) to 36 (83.72%). Among the 250 accessions, 69 had more than 22 (50%) SW increasing alleles, while 181 had less than 21 (50%) SW increasing alleles. Interestingly, eight accessions, Yixingwuhuangdou, Yafanzaodou, Nannong 95C-5, Ribendaheidou, Fujiandadou, Quxiandahuangdou, Bayueqing, and Nanchengqingpidadou had [bib_ref] The transcriptomic signature of developing soybean seeds reveals the genetic basis of..., Lu [/bib_ref] [bib_ref] Transcriptional and hormonal signaling control of Arabidopsis seed development, Sun [/bib_ref] [bib_ref] Short Hypocotyl Under Blue1 associates with Miniseed3 and Haiku2 promoters in vivo..., Zhou [/bib_ref] [bib_ref] Genetic dissection of yield-related traits via genome-wide association analysis across multiple environments..., Hu [/bib_ref] [bib_ref] Genome-wide associations and epistatic interactions for internode number, plant height, seed weight..., Assefa [/bib_ref] , and 27 SW increasing alleles, respectively. Six accessions Qinyan 1, Baihuadou, Mayidan, Mingshanhongxingjiroudou, Heibidou, and Qingcha 1 had 11, [bib_ref] Soybeans: Improvement, production, and uses, Russell [/bib_ref] [bib_ref] Assessment of phenotypic variations and correlation among seed composition traits in mutagenized..., Zhou [/bib_ref] , and 8 SW increasing alleles, respectively, while these accessions had 30, 28, 33, 30, 25, and 32 SW decreasing alleles, respectively. All the accessions can be used for the soybean breeding program by increasing or decreasing the number of SW increasing alleles in one cultivar. For example, the cross between Yafanzaodou (36 SW increasing alleles) and Ribendaheidou (34 SW increasing alleles) may produce the offspring with 41 SW increasing alleles. Thus, the best five cross combinations of large and small seeds were predicted and listed in . It should be noted that some parents were repeatedly present in these predicted combinations. For example, Ribendaheidou was used as parent in three cross combinations for larger seed, while Heibidou, Mayidan, and Qinyan1 were used as parents in at least two combinations for small seeds . . The best parental combinations predicted from genome-wide association studies for 100seed weight in soybean.
## Breeding objective predicted parental combinations
## Candidate genes underlying the stable qtns for 100-sw in soybean
A search for putative candidate genes resulted in 774 potential candidate genes located between 50 kb-100 kb up-or downstream of the above 43 stable QTNs. Among the 774 genes, 205 exhibited high expression levels at seven seed development stages. Among the 205 genes, 175 were found to have homologs in Arabidopsis. Among these homologs, 36 genes were found to be related to seed weight [fig_ref] Table 1: Phenotypic analysis of soybean 100-seed weight in 250 accessions in four environments [/fig_ref]. Kyoto encyclopedia of genes and genomes (KEGG, http://kobas.cbi. pku.edu.cn/kobas3) analysis from the above 36 genes indicated that nine genes were involved in eleven seed-weight-related pathways [fig_ref] Table 1: Phenotypic analysis of soybean 100-seed weight in 250 accessions in four environments [/fig_ref]. In RNA-seq analysis, twenty-nine genes had two times higher gene expression, as compared with Glyma03g29431, Glyma04g33610, Glyma06g07940, Glyma07g03810, Glyma07g05260, Glyma11g07523, and Glyma20g21082 in seed development stages . Between the above nine and twenty-nine genes, there were six common genes, which were considered as candidate genes in this study [fig_ref] Table 5: Predicted potential candidate genes for 100-seed weight near the stable QTNs in... [/fig_ref] ;. Among these candidate genes, Glyma05g34120, Glyma06g10830 (GmCRY1), and Glyma06g16920 (GmCPK11) had known functions for seed size/weight in soybean [fig_ref] Table 5: Predicted potential candidate genes for 100-seed weight near the stable QTNs in... [/fig_ref] ;bold text), and the others were homologous to the known genes for seed size and seed development in Arabidopsis, for example, candidate gene Glyma07g07850 underlying the stable QTN qcHSW-7-3 were homologous to AT4G00710 (BSK3), which annotate BR-signaling kinase 3 in Arabidopsis. Glyma10g03440 and Glyma10g36070 underlying the stable QTNs qcHSW-10-1 and qcHSW-10-4 were homologous to, respectively, the Arabidopsis genes AT1G03090 (MCCA) and AT1G35680 (RPL21), which are related to seed weight or development [fig_ref] Table 5: Predicted potential candidate genes for 100-seed weight near the stable QTNs in... [/fig_ref] ;. . The expressional levels [log2(RPKM + 1)] of potential candidate genes associated with seed weight in seven soybean tissues. Among the nine genes, three were previously reported seedweight genes (bold text) in soybean, three were lowly expressed genes (grey text), and three were newly identified as candidate genes (red text) to be related to seed size/weight and development.
## Figure 5.
Manhattan plot in the detection of QTNs for 100-seed weight in soybean using multi-locus genome-wide association study approaches. The blue color dots were used to represent the QTNs near previously reported genes (Glyma05g34120, GmCRY1, and GmCPK11), and the red color dots were used to represent the QTNs near predicted candidate genes in this study, whereas light blue and light green color dots were used to indicate the negative log10(p-value) of each marker on the adjacent chromosomes in the first step of multi-locus approaches. The bold text candidate genes were previously reported in soybean and the remaining newly identified in this study.
# Discussion
To dissect the genetic basis of 100-SW and provide SW increasing alleles for molecular breeding in soybean, the 100-SW phenotypes of 250 soybean accessions in four environments were used to be associated with 43,834 SNP markers using seven GWAS approaches in this study. As a result, we obtained some valuable results in two aspects. On one hand, 43 stable QTNs were identified, and showed significant differences of 100-SW between the two alleles (Figure S1-S6; Tables 3 and S8). Using the above 43 stable QTN information, new cross combinations were predicted , and a number of SSR markers were obtained from overlapping and previously published QTLs and comparative genomics analysis [fig_ref] Table 6: Stable QTNs of soybean 100-seed weight in this study that are reported... [/fig_ref]. Thus, these SSR markers can be used to conduct marker-assisted selection in soybean breeding. Based on the above 43 stable QTNs, on the other hand, multi-omics analyses were used to mine candidate genes. As a result, six candidate genes were obtained in this study. Among the six candidate genes, Glyma05g34120, Glyma06g10830 (GmCRY1), and Glyma06g16920 (GmCPK11) were found to be associated with soybean 100-SW, respectively, in Li et al. [bib_ref] Genomewide association study of four yield-related traits at the R6 stage in..., Li [/bib_ref] , Du et al. [bib_ref] Identifcation of regulatory networks and hub genes controlling soybean seed set and..., Du [/bib_ref] , and Aghamirzaie et al. [bib_ref] Transcriptome-wide functional characterization reveals novel relationships among differentially expressed transcripts in developing..., Aghamirzaie [/bib_ref] , and Glyma07g07850, Glyma10g03440, and Glyma10g36070 were new in soybean [fig_ref] Table 5: Predicted potential candidate genes for 100-seed weight near the stable QTNs in... [/fig_ref]. These new 100-SW-related candidate genes are valuable in soybean molecular biology research.
## Comparison of stable qtns in this study with previously reported qtls
Up to now, more than 200 QTLs/QTNs were identified by QTL mapping and/or GWAS to be associated with seed weight in soybean (http://soybase.org/). Thus, it is possible to compare these results with 43 stable QTNs in this study. As a result, 33 stable QTNs were found to be located in the genomic regions of previously reported QTLs underlying seed weight [fig_ref] Table 6: Stable QTNs of soybean 100-seed weight in this study that are reported... [/fig_ref]. For example, two stable QTNs qHSW-2-1 and qHSW-2-2 were located simultaneously in the genomic region of one known QTL SW 49-8 [bib_ref] QTL analyses of seed weight during the development of soybean, Teng [/bib_ref] , whereas one stable QTN qcHSW-1-1 was overlapped with previously reported QTL SW 15-2 [bib_ref] Seed quality QTL in a prominent soybean population, Hyten [/bib_ref]. Two stable QTNs qHSW-7-2 and qHSW-11-1 were found in the genomic region of known QTLs SW 49-15 [bib_ref] QTL analyses of seed weight during the development of soybean, Teng [/bib_ref] and SW 37-9 [bib_ref] Multi-environment mapping and meta-analysis of 100-seed weight in soybean, Sun [/bib_ref] , respectively. Interestingly, five stable QTNs qHSW4-3, qcHSW-6-3, qHSW-10-3, qHSW-17-1, qHSW-20-2 were identified in the genomic region of previously reported QTLs SW 45-3 [bib_ref] Identification and validation of an over-dominant QTL controlling soybean seed weight using..., Yan [/bib_ref] , SW 4-1, SW 25-4 [bib_ref] Molecular-marker analysis of seed-weight: Genomic locations, gene action, and evidence for orthologous..., Maughan [/bib_ref] , SW 23-1 [bib_ref] QTL analysis of major agronomic traits in soybean, Chen [/bib_ref] , SW 42-2 [bib_ref] Exploration of presence/absence variation and corresponding polymorphic markers in soybean genome, Wang [/bib_ref] , and SW 50-16 [bib_ref] A major and stable QTL associated with seed weight in soybean across..., Kato [/bib_ref] , respectively. In previous study, Teng et al. [bib_ref] QTL analyses of seed weight during the development of soybean, Teng [/bib_ref] identified QTL SW 49-10 underlying seed weight, and this QTL was consistent with our two stable QTNs qHSW-17-5 and qHSW-17-6 [fig_ref] Table 6: Stable QTNs of soybean 100-seed weight in this study that are reported... [/fig_ref]. In addition, the QTLs reported in Han et al. [bib_ref] QTL analysis of soybean seed weight across multigenetic backgrounds and environments, Han [/bib_ref] , Hyten et al. [bib_ref] Seed quality QTL in a prominent soybean population, Hyten [/bib_ref] , Kato et al. [bib_ref] A major and stable QTL associated with seed weight in soybean across..., Kato [/bib_ref] , Li et al. [bib_ref] Identification of QTL underlying soluble pigment content in soybean stems related to..., Li [/bib_ref] , Teng et al. [bib_ref] QTL analyses of seed weight during the development of soybean, Teng [/bib_ref] , Yan et al. [bib_ref] Identification and validation of an over-dominant QTL controlling soybean seed weight using..., Yan [/bib_ref] , and Yao et al. [bib_ref] Analysis of quantitative trait loci for main plant traits in soybean, Yao [/bib_ref] were found as well to be consistent with our stable QTNs in this study. Therefore, 33 stable QTNs were overlapped with previously reported QTLs, indicating the accuracy of our QTN detection. More importantly, 10 stable QTNs (qHSW-3-1, qHSW-3-2, qHSW-3-3, qHSW-4-1, qHSW-7-4, qcHSW-10-1, qHSW-14-2, qHSW-15-1, qHSW-16-1, and qHSW- were newly identified in this study.
## Reliability of qtns and application of sw increasing allele in soybean breeding
In this study, 218 significant QTNs were identified to be associated with 100-SW in soybean. Among them, a total of 43 QTNs were repeatedly identified in more than three environments and/or methods, and viewed as stable QTNs. Of these stable QTNs, 36 QTNs were identified in at least three environments/BLUP by multiple methods and 22 QTNs were detected by at least three methods in multiple environments. Among them, eight were detected in one environment by at least three ML-GWAS methods, whereas three QTNs were detected by one ML-GWAS method in at least three environments/BLUP [fig_ref] Table 3: Stable QTNs for soybean 100-seed weight identified in multiple environments and the... [/fig_ref]. The QTNs found across different environments are reliable, i.e., Zhou et al.repeatedly detected 31 QTNs associated with 100-SW in 185 soybean accessions in multiple environments, while Li et al. [bib_ref] Genomewide association study of four yield-related traits at the R6 stage in..., Li [/bib_ref] identified 35 QTNs associated with soybean yield traits in at least three environments. Likewise, the QTNs identified by multiple methods are also reliable when several multi-locus approaches are used to evaluate the same dataset [bib_ref] Editorial: The applications of new multi-locus gwas methodologies in the genetic dissection..., Zhang [/bib_ref]. For example, 58 QTNs associated with embryonic callus-related traits have been detected by at three multi-locus methods in Ma et al. [bib_ref] Genetic dissection of maize embryonic callus regenerative capacity using multi-locus genome-wide association..., Ma [/bib_ref] , seven QTNs associated with starch pasting properties-relate traits in maize have been identified by more than one method in Xu et al. [bib_ref] Genome-wide association mapping of starch pasting properties in maize using single-locus and..., Xu [/bib_ref] , and all the 56 QTNs associated with seven salt tolerance-related traits have been determined by at least three multilocus methods in Cui et al. [bib_ref] The application of multi-locus GWAS for the detection of salt-tolerance loci in..., Cui [/bib_ref].
To verify the reliability of each stable QTN, we divided the 250 accessions into two groups based on their allelic types and compared the mean phenotypic values of both alleles. As a result, forty-three QTNs exhibited significant differences of 100-SW between the two alleles (Figures S1-S6), suggesting the reliability of QTNs identified in this study. More importantly, these SW increasing alleles can be utilized in molecular breeding [bib_ref] Genomewide association study of four yield-related traits at the R6 stage in..., Li [/bib_ref] [bib_ref] Association mapping for seed size and shape traits in soybean cultivars, Niu [/bib_ref] [bib_ref] Identification of domestication-related loci associated with flowering time and seed size in..., Zhou [/bib_ref] [bib_ref] Genome-wide association mapping for seed protein and oil contents using a large..., Li [/bib_ref].
In this study, the average number of SW increasing alleles per accession was 18.42, indicating the predominance of SW increasing alleles in cultivars after the disappearance of some alleles during artificial or natural selection process. Based on these SW increasing and decreasing alleles, we also predict some parental combinations. In these combinations, the cultivars Ribendaheidou, Mayidan, and Heibidou are repeatedly present. These predictions might be valuable for the following reasons. Firstly, three selected parents, Yixingwuhuangdou, Quxiandahuangdou, and Nannong 95C-5, were also predicted as parents for seed size traits based on the effects of elite alleles in Niu et al. [bib_ref] Association mapping for seed size and shape traits in soybean cultivars, Niu [/bib_ref]. Secondly, some selected parents have been widely planted in some areas owing to their high yield, e.g., Fujiandadou, Nannong 95C-5, Yixingwuhuangdou, Quxiandahuangdou, and Ribendaheidou. The similar idea can be found in rice breeding, e.g., Wang et al. [bib_ref] QTL mapping of grain length in rice (Oryza sativa L.) using chromosome..., Wang [/bib_ref] developed a japonica cultivar (chromosome segment substitution line) for large grain (>8.5 mm grain length × 3.2 mm grain width) by molecular breeding, demonstrating that elite alleles from different cultivars can be pyramided into a new cultivar.
## Candidate genes underlying stable qtns for seed weight
Identification of candidate genes underlying stable QTNs is of great interest for practical plant breeding and is necessary for further gene cloning and functional verification. To date, only a few seed-weight-related genes have been identified based on association mapping in soybean. Based on functional annotations, available literature, comparative genome analysis, KEGG pathways, and gene expression data, the present study mined candidate genes regulating seed size/weight and development in soybeans. Among 774 genes within the physical regions of 43 stable QTNs, therefore, 36 genes were considered as candidate genes to be involved in seed size/weight and development. Among the 36 candidate genes [fig_ref] Table 1: Phenotypic analysis of soybean 100-seed weight in 250 accessions in four environments [/fig_ref] , nine were found in KEGG pathway analysis, 29 had significantly higher expressed at seed developmental stages, and there were six common genes between the nine and twenty-nine genes (Table 5; . Among the six candidate genes, Glyma05g34120, Glyma06g16920 (GmCPK11), and Glyma06g10830 (GmCRY1) have been reported to directly control seed weight in soybean and to have seed size/weight related functions [bib_ref] Genomewide association study of four yield-related traits at the R6 stage in..., Li [/bib_ref] [bib_ref] Identifcation of regulatory networks and hub genes controlling soybean seed set and..., Du [/bib_ref] [bib_ref] Transcriptome-wide functional characterization reveals novel relationships among differentially expressed transcripts in developing..., Aghamirzaie [/bib_ref]. Therefore, these candidate genes are very reliable and useful in the improvement of 100-SW in soybean.
The other three candidate genes are newly identified in this study. Glyma07g07850 is homologous to BSK3, one brassinosteroid (BR) biosynthesis or signaling gene in Arabidopsis. The BSK3 gene has a decisive role in the initial steps of the BR signal transduction pathway [bib_ref] Brassinosteroid signal transduction: From receptor kinase activation to transcriptional networks regulating plant..., Clouse [/bib_ref] , and mutant bsk3 has been known to play an important role in seed size. BRs are plant hormones that regulate plant growth and development, and the deficiency of this hormone causes abnormal plant growth and hence yield reduction [bib_ref] C-23 hydroxylation by Arabidopsis CYP90C1 and CYP90D1 reveals a novel shortcut in..., Ohnishi [/bib_ref]. Physiological, cellular, and molecular mechanisms influencing plant growth and yield production also indicate the diverse role of BR in plant growth and development [bib_ref] The mechanisms of brassinosteroids' action: From signal transduction to plant development, Yang [/bib_ref]. In addition, some factors have been found to affect the soybean 100-SW, such as seed size, hormones (ABA, BRs, GA3, and IAA), enzymes, silique development, cell growth rate, cotyledon cell number, pollen development and cell volume [bib_ref] Molecular markers associated with seed weight in two soybean populations, Mian [/bib_ref] [bib_ref] Cotyledon cell number and cell size in relation to seed size and..., Hirshfield [/bib_ref].
Glyma10g03440 encodes 3-Methylcrotonyl CoA carboxylase (MCCA), which is a nuclearencoded biotin-localized enzyme and also plays an important role in leucine and isoprenoids catabolism. In Arabidopsis, knockout alleles of the MCCA gene and metabolite study suggest that MCCA mutations block mitochondrial leucine catabolism, which is associated with reduced reproductive growth phenotype, including abnormal flower and silique development [bib_ref] Genetic dissection of methylcrotonyl CoA carboxylase indicates a complex role for mitochondrial..., Ding [/bib_ref]. Glyma10g36070 encodes ribosomal protein L21 (RPL21) that is required for chloroplast and pollen development and embryogenesis in Arabidopsis [bib_ref] The chloroplast ribosomal protein L21 gene is essential for plastid development and..., Yin [/bib_ref]. However, these candidate genes require further functional validation/cloning to determine their actual role in seed weight in soybean.
## Statistical power of multi-locus gwas approaches
In this study, 218 significant QTNs for 100-SW in soybean were detected from six ML-GWAS approaches. These significant QTNs were divided into four groups. In the first group, all the QTNs are both msQTNs and esQTNs. All the QTNs in the second group are esQTNs rather than msQTNs, while all the QTNs in the third group are msQTNs rather than esQTNs. In the last group, all the QTNs are neither esQTNs nor msQTNs. Thus, we summarized their characteristics of the above four groups, such as the number of significant QTNs, the average of absolute effects, LOD score and r 2 , and the proportion of previously reported QTNs in . As a result, it is easy (the highest proportion of previously reported QTNs) to identify the QTNs in the first group and these QTNs have the largest values for QTN effects, LOD scores and r 2 , while it is relatively difficult (the lowest proportion of previously reported QTNs) to detect the QTNs in the last group and these QTNs have relatively small values for QTN effects, LOD scores, and r 2 . The above results show the advantage of our multi-locus GWAS approaches in detecting small-effect QTNs. The results support our previous recommendation that the QTNs identified by individual approaches or in individual environments are valuable in mining the genes for the trait of interest [bib_ref] Editorial: The applications of new multi-locus gwas methodologies in the genetic dissection..., Zhang [/bib_ref]. . Comparison of four kinds of QTNs for soybean 100-seed weight in this study.
## Group
No. of QTNs Absolute Effect LOD Score r 2 (%) % Known QTNs 1 In addition, we also found the gap between the trait heritability (83.23-93.70%) and the sum of r 2 (24.13-35.52%) for all the QTNs identified by each approach in one environment or BLUP. This is the heritability missing in GWAS [bib_ref] Editorial: The applications of new multi-locus gwas methodologies in the genetic dissection..., Zhang [/bib_ref]. The possible reasons are the exclusion of QTN-byenvironment and QTN-by-QTN interactions in this study. Thus, it is necessary to develop the methodologies for detecting QTN-by-environment and QTN-by-QTN interactions in the near future.
# Conclusions
In this study, 43 stable QTNs were detected in at least three environments/BLUP and/or by at least three ML-GWAS methods, and they showed significant differences of 100-SW between the two alleles in the GWAS population. Using these SW increasing or decreasing alleles of stable QTNs, the best five cross combinations were predicted in large or small seed directions. Among the 36 potential candidate genes from multi-omics analysis, Glyma05g34120, GmCRY1, and GmCPK11 are the known seed-size-related genes in soybean, and Glyma07g07850, Glyma10g03440, and Glyma10g36070 were identified to be candidate genes in this study.
Supplementary Materials: The following are available online at www.mdpi.com/2073-4425/11/7/714/s1, Figures S1 to S6: Phenotypic differences of 100-seed weight between accessions carrying different alleles of each QTN. [fig_ref] Figure 1: Phenotypic distributions of 100-seed weight in 250 soybean accessions in four environments... [/fig_ref] : these QTNs include qcHSW-1-1 (A), qHSW-2-1 (B), qHSW-2-2 (C), qHSW-3-1 (D), qHSW-3-2 (E), qHSW-3-3 (F), qHSW-4-1 (G), and qHSW-4-2 (H). [fig_ref] Figure 2: Population structure of 250 soybean accessions using 16,174 SNPs on soybean genome [/fig_ref] : these QTNs include qHSW-4-3 (A), qHSW-5-1 (B), qHSW-5-2 (C), qHSW-6-1 (D), qHSW-6-2 (E), qcHSW-6-3 (F), and qHSW-6-4 (G).: these QTNs include qHSW-7-1 (A), qHSW-7-2 (B), qcHSW-7-3 (C), qHSW-7-4 (D), qHSW-8-1 (E), qcHSW-10-1 (F), qcHSW-10-2 (G), and qHSW-10-3 (H).: these QTNs include qcHSW-10-4 (A), qHSW-11-1 (B), qHSW-11-2 (C), qHSW-11-3 (D), qHSW-11-4 (E), qHSW-14-1 (F), qHSW-14-2 (G). : The expressional levels [log2(RPKM + 1)] of candidate genes associated with seed weight in seven soybean tissues. Tables S1: Significant QTNs for 100-seed weight detected in four environments and BLUP model using mrMLM (S1). [fig_ref] Table 2: Summary of QTNs identified in four environments and their BLUP values using... [/fig_ref] : Significant QTNs for 100-seed weight detected in four environments and BLUP model by using FASTmrMLM. [fig_ref] Table 3: Stable QTNs for soybean 100-seed weight identified in multiple environments and the... [/fig_ref] : Significant QTNs for 100-seed weight detected in four environments and BLUP model by using FASTmrEMMA. : Significant QTNs for 100-seed weight detected in four environments and BLUP model by using pLARmEB. [fig_ref] Table 5: Predicted potential candidate genes for 100-seed weight near the stable QTNs in... [/fig_ref] : Significant QTNs for 100-seed weight detected in four environments and BLUP model by using pKWmEB. [fig_ref] Table 6: Stable QTNs of soybean 100-seed weight in this study that are reported... [/fig_ref] : Significant QTNs for 100-seed weight detected in four environments and BLUP model by using ISIS EM-BLASSO. : Significant QTNs for 100-seed weight detected in four environments and BLUP model by using MLM. . Stable QTNs of soybean 100-seed weight identified in multiple environments and/or by multiple methods. : Distribution of SW increasing alleles in stable QTNs among 250 soybean accessions. [fig_ref] Table 1: Phenotypic analysis of soybean 100-seed weight in 250 accessions in four environments [/fig_ref]. Predicted candidate genes for seed weight near the stable QTNs in soybean.
## Conflicts of interest:
The authors declare that they have no conflict of interest.
[fig] Figure 1: Phenotypic distributions of 100-seed weight in 250 soybean accessions in four environments and BLUP. SD: standard deviation. The significant differences of 100-seed weight among four environments/BLUP are tested by the LSD method at the 0.05 level of significance. [/fig]
[fig] Figure 2: Population structure of 250 soybean accessions using 16,174 SNPs on soybean genome. (A) population structure (k = 3). Each vertical column represents one individual, and the red, green, and blue color segments in each column represents the percentages of cultivated, landrace, and mixture subgroups, respectively; (B) the determination for the number of subgroups via the Delta method of Evanno et al.[55]. [/fig]
[fig] Figure 3: Stable QTNs of soybean 100-seed weight identified in four environments and BLUP values by using six ML-GWAS methods. The black color dots were the stable QTNs near previously reported genes (Glyma05g34120, GmCRY1, and GmCPK11), while the red color dots were the stable QTNs near predicted candidate genes in this study. The sky blue color dots were other stable QTNs, while the light green color dots were remaining significant QTNs. [/fig]
[fig] 1: 号) × Mayidan (蚂蚁蛋) Qingcha 1 (庆茶-1) × Mayidan (蚂蚁蛋) Qinyan 1 (勤研 1 号) × Heibidou (黑鼻青) Mingshanhongxingjiroudou (名山红星鸡肉豆) × Mayidan (蚂蚁蛋) [/fig]
[fig] Figure 4: The expressional levels [log2(RPKM + 1)] of potential candidate genes associated with seed weight in seven soybean tissues. Among the nine genes, three were previously reported seedweight genes (bold text) in soybean, three were lowly expressed genes (grey text), and three were newly identified as candidate genes (red text) to be related to seed size/weight and development. [/fig]
[fig] Funding: This work was supported by the National Natural Science Foundation of China (31871242, U1602261, and 31571268), Huazhong Agricultural University Scientific & Technological Self-Innovation Foundation (2014RC020), and State Key Laboratory of Cotton Biology Open Fund (CB2019B01). [/fig]
[table] Table 2: Summary of QTNs identified in four environments and their BLUP values using six ML-GWAS methods. [/table]
[table] Table 3: Stable QTNs for soybean 100-seed weight identified in multiple environments and the BLUP model. [/table]
[table] Table 5: Predicted potential candidate genes for 100-seed weight near the stable QTNs in soybean. [/table]
[table] Table 6: Stable QTNs of soybean 100-seed weight in this study that are reported in previous studies. [/table]
|
Hepatobiliary case report and literature review of hepatic reactive lymphoid hyperplasia with positive anti-smooth muscle antibody and anti-nuclear antibody tests
Hepatic reactive lymphoid hyperplasia (RLH) is a rare benign tumor of the liver that exhibits similar imaging characteristics to those of other hepatic malignant tumors; therefore, it requires novel biomarkers to be differentiated from the other tumors. A 69-year-old female was found to have a mass in the liver and was admitted to Peking Union Medical College Hospital for further evaluation. Most laboratory tests, including liver function tests, tumor biomarkers, and autoimmune markers were within normal range, except for positive antinuclear antibody and anti-smooth muscle antibody (SMA) tests. Diagnostic imaging, including ultrasonography, computerized tomography and magnetic resonance imaging displayed a small hepatic mass suggestive of hepatocellular carcinoma (HCC). Partial hepatectomy was performed, and histological diagnosis suggested RLH. Postoperative treatments included anti-infection, nutritional enhancement and liver protection. The suspicion of autoimmune hepatitis could not be confirmed. No recurrence or autoimmune disease was observed over 6-month follow-up. Positive anti-nuclear antibody (ANA) and anti-SMA may be potential biomarkers for hepatic RLH. Sang XT, Du SD. Hepatobiliary case report and literature review of hepatic reactive lymphoid hyperplasia with positive antismooth muscle antibody and anti-nuclear antibody tests. Transl
# Introduction
Reactive lymphoid hyperplasia (RLH) is a benign nodular lesion, clinically recognized as a pseudolymphoma, affecting various organs including the gastrointestinal tract, orbit, lung, skin and thyroid [bib_ref] Reactive lymphoid hyperplasia of the liver and pancreas. A report of two..., Amer [/bib_ref]. Histopathological characteristics of RLH include marked proliferation of nonneoplastic, polyclonal lymphocytes that form follicles with active germinal centers. Most reported RLH cases are associated with other malignant gastrointestinal tumors, hepatic virus infections or autoimmune diseases. This case report discusses a case of RLH in liver that was characterized by mostly normal laboratory tests except for positive antismooth muscle antibody (SMA) and anti-nuclear antibody (ANA) tests and that had no direct complications.
## Case presentation
## Chief complaints
A 69-year-old female was admitted for further evaluation of an evident hepatic mass discovered one month earlier.
The patient initially complained of intermittent abdominal distention without other discomfort. distention without other discomfort for two months, and CT scan showed an evident hepatic mass discovered one month before she was admitted into hospital.
## History of past illness
The patient had a past history of hypertension for 40 years and type II diabetes mellitus for 10 years. She had no history of HBV or tuberculosis infection.
## Physical examination
Physical examination revealed no abnormalities.
## Laboratory examinations
Routine laboratory tests revealed normal whole blood cell counts and albumin/globulin ratio 0.9 (normal range, 1.5-2.5). All liver functional tests were normal, including serum aspartate aminotransferase (AST), alanine transaminase (ALT), total bilirubin (TBil) and prothrombin time (PT). An antinuclear antibody-immunofluorescence (ANA-IF) test was positive for nucleolar type at 1:640 dilution and for cytoplasmic type at 1:80 dilution. Tumor marker titers, including α-fetoprotein (AFP), carbohydrate antigen 242 (CA242), carcinoembryonic antigen (CEA) and carbohydrate antigen 19-9 (CA19-9) were within the normal range. Hepatopathic autoimmune antibodies (SLA, AMA, PCA, LKM, HRA, AMA-M2, and ACA) were negative, while SMA was positive at 1:320 dilution, and serum immunoglobulin IgG (18.2 g/L, normal 7-17 g/L) and IgA (4.58 g/L, normal 0.7-3.8 g/L) were slightly elevated. Antibodies against extractable nuclear antigens (ENA) including Smith, RNP, Ro (SSA), La (SSB), Scl-70, Jo-1, rRNP, and antibodies against neutrophil cytoplasmic antibodies (ANCA) including RP3-ANCA, IF-ANCA and MPO-ANCA, were negative. Thyroid parameters were within normal range.
## Imaging examinations
The details of this neoplasm are displayed in dynamic CT scan and MRI [fig_ref] Figure 1: Dynamic CT scan revealing a solid, finely delineated mass in segment II [/fig_ref]. The nodule was significantly enhanced in the arterial phase, followed by a slight attenuation in the portal phase. MRI suggested differentiation of this small hypervascular lesion from hepatocellular carcinoma (HCC), focal nodular hyperplasia (FNH) and adenoma. Based on radiographic observation indicating malignancy, diagnosis of HCC was considered.
## Diagnosis before surgery
The diagnosis of the presented case was HCC before surgery.
## Treatment
The patient underwent partial resection of the left lateral hepatic lobe with cholecystectomy.
## Outcome and follow-up
The resected liver tissue exhibited a normal smooth surface appearance. A gray-white, well-circumscribed firm nodule 1.2×0.7×1 cm 3 in size was located within the removed section. The remaining hepatic parenchyma presented a
## A b c
gray-brown appearance, solid but relatively soft in quality. The S2 nodule was composed of many lymphoid follicles with large germinal centers containing small polymorphic lymphocytes. Peri-nodular hepatocytes in the peripheral hepatic tissue were partially swollen. Infiltration of a few lymphocytes and local bile ductile hyperplasia were observed in the portal area. A gallstone was found in the gall bladder along with multiple cholesterol polyps, and chronic cholecystitis.
Immunohistochemical studies indicated intrafollicular lymphocytes were Bcl-2-negative, CD10-positive, CD68positive, CD3-positive and CD20-positive. The germinal center Ki-67 index was 90%. These findings suggested polyclonal origins of the infiltrating lymphocytes.
The patient received routine postoperative anti-infection agents, nutritional support and liver protection treatments. No recurrence or other combination were observed during 6-months of follow-up.
# Discussion
RLH is a small, benign, tumorous lesion rarely found in the liver, but more often in other organs, including the gastrointestinal tract, orbit, lung and skin [bib_ref] Reactive lymphoid hyperplasia of the liver and pancreas. A report of two..., Amer [/bib_ref]. From 1981 to the end of 2015, there have been fewer than 60 cases of hepatic RLH reported in the literature (2,3), suggesting its extreme rarity. Nevertheless, the real incidence of hepatic RLH may be higher than is reported in the literature, due to insignificant symptoms and failure to detect or diagnose. Interestingly, increasing numbers of RLH reports have been published in recent years, suggesting an improvement in the differential diagnosis of hepatic tumor mass and the advancement of laboratory and imaging examination.
RLH usually presents as a single small nodule, ranging from 4-27 mm in size, and mostly occurs in middle-aged women with an average age of 56.8 (range, 15-77) years old in East Asian countries. Hepatic RLH resembles HCC or metastatic neoplasm in liver on imaging findings. Various imaging methods, including US, CT, MRI and PET have been commonly used for diagnosis and differentiation from HCC or metastasis [bib_ref] Primary lymph node ratio and hepatic resection for colorectal metastases, Ansari [/bib_ref] [bib_ref] Association between the lymph node ratio and hepatic tumor burden: importance for..., Brudvik [/bib_ref]. RLH lesions often appear as hypoechoic masses on abdominal US (7-9). Plain CT generally reveals areas of low-density compared to the surrounding hepatic tissue. Contrast CT generally reveals tumor enhancement in the arterial phase with substantial attenuation in sequential phases, or no enhancement in any phase. A majority of MRI results reveal a low signal intensity lesion in the T1weighted image and a high signal intensity lesion in the T2weighted image [bib_ref] MR imaging of reactive lymphoid hyperplasia of the liver, Kobayashi [/bib_ref] [bib_ref] Reactive lymphoid hyperplasia of the liver: literature review and 3 case reports, Yoshikawa [/bib_ref]. Hepatic artery angiography reveals a tumor staining in the corresponding area, indicating its hypervascularity [bib_ref] MR imaging of reactive lymphoid hyperplasia of the liver, Kobayashi [/bib_ref] [bib_ref] Reactive lymphoid hyperplasia of the liver: literature review and 3 case reports, Yoshikawa [/bib_ref]. Some patients are considered for CT arterial portography (CTAP), during which a perfusion defect is often identified [bib_ref] Reactive lymphoid hyperplasia of the liver: a case report and review of..., Machida [/bib_ref]. A tumor-rim was found in 6 cases via various radiographic modalities [bib_ref] MR imaging of reactive lymphoid hyperplasia of the liver, Kobayashi [/bib_ref] [bib_ref] Reactive lymphoid hyperplasia of the liver: literature review and 3 case reports, Yoshikawa [/bib_ref]. Radiological findings are also comparable with those of vascular hepatic tumors, including HCC or metastasis. Therefore, patients are often misdiagnosed as having HCC or metastatic tumors [bib_ref] MR imaging of reactive lymphoid hyperplasia of the liver, Kobayashi [/bib_ref] [bib_ref] Reactive lymphoid hyperplasia of the liver: literature review and 3 case reports, Yoshikawa [/bib_ref] [bib_ref] Reactive lymphoid hyperplasia of the liver associated with primary biliary cirrhosis, Fukuo [/bib_ref]. There is a lack of efficiency in hepatic RLH differentiation from malignancy using current imaging approaches.
The pathogenesis of RLH remains unclear. An association between the development of hepatic RLH and systemic or local immunological abnormalities has been suggested in earlier reports [bib_ref] Pseudolymphoma (reactive lymphoid hyperplasia) of the liver: A clinical challenge, Kwon [/bib_ref] [bib_ref] REPORT: Reactive lymphoid hyperplasia of liver coexisting with chronic thyroiditis: Radiographical characteristics..., Nagano [/bib_ref]. In fact, some cases have been complicated with immunologic abnormalities such as autoimmune thyroiditis, Sjogren's syndrome, CREST and primary biliary cirrhosis A B [bib_ref] MR imaging of reactive lymphoid hyperplasia of the liver, Kobayashi [/bib_ref] [bib_ref] Reactive lymphoid hyperplasia of the liver: literature review and 3 case reports, Yoshikawa [/bib_ref] [bib_ref] Reactive lymphoid hyperplasia of the liver: a case report and review of..., Machida [/bib_ref] , while other reported cases have been complicated with cancer, such as gastric cancer, renal cell cancer and colon cancer, suggesting a potential association between RLH and malignant tumors [bib_ref] Histiocyte-rich reactive lymphoid hyperplasia of the liver: unusual morphologic features, Park [/bib_ref] [bib_ref] Reactive lymphoid hyperplasia of the liver in a patient with colon cancer:..., Takahashi [/bib_ref]. Furthermore, 4 cases of hepatic RLH were diagnosed in hepatitis patients (ADIH, CHB, and CHC). Lymphoproliferative diseases have been reported to be associated with chronic liver disease. Lymphocyte infiltration in the portal tracts is frequently observed in chronic liver diseases such as chronic active hepatitis (CAH) and liver cirrhosis, thus supporting the hypothesis of continuity between CAH and pseudolymphoma in RLH development and progression [bib_ref] A case of pseudolymphoma of the liver with chronic hepatitis C, Kim [/bib_ref]. The unique properties of this case, compared with previous reports, include negative history of chronic hepatitis or other liver disease, positive ANA and SMA test, negative laboratory and histologic findings that did not indicate autoimmune hepatitis, primary biliary cirrhosis or other systemic autoimmune disease. Nevertheless, considering the fact that positive ANA and SMA results are indeed indicating factors for diagnosis of autoimmune hepatitis, it is also possible that the patient was in a very early asymptomatic stage of autoimmune hepatitis development; therefore, longterm follow up is required. However, within 6 months of the surgery, the patient did not exhibit any symptoms relevant to autoimmune hepatitis; further follow-up is underway. Another possibility is that the positive ANA and SMA tests could be specific autoimmune biomarkers for some subtype of RLH in liver, perhaps useful for diagnosis and differentiation of hepatic RLH. Due to the lack of data on ANA and SMA in hepatic RLH in previous studies, more tests on ANA and SMA are needed to further validate their correlation with hepatic RLH.
In respect to treatment, partial hepatectomy is often considered in most cases. If established RLH diagnosis is obtainable before surgery, conservative treatments, including attempted treatment of primary disease (autoimmune abnormality or chronic hepatitis) with close observation of disease progression from RLH to malignancy is plausible.
The differential diagnosis includes inflammatory pseudotumor (IPT) and mucosa-associated lymphoid tissue (MALT) lymphoma. RLH usually lacks typical clinical and pathologic features of IPT, including fever, myofibroblastic proliferation, xanthogranulomatous change and occlusive portal phlebitis [bib_ref] Reactive lymphoid hyperplasia of the liver in a patient with multiple carcinomas:..., Sato [/bib_ref]. Primary MALT lymphoma of the liver is also rare, most characteristic of clonal B-lymphoproliferative disorder. Evidence of B cell clonality is based on immunoglobulin light chain restriction detected through immunohistochemistry. Only in a single case was clonality of lymphocytes assessed through molecular gene rearrangement studies [bib_ref] Nodular lymphoid lesion of the liver with simultaneous focal nodular hyperplasia and..., Willenbrock [/bib_ref]. Our immunohistochemical tests indicated germinal centers to be negative for Bcl-2, with mixed infiltrates of B-and T-lymphocytes, indicating that they were reactive and nonneoplastic. Polymerase chain reaction (PCR) analysis for the immunoglobulin gene is useful for detecting clonal B-cell populations. PCR analysis may be considered if immunohistochemistry does not provide enough evidence to prove polyclonality.
Generally, RLH is difficult to diagnose before surgery and without histological testing. Autoimmune antibodies, including ANA and SMA, can be tested before surgery and may provide some useful information regarding diagnosis and treatment. RLH is considered a benign process; patients generally experience smooth uneventful postoperative recovery with no recurrence. Nevertheless, the risk of malignant transformation to lymphoma has been speculated. Therefore, regular followup of hepatic RLH patients is crucial for the prevention of potential development of lymphoma.
# Conclusions
Hepatic RLH is a rare disease most commonly seen in middleaged females, characterized by small lesion size, concomitant immunological disorder or carcinoma, and malignant tumorlike radiographic findings. Comprehensive examinations, including autoimmune antibody tests and differential diagnosis should be performed to achieve the diagnosis of RLH. Surgical intervention and histological diagnosis are currently the standard means of accurate treatment and diagnosis.
[fig] Figure 1: Dynamic CT scan revealing a solid, finely delineated mass in segment II (A, arrow), 1.0 cm in diameter, with slight arterial phase enhancement (B, arrow) followed by an obvious washout in the venous phase (C, arrow). Abdominal ultrasonography indicating a hypoechoic liver mass of 1.3 cm × 1.2 cm in size on the left lobe without blood signal, which was evenly and rapidly enhanced in the arterial phase through contrast but hypoechoic in the delayed phase. Dynamic CT scan revealing a solid, clearly-delineated mass of 1.0 cm in diameter, with slight arterial phase enhancement followed by an obvious washout in the venous phase. [/fig]
[fig] Figure 2: MRI displaying a round-like nodule in segment II (S2), 9.6 mm × 8.9 mm in size, with a slightly hyperintense signal on T1-weighted imaging (A) and an iso-intense signal on T2-weighted imaging (B). Moreover, a hyperdense nodule in segment VIII compatible with calcification, and a hyperdense mass in the cholecyst compatible with gallbladder stone, were found. MRI displayed a round nodule in segment II (S2), 9.6 mm × 8.9 mm in size, with a slightly hyperintense signal on T1-weighted imaging and an iso-intense signal on T2-weighted imaging. [/fig]
|
Postoperative radiotherapy is effective in improving survival of patients with stage pIII-N2 non-small-cell lung Cancer after pneumonectomy
Background: There were few reports of postoperative radiotherapy (PORT) in stage pIII-N2 Non-small-cell Lung Cancer (NSCLC) patients receiving pneumonectomy followed by adjuvant chemotherapy. This study aims to evaluate safety and efficacy of PORT among these patients.Methods: Between Jan. 2004 and Dec. 2015, stage pIII-N2 NSCLC patients receiving pneumonectomy and adjuvant chemotherapy with or without PORT in our institution were retrospectively reviewed.Results: Totally 119 patients were included, 32 patients receiving adjuvant chemotherapy and PORT (PORT group) and 87 receiving adjuvant chemotherapy alone (Control group). There were more patients with non-R0 resection in PORT group than Control group (25% vs. 8%, p = 0.031). In PORT group, ≥Grade 2 radiation-induced pneumonitis was 2/32. No severe radiation-related heart injury was observed. There was no PORT-related death. Of all patients, median follow-up time was 25 months. Median overall survival time (mOS) and median disease-free survival time (mDFS) were 46 months and 15 months, respectively. The PORT group had significantly better OS (not reached vs. 34 months, p = 0.003), DFS (19 months vs. 13 months, p = 0.024), local recurrence free survival (LRFS, p = 0.012), and distant metastasis free survival (DMFS, p = 0.047) than the Control group. As for failure pattern, PORT significantly reduced local regional failure rate (39.1% vs. 15.6%, p = 0.016). In subgroup analysis, patients with R0 resection (n = 104), OS and LRFS in PORT group were significantly longer, and PORT tended to increase DFS and DMFS.Conclusion: For patients with stage pIII-N2 NSCLC after pneumonectomy and adjuvant chemotherapy, PORT can improve OS, DFS, LRFS and DMFS with tolerable toxicity.
# Background
For patients with stage pIII-N2 non-small-cell lung cancer (NSCLC) after surgery followed by adjuvant chemotherapy, postoperative radiotherapy (PORT) is a major treatment modality. Recently, a large number of retrospective studies have confirmed that 5-year overall survival (OS) could be improved by 5-6% with the application of modern radiation techniques such as three-dimensional conformal radiotherapy (3D-CRT) and intensity modulated radiation therapy (IMRT) [bib_ref] Re-evaluation of the role of postoperative radiotherapy and the impact of radiation..., Corso [/bib_ref] [bib_ref] Lymph node ratio may predict the benefit of postoperative radiotherapy in non-small-cell..., Urban [/bib_ref] [bib_ref] Postoperative radiotherapy is associated with better survival in non-small cell lung cancer..., Mikell [/bib_ref] [bib_ref] Postoperative radiotherapy for pathologic N2 non-small-cell lung cancer treated with adjuvant chemotherapy:..., Robinson [/bib_ref]. However, most patients in these studies received lobectomy. At present, pneumonectomy is still an important therapeutic option for central, giant mass or tumors involving main bronchus or large blood vessels [bib_ref] Pneumonectomy for bronchogenic carcinoma: analysis of factors predicting survival, Ramnath [/bib_ref] [bib_ref] The prevention of spread during pulmonary resection by the use of a..., Bjork [/bib_ref] [bib_ref] Pneumonectomy for nonsmall cell lung Cancer, Sugarbaker [/bib_ref] , which may bring benefits to survival. As respiratory and circulatory system load (such as right ventricle) increases after pneumonectomy [bib_ref] Pneumonectomy for bronchogenic carcinoma: analysis of factors predicting survival, Ramnath [/bib_ref] [bib_ref] Mortality, complications and loss of pulmonary function after pneumonectomy vs. sleeve lobectomy..., Melloul [/bib_ref] , tolerance to radiation-induced pneumonitis, esophagitis and cardiac toxicity decreases after radiotherapy. Therefore, for patients with stage pIII-N2 NSCLC after pneumonectomy and adjuvant chemotherapy, safety and efficacy of PORT needs to be evaluated. The study aimed to evaluate the safety and efficacy of PORT for patients with stage pIII-N2 NSCLC after pneumonectomy and adjuvant chemotherapy.
# Methods
## Patients
Between Jan. 2004 and Dec. 2015, patients with stage pIII-N2 NSCLC who underwent pneumonectomy followed by adjuvant chemotherapy with or without PORT were retrospectively analyzed. Clinical features included: age, performance status, sex, smoking index, lung function, etc. Smoking index was defined as the number of cigarettes smoked per day plus years of smoking; preoperative pulmonary function was assessed by the absolute value of FEV1 and the percentage of FVC. Pathological information included: site of primary tumor, pathological type, tumor size, pathological stage, invasion of large vessels, lymphatic vascular invasion (LVSI), number of (positive) lymph nodes dissected and margin status. Invasion of large vessels was defined as the invasion of aorta, trunk of the pulmonary artery, superior and inferior vena cava, and any other intra-pericardial vessels.
## Treatment approaches
All patients received pneumonectomy and mediastinal lymph node dissection. Adjuvant chemotherapy consisted of 4 cycles of platinum-based doublet chemotherapy. PORT was performed according to resection status, KPS and patients' wills with 6MV x-ray linear accelerator using IMRT or 3D-CRT techniques. Clinical target volume (CTV) included bronchial stump, ipsilateral hilum, mediastinum with or without supraclavicular area. Radiation dose was 50-60 Gy given by 1.8-2.0 Gy per fraction. Percentage of contralateral lung volume receiving at least 20Gy (V20) was restricted as equal to or less than 10% and mean dose of the contralateral lung was limited as equal to or less than 8Gy.
## Toxicity and outcome assessment
Radiation toxicities were evaluated by Common Terminology Criteria for Adverse Events Version 4.0(CTCAE). Survival durations were defined as the time from surgery to the date of death due to any cause (overall survival, OS), to treatment failure or death due to any cause (disease free survival, DFS); to locoregional failure or death due to any cause (local recurrence free survival, LRFS) or to distant metastasis or death due to any cause (distant metastasis free survival, DMFS).
# Data analysis
Data analysis was performed by SPSS statistical software (version 20.0; SPSS Inc., Chicago, IL). Chi-square test was used for categorical variables. Normality test for continuous variables was performed with Kolmogorov-Smirnov method. Student-t test was used to assess normal distribution variables, and Mann-Whitney U test was used for non-normal distribution variables. Kaplan-Meier method was performed to calculate survival data, and differences between groups were determined by log-rank test. A p-value of < 0.05 was considered as statistical significance.
# Results
## Patients and tumor characteristics
A total of 119 patients were enrolled. Characteristics were presented in [fig_ref] Table 1: Baseline Clinical Characteristics for Patients Treated with and without PORT Abbreviation [/fig_ref]. All patients had a Karnofsky performance score over 80. The median age was 53 years (range, 27 to 73 years), and 84% (100/119) were male. 78.2% (93/119) underwent left lung resection, and squamous cell carcinoma (74/119, 62.2%) was the main histologic type. According to AJCC 7th Staging System, most patients were in pathological stage IIIA (104/119, 87.4%). All patients received pneumonectomy and mediastinal lymph node dissection. R0-, R1-and R2-resection were 104/119 (87.4%), 9/119 (7.6%) and 6/ 119 (5%), respectively.
After pneumonectomy, 32 patients (26.9%) received adjuvant chemotherapy and radiotherapy (PORT group), while 87 patients (73.1%) underwent adjuvant chemotherapy alone (Control group). Among 32 patients in PORT group, 27 received chemotherapy followed by PORT. The other 5 received PORT followed by chemotherapy. Baseline characteristics between the two groups were comparable (p > 0.05), except that the rate of non-R0 resection (R1 and R2) was significantly higher in PORT group than in Control group (25% vs. 8%, p = 0.031). In PORT group, 25 patients (78.1%) had IMRT and 7 patients (21.9%) had 3D-CRT. The median prescription dose was 54 Gy (range, 30 to 66Gy), and median percentage of contralateral lung volume receiving at least 20Gy (V20) was 4.75% (range, 1-11.5%).
## Toxicities of radiotherapy
93.75% (30/32) patients completed PORT (completing dose above 50Gy) after pneumonectomy. One patient terminated radiotherapy at 30Gy due to chest pain, and the other patient terminated radiotherapy at 40Gy due to brain metastases. Grade 2 and grade 3 radiation pneumonitis was observed in one patient each, and the
## Survival
For all patients, the median follow-up time was 25 months (range, 2-160 months). The median survival time (MST) was 46 months, and the 1-, 3-and 5-year OS were 87, 53.8 and 46.3%, respectively. The median DFS was 15 months, and the 1-, 3-and 5-year DFS were 55, 28.5 and 24.5%, respectively. Besides, the median LRFS and DMFS were 25 months and 19 months, respectively [fig_ref] Table 2: Impact of Different Treatment Models on SurvivalAbbreviation [/fig_ref]. The median OS in PORT group was significantly improved compared with Control group [fig_ref] Figure 1: Effect of postoperative radiotherapy [/fig_ref] , which was not reached and 34 months, respectively (p = 0.003). The [fig_ref] Figure 1: Effect of postoperative radiotherapy [/fig_ref] shows the DFS curves for the both groups. The median DFS were 19 months in PORT group and 13 months in Control group, respectively (p = 0.024). There were significant differences in LRFS and DMFS between PORT group and Control group as shown in [fig_ref] Figure 1: Effect of postoperative radiotherapy [/fig_ref] and d. The median LRFS were 95 months and 22 months (p = 0.012), respectively. The median DMFS were 47 months and 17 months (p = 0.047), respectively. In subgroup analysis, statistically improved OS and LRFS were also found between the two groups in patients with R0 resection. The median OS was not reached versus 34 months (p = 0.032) and the median LRFS was 95 months versus 22 months (p = 0.033), respectively. PORT tended to improve DFS (p = 0.111) and DMFS (p = 0.179), with no crossing of the curves. For patients with non-R0 resection (8 in PORT group, and 7 in Control group), the OS, DFS and LRFS in PORT group was also significantly improved. While there was no significant difference of DMFS between the two groups (p = 0.087).
## Failure pattern
Eighty-two relapses (68.9%) were found. Of all the relapses, 58 (48.7%) were distant metastasis, and 39 (32.8%) were local-regional recurrence. Distant metastasis was the main failure pattern in both PORT group and Control group (50.0% vs. 48.3%, p = 0.867). However, failure occurred in local-regional site in PORT group was significantly lower than in Control group (15.6% vs. 39.1%, p = 0.016) [fig_ref] Table 3: Failure Pattern for Patients Treated with and without PORT [/fig_ref]. Of the 5 patients with local-regional failure in PORT group, 4 were out of radiation field, and 1 (disrupted at 30Gy because of chest pain) within radiation field.
# Discussion
For stage pIII-N2 NSCLC after pneumonectomy followed by adjuvant chemotherapy, the results showed good compliance and safety of PORT without any severe radiation pneumonitis and esophagitis. Compared with the control, PORT significantly improved OS, DFS, LRFS and DMFS, although there were more patients with R1/ R2 resection in PORT group. To our knowledge, this is the first study focusing on the safety and efficacy of PORT for stage pIII-N2 NSCLC after pneumonectomy followed by chemotherapy.
Safety of radiotherapy after pneumonectomy is the most important concern in clinical practice. Pneumonectomy leads to great change on cardiopulmonary function. The removal of an entire lung decreases the tolerance to radiotherapy and amplifies potential risks, leading to few application of radiotherapy. In our study, 32 patients (26.9%) received PORT. In 104 patients with R0 resection, 24 cases (23.1%) were in the PORT group, which was consistent with previous reports (less than 25%) [bib_ref] Pneumonectomy for bronchogenic carcinoma: analysis of factors predicting survival, Ramnath [/bib_ref] [bib_ref] For non-small cell lung cancer with T3 (central) disease, sleeve lobectomy or..., Ma [/bib_ref] [bib_ref] Sleeve pneumonectomy for central non-small cell lung cancer: indications, complications, and survival, Eichhorn [/bib_ref] , and was significantly less than patients receiving PORT after lobectomy (43.4%) [bib_ref] Selection of proper candidates with resected pathological stage IIIA-N2 non-small cell lung..., Hui [/bib_ref]. The postoperative residual tumor in gross or positive margin in microscopy was an important indication of PORT. In this study, although more patients (8/15, 53.3%) with non-R0 resection received PORT comparing with those with R0 resection (23.1%), there were still nearly half of the patients not receiving PORT, which revealed the concern on safety of PORT in both physicians and patients.
Contrary to concerns on the risks of PORT after pneumonectomy, this study showed that PORT was well tolerant with high compliance (completion rate of 93.8%). Only one patient (3.1%) suffered grade 2 radiation pneumonitis and one patient (3.1%) suffered grade 3 radiation pneumonitis, while no severe radiation pneumonitis (≥ grade 4) occurred. The incidence of ≥ grade 2 radiation pneumonitis in this study was significantly lower compared with the toxicities in previous studies of radical chemoradiotherapy (10-35%) [bib_ref] Etoposide and cisplatin versus paclitaxel and carboplatin with concurrent thoracic radiotherapy in..., Liang [/bib_ref] [bib_ref] The predictive role of plasma TGF-beta1 during radiation therapy for radiation-induced lung..., Zhao [/bib_ref] , and previous studies of radiotherapy or chemoradiotherapy after lobectomy [bib_ref] Phase II trial of postoperative adjuvant paclitaxel/carboplatin and thoracic radiotherapy in resected..., Bradley [/bib_ref] [bib_ref] Risk factors for radiation-induced lung toxicity in patients with non-small cell lung..., Zhao [/bib_ref]. Bradly [bib_ref] Phase II trial of postoperative adjuvant paclitaxel/carboplatin and thoracic radiotherapy in resected..., Bradley [/bib_ref] reported ≥grade 3 radiation pneumonitis with the incidence rate of 6% in 88 patients receiving adjuvant chemotherapy and concurrent radiotherapy after surgery, while pneumonectomy accounted for 14%. Besides, Zhao [bib_ref] Risk factors for radiation-induced lung toxicity in patients with non-small cell lung..., Zhao [/bib_ref] demonstrated that the incidence of ≥grade 2 radiation pneumonitis was 13% after lobectomy, and 0% after pneumonectomy, which verified the safety of radiotherapy after pneumonectomy. Zhao believed that strict lung dose constraint and experienced beam arrangement (outside pulmonary parenchyma) were all protective and beneficial factors, which might be the reason for no serious radiation pneumonitis after pneumonectomy. In addition, in our research, low prescription dose (median dose 54Gy) and modern techniques could effectively reduce the volume and dose of the contralateral lung (median V20 4.75%), which was also an important reason for the low occurrence of radiation pneumonitis. Radiation esophagitis was another common side effect of radiotherapy. However, large-scaled analysis showed that the rate of ≥grade 3 radiation esophagitis was only 4% in NSCLC patients after postoperative chemotherapy and radiotherapy [bib_ref] Sequencing of postoperative radiotherapy and chemotherapy for locally advanced or incompletely resected..., Francis [/bib_ref]. Although there were nearly 1/3 of patients suffering grade 2 radiation esophagitis in our study, no severe radiation esophagitis was noted, which may be related to the precise techniques and careful protection of normal tissue in radiotherapy after pneumonectomy. Therefore, with careful assessment of pulmonary function, reasonable prescription dose, contralateral lung dose limitation and beam arrangement, radiotherapy after pneumonectomy was safe and feasible for stage pIII-N2 patients.
In recent years, a number of retrospective studies have shown survival benefits brought by PORT in stage pIII-N2 NSCLC patients after surgery. However, most patients in these studies received lobectomy. As reported, only a few patients received pneumonectomy with the rate of 16% in Corso's study [bib_ref] Re-evaluation of the role of postoperative radiotherapy and the impact of radiation..., Corso [/bib_ref] , 8.3% in Robinson's study [bib_ref] Postoperative radiotherapy for pathologic N2 non-small-cell lung cancer treated with adjuvant chemotherapy:..., Robinson [/bib_ref] , 10.8% in Hui's study [bib_ref] Selection of proper candidates with resected pathological stage IIIA-N2 non-small cell lung..., Hui [/bib_ref] , 27.4% in Shen's study [bib_ref] Comparison of efficacy for postoperative chemotherapy and concurrent radiochemotherapy in patients with..., Shen [/bib_ref] , and highest in the ANITA study, which was only 36.9% [bib_ref] Impact of postoperative radiation therapy on survival in patients with complete resection..., Douillard [/bib_ref]. Therefore, the benefit of PORT to survival in patients receiving pneumonectomy is not clear and still needs to be evaluated. Our study showed that PORT significantly improved OS, DFS, LRFS and DMFS in stage pIII-N2 NSCLC patients receiving pneumonectomy and adjuvant chemotherapy, although there were more patients with R1/R2 resection in PORT group. In subgroup analysis of patients with R0 resection, PORT was significantly related to better OS and LRFS, and DFS and DMFS in the PORT group trended to be improved. Even in patients with R1/R2 resection, there was a positive correlation between OS and PORT. Local-regional recurrence is common in stage pIII-N2 NSCLC patients receiving surgery alone [bib_ref] Postoperative radiotherapy in localized non-small cell lung cancer, Rube [/bib_ref]. PORT could increase local-regional control. However, the important premise of transforming improved local-regional control into better OS was the effective prevention of distant metastasis by adjuvant chemotherapy. Thus, we enrolled patients with adjuvant chemotherapy. A meta-analysis showed that some prospective studies failed to show efficacy of PORT in improving overall survival without adjuvant chemotherapy. On the other hand, a large sample retrospective analysis from national cancer database in the United States showed that the implementation of PORT on the basis of adjuvant chemotherapy, could prolong patients' median OS by 5 months, and could increase the absolute 5-year survival rate by 5% (39.3% vs 34.8%, p = 0.014, 4]. Additionally, less side effects of chemotherapy and radiotherapy in patients with both R0 resection and non-R0 resection had positive impact on survival. Previous randomized trials showed that ≥grade 3 adverse events occurred in 91% of patients with postoperative concurrent chemoradiotherapy [bib_ref] Sequential vs. concurrent chemoradiation for stage III non-small cell lung cancer: randomized..., Curran [/bib_ref] , which was significantly higher than that of patients treated with chemotherapy and radiotherapy. A recent retrospective study revealed that the OS in patients, with R0, R1 or R2 resection, receiving chemotherapy and radiotherapy was higher than those receiving concurrent chemoradiotherapy, although the survival benefits were not statistically significant in patients with R1/ R2 resection [bib_ref] Sequencing of postoperative radiotherapy and chemotherapy for locally advanced or incompletely resected..., Francis [/bib_ref]. The researchers believed that toxicities of concurrent chemoradiotherapy counteract the benefits. Therefore, pneumonectomy followed by chemotherapy and radiotherapy was a suitable mode of treatment for stage pIII-N2 NSCLC patients in order to ensure safety of surgery and adjuvant treatment.
As for failure mode, this study showed that distant metastasis was still the main reason (48.7%) of failure after pneumonectomy, followed by local-regional recurrence (32.8%). The result was consistent with the failure patterns in previously published studies, which showed that distant metastasis being the main cause of failure in stage pIII-N2 NSCLC after surgery (about 37-51%), and local-regional recurrence rate accounting for 10-35% [bib_ref] Impact of postoperative radiation therapy on survival in patients with complete resection..., Douillard [/bib_ref] [bib_ref] Radiotherapy plus chemotherapy with or without surgical resection for stage III non-small-cell..., Albain [/bib_ref] [bib_ref] Radiation therapy-first strategy after surgery with or without adjuvant chemotherapy in stage..., Lee [/bib_ref] [bib_ref] Postoperative radiotherapy for resected pathological stage IIIA-N2 non-small cell lung cancer: a..., Dai [/bib_ref]. In several case-controlled studies enrolling stage pIII-N2 patients, PORT significantly reduced local and regional recurrence rates, compared with non-radiotherapy groups. In the ANITA study [bib_ref] Impact of postoperative radiation therapy on survival in patients with complete resection..., Douillard [/bib_ref] , local-regional recurrence occurred in 25.7% of patients, which was significantly reduced to 14.6% after combined radiotherapy. In the previous study in our center [bib_ref] Postoperative radiotherapy for resected pathological stage IIIA-N2 non-small cell lung cancer: a..., Dai [/bib_ref] , 30.2% of locoregional recurrence occurred in the PORT group while 39.2% in patients without PORT (p = 0.025). In this study, for patients receiving pneumonectomy, PORT also achieved a significant reduction in local-regional failure compared with patients received adjuvant chemotherapy alone (15.6% vs. 39.1%, p = 0.016). Moreover, PORT could effectively reduce local-regional recurrence in patients after R0 resection or non-R0 resection, although the difference was not statistically significant (57.1%vs. 37.5%, p = 0.405) in those with non-R0 resection as small sample size (15 cases). In addition, local-regional recurrence in PORT group mostly occurred out of the radiation field (in-filed 1/5 vs. out-of filed 4/5), which also suggested the effective local-regional control of PORT. In summary, distant control still needed to be highlighted for stage pIII-N2 patients as NSCLC is a highly malignant tumor. On condition of effective prevention of chemotherapy in distant metastasis, radiotherapy could reduce local-regional recurrence and further improve survival.
Lobectomy combining with adjuvant chemotherapy and radiotherapy had showed median OS of 20-48 months, and 5-year OS of 25-61.3% for stage pIII-N2 NSCLC [bib_ref] Postoperative radiotherapy is associated with better survival in non-small cell lung cancer..., Mikell [/bib_ref] [bib_ref] Impact of postoperative radiation therapy on survival in patients with complete resection..., Douillard [/bib_ref] [bib_ref] Radiation therapy-first strategy after surgery with or without adjuvant chemotherapy in stage..., Lee [/bib_ref] [bib_ref] Postoperative radiotherapy for resected pathological stage IIIA-N2 non-small cell lung cancer: a..., Dai [/bib_ref] [bib_ref] A multicenter retrospective analysis of survival outcome following postoperative chemoradiotherapy in non-small-cell..., Zou [/bib_ref] [bib_ref] Long-term survival of patients with pN2 lung cancer according to the pattern..., Legras [/bib_ref]. In the INT0139 study, stage pIII-N2 patients receiving chemotherapy and radiotherapy after pneumonectomy approached the median OS of 18.9 months, and the 5-year OS of 21.9% [bib_ref] Radiotherapy plus chemotherapy with or without surgical resection for stage III non-small-cell..., Albain [/bib_ref] , while 5-year OS was usually lower than 20% in other studies [bib_ref] Sleeve pneumonectomy for central non-small cell lung cancer: indications, complications, and survival, Eichhorn [/bib_ref] [bib_ref] Postoperative radiotherapy for resected pathological stage IIIA-N2 non-small cell lung cancer: a..., Dai [/bib_ref] [bib_ref] Therapeutic options following pneumonectomy in non-small cell lung cancer, Petrella [/bib_ref] [bib_ref] Analysis of patients unable to undergo adjuvant chemotherapy after surgery in stage..., Graeter [/bib_ref]. In this study, the median OS was 46 months and 5-year OS was 46.3%, which was similar to the prognosis of patients after lobectomy and significantly better than patients after pneumonectomy. The reasons for better prognosis in this study were as follows. Firstly, patients who died during peri-operative period were excluded. In previously reported studies, the mortality rate was high after pneumonectomy, which was about 6-20% within 30 days, and up to 25% within 6 months in some studies [bib_ref] Pneumonectomy for bronchogenic carcinoma: analysis of factors predicting survival, Ramnath [/bib_ref] [bib_ref] Mortality, complications and loss of pulmonary function after pneumonectomy vs. sleeve lobectomy..., Melloul [/bib_ref]. In the INT0139 study, 14/54 died of postoperative complications after pneumonectomy [bib_ref] Radiotherapy plus chemotherapy with or without surgical resection for stage III non-small-cell..., Albain [/bib_ref]. Secondly, only patients receiving adjuvant chemotherapy after pneumonectomy were enrolled in this study. Thirdly, in this study, KPS scores in all patients were more than 80, and the preoperative pulmonary function approached a median FEV1 of 2.36 L. Assessment of the cardiopulmonary function was necessary before pneumonectomy. Some studies have demonstrated that patients with at least 2 L of preoperative FEV1 could tolerate pneumonectomy [bib_ref] Pneumonectomy for nonsmall cell lung Cancer, Sugarbaker [/bib_ref].
There were some limitations in this study. Firstly, as a retrospective, single-center based study, there might be selective bias in enrollment. Secondly, radiation induced heart injury was evaluated by heart related symptoms instead of routine ECG and echocardiography, which might lead to underestimating side effects of the heart. Thirdly, in this article, only patients who could tolerate pneumonectomy and some cycles of chemotherapy were included, which might underestimate the potential harm of pneumonectomy. However, we aimed to focus to this group of people with high KPS score and survived chemotherapy.
Finally, although the clinical characteristics between PORT group and Control group were comparable except that non-R0 resection rate was higher in the former, the number of patients in the PORT group was much lower than the control group (32 patients vs. 87 patients), which might greatly weaken our findings and conclusions. Despite these limitations, this is the largest population based study exploring safety and efficacy of PORT in stage pIII-N2 NSCLC patients after pneumonectomy and adjuvant chemotherapy, which can provide high reference and guidance for PORT of stage pIII-N2 NSCLC after pneumonectomy.
# Conclusion
For patients with stage pIII-N2 NSCLC after pneumonectomy followed by chemotherapy, PORT can improve OS, DFS, LRFS and DMFS with tolerable toxicity, which makes PORT be a preferred adjuvant treatment option. Further validation is expected in the future.
[fig] Figure 1: Effect of postoperative radiotherapy (PORT) on survival for stage pIIIA-N2 non-small cell lung cancer after pneumonectomy. (a): OS; (b) DFS; (c) LRFS; (d) DMFS. PORT group; Control group. Abbreviations: OS, overall survival; DFS, disease free survival; LRFS, local recurrence free survival; DMFS, distant metastasis free survival [/fig]
[table] Table 1: Baseline Clinical Characteristics for Patients Treated with and without PORT Abbreviation: PORT postoperative radiotherapy, SCC squamous cell carcinoma, LVSI lymph vascular space invasion, LN lymph node [/table]
[table] Table 2: Impact of Different Treatment Models on SurvivalAbbreviation: PORT postoperative radiotherapy, MST median survival time, mDFS median disease free survival, mLRFS median local recurrence free survival, mDMFS median distant metastasis free survival, mos months [/table]
[table] Table 3: Failure Pattern for Patients Treated with and without PORT [/table]
|
Identification of benign and malignant thyroid nodules by in vivo iodine concentration measurement using single-source dual energy CT
This study proposed to determine whether in vivo iodine concentration measurement by single-source dual energy (SSDE) CT can improve differentiation between benign and malignant thyroid nodules. In total, 53 patients presenting with thyroid nodules underwent SSDE CT scanning. Iodine concentrations were measured for each nodule and normal thyroid tissue using the GSI-viewer image analysis software. A total of 26 thyroid nodules were malignant in 26 patients and confirmed by surgery; 33 nodules from 27 patients were benign, with 10 confirmed by surgery and others after follow-up. Iodine concentrations with plain CT were significantly lower in malignant than benign nodules (0.47 ± 0.20 vs 1.17 ± 0.38 mg/mL, P = 0.00). Receiver operating characteristic (ROC) curve showed an area under the curve (AUC) of 0.93; with a cutoff of 0.67, iodine concentration showed 92.3% sensitivity and 88.5% specificity in diagnosing malignancy. Iodine concentration obtained by enhanced and plain CT were significantly higher in malignant than benign nodules (9.05 ± 3.35 vs 3.46 ± 2.24 mg/mL, P = 0.00). ROC curve analysis showed an AUC of 0.93; with a cutoff value of 3.37, iodine concentration displayed 78% sensitivity, 95% specificity in diagnosing malignancy. Combining unenhanced with enhanced iodine concentrations, the diagnostic equation was: Y = -8.641 Â unenhanced iodine concentration + 0.663 Â iodine concentration. ROC curve showed an AUC of 0.98 (95% CI, 0.94, 1.00). With Y ≥ -2 considered malignancy, diagnostic sensitivity and specificity were 96%, 96.3%, respectively. This study concluded that SSDE CT can detect the differences in iodine uptake and blood supply between benign and malignant thyroid lesions.Abbreviations: AUC = area of the curve, FNA = fine needle aspiration, HMW = high molecular weight, HU = Hounsfield unit, NIS = Na(+)/I(-) symporter, ROC curve = receiver operation characteristic curves, ROI = region of interest, SSDE CT = single-source dual energy CT.
# Introduction
Thyroid cancer is the most commonly diagnosed cancer for Chinese women before the age of 30 years. The largest increase in incidence was seen for thyroid cancers for women from 2000 to 2011, except prostate and cervix cancer. [bib_ref] Cancer Statistics in China, Wanqing [/bib_ref] The dramatic rise in thyroid cancer among women is consistent with that observed in other countries. [bib_ref] Pathways to the diagnosis of thyroid cancer in New South Wales: a..., Kahn [/bib_ref] [bib_ref] International patterns and trends in thyroid cancer incidence, Kilfoy [/bib_ref] [bib_ref] The increasing incidence of thyroid cancer: the influence of access to care, Morris [/bib_ref] [bib_ref] Increasing thyroid cancer incidence in Queensland, Australia 1982-2008-true increase or overdiagnosis?, Pandeya [/bib_ref] Although thyroid cancer is one of the most curable malignancies and only 7% of thyroid nodules are malignant, [bib_ref] Clinical practice. The thyroid nodule, Hegedus [/bib_ref] differential diagnosis cannot be certain. [bib_ref] Identification of genes differentially expressed in benign versus malignant thyroid tumors, Prasad [/bib_ref] [bib_ref] Differentiation of benign and malignant thyroid nodules based on the proportion of..., Kim [/bib_ref] The surgical management of benign and malignant thyroid nodules differs substantially. So it is important to diagnose malignancy pre-operatively and to limit unnecessary surgery in the vast majority with benign nodules. Fine needle aspiration (FNA) biopsy is promoted as the tools of choice for diagnosis. But, still with some limitations, [bib_ref] Biopsy of thyroid nodules: comparison of three sets of guidelines, Ahn [/bib_ref] [bib_ref] Cytopathologic diagnosis of fine needle aspiration biopsies of thyroid nodules, Misiakos [/bib_ref] [bib_ref] Novel approaches to thyroid cancer treatment and response assessment, Grewal [/bib_ref] [bib_ref] Biopsy of thyroid nodules: comparison of three sets of guidelines, Ahn [/bib_ref] [bib_ref] Thyroid cytology and the risk of malignancy in thyroid nodules: importance of..., Kelman [/bib_ref] for example (i) relatively high falsenegative rate especially for posterior nodules; (ii) sampling error, because of necrosis in the nodule or the sample cannot represent the whole nodule; (iii) for a patient with multiple nodules in the thyroid gland, FNA biopsy of every nodule is neither costeffective nor feasible; [bib_ref] Identification of genes differentially expressed in benign versus malignant thyroid tumors, Prasad [/bib_ref] (iiii) 15% to 30% of FNAs cannot be diagnosed by cytological review with enough certainty. [bib_ref] Thyroid cancer, Lebastchi [/bib_ref] As well as we know, nodule formation in the thyroid gland can influence iodine uptake and utilization. A study by Doban [bib_ref] Rapid communication: predominant intracellular overexpression of the Na(+)/I(À) symporter (NIS) in a..., Dohan [/bib_ref] assessing a large number of samples showed decreased iodine uptake in thyroid cancer cells. Multiple reports have shown that the functional differences between benign and malignant thyroid nodules are associated with Na(+)/I(-) symporter (NIS) protein dysfunction in most thyroid cancer cells. [bib_ref] Rapid communication: predominant intracellular overexpression of the Na(+)/I(À) symporter (NIS) in a..., Dohan [/bib_ref] [bib_ref] The Na+/IÀ symporter (NIS): mechanism and medical impact, Portulano [/bib_ref] [bib_ref] The BRAFV600E oncogene induces transforming growth factor beta secretion leading to sodium..., Riesco-Eizaguirre [/bib_ref] [bib_ref] Hypermethylation of a new distal sodium/iodide symporter (NIS) enhancer (NDE) is associated..., Galrão [/bib_ref] [bib_ref] Differential regulation of the human sodium/iodide symporter gene promoter in papillary thyroid..., Kogai [/bib_ref] [bib_ref] Detection of mRNA of sodium iodide symporter in benign and malignant human..., Liou [/bib_ref] [bib_ref] Expression, exon-intron organization, and chromosome mapping of the human sodium iodide symporter, Smanik [/bib_ref] Physiologically, thyroid glandular epithelial cells actively transport iodine through the NIS protein for thyroid hormone synthesis. Reduced expression of the NIS gene has been reported in thyroid carcinoma specimens compared with normal thyroid tissues. [bib_ref] Expression, exon-intron organization, and chromosome mapping of the human sodium iodide symporter, Smanik [/bib_ref] [bib_ref] Retinoic acid increases sodium/iodide symporter mRNA levels in human thyroid cancer cell..., Schmutzler [/bib_ref] [bib_ref] Iodide symporter gene expression in human thyroid tumors, Arturi [/bib_ref] [bib_ref] Development of reverse transcription-competitive polymerase chain reaction method to quantitate the expression..., Ryu [/bib_ref] Consequently, NIS was found in higher amounts in normal and benign thyroid tissues compared with malignant tissues. Another study reported decreased NIS expression as the tissues mutated from normal to benign and malignant tissues. All the above data were obtained from resected specimens.
The currently recommended in vivo iodine measurement methods include urinary iodine concentration assessment [bib_ref] Methods to assess iron and iodine status, Zimmermann [/bib_ref] and thyroid 131 I absorption rate determination, both of which only indirectly reflect the iodine concentration with certain limitations; indeed, the effect of dietary iodine can result in iodine measurement inaccuracy. In radiology, Zhang et al [bib_ref] The research of iodine measurement in thyroid gland by CT and monitoring..., Xin-Chuan Zhang [/bib_ref] attempted to measure CT values of the thyroid gland by using mixed energy CT imaging; afterward, they converted the obtained CT values into iodine concentration, which affects accurate quantitative measurement of iodine concentration.
We wondered whether in vivo measurement of iodine concentration using a noninvasive method could improve the differentiation between benign and malignant thyroid nodules. Indeed, recent technological advances in radiology have made this possible. Single-source dual energy (SSDE) CT is a relatively new CT technique. A generator electronically switches the tube, pulsing rapidly between low energy (80 kVp) and high energy (140 kVp) with each exposure time of 0.5 ms, to produce dual energy images. This scanning mode enables precise registration of data sets for the creation of material decomposition images (e.g., water-based and iodine-based material decomposition images).
The purpose of this study was to determine the value of in vivo iodine concentration measurement in differentiating benign from malignant thyroid nodules by SSDE CT.
# Materials and methods
## In vitro study
An in vitro experiment was carried out to evaluate the accuracy of iodine concentration measurements. A set of 8 test tubes containing iodine at 0.5 -30.0 mg/mL were scanned by the SSDE CT protocol using fast tube voltage alternating between 80 and 140 kVp with collimation thickness of 0.625 mm, rotation speed of 0.6 s, and helical pitch of 0.983:1. The iodine concentrations were measured using a circular region of interest that covered 80% of the test tube's cross section on the iodine-based material decomposition images. Then, the iodine level measured in each test tube was compared with the known concentration. Finally, a radiologist ( * * ) undertook the quantitative measurements with dedicated imaging viewer (GSI Viewer; GE Healthcare) to analyze the material decomposition images.
## In vivo study
Study population: We retrospectively searched our department's computerized clinical database for all thoracic and cervical SSDE plain and contrast-enhanced CT scanning in which a new abnormality was incidentally identified in the thyroid gland from July 2010 to May 2011. Ethical approval was not necessary. A total of 92 patients were obtained, among which 12 were excluded for diffuse diseases in the thyroid gland (no normal gland found) or with a history of thyroid surgery. A list of patients whose pathologic examination or further imaging and clinical evaluation was scheduled after CT was then extracted. This yielded a final study population of 53 patients (38 women and 15 men) aged 54 ± 9 (range, 38-66) years.
CT protocol: All CT examinations were performed on an SSDE CT scanner (GE Discovery CT750 HD CT scanner; GE Healthcare, Milwaukee, WI) for neck or thorax. The scanner was calibrated regularly with air and water phantom to ensure reliable measurements. Each patient underwent plain and contrast-enhanced CT examinations. Plain CT scanning was carried out in the conventional helical mode after scout CT scanning (tube peak voltage, 120 kVp). Patients were injected with nonionic contrast material via antecubital vein at a rate of 3-4 mL/s (1.5 mL/kg body weight; Omnipaque 300 mg I/mL, GE Healthcare). Other parameters included: collimation thickness, 5 mm; reconstructed thickness, 0.625 mm; tube current, 640 mA; rotation speed, 0.6 s; helical pitch, 1.375; CT dose index volume, 17.65 mGy. The CT images were reconstructed using GSI Viewer. The iodine-based material decomposition images were obtained with pseudo-monochromatic images ranging from 40 to 140 keV.
CT Image Quantitative Analysis: CT images were analyzed by 2 radiologists ( * * and * * with 5 and 15 years experience in clinical CT, respectively) working together with AW4.4 workstation (GE Medical Systems, Milwaukee, WI) and blinded to the clinical and pathological information. Quantitative analysis of iodine concentration was carried out with GSI Viewer.
ROI drawing: For single isolated nodule, the region of interest (ROI) was placed on its solid regions, avoiding cystic or necrotic regions. For multiple nodules, the ROI was placed on the solid regions of the biggest nodule. For comparison, an additional ROI was placed in the normal thyroid tissue.
Quantitative parameters: Iodine concentrations (mg/mL) in nodules and normal thyroid tissues, respectively, were determined by plain and enhanced CT scanning. Iodine concentrations obtained from plain CT scanning represented the iodine uptake function of the thyroid tissue; Iodine concentration characterized the blood supply in the thyroid tissue.
Iodine concentration = iodine concentration (enhanced CT scanning)-iodine concentration (plain CT scanning).
To ensure consistency, all measurements were done twice at 2-week intervals and mean values were calculated. Prominent artifacts were carefully avoided. For all measurements, the size, shape, and position of the ROI were consistent between the softtissue window images and iodine-based material decomposition images as judged by the copy-and-paste function [fig_ref] Figure 1: Plain CT image [/fig_ref].
## Clinical and pathological results
The final diagnosis was based on pathological or follow-up information. Thirty-six patients underwent thyroid surgery at our clinic within 1 week after SSDE CT with pathological confirmation. They included 10 patients with benign lesions (4 macrofollicular adenomas, 1 mixed but predominantly macrofollicular adenomas, and 5 adenomatous goiter) and 26 with malignant tumors (25 papillary and 1 follicular carcinomas, respectively).
In total, 17 patients not referred for surgery or FNA biopsy were monitored by clinical and radiological examinations for a median period of 41 months (36-52 months), and no sign of malignancy was observed.
# Results
## In vitro study
An excellent correlation was obtained between measured and known iodine concentrations used in the in vitro experimentIodine concentrations were also obtained via material decomposition imaging from enhanced CT scans. There was no significant difference in iodine concentrations of normal thyroid tissues between benign and malignant nodule groups (P = 0.6). However, iodine concentrations of 9.05 ± 3.35, 3.46 ± 2.24, www.md-journal.com and 13.97 ± 3.26 mg/mL were obtained for malignant nodules, benign nodules, and normal thyroid gland tissues, representing statistically significant differences (F = 165.12, P = 0.00, < 0.05). The ROC curve yielded an AUC value of 0.93 (0.82-1.00). With a cutoff value of 3.37, the iodine concentrations showed 78% sensitivity and 95% specificity in diagnosing malignancy.
Combining the unenhanced iodine concentrations and iodine concentrations, the following fitted diagnostic equation was obtained: Y = À8.641 Â unenhanced iodine concentration +0.663 Â iodine concentrations. This can improve the differentiation between benign and malignant thyroid nodules. ROC curve yielded an AUC of 0.98 (95% CI, 0.94, 1.00). Y ≥ À2 represented malignancy, whereas Y <À2 was considered as benignity. In these conditions, diagnostic sensitivity, specificity of 96% and 96.3%, respectively, were obtained.
# Discussion
Current in vivo iodine measurement methods, including urinary iodine examination and thyroid 131 I absorption rate assessment, can indirectly reflect the iodine concentration in the human body, but with limitations such as interference factors, complex preexamination requirements, inaccurate measurement results, and sometimes radiation hazard. Studies have explored the value of CT imaging in iodine measurement. Imanishi et al [bib_ref] Measurement of thyroid iodine by CT, Imanishi [/bib_ref] [bib_ref] Correlation of CT values, iodine concentration, and histological changes in the thyroid, Imanishi [/bib_ref] compared the CT values of different thyroid lesions, and found no significant statistical difference. Zhang et al [bib_ref] The research of iodine measurement in thyroid gland by CT and monitoring..., Xin-Chuan Zhang [/bib_ref] obtained iodine concentrations of thyroid lesions by using the CT value conversion method, and assessing various thyroid lesions, they found no obvious differences in iodine concentrations among different lesions. A series of contradictory results from pathology and basic medical experiments have demonstrated that among different thyroid lesions there are real differences in iodine concentrations, which cannot be detected by conventional CT scanning. We hypothesized that these discrepancies may be due to the mixed energy x-ray used in conventional CT imaging, which result in inaccurate CT value measurement. The mixed energy xray CT imaging only reflects the average effect in the mixed energy and cannot be used to further distinguish different materials, for example, in a phenomenon called "different objects with same image characteristics."
Recently a series of papers focused on SSDE CT have demonstrated the usefulness of this new technological advancement in distinguishing one substance from another. [bib_ref] Differentiation of small hepatic hemangioma from small hepatocellular carcinoma: recently introduced spectral..., Lv [/bib_ref] [bib_ref] The role of dual energy CT in differentiating between brain haemorrhage and..., Tijssen [/bib_ref] [bib_ref] Differentiating calcium oxalate and hydroxyapatite stones in vivo using dual-energy CT and..., Liu [/bib_ref] [bib_ref] Material differentiation in forensic radiology with single-source dual-energy computed tomography, Ruder [/bib_ref] [bib_ref] Urinary stone differentiation in patients with large body size using dual-energy dualsource..., Qu [/bib_ref] The fundamentals of physics of SSDE CT mainly states: the substance's absorption capacity varies with changes of x-ray energy, the changes of the absorption capacity of different substances vary with changes of x-ray energy. High-molecularweight (HMW) substances such as bone and iodine vary a lot with changes of x-ray energy. It is this change with different energy x-ray, which enables further distinction of different substances. This study used the material decomposition technique of the SSDE CT imaging to measure iodine concentrations of the nodule's solid composition and normal thyroid gland tissue. We found overtly higher iodine concentrations in benign nodules compared with malignant ones. Our results were consistent with other reports using basic medical experiments: [bib_ref] Rapid communication: predominant intracellular overexpression of the Na(+)/I(À) symporter (NIS) in a..., Dohan [/bib_ref] [bib_ref] The Na+/IÀ symporter (NIS): mechanism and medical impact, Portulano [/bib_ref] [bib_ref] The BRAFV600E oncogene induces transforming growth factor beta secretion leading to sodium..., Riesco-Eizaguirre [/bib_ref] [bib_ref] Hypermethylation of a new distal sodium/iodide symporter (NIS) enhancer (NDE) is associated..., Galrão [/bib_ref] [bib_ref] Differential regulation of the human sodium/iodide symporter gene promoter in papillary thyroid..., Kogai [/bib_ref] [bib_ref] Detection of mRNA of sodium iodide symporter in benign and malignant human..., Liou [/bib_ref] [bib_ref] Expression, exon-intron organization, and chromosome mapping of the human sodium iodide symporter, Smanik [/bib_ref] [bib_ref] Retinoic acid increases sodium/iodide symporter mRNA levels in human thyroid cancer cell..., Schmutzler [/bib_ref] [bib_ref] Iodide symporter gene expression in human thyroid tumors, Arturi [/bib_ref] [bib_ref] Development of reverse transcription-competitive polymerase chain reaction method to quantitate the expression..., Ryu [/bib_ref] some thyroid follicular cells still existed in adenoma, nodular goiter, and other benign lesions showed normal iodine uptake function; however, thyroid follicular cells were almost totally replaced by cancer cells and fibrotic connective tissues in thyroid papillary carcinoma, which had lost normal iodine uptake function. In our results, the iodine concentration differences detected by SSDE CT were found between benign and malignant thyroid lesions. With a cut-off value of iodine concentration obtained by plain CT scanning set at 0.67 mg/mL, 5 cases were misdiagnosed, including 2 cases misdiagnosed as malignancy, and 3 as benign lesions. These data indicated a high sensitivity (92.3%), but a low specificity (88.5%).
It is known that iodine concentrations measured by plain CT scanning reflect the iodine uptake function of thyroid nodules; meanwhile, iodine levels measured by contrast enhanced CT scanning better reflect blood supply of thyroid nodules. In addition, for thyroid nodules with fuzzy boundary or of density difference not so obvious with the surrounding normal thyroid tissues, ROI can be easily obtained by using contrast enhanced CT imaging. Our results showed overtly higher iodine concentrations in malignant thyroid nodules compared with benign nodules after contrast agent injection, in agreement with previous histopathological reports [bib_ref] Thyroid cytology and the risk of malignancy in thyroid nodules: importance of..., Kelman [/bib_ref] [bib_ref] Rapid communication: predominant intracellular overexpression of the Na(+)/I(À) symporter (NIS) in a..., Dohan [/bib_ref] [bib_ref] The Na+/IÀ symporter (NIS): mechanism and medical impact, Portulano [/bib_ref] [bib_ref] The BRAFV600E oncogene induces transforming growth factor beta secretion leading to sodium..., Riesco-Eizaguirre [/bib_ref] [bib_ref] Hypermethylation of a new distal sodium/iodide symporter (NIS) enhancer (NDE) is associated..., Galrão [/bib_ref] [bib_ref] Differential regulation of the human sodium/iodide symporter gene promoter in papillary thyroid..., Kogai [/bib_ref] [bib_ref] Detection of mRNA of sodium iodide symporter in benign and malignant human..., Liou [/bib_ref] [bib_ref] Expression, exon-intron organization, and chromosome mapping of the human sodium iodide symporter, Smanik [/bib_ref] [bib_ref] Retinoic acid increases sodium/iodide symporter mRNA levels in human thyroid cancer cell..., Schmutzler [/bib_ref] showing that thyroid carcinoma tissues are rich in blood vessels. Using a cutoff value of 3.37 for iodine concentrations obtained by enhanced CT scanning alone, 7 cases were misdiagnosed as benign lesions, and 1 case as malignancy. These data indicated a low sensitivity (74%), but a high specificity (97%). Considering both sensitivity and specificity, iodine concentrations obtained from plain-and enhanced CT data were assigned different weights to establish a fitting diagnosis equation. Based on this equation, 1 case was misdiagnosed as benign lesion, and 3 cases were misdiagnosed as malignancy. The diagnostic sensitivity and specificity were 96% and 96.2%, respectively. The overall diagnostic accuracy was 96.2%. These data indicate that combining iodine concentrations from plain and iodine concentration from enhanced CT is useful to evaluate the thyroid nodular function and blood supply, providing a better assessment of the thyroid lesion's nature.
In this study, we found an excellent association between the measured and known iodine concentrations in vitro, suggesting the differences in iodine concentrations detected in vivo cannot be attributed to systematic errors. In addition, the iodine concentrations were measured in normal thyroid tissues from the benign and malignant nodule groups, and served as baseline values in this study. No differences were obtained between these groups, indicating that the differences found in the lesions reflect actual characteristics of the lesions and were not the result of systematic errors. There are some limitations of this study. First, we did not compare the results presented here with those obtained by ultrasound, MRI, or nuclear medicine. Second, in this limited number of patients, most cases of malignant nodules were papillary carcinomas, the most common type of thyroid malignancy, which could bias the analysis. Third, multiple benignancy cases were decided based on clinical data and followup; since many thyroid carcinomas are slow growing, no perceptible change might be seen within 3 years or even more. Therefore, the number of benign nodules in this study may be inaccurate. Fourth, the patient population in this study is too small to make a strong case. And as the same reason, we did not do cross-validated analysis for the fitted diagnostic equation. Further studies (more sample size) are therefore required to explore its clinical application. Fifth, the patients in this study received CT scanning due to other tumors or for lymph node evaluation in thyroid cancer patients, the patients did not receive extra or excessive radiation, but in clinical application, we still need to consider carefully avoiding unnecessary radiation.
This study is the first in vivo iodine measurement focusing on thyroid nodules by the material decomposition technology of SSDE CT, which confirmed the feasibility of in vivo iodine measurement. In addition, our conclusions were consistent with known pathological and basic experimental data, indicating the accuracy of in vivo iodine measurement. Therefore, SSDE CT can detect the differences in iodine uptake function and blood supply between benign and malignant thyroid lesions. These findings demonstrate the potential value of SSDE CT imaging to combine anatomical and functional imaging evaluations.
[fig] Figure 1: Plain CT image (A) and iodine-base material decomposition map (B), ROI (red circle) in A represent the CT value of the thyroid nodule, ROI (red circle) in B represent the corresponding iodine concentration of the same thyroid nodule. The size, shape, and position of the ROI were consistent between the soft-tissue window images (A) and iodine-based material decomposition images (B) by the copy-and-paste function. CT = computed tomography, ROI = region of interest. [/fig]
[fig] Figure 2: (A) CT image of the phantom contained of different iodine concentrations. (B) The corresponding water-based material-decomposition image. (C) The corresponding iodine-based material-decomposition image. CT = computed tomography. [/fig]
[fig] Figure 3: (A) The box plot showed the comparison of iodine concentration (on Y-axis) between benign and malignant thyroid nodules (on X-axis) obtained by plain CT. (B) The ROC curve (with sensitivity on Y-axis and 1-specificity on X-axis) showed the diagnostic performance of iodine concentration for assessing malignant nodules. CT = computed tomography, ROC curve = receiver operating characteristic curves. [/fig]
[fig] Figure 4: (A) The box plot showed the comparison of iodine concentrations (on Y-axis) between benign and malignant thyroid nodules (on X-axis) obtained by enhanced CT. (B) The ROC curve (with sensitivity on Y-axis and 1-specificity on X-axis) showed the diagnostic performance of iodine concentrations for assessing malignant nodules. CT = computed tomography, ROC curve = receiver operating characteristic curves. [/fig]
[fig] Figure 5: (A) The box plot showed the comparison of Y value (on Y-axis) between benign and malignant thyroid nodules (on X-axis) combining the unenhanced (plain) iodine concentrations and iodine concentrations. (B) The ROC curve (with sensitivity on Y-axis and 1-specificity on X-axis) showed the diagnostic performance of Y value for assessing malignant nodules. Y = À8.641 Â unenhanced iodine concentration +0.663 Â iodine concentrations. ROC curve = receiver operating characteristic curves. [/fig]
[table] Table 1: Actual versus measured iodine concentrations at in vitro experiment with SSDE CT. [/table]
|
Design and Implementation of a Distant-Learning Clinical Research Mentorship Program: The Accra-Toronto Collaboration
PURPOSE For many oncology training programs in low-and middle-income countries, dedicated time for research education and mentorship of trainees is limited. Here, we report a 1-year-long collaboration between a cancer center in Canada and one in Ghana with the aim of imparting clinical research skills and mentoring the research of radiation oncology residents.METHODS On the basis of a needs assessment conducted in Ghana, we designed a curriculum consisting of 13 weekly seminars delivered via videoconference, followed by a 1-year-long mentorship program to support research projects. The primary outcome was the feasibility of the program from seminars to manuscript preparation. We used multiple secondary outcomes to capture the learning experience with study-specific questionnaires. We evaluated critical thinking ability using the Berlin questionnaire. Funding was made available for research and travel to international conferences.RESULTS Five Ghanaian trainees submitted research proposals. Nine Canadian faculty members delivered the seminars and two served as methodology mentors, and two Ghanaian faculty acted as local supervisors. Feedback questionnaires from all participants showed that they agreed strongly that they would recommend the sessions to another resident (75%), that the objectives were clear (71%), and that the topics were useful for their training (73%). At the end of the program, two Ghanaian trainees finalized their manuscripts and one was published.CONCLUSION Here, we report on the implementation of a mentorship program focused on research methods and evidence-based medicine in sub-Saharan Africa. The program was successful in the drafting and publication of abstracts and manuscripts by local trainees.JCO Global Oncol 6:919-928.
# Introduction
In 2013, an International Atomic Energy Agency (IAEA) report described only 23 of 52 African countries with access to radiotherapy.This is slowly improving 2 as governments and policymakers respond to the growing crisis through the implementation of long-term cancer strategies and capital investment in radiotherapy facilities and human resources. [bib_ref] Changing the global radiation therapy paradigm, Pistenmaa [/bib_ref] Training goes beyond the fundamentals of radiation medicine and the safe use of equipment. It must be complimented by critical appraisal and research skills, as well as a dedication to lifelong learning, to achieve the best possible outcomes for our patients. [bib_ref] Improving quality and access to radiation therapy: An IAEA perspective, Zubizarreta [/bib_ref] The National Center for Radiotherapy and Nuclear Medicine (NCR) in Accra, Ghana, and the Princess Margaret Cancer Centre in Toronto, ON, Canada, entered into a partnership to support training needs after bilateral academic exchanges. We hypothesized that a distance-learning mentorship program can be effective in supporting the development of clinical research skills in radiation oncology trainees. Ultimately, such collaborations are meant to improve clinical care and patient outcomes. In this report, we present the experience of the Accra-Toronto Clinical Research Mentorship Program (CRMP). We describe our initial needs assessment followed by the program's design, evaluation, and early impact.
# Methods
## Needs assessment
We held a focus group in 2014 at the NCR with Ghanaian residents and fellows. The focus group was facilitated by one of the authors (H.V.), a resident from Toronto undertaking an elective in Accra at the time. The group was guided by open-ended questions that explored the level of interest in a collaborative mentorship program, potential research topics, and format Author affiliations and support information (if applicable) appear at the end of this article.
Accepted on May 5, 2020 and published at ascopubs.org/journal/ go on June 30, 2020: DOI https://doi.org/10. 1200/JGO.19.00240 of the program, including scheduling, teleconference capability, and class structure.
## Program design
In designing our program, we were guided by our needs assessment in Accra and iterative input from trainees and faculty in both Accra and Toronto, including faculty with expertise in education and research methodology.
The final design took the shape of a 1-year-long program summarized in . The program has four key features.
Selecting a research question. To ensure learning through practice and application, all trainees who were enrolled in the radiation oncology training program (N = 5) were invited to submit a short research proposal (2 to 3 pages). The trainee was asked to identify one of the two Ghana consultants as their supervisor for the proposal. This was intended to ensure that the proposal had the support of the program for feasibility and scientific interest. Using a standardized rubric, a panel of assessors-2 faculty and 1 trainee-evaluated proposals on the basis of the project's feasibility, prior knowledge of clinical research methods as demonstrated in the proposal, and anticipated impact on the clinical or training environment .
Defining a 1-year supervisor, student, and mentor relationship. For this pilot project, to focus our efforts and faculty commitment, only the top 2 trainee proposals were selected for a mentor assignment, although all trainees were invited to participate in the rest of the program. The supervisor based in Accra provided clinical expertise on the research topic, facilitated problem solving from a systems perspective, and played a central role in ensuring the success of the international collaboration. The mentor based in Toronto was selected for their clinical epidemiology expertise and interest in global health. The trainee, supervisor, and mentor worked together through the year-long program with the goal of translating the research question to a viable research protocol, collecting data, interpreting results, and preparing an abstract and manuscript for knowledge dissemination. We deliberately used this terminology to reinforce the collaborative relationship. Trainees without an assigned mentor at Princess Margaret Cancer Centre were invited to participate in the seminar series and expected to work with their local supervisors to complete their research projects as part of their training program.
Clinical research methods teaching on the basis of a flipped classroom design. All trainees were invited to participate in a seminar series at the start of the program. This consisted in 13 1.5-hour seminars delivered weekly through a Webbased video connection-to maximize interactivity-in the first 3 months of the program. These seminars used a flipped classroom design in which learners were provided with reading material, learning objectives, and practice assignments before each session. The goal was to allocate 50% of the seminar time for the facilitator to review key concepts, whereas the second half was dedicated to group discussion and applying the concepts to the learner's research project. The curriculum included topics from framing a research question, clinical trial design, research ethics, biostatistics, and abstract and manuscript preparation.
Funding support. A modest research grant was allocated to each learner ($2,000 Canadian). These funds were designed to cover conference fees, transportation costs, and lodging costs for trainees. All trainees who participated in the seminars (N = 5) were eligible to access these funds upon successful acceptance of their abstract by an international conference-for example, the American Society for Radiation Oncology or the Canadian Association of Radiation Oncology. Only 80% of the funds were accessible to support travel, while the remaining 20% was accessible upon preparation of a manuscript. The National Center for Radiotherapy-Princess Margaret CRMP was supported by an unrestricted grant from the University Health Network Global Capacity Fund to support the research and travel stipend for students.
## Context key objective
Dedicated time for clinical research education and mentorship is challenging to find, especially in training programs in developing countries. We present a collaboration between the Princess Margaret Cancer Center in Toronto, ON, Canada, and the National Center for Radiotherapy in Accra, Ghana, focused on imparting clinical research skills and mentoring a group of trainees.
## Knowledge generated
We describe our needs assessment, program design, evaluation, and early impact, and find it to be a successful and worthwhile model to pursue.
## Relevance
Our paper provides a blueprint for international partnerships to support clinical research for oncology trainees through distantlearning and longitudinal mentorship principles.
## Outcomes of interest
The primary outcome of the project was feasibility, as measured by the number of learners who completed the seminar component and the rest of the mentorship program, up until manuscript preparation. We used multiple secondary outcomes to capture the learning experience aligned with the Kirkpatrick levels of evaluation-that is, reaction, learning, behavior change, and results. 5 A 13item study-specific questionnaire using a 5-point Likert scale (1 = strongly disagree; 5 = strongly agree) was designed to evaluate each seminar topic and the effectiveness of faculty facilitators . Qualitative feedback was solicited at the end of the seminar series through a videoconference and e-mail exchange from learners and the faculty in Ghana and Toronto. We evaluated the proficiency of learners with critical appraisal skills using the Berlin questionnaire, a validated instrument administered before and after the completion of the program. [bib_ref] Do short courses in evidence based medicine improve knowledge and skills? Validation..., Fritsche [/bib_ref] We reported the completion of research projects, abstract submissions, acceptance at international conferences, manuscript preparation, and publications. Finally, we explored interest in this kind of mentorship collaboration with other oncology training programs in Africa through personal introductions by our collaborator (V.V. & J.Y.).
# Results
The program was designed, funding secured, and the call for proposals first launched in 2015. Five Ghanaian trainees responded to the call for research proposals. Three faculty members in Toronto (J.R., Z.K., and R.K.S.W.) with Masterslevel clinical epidemiology training served as research mentors. Both oncologists at the NCR (V.V. and J.Y.) served as supervisors and supported the learners. As described in Methods, two learners were assigned a mentor in Toronto, although all trainees were invited to attend the seminars, and were eligible to receive travel funding upon acceptance of their results at the end of the program. Nine Canadian faculty members delivered the initial weekly seminars.
The research proposals submitted by the 5 trainees are summarized in [fig_ref] TABLE 1: Research Proposals [/fig_ref]. Feedback that was obtained using the study-specific questionnaire (N = 5) showed participants agreed strongly that they would recommend the sessions to another resident (75%), the objectives were clear (71%), the topics were useful for their training (73%), and the discussions were helpful to understand the topics (70%). Audio-video connection functionality scored the lowest (32% strongly agreed; Figs 4A and 4B). Qualitative feedback from faculty highlighted areas that worked well and those that did not. From the trainees, the strength of the teamwork that went into supporting the teaching was highlighted. Topics selected for teaching were felt to be highly appropriate. The relevance of "Design of quality assurance projects" was less obvious. Practical biostatistics sessions were particularly well received. The ability to apply the concepts to protocol design was useful to reinforce learning. Participants generally acknowledged that videoconferencing was effective in providing interactivity, and the ability of participants to see each other was important for building relationships. However, technology limitations were noted. Setting up the connection reliably was a consistent issue, and technical support was required for Vulpe et al efficiency and to avoid frustration. Preclass preparation and learner participation was heterogeneous and the reasons are multifactorial. Preclass reading material and tasks needed to be received with more lead time-for example, more than 1 week. Cultural differences in learning and interaction must be recognized. In general, it was perceived by the Toronto faculty that the concept of the flipped classroom style of learning and engagement in discussion by trainees was less familiar. Trainees had limited time to take advantage of the preclass reading material. Clinical commitments also made it difficult to commence sessions as scheduled on occasion. It was suggested that the reading material be more selective and made available well in advance.
Five trainees, including the 2 trainees who were assigned mentors, completed the baseline Berlin questionnaire, and Please rate the session on the following items (1 = Strongly agree, 2 = Agree, 3 = Neither agree or disagree, 4 = Disagree, 5 = Strongly disagree)
[formula] 1 2 3 4 5 [/formula]
The objectives of the session were clear
The assignment was helpful in learning about the topic
The reading material provided was helpful in preparing for the session
The session was well organized
The discussion was helpful in understanding the topic There was enough time to ask questions I found this topic useful for my training
The audio-video connection was adequate
Overall I would recommend this session to a fellow resident Please rate the presenter on the following items (1 = Strongly agree, 2 = Agree, 3 = Neither agree or disagree, 4 = Disagree, 5 = Strongly disagree)
[formula] 1 2 3 4 5 [/formula]
The presenter spoke clearly and was easy to understand
The presenter was knowledgeable about the topic
The presenter made good use of the allotted time
## No. of evaluations
The presenter spoke clearly and was easy to understand
The presenter was knowledgeable about the topic
The presenter made good use of the allotted time 4 completed the postseminar questionnaire. Results for 3 trainees showed an average 20% absolute improvement, whereas the score was unchanged for 1 trainee.
Two trainees have seen their research abstracts accepted at international conferences, including the Canadian Association of Radiation Oncology Annual Scientific Meeting and the International Gynecologic Cancer Society meeting. Two original research manuscripts have been finalized and one has been published in a major radiation oncology journal, with a second submitted. The first is a large retrospective review on cervical cancer outcomes at the NCR (no mentor assigned) and the other a review of the management and outcomes of pediatric patients with Wilms tumors in Ghana (mentor assigned). A third project investigating pain scores during brachytherapy for cervical cancer was initially postponed because of equipment delay but is currently underway (mentor assigned). Since the completion of the CRMP in Ghana, the program was offered to the oncology training program in Harare, Zimbabwe, where it is ongoing.
# Discussion
Our distant-learning clinical research mentorship program was effective in enabling the completion of 2 clinical research projects-with manuscripts documenting their findings-yielding novel clinical data with the potential to improve patient care. The seminar series was successfully delivered despite some technical limitations using a videoconferencing platform in an environment with limited Internet bandwidth and a 5-hour time difference. Learners successfully demonstrated an improvement in their critical appraisal skills after the seminar series, as evaluated using the Berlin questionnaire. The quality of the collaboration overall was strong and the program well received, garnering interest from other training programs. It is currently in its second iteration in another training program in sub-Saharan Africa.
Lessons learned from our assessment have highlighted specific areas for improvement. These include focusing the amount of preclass reading material, including trainees and local faculty in the initial scoring task to enhance learning, and modifying content in response to the research method need of different groups.
Not all trainees completed their research project. Reasons for incomplete projects were multifactorial. Participation in the research mentorship program was voluntary. In one mentorassigned project, the research topic had a prospective design. Delays in commencing the clinical program of interest meant the project needed to be deferred. Resources to support residents extracting data, creating and managing databases, performing analyses, securing ethics approvals, and performing additional study specific investigations were limited. These limitations are inherent to the process of conducting research, particularly in resource-constrained settings.
Resources needed to enable research can be considered in terms of infrastructure and human resources. These include faculty time to prepare and teach seminars, as well as to provide supervision (Ghana) and mentorship (Toronto) to selected residents throughout the year; specific content expertise on each of the topics and research methodology; technical support to ensure videoconference connectivity from both sites; and a coordinator function to ensure a smooth experience for both trainee and faculty in coordinating the needs assessment, submitting the proposal, providing evaluation and feedback, designing curriculum, scheduling sessions, distributing reading materials, assignments, evaluation surveys, and collating responses [fig_ref] TABLE 2: Resource Requirements [/fig_ref]. In the spirit of collaboration, all of these resources were provided in kind. Funding was preferentially used to facilitate the trainees' ability make presentations at conferences and disseminate their findings.
The NCR was established in 1997 in collaboration with the IAEA, serves as a large catchment area, and is a resource to other West African countries, including Togo, Benin, Côte D'Ivoire, Burkina Faso, Liberia, and Sierra Leone. 7 Common to many other areas that are undergoing stages of rapid change, dedicated time for research education and mentorship is challenging to find. Relational partnerships between resource-rich and -limited settings can provide a particularly effective solution in this setting. [bib_ref] The international partner as invited guest: Beyond colonial and import-export models of..., Whitehead [/bib_ref] Recent medical literature has acknowledged the strong need for mentorship in lowand middle-income countries for the advancement of global health. [bib_ref] Conceptual framework of mentoring in low-and middle-income countries to advance global health, Prasad [/bib_ref] [bib_ref] A scoping review of mentorship of health personnel to improve the quality..., Schwerdtle [/bib_ref] The theoretical underpinning of our mentorship program is in direct alignment with Prasad and colleagues' 9 conceptual framework for mentoring in global health. This framework highlights an ecologic systems model whereby the interaction between the mentor and mentee is securely embedded within several concentric circles, including a mesosystem (eg, interpersonal interactions and alignment of goals and objectives), exosystem (eg, institutional factors, such as policy and resources), macrosystem (eg, national economy and sociopolitical movements), and supramacrosystem (eg, global economy and global politics). [bib_ref] Conceptual framework of mentoring in low-and middle-income countries to advance global health, Prasad [/bib_ref] Ongoing consideration within this framework facilitates a holistic approach to mentorship interactions.
When reflecting on enablers of and barriers to our collaboration, building a strong relational partnership is key. Strengthening research capacity has been touted as a powerful and cost-effective way to improve health and development.The paucity of data on operational guidance, monitoring, and evaluation may be one of the reasons why progress is slower than desired. Well-conceptualized programs with a priori deliverables remain the minority, but are growing in number. [bib_ref] Health research capacity development in low and middle income countries: Reality or..., Franzen [/bib_ref] The ESSENCE Good Practice framework for research capacity strengthening provides useful principles and practical considerations. 15,16 Our program design included many of the principles and deliverables recommended.
Our experience highlights enabling strategies for building a successful relational partnership in education collaboration, which we have operationalized into 8 steps. These include the following: identify potential partners, perform a needs assessment of learners, define the knowledge gap, design and refine the curriculum, select evaluation tools, define short-and long-term deliverables, consider the budget, and repeat the process for each successful group of learners or collaborating partners. Identifying appropriate partners aligns with the ESSENCE principles. The needs assessment of learners is a discrete step well described in educational design 5 to understand the learning needs and perspective of the target audience, intentionally refining educational offerings to align with different learning styles and expectations. The choice of evaluation tools should be focused on instruments that are validated and aligns with the Kirkpatrick model 17 of program evaluation. Adequate consideration of budgetary requirements to support the goals of learners and partnering institutions is also important. We expect this simple action cycle will enable experience to be gained, evidence to be built, and best practices to be followed, doing so with improved efficiency, sustainability, and growth. This is particularly important for academic institutions designing education collaborations with partners from different settings .
There are limitations to our project. Our report represents our first experience with a single institution and a modest cohort of 5 learners. Continued sustainability is required to demonstrate real success with our approach. There is no easy way to quantify the long-term impact of our program as incorporating the skills that have been learned into clinical practice is subject to multiple environmental factors. We expect that, through continuation of our project with successive cohorts of learners, a network of alumni from our mentorship program will provide a greater experience as well as opportunities for longer-term follow up. A qualitative study of trainees who have completed the program from Ghana and Zimbabwe is underway. We are optimistic that building bridges and promoting research capacity can ultimately improve the quality of care and meaningful clinical outcomes for patients in low-and middle-income countries.
In conclusion, we report here on the implementation of a mentorship program focused on research methods and evidence-based medicine in sub-Saharan Africa. To our knowledge, this is a unique model that was established in close collaboration with its intended recipients. It served to build bridges between institutions and fostered friendships and cooperation. The program was successful in the drafting and publication of abstracts and manuscripts by local trainees.
## Affiliations
[fig] 1 2 3 FIG 3: I felt encouraged to ask questions and participate in the discussionFor my experience level, this session wasToo basic (1) , About right (2), Too advanced (3Study-specific questionnaire. [/fig]
[fig] FIG 4: Evaluations of (A) seminars and (B) seminar leads (13-item questionnaire using a 5-point scale). [/fig]
[table] TABLE 1: Research Proposals [/table]
[table] TABLE 2: Resource Requirements [/table]
|
Hemorrhagic Bullous Henoch-Schönlein Purpura: Case Report and Review of the Literature
Henoch-Schönlein Purpura (HSP) or IgA vasculitis is the most common systemic vasculitis of childhood and may affect skin, joints, gastrointestinal tract, and kidneys. Skin manifestations of HSP are characteristic and include a non-thrombocytopenic palpable purpura of the lower extremities and buttocks. Rarely, HSP may initially present as or evolve into hemorrhagic vesicles and bullae. We present an otherwise healthy 5-year-old boy with an acute papulovesicular rash of both legs and intermittent abdominal pain. After a few days the skin lesions rapidly evolved into palpable purpura and hemorrhagic bullous lesions of variable size and severe hemorrhagic HSP was suspected. A histological examination of a skin biopsy showed signs of a small vessel leukocytoclastic vasculitis limited to the upper dermis and direct immunofluorescence analysis revealed IgA deposits in vessel walls, compatible with HSP. To further characterize the clinical picture and treatment options of bullous HSP we performed an extensive literature research and identified 41 additional pediatric patients with bullous HSP. Two thirds of the reported patients were treated with systemic corticosteroids, however, up to 25% of the reported patients developed skin sequelae such as hyperpigmentation and/or scarring. The early use of systemic corticosteroids has been discussed controversially and suggested in some case series to be beneficial by reducing the extent of lesions and minimizing sequelae of disease. Our patient was treated with systemic corticosteroids tapered over 5 weeks. Fading of inflammation resulted in healing of most erosions, however, a deep necrosis developing from a large blister at the dorsum of the right foot persisted so that autologous skin transplantation was performed. Re-examination 11 months after disease onset showed complete clinical remission with re-epithelialization but also scarring of some affected areas.
# Introduction
Henoch-Schönlein Purpura (HSP) or IgA vasculitis is the most common systemic vasculitis of childhood and may affect skin, joints, gastrointestinal tract, and kidneys (1). Skin manifestations of HSP are characteristic and include a non-thrombocytopenic palpable purpura of the lower extremities and buttocks. Rarely, HSP may initially present as or evolve into hemorrhagic vesicles and bullae resulting in cutaneous necrosis (2). We present the case of a HSP patient with hemorrhagic-bullous skin lesions and by summarizing additional reported cases aimed at characterizing the clinical picture of bullous HSP and discussing treatment options.
## Case report
An otherwise healthy 5-year-old boy presented with an acute papulovesicular rash of both legs [fig_ref] FIGURE 1 |: An acute papulovesicular rash of both legs [/fig_ref] and intermittent abdominal pain. The patient did not have fever. Differential blood count (white blood cell count of 12,900/µl with 60% granulocytes and 29% lymphocytes), C-reactive protein (0.45 mg/dl), erythrocyte sedimentation rate (20 mm/h) and global blood clotting tests (INR 0.98, PTT 31.5 s) were normal. Serum IgA (147 mg/dl) and IgM (66 mg/dl) levels were within age matched reference ranges whereas IgG levels were slightly decreased (557 mg/dl, reference range 640-1420). No hematuria or fecal occult blood could be detected. Abdominal ultrasound could exclude intussusception but revealed thickened bowel wall at the ileocecal junction. The abdominal symptoms resolved spontaneously within 2 days but arthralgia appeared thereafter.
Nine days after the onset of disease the skin lesions at the arms, legs, feet and ankles rapidly evolved into palpable purpura and hemorrhagic-bullous lesions of variable size ranging from 5 to 40 mm [fig_ref] FIGURE 1 |: An acute papulovesicular rash of both legs [/fig_ref]. Some of the blisters spontaneously ruptured and disclosed hemorrhagic fluid which remained sterile in the microbiological work-up. The patient was given cefuroxime as antibiotic prophylaxis. Severe hemorrhagicbullous HSP was suspected but differential diagnoses included septicemia/septic emboli and autoimmune blistering disease. Absence of fever and leukocytosis and sterile blood cultures argued against an infectious etiology. Neither circulating antibodies directed against structural proteins of the basementmembrane zone nor ANAs or ANCAs could be detected in the patient's serum. C4 levels were in the normal range while C3c levels were slightly elevated (152 mg/dl, reference range 80-120). A skin biopsy was performed and histological examination showed signs of a small vessel leukocytoclastic vasculitis limited to the upper dermis [fig_ref] FIGURE 2 |: Skin biopsy of one lesion in H&E staining revealing signs of a... [/fig_ref] , and direct immunofluorescence analysis revealed IgA and C3 deposits in vessel walls, compatible with HSP.
The patient was treated with oral corticosteroids (prednisolone 1 mg/kg/day) for 7 days, then subsequently tapered over 39 additional days. Although fading of inflammation paralleled healing of most erosions, a deep necrosis resulting from a large blister at the dorsum of the right foot persisted [fig_ref] FIGURE 3 |: Hemorrhagic bullae developed on both feet an lower legs [/fig_ref] so that autologous skin transplantation was performed. Re-examination 11 months after disease onset showed complete clinical remission of disease with re-epithelialization but also scarring of some affected areas [fig_ref] FIGURE 3 |: Hemorrhagic bullae developed on both feet an lower legs [/fig_ref]
## Literature search
Skin manifestations of HSP are characteristic and include a nonthrombocytopenic palpable purpura of the lower extremities and buttocks (1). Rarely, HSP in children may initially present as or evolve into hemorrhagic vesicles and bullae which may result in necrosis of the skin and subsequently scarring (2). To characterize bullous HSP in children we performed a literature search and searched NCBI PubMed for publications using "bullous purpura" as well as the combination of "bullous" and "purpura." In addition to our patient we identified 41 pediatric HSP patients with bullous skin lesions (Table S1) (3-32). Included were 23 isolated case reports and 6 case series with between 2 and 6 patients. We excluded two of these patients from further analysis since both were later on diagnosed with epidermolysis bullosa and might therefore have dual pathology. We compared the clinical and demographic characteristics of our and the remaining 39 patients (in total 40 patients) with those reported in 4 separate cohorts of unselected pediatric HSP patients using unpaired t-test for continuous variables and Fisher's exact test for categorial variables [fig_ref] TABLE 1 |: Comparison of reported patients with bullous Henoch-Schönlein purpura with unselected cohorts [/fig_ref] [bib_ref] Report of 100 patients and review of the literature, Saulsbury [/bib_ref] [bib_ref] Henoch-Schonlein purpura in children from northwestern Spain: a 20-year epidemiologic and clinical..., Calvino [/bib_ref] [bib_ref] Risk factors of renal involvement and significant proteinuria in Henoch-Schonlein purpura, Sano [/bib_ref]. These cohorts represent the complete clinical range of HSP, however, due to their low prevalence, patients with bullous skin lesions do not significantly contribute to the overall characteristics of these cohorts.
The mean age of patients with bullous HSP patients at disease onset was significantly higher compared with each of the four cohorts of unselected HSP patients [fig_ref] Abbreviations [/fig_ref]. The prevalence of joint, abdominal and renal involvement in bullous HSP patients was comparable to that of the unselected HSP patient cohorts [fig_ref] Abbreviations [/fig_ref].
In all except two cohorts, bullous HSP patients were treated more often with systemic corticosteroids than the average group of HSP patients [fig_ref] TABLE 1 |: Comparison of reported patients with bullous Henoch-Schönlein purpura with unselected cohorts [/fig_ref]. In addition, two (5.0%) out of the 40 bullous HSP patients were treated with colchicine, 3 (7.5%) with azathioprine and one (2.5%) with dapsone [fig_ref] Abbreviations [/fig_ref]. Three (7.5%) patients were treated with topical steroids and 11 (27.5%) did not receive any anti-inflammatory medical treatment [fig_ref] Abbreviations [/fig_ref].
Bullous lesions resolved in all 11 reported patients who did not receive anti-inflammatory medical treatment and lesions did not re-occur. Hyperpigmentation and/or scar formation has been reported in 11 patients, in 10 of these despite medical treatment. However, detailed information regarding skin sequelae including explicit exclusion of scarring and/or hyperpigmentation is missing in many of the reported patients [fig_ref] Abbreviations [/fig_ref]. Interestingly, in 3 out of the 26 (11.5%) corticosteroid treated bullous HSP patients, corticosteroid treatment was already initiated up to 10 days before appearance of hemorrhagicbullous lesions, mainly due to abdominal involvement [fig_ref] TABLE 1 |: Comparison of reported patients with bullous Henoch-Schönlein purpura with unselected cohorts [/fig_ref].
# Discussion
Palpable purpura mainly affecting the buttocks and legs is the characteristic skin manifestations of HSP. The differential diagnosis of these lesions include congenital (e.g., thrombocytopenia, hemophilia) or acquired (purpura fulminans, disseminated intravascular coagulation) forms of bleeding diathesis and other types of vasculitis (e.g., granulomatosis with polyangiitis, eosinophilic granulomatosis with polyangiitis). However, HSP may initially present as or evolve into hemorrhagic vesicles and bullae progressing to cutaneous necrosis which may lead to a diagnostic dilemma (2). The differential diagnosis of blistering skin lesions includes infections (e.g., HSV), toxins (e.g., staphylococcal scalded skin syndrome), autoimmune diseases (e.g., pemphigus), genetic diseases (e.g., epidermolysis bullosa) as well as physical pressure or trauma [bib_ref] Blistering disorders in children, Yun [/bib_ref].
Whereas, hemorrhagic-bullous lesions are rather common in adult HSP patients, children with HSP less often develop this type of lesions (<2%) [bib_ref] Schonlein-Henoch purpura in adult patients. Predictive factors for IgA glomerulonephritis in a..., Tancrede-Bohin [/bib_ref]. Due to the wide differential diagnosis, a skin biopsy was performed in our presented case as it has been done in more than half of the reported patients with bullous HSP [fig_ref] TABLE 1 |: Comparison of reported patients with bullous Henoch-Schönlein purpura with unselected cohorts [/fig_ref]. Histological analysis revealed signs of leucocytoclastic vasculitis in all reported patients, however, IgA deposition could not be detected in all patients [fig_ref] TABLE 1 |: Comparison of reported patients with bullous Henoch-Schönlein purpura with unselected cohorts [/fig_ref]. This might be explained by the time of biopsy, since deposition of IgA is more difficult to detect in older lesions. Therefore, the biopsy should be obtained from the border of a fresh non-bullous and non-necrotized lesion where proteolytic degradation of IgA is less advanced. Skin biopsy in our patient was performed 10 days after onset of the first bullous lesion but targeted a lesion that evolved <2 days before biopsy. Histological analysis of this skin biopsy revealed the characteristic signs of leucocytoclastic vasculitis with IgA deposition thereby confirming our tentative diagnosis of HSP in or patient.
It has been speculated whether appearance of bullous lesions reflects a more severe disease course or even a distinct disease entity of HSP. The long-term prognosis in HSP is mainly determined by the extent of renal involvement [bib_ref] Henoch-Schonlein purpura nephritis in children, Davin [/bib_ref]. However, our literature review on bullous lesions in pediatric HSP revealed rather less renal involvement in this patient group and seems to contradict this assumption. In most of the published patients the bullous lesions resolved within a few weeks and did not re-occur which led to the assumption of a rather benign and self-limiting course of bullous HSP (2). However, remaining hyperpigmentation and/or scarring have been reported in at least 25% of the reported patients [bib_ref] Bullous lesions in Henoch Schonlein Purpura as indication to start systemic prednisone, Boer [/bib_ref] [bib_ref] Severe hemorrhagic bullous lesions in Henoch Schonlein purpura: three pediatric cases and..., Trapani [/bib_ref] [bib_ref] Bullous Henoch-Schonlein purpura in children, Chen [/bib_ref]. Since information regarding skin sequelae in the disease course of the reported patients is scarce and scarring has not explicitly been excluded in many of the reported patients, the extent of reported skin sequelae rather might be underestimated. In contrast, re-occurrence of active disease has only been described in very few individual cases of bullous HSP and most patients followed a monophasic disease course. However, two patients described to have clinical findings compatible with bullous HSP suffered from re-occurrence of bullous skin lesions 1 and 2.5 years after onset of the first symptoms [bib_ref] Hemorrhagic, bullous Henoch Schonlein purpura in a 16-year-old girl with previously undiagnosed..., Abdulla [/bib_ref] [bib_ref] Bullous Henoch-Schonlein purpura in children: a report of 6 cases and review..., Maguiness [/bib_ref] and were later diagnosed with dominant dystrophic epidermolysis bullosa (DDEB) and recessive-dystrophic epidermolysis bullosa (RDEB), respectively. Histological analysis of a skin biopsy performed at first onset of bullous lesions in the patient with DDEB revealed a non-inflammatory subepidermal bulla in addition to the characteristic signs of leukocytoclastic vasculitis. The pathophysiology of blistering skin lesions in HSP still remains unclear. Matrix metalloproteinase (MMP)-2 and MMP-9 have been detected in blister fluid and suggested to cause proteolysis of collagen (6). However, the triggers that drive the secretion of these enzymes are not well-understood. Therefore, future research should address the inflammatory mediators that are involved in blister formation in HSP and might be targets of more specific therapies. It might be speculated that bullous skin lesions preferentially develop in HSP patients with intrinsic (e.g., genetic predisposition, general skin fragility) or extrinsic (e.g., physical skin pressure) risk factors but that the appearance of bullous skin lesions does not mirror a distinct pathophysiology of HSP itself.
The optimal treatment regime of patients with bullous HSP is still controversial. The use of systemic corticosteroids in HSP seems to shorten the abdominal pain and arthralgia and decrease the prevalence of purpura [bib_ref] Clinical course of extrarenal symptoms in Henoch-Schonlein purpura: a 6-month prospective study, Jauhola [/bib_ref] [bib_ref] Early prednisone therapy in Henoch-Schonlein purpura: a randomized, double-blind, placebo-controlled trial, Ronkainen [/bib_ref]. Although corticosteroids do not seem to hinder renal manifestation during the disease course, they may reduce the risk of end-stage renal disease [bib_ref] Renal manifestations of Henoch-Schonlein purpura in a 6-month prospective study of 223..., Jauhola [/bib_ref] [bib_ref] Effects of corticosteroid on Henoch-Schonlein purpura: a systematic review, Weiss [/bib_ref]. There is no clear evidence for a beneficial use of systemic corticosteroids in treating non-bullous skin lesions in HSP. Some authors of published cases with bullous HSP suggested that early systemic treatment with corticosteroids may be beneficial in bullous HSP by promoting faster resolution of skin symptoms and reducing scar formation [bib_ref] Bullous lesions in Henoch Schonlein Purpura as indication to start systemic prednisone, Boer [/bib_ref]. In contrast, the authors of another systematic case review of patients with bullous HSP concluded that patients with bullous HSP managed expectantly and those managed with systemic steroids display a similar disease course (2).
Whereas, resolution and absence of re-occurrence of lesions has mainly been used as parameters to describe the diseases course and response to therapy in most of the reported patients, less attention has been paid to the potential sequelae of necrotic skin lesions (scarring, hyperpigmentation). Of note, at least 25% of the reported patients with bullous HSP developed some kind of sequelae ranging from hyperpigmented skin areas to deep necrosis necessitating autologous skin transplantation as in our patient [fig_ref] Abbreviations [/fig_ref].
Despite missing of strong evidence due to lack of randomizedcontrolled trials in bullous HSP, it might be speculated that corticosteroid treatment initiated very early in the disease course of bullous HSP may reduce the severity of necrosis and thereby the extent of skin sequelae. In contrast to this assumption, evolution of bullous skin lesions in HSP patients already receiving high-dose systemic steroids due to abdominal involvement has been documented in 3 cases (7, 21, 22). Additionally, 12 published patients with bullous HSP who showed resolution of skin lesions did not receive any antiinflammatory medical treatment (3, 4, 6, [bib_ref] Medical images. An unusual manifestation of Henoch-Schonlein purpura: haemorrhagic bullous lesions, Aydinoz [/bib_ref] [bib_ref] Management of haemorrhagic bullous lesions in Henoch-Schonlein purpura: is there any consensus?, Kausar [/bib_ref] [bib_ref] Bullous Henoch-Schonlein purpura in children: a report of 6 cases and review..., Maguiness [/bib_ref] [bib_ref] Severe hemorrhagic bullous lesions in Henoch Schonlein purpura: three pediatric cases and..., Trapani [/bib_ref] [bib_ref] Self-limiting recurrent bullous Henoch-Schonlein purpura with lupus anticoagulant, Gration [/bib_ref] [bib_ref] Case report: bullous Henoch-Schonlein purpura, Hooper [/bib_ref]. Although some of these did not develop scarring and/or hyperpigmentation, explicit information about exclusion of skin sequelae is missing in many of these patients. Of note, side effects of systemic corticosteroid treatment have not been described in any of the reported patients, which in our opinion additionally favors the attempt of ameliorating the disease course by using systemic corticosteroids early in the diseases course in bullous HSP.
Steroid-sparing medication has been introduced in patients with long-lasting disease course or inadequate responsiveness to corticosteroids. These treatment regimens included colchicine, azathioprine, or dapsone and have been described in only a few cases [bib_ref] Bullous lesions in Henoch-Schonlein purpura, Chan [/bib_ref] [bib_ref] Severe hemorrhagic bullous lesions in Henoch Schonlein purpura: three pediatric cases and..., Trapani [/bib_ref] [bib_ref] Hemorrhagic bullous lesions in a girl with Henoch Schonlein purpura, Mehra [/bib_ref] [bib_ref] Bullous Henoch-Schonlein purpura in children, Chen [/bib_ref]. Despite the reported benefits in these patients during treatment, the small number of published case does not allow any general conclusion.
## Summary and concluding remarks
-Hemorrhagic bullous skin lesions may rarely occur in childhood HSP -A diagnostic skin biopsy is helpful in the differential work-up of these patients -Blistering skin lesions in HSP do not generally seem to be associated with more severe extracutaneous/systemic involvement and might not have any prognostic value -Most of the bullous skin lesions resolve within a few weeks but may lead to necrosis resulting in scarring and/or hyperpigmentation; re-occurrence of bullous lesions may hint at an underlying disease -The early use of systemic corticosteroids has been suggested in bullous HSP patients but their beneficial effect on the extent of lesions and/or the sequelae of disease has not been analyzed in randomized controlled trials yet -Supportive care (e.g., avoiding skin pressure, pain control, antibiotic prophylaxis/treatment) is beneficial for all patients
# Ethics statement
The authors state that written informed consent was obtained from the parents of the patient for the publication of this case report.
# Author contributions
HM coordinated the writing group. MN and HM performed the literature review. All authors were involved in the patient care, critically reviewed the manuscript and approved the final version.
# Funding
This publication was funded by the German Research Foundation (DFG) and the University of Wuerzburg in the funding programme Open Access Publishing.
# Supplementary material
The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fped. 2018.00413/full#supplementary-material
[fig] FIGURE 1 |: An acute papulovesicular rash of both legs (a) rapidly evolved into palpable purpura and hemorrhagic-bullous lesions of variable size ranging from 5 to 40 mm (b,c). [/fig]
[fig] FIGURE 2 |: Skin biopsy of one lesion in H&E staining revealing signs of a small vessel leukocytoclastic vasculitis limited to the upper dermis. [/fig]
[fig] FIGURE 3 |: Hemorrhagic bullae developed on both feet an lower legs (a-c). A deep necrosis resulting from a large blister at the dorsum of the right feet evolved (a,b) neccessitating autologous skin transplantation. Re-examination 11 months after disease onset showed complete clinical remission of disease with re-epithelialization of affected areas (d-f). [/fig]
[table] TABLE 1 |: Comparison of reported patients with bullous Henoch-Schönlein purpura with unselected cohorts. [/table]
|
Doctor Referral of Overweight People to Low Energy total diet replacement Treatment (DROPLET): pragmatic randomised controlled trial
[bib_ref] From quantum dynamics to the canonical distribution: general picture and rigorous example, Tasaki [/bib_ref] [bib_ref] Entanglement and the foundations of statistical mechanics, Popescu [/bib_ref] [bib_ref] Thermalization and its mechanism for generic isolated quantum systems, Rigol [/bib_ref] [bib_ref] Quantum many-body systems out of equilibrium, Eisert [/bib_ref] [bib_ref] Thermalization and its mechanism for generic isolated quantum systems, Rigol [/bib_ref] [bib_ref] Quantum statistical mechanics in a closed system, Deutsch [/bib_ref] [bib_ref] Chaos and quantum thermalization, Srednicki [/bib_ref] [bib_ref] Thermalization of interacting fermions and delocalization in Fock space, Neuenhahn [/bib_ref] [bib_ref] Alternatives to eigenstate thermalization, Rigol [/bib_ref] [bib_ref] Finite-size scaling analysis of the eigenstate thermalization hypothesis in a one-dimensional interacting..., Ikeda [/bib_ref] [bib_ref] Finite-size scaling of eigenstate thermalization, Beugeling [/bib_ref] [bib_ref] Pushing the limits of the eigenstate thermalization hypothesis towards mesoscopic quantum systems, Steinigeweg [/bib_ref] [bib_ref] Thermal equilibrium of a macroscopic quantum system in a pure state, Goldstein [/bib_ref] [bib_ref] Exploring local quantum many-body relaxation by atoms in optical superlattices, Cramer [/bib_ref] [bib_ref] Probing the relaxation towards equilibrium in an isolated strongly correlated one-dimensional Bose..., Trotzky [/bib_ref] [bib_ref] Relaxation and prethermalization in an isolated quantum system, Gring [/bib_ref] [bib_ref] Relaxation dynamics of a Fermi gas in an optical superlattice, Pertot [/bib_ref] [bib_ref] Thermalization and its mechanism for generic isolated quantum systems, Rigol [/bib_ref] [bib_ref] Breakdown of thermalization in finite one-dimensional systems, Rigol [/bib_ref] [bib_ref] Quantum quenches and thermalization in one-dimensional fermionic systems, Rigol [/bib_ref] [bib_ref] Onset of quantum chaos in one-dimensional bosonic and fermionic systems and its..., Santos [/bib_ref] [bib_ref] Many-body localization phase transitions, Pal [/bib_ref] [bib_ref] Effect of rare fluctuations on the thermalization of isolated quantum systems, Brioli [/bib_ref] [bib_ref] Absence of thermalization in nonintegrable systems, Gogolin [/bib_ref] [bib_ref] Strong and weak thermalization of infinite non-integrable quantum systems, Banuls [/bib_ref] [bib_ref] From quantum dynamics to the canonical distribution: general picture and rigorous example, Tasaki [/bib_ref] [bib_ref] English translation by Tumulka, R. Proof of the ergodic theorem and the..., Von Neumann [/bib_ref] [bib_ref] Long-time behavior of macroscopic quantum systems: commentary accompanying the english translation of..., Goldstein [/bib_ref] [bib_ref] Approach to thermal equilibrium of macroscopic quantum systems, Goldstein [/bib_ref] [bib_ref] Normal typicality and von Neumann's quantum ergodic theorem, Goldstein [/bib_ref] [bib_ref] Generalization of von Neumann's approach to thermalization, Reimann [/bib_ref] [bib_ref] Exact relaxation in a class of non-equilibrium quantum lattice systems, Cramer [/bib_ref] [bib_ref] Foundation of statistical mechanics under experimentally realistic conditions, Reimann [/bib_ref] [bib_ref] Quantum mechanical evolution towards equilibrium, Linden [/bib_ref] [bib_ref] Canonical thermalization, Reimann [/bib_ref] [bib_ref] Equilibration of quantum systems and subsystems, Short [/bib_ref] [bib_ref] Equilibration of macroscopic quantum systems, Reimann [/bib_ref] [bib_ref] Equilibration of isolated macroscopic quantum systems under experimentally realistic conditions, Reimann [/bib_ref] [bib_ref] Quantum equilibration in finite time, Short [/bib_ref] [bib_ref] Quantum systems equilibrate rapidly for most observables, Malabarba [/bib_ref] [bib_ref] Extremely quick thermalization in a macroscopic quantum system for a typical nonequilibrium..., Goldstein [/bib_ref] [bib_ref] Long-time behavior of macroscopic quantum systems: commentary accompanying the english translation of..., Goldstein [/bib_ref] [bib_ref] Approach to thermal equilibrium of macroscopic quantum systems, Goldstein [/bib_ref] [bib_ref] Normal typicality and von Neumann's quantum ergodic theorem, Goldstein [/bib_ref] [bib_ref] Generalization of von Neumann's approach to thermalization, Reimann [/bib_ref]
# Results
Set-up. Employing textbook quantum mechanics, we consider time-independent Hamiltonians H with eigenvalues E n and eigenvectors |ni on a Hilbert space H of large (but finite) dimensionality D ) 1. As usual, system states (pure or mixed) are described by density operators r : H ! H and observables by Hermitian operators A : H ! H with matrix elements r mn : ¼ hm|r|ni and A mn : ¼ hm|A|ni, respectively. Expectation values are given by hAi r : ¼ Tr{rA} and the time evolution by
[formula] rðtÞ ¼ U t rð0ÞU w t with propagator U t :¼ e À iHt=' , yielding A h i rðtÞ ¼ X D m;n¼1 r mn ð0ÞA nm e iðE n À E m Þt=' :ð1Þ [/formula]
The main examples are closed many-body systems with a macroscopically well-defined energy, that is, all relevant eigenvalues E 1 , ..., E D are contained in some microcanonical energy window [E À DE, E], where DE is small on the macroscopic but large on the microscopic scale. For systems with f ) 1 degrees of freedom, D is then exponentially large in f . Accordingly, the relevant Hilbert space H is spanned by the eigenvectors n j i f g D n¼1 and is sometimes also named energy shell or active Hilbert space, see, for example, refs 8-12 and Supplementary Note 2 for more details.
Analytical results. Our main players are the three Hermitian operators H (Hamiltonian), A (observable) and r(0) (initial state), each with its own eigenvalues (spectrum) and eigenvectors (basis of H). In the following, the three spectra will be considered as arbitrary but fixed, while the eigenbases will be randomly varied relatively to each other. More precisely, all unitary transformations U : H ! H between the eigenbases of H and A are considered as equally likely (Haar distributed [bib_ref] English translation by Tumulka, R. Proof of the ergodic theorem and the..., Von Neumann [/bib_ref] [bib_ref] Long-time behavior of macroscopic quantum systems: commentary accompanying the english translation of..., Goldstein [/bib_ref] [bib_ref] Approach to thermal equilibrium of macroscopic quantum systems, Goldstein [/bib_ref] [bib_ref] Normal typicality and von Neumann's quantum ergodic theorem, Goldstein [/bib_ref] , while the basis of r(0) relatively to that of A is arbitrary but fixed. (Equivalently, we could let 'rotate' H relatively to r(0) while keeping A fixed relatively to r(0).) In particular, the initial expectation value hAi r(0) can be chosen arbitrary but then remains fixed (U independent). It is only for times t40 that the randomness of the unitary U also randomizes (via H) the further temporal evolution of r(t) and thus of hAi r(t) .
The basic idea behind this randomization of U is akin to random matrix theory 9-12 , namely, to derive an approximation for hAi r(t) , which applies to the overwhelming majority of all those randomly sampled U's, hence it typically should apply also to the particular (non-random) U of the actual system of interest. A more detailed justification of this 'typicality approach' will be provided in the section 'Typicality of thermalization'.
Since A mn refers to the basis of H, these matrix elements depend on U, and likewise for r mn (0) (the explicit formulae are provided in the section 'Basic matrices'). Indicating averages over U by the symbol [?] U and exploiting that all basis transformations U are equally likely, it follows for symmetry reasons that [r nn (0)A nn ] U must be independent of n. Likewise, [r mn (0)A nm ] U must be independent of m and n for all man. We thus can conclude that for any n
[formula] D r nn ð0ÞA nn ½ U ¼ X D k¼1 r kk ð0ÞA kk " # Uð2Þ [/formula]
and that for any man
[formula] DðD À 1Þ r mn ð0ÞA nm ½ U ¼ X j = ¼k r jk ð0ÞA kj 2 4 3 5 U ¼ X D j;k¼1 r jk ð0ÞA kj 2 4 3 5 U À X D k¼1 r kk ð0ÞA kk " # U :ð3Þ ARTICLE½ U ¼ Trfr av Ag=D ¼ A h i r av =Dð4Þ [/formula]
for arbitrary n. Likewise, equation (3) yields
[formula] r mn ð0ÞA nm ½ U ¼ A h i rð0Þ À A h i r av DðD À 1Þð5Þ [/formula]
for arbitrary man. Upon separately averaging in equation (1), the summands with m ¼ n and those with man over U, and then exploiting equations (4) and (5) one readily finds that
[formula] A h i rðtÞ h i U ¼ A h i r av þ FðtÞ A h i rð0Þ À A h i r av n oð6Þ [/formula]
FðtÞ :
[formula] ¼ D D À 1 fðtÞ j j 2 À 1 D ;ð7Þ [/formula]
where f(t) is the Fourier transform of the spectral density from ref. [bib_ref] Thermalization under randomized local Hamiltonians, Cramer [/bib_ref] (see also refs 51-53)
[formula] fðtÞ :¼ 1 D X D n¼1 e iEnt=' :ð8Þ [/formula]
The following results can be derived in principle along similar lines (symmetry arguments being one key ingredient), but since the actual details are quite tedious, they are postponed to Methods. As a first result, one obtains
[formula] A h i r av ¼ A h i r mc þ A h i rð0Þ À A h i r mc D þ 1 ;ð9Þ [/formula]
where r mc : ¼ I/D is the microcanonical density operator and I the identity on H. As a second result, one finds for the statistical fluctuations
[formula] xðtÞ :¼ A h i rðtÞ À A h i rðtÞ h i Uð10Þ [/formula]
the estimate
[formula] x 2 ðtÞ Â Ã U ¼ O D 2 A Tr r 2 ð0Þ È É =D À Áð11Þ [/formula]
for arbitrary t, where D A is the range of A, that is, the difference between the largest and smallest eigenvalues of A. Since averaging over U and integrating over t are commuting operations, equation [bib_ref] Normal typicality and von Neumann's quantum ergodic theorem, Goldstein [/bib_ref] implies that 1
[formula] t 2 À t 1 Z t2 t 1 x 2 ðtÞdt ! U ¼ O D 2 A Tr r 2 ð0Þ f g Dð12Þ [/formula]
for arbitrary t 2 4t 1 . Considering t in equation [bib_ref] Normal typicality and von Neumann's quantum ergodic theorem, Goldstein [/bib_ref] as arbitrary but fixed, equation [bib_ref] Approach to thermal equilibrium of macroscopic quantum systems, Goldstein [/bib_ref] and D ) 1 imply (obviously or by exploiting Chebyshev's inequality [bib_ref] From quantum dynamics to the canonical distribution: general picture and rigorous example, Tasaki [/bib_ref] [bib_ref] Long-time behavior of macroscopic quantum systems: commentary accompanying the english translation of..., Goldstein [/bib_ref] [bib_ref] Quantum mechanical evolution towards equilibrium, Linden [/bib_ref] [bib_ref] Equilibration of macroscopic quantum systems, Reimann [/bib_ref] [bib_ref] Quantum equilibration in finite time, Short [/bib_ref] that hAi r(t) is practically indistinguishable from the average in equation [bib_ref] Quantum many-body systems out of equilibrium, Eisert [/bib_ref] for the vast majority of all unitaries U. Indeed, the fraction (normalized Haar measure) of exceptional U's is unimaginably small for typical macroscopic systems with, say, fE10 23 degrees of freedom, since D in equation [bib_ref] Normal typicality and von Neumann's quantum ergodic theorem, Goldstein [/bib_ref] is exponentially large in f (see below equation (1)). Likewise, considering an arbitrary but fixed time interval [t 1 ,t 2 ] in equation [bib_ref] Generalization of von Neumann's approach to thermalization, Reimann [/bib_ref] , it follows for all but a tiny fraction of U's that the time average over x 2 (t) on the left-hand side of equation [bib_ref] Generalization of von Neumann's approach to thermalization, Reimann [/bib_ref] must be unimaginably small, and hence also the integrand x 2 (t) itself must be exceedingly small for the overwhelming majority of all tA[t 1 ,t 2 ]. Accordingly, hAi r(t) must remain extremely close to equation (6) simultaneously for all those tA[t 1 ,t 2 ].
Due to equation [bib_ref] Long-time behavior of macroscopic quantum systems: commentary accompanying the english translation of..., Goldstein [/bib_ref] and D ) 1, we furthermore can safely approximate A h i r av in equation (6) by A h i r mc . Altogether, we thus can conclude that in very good approximation
[formula] A h i rðtÞ ¼ A h i r mc þ FðtÞ A h i rð0Þ À A h i r mc n oð13Þ [/formula]
for the vast majority of unitaries U and times t. As detailed in Methods, the neglected corrections in equationconsist of a systematic (U independent) part, which is bounded in modulus by D A /(D 2 À 1) for all t, and a random (U dependent) part (namely x(t)), whose typical order of magnitude is D A ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Tr r 2 ð0Þ f g=D p (for most U and t, cf. equations [bib_ref] Normal typicality and von Neumann's quantum ergodic theorem, Goldstein [/bib_ref] and [bib_ref] Generalization of von Neumann's approach to thermalization, Reimann [/bib_ref] , that is, x(t) is dominating by far (note that 1 ! Tr r 2 ð0Þ f g! Tr r 2 mc È É ¼ 1=D). Moreover, the correlations of x(t) decay on timescales comparable to those governing F(t).
These are our main formal results. In the rest of the paper, we discuss their physical content.
Basic properties of F(t). Equation [bib_ref] English translation by Tumulka, R. Proof of the ergodic theorem and the..., Von Neumann [/bib_ref] implies that f(0) ¼ 1, f( À t) ¼ f*(t) and |f(t)|r1. With equation (7) and D ) 1, it follows that in very good approximation
[formula] FðtÞ ¼ fðtÞ j j 2 ;ð14Þ [/formula]
and thus
Fð0Þ ¼ 1; 0 FðtÞ 1; Fð À tÞ ¼ FðtÞ:
Indicating averages over all tZ0 by an overbar, one can infer from equations [bib_ref] English translation by Tumulka, R. Proof of the ergodic theorem and the..., Von Neumann [/bib_ref] and (14) that
[formula] FðtÞ ¼ P k d 2 k =D 2 , [/formula]
where k labels the eigenspaces of H with mutually different eigenvalues and d k denotes their dimensions. Since P k d k ¼ D, we thus obtain FðtÞ max k ðd k =DÞ. Excluding extremely large multiplicities (degeneracies) of energy eigenvalues, it follows that the time average FðtÞ is negligibly small, and hence [bib_ref] From quantum dynamics to the canonical distribution: general picture and rigorous example, Tasaki [/bib_ref] [bib_ref] Long-time behavior of macroscopic quantum systems: commentary accompanying the english translation of..., Goldstein [/bib_ref] [bib_ref] Quantum mechanical evolution towards equilibrium, Linden [/bib_ref] [bib_ref] Equilibration of macroscopic quantum systems, Reimann [/bib_ref] [bib_ref] Quantum equilibration in finite time, Short [/bib_ref] that F(t) itself must be negligibly small for the overwhelming majority of all sufficiently large t, symbolically indicated as Fðt ! 1Þ*0: ð16Þ
Note that there still exist arbitrarily large exceptional t's owing to the quasi-periodicity of f(t) implied by equation [bib_ref] English translation by Tumulka, R. Proof of the ergodic theorem and the..., Von Neumann [/bib_ref]. We also emphasize that our main result (13) itself admits arbitrary degeneracies of H.
As an example, we focus on the microcanonical set-up introduced below equation (1) and on not too large times, so that equation (8) is well approximated by
[formula] fðtÞ ¼ Z E E À DE rðxÞe ixt=' dx;ð17Þ [/formula]
where r(x) represents the (smoothened and normalized) density of energy levels E n in the vicinity of the reference energy x. If the level density is constant throughout the energy window [E À DE, E], we thus obtain with equation (14)
[formula] FðtÞ ¼ sin 2 ðDEt=2' Þ ðDEt=2' Þ 2 :ð18Þ [/formula]
Next, we recall Boltzmann's entropy formula S(x) ¼ k B ln (O(x)), where O(x) counts the number of E n 's below x and k B is Boltzmann's constant. Hence, O 0 (x) must be proportional to the level density r(x) from above. Furthermore, T: ¼ 1/S 0 (E) is the usual microcanonical temperature of a system with energy E at thermal equilibrium. A straightforward expansion then yields the approximation rðE À yÞ ¼ ce À y=k B T for yZ0, where c is fixed via
[formula] R E E À DE rðxÞdx ¼ 1. [/formula]
The omitted higher-order terms are safely negligible for all yZ0 and systems with f ) 1 degrees of freedom, see also ref. [bib_ref] General relaxation time of the fidelity for isolated quantum thermodynamic systems, Monnai [/bib_ref]. With equations [bib_ref] Concentration of measure for quantum states with a fixed expectation value, Müller [/bib_ref] and (17), one thus finds
[formula] FðtÞ ¼ 1 À 2a cosðDE t=' Þ þ a 2 ð1 À aÞ 2 1 þ ðk B Tt=' Þ 2 Â Ã;ð19Þ [/formula]
where a :¼ e À DE=kBT . For DE ( k B T, one recovers equation (18), and for DE ) k B T, one obtains
[formula] FðtÞ ¼ 1 1 þ k B Tt=' ð Þ 2 :ð20Þ [/formula]
Typicality of thermalization. Equations (13) and (16) imply thermalization in the sense that the expectation value hAi r(t) becomes (for most U) practically indistinguishable from the microcanonical average A h i r mc for the overwhelming majority of all sufficiently large t. Exceptional t's are, for instance, due to quantum revivals, which, in turn, are apparently closely related to the quasi-periodicities of F(t).
Our assumption that energy eigenvalues must not be extremely highly degenerate (see above equation (16)) is similar to refs 46,47,49-51, but considerably weaker than the corresponding premises in most other related works [bib_ref] From quantum dynamics to the canonical distribution: general picture and rigorous example, Tasaki [/bib_ref] [bib_ref] English translation by Tumulka, R. Proof of the ergodic theorem and the..., Von Neumann [/bib_ref] [bib_ref] Long-time behavior of macroscopic quantum systems: commentary accompanying the english translation of..., Goldstein [/bib_ref] [bib_ref] Approach to thermal equilibrium of macroscopic quantum systems, Goldstein [/bib_ref] [bib_ref] Normal typicality and von Neumann's quantum ergodic theorem, Goldstein [/bib_ref] [bib_ref] Generalization of von Neumann's approach to thermalization, Reimann [/bib_ref] [bib_ref] Foundation of statistical mechanics under experimentally realistic conditions, Reimann [/bib_ref] [bib_ref] Quantum mechanical evolution towards equilibrium, Linden [/bib_ref] [bib_ref] Canonical thermalization, Reimann [/bib_ref] [bib_ref] Equilibration of quantum systems and subsystems, Short [/bib_ref] [bib_ref] Equilibration of macroscopic quantum systems, Reimann [/bib_ref] [bib_ref] Equilibration of isolated macroscopic quantum systems under experimentally realistic conditions, Reimann [/bib_ref] [bib_ref] Quantum equilibration in finite time, Short [/bib_ref].
The usual time inversion invariance on the fundamental, microscopic level 7 is maintained by equation (13) due to equation [bib_ref] On the basis of quantum statistical mechanics, Sugita [/bib_ref]. Surprisingly, and in accordance with the second law of thermodynamics, the latter symmetry persists even if it is broken in the microscopic quantum dynamics, for example, by an external magnetic field.
By propagating r(0) backward in time (with respect to one particular U) and taking the result as new initial state, one may easily tailor 41 examples of the very rare U's and t's, which notably deviate from the typical behaviour. Equivalently, one may back-propagate A instead of r(0) (Heisenberg picture).
Note that S and T were introduced below equation (18) not in the sense of associating some entropy and temperature to the non-equilibrium states r(t) but rather as a convenient levelcounting tool. However, we now can identify them a posteriori with the pertinent entropy and temperature after thermalization.
The randomization via U (see the section 'Analytical results') can be viewed in two ways: either one considers r(0), A and the spectrum of H as arbitrary but fixed, while the eigenbasis of H is sampled from a uniform distribution (Haar measure). Or one considers H and the spectra of r(0) and A as arbitrary but fixed and randomizes the eigenvectors of A and r(0). In doing so, a key point is that the relative orientation of the eigenbases of r(0) and A can be chosen arbitrarily but then is kept fixed. Indeed, it is well known 12,49 that for 'most' such orientations the expectation values hAi r(0) and A h i r mc are practically indistinguishable, that is, an initial hAi r(0) far from equilibrium requires a careful fine-tuning of r(0) relatively to A.
In reality, there is usually nothing random in the actual physical systems one has in mind. Hence, results such as equation, which (approximately) apply to the overwhelming majority of unitaries U, should be physically interpreted according to the common lore of random matrix theory 9,10,12 , namely, as to apply practically for sure to a concrete system under consideration, unless there are particular reasons to the contrary.
Such reasons arise, for instance, when A is known to be a conserved quantity, implying a common eigenbasis of A and H, that is, the basis transformations U must indeed be very special. Furthermore, this non-typicality is structurally stable against sufficiently small perturbations of A and/or H so that the eigenvectors remain 'almost aligned' (each eigenvector of A mainly overlaps with one or a few eigenvectors of H), and hence A remains 'almost conserved' (almost commuting with H).
Analogous non-typical U's are expected when r(0) is known to be (almost) conserved (commuting with H).
Further well-known exceptions are integrable systems, for which thermalization in the above sense may be absent for certain r(0) and A 4,32 (but not for others [bib_ref] Alternatives to eigenstate thermalization, Rigol [/bib_ref] , systems exhibiting many-body localization [bib_ref] Many-body localization phase transitions, Pal [/bib_ref] [bib_ref] Absence of thermalization in nonintegrable systems, Gogolin [/bib_ref]
## Or trivial cases with non-interacting subsystems (supplementary note 2).
Our present focus is different: taking thermalization for granted is the temporal relaxation well approximated by equation (13)?
Typical fast relaxation and prethermalization. Equation (20) is governed by the Boltzmann time t B : ¼ h/k B T, amounting to t B E10 À 13 s at room temperature. Equation [bib_ref] Quantum statistical mechanics in a closed system, Deutsch [/bib_ref] gives rise to comparably short timescales, unless the temperature is exceedingly low or the energy window DE is unusually small. Such relaxation times are much shorter than commonly observed in real systems [bib_ref] Thermalization under randomized local Hamiltonians, Cramer [/bib_ref] [bib_ref] Quantum systems equilibrate rapidly for most observables, Malabarba [/bib_ref] [bib_ref] Extremely quick thermalization in a macroscopic quantum system for a typical nonequilibrium..., Goldstein [/bib_ref]. Moreover, the temporal decay is typically non-exponential (see, for example, equations (18)-(20)), again in contrast to the usual findings.
This seems to imply that typical experiments correspond to non-typical unitaries U. Plausible explanations are as follows: to begin with, the above-predicted typical relaxation times are so short that they simply could not be observed in most experiments. Second (or as a consequence), the usual initial conditions and/or observables are indeed quite 'special' with respect to the prominent role of almost conserved quantities (see previous section), in particular, 'local descendants' of globally conserved quantities such as energy, charge, particle numbers and so on: examples are the amount of energy, charge and so on within some subdomain of the total system or, more generally, local densities, whose content within a given volume can only change via transport currents through the boundaries of that volume. As a consequence, the global relaxation process becomes 'unusually slow' if the densities between macroscopically separated places need to equilibrate (small surface-to-volume ratio) or if there exists a natural 'bottleneck' for their exchange (weakly interacting subsystems).
Put differently, our present theory is meant to describe the very rapid relaxation towards local equilibrium, but not any subsequent global equilibration. Only if there exists a clear-cut timescale separation between these two relaxation steps (or if there is no second step at all), can we hope to quantitatively capture the first step by our results. Conversely, the timescale separation usually admits some Markovian approximation for the second step, yielding an exponential decay, whose timescale still depends on many details of the system. Natural further generalizations include the closely related concepts of hindered equilibrium, quasi-equilibrium (metastability) and, above all, prethermalization 29,54,55 , referring, for example, to a fast partial thermalization within a certain subset of modes, (quasi-)particles or other generalized degrees of freedom. (Like in ref., we do not adopt here the additional requirement 55 that the almost conserved quantities originate from a weak perturbation of an integrable system.)
In short, our working hypothesis is that the theory (13) describes the temporal relaxation of hAi r(t) for any given pair (r(0), A) unless one of them is exceptionally close to or in some other way slowed down by an (almost) conserved quantity.
Comparison with experimental results. We focus on experiments in closed many-body systems in accordance with the above general requirements. In comparing them with our theory (13), we furthermore assume that the (pre-)thermalized system occupies a microcanonical energy window with some (effective) ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms10821 temperature T and DE ) k B T, so that equation [bib_ref] Chaos and quantum thermalization, Srednicki [/bib_ref] applies. Finally, the asymptotic values hAi r(0) and A h i r mc in equation (13) are either obvious or will be estimated from the measurements, hence no further knowledge about the often quite involved details of the experimental observables will be needed! [fig_ref] Figure 1 |: Prethermalization of ultracold atoms [/fig_ref] demonstrates the very good agreement of the theory with the rapid initial prethermalization of a coherently split Bose gas, observed by the Schmiedmayer group in ref. [bib_ref] Relaxation and prethermalization in an isolated quantum system, Gring [/bib_ref].
In [fig_ref] Figure 2 |: Ultrafast relaxation of hot electrons [/fig_ref] , the theory is compared with the pump-probe experiment by the Bigot group from ref. [bib_ref] Magneto-optics in the ultrafast regime: Thermalization of spin populations in ferromagnetic films, Guidoni [/bib_ref]. The finite widths of the pump and the probe laser pulses are roughly accounted for by convoluting equation (13) with a Gaussian of 35 fs FWHM (full width at half maximum). In ref. [bib_ref] Magneto-optics in the ultrafast regime: Thermalization of spin populations in ferromagnetic films, Guidoni [/bib_ref] , the FWHM of the pump pulse is estimated as 20 fs and the combined FWHM for both pulses as 22 fs, implying a FWHM of 9 fs for the probe pulse. The latter value seem quite optimistic to us. A second 'excuse' for our slightly larger FWHM value of 35 fs is that the tails of the experimental pulse shape may be considerably broader than those of a Gaussian with the same FWHM (see, for example, [fig_ref] Figure 2 |: Ultrafast relaxation of hot electrons [/fig_ref] in the [fig_ref] Figure 5 |: Prethermalization in a one-dimensional electron gas [/fig_ref]. Finally, the convolution of equation (13) with a Gaussian represents a rather poor 'effective description' in the first place: our entire theoretical approach becomes strictly speaking invalid when the duration of the perturbation becomes comparable to the thermalization time.
A similar comparison with the pump-probe experiments from ref. 58 is presented in [fig_ref] Figure 3 |: Temperature-dependent relaxation of hot electrons [/fig_ref]. As before, we adopted a slightly larger FWHM of 100 fs than the estimate of 76 fs in ref. [bib_ref] Direct observation of electron thermalization and electronphonon coupling in photoexcited bismuth, Faure [/bib_ref]. Due to the above-mentioned fundamental limitations of our theory for such rather large FWHM values, the temperatures adopted in [fig_ref] Figure 3 |: Temperature-dependent relaxation of hot electrons [/fig_ref] should still be considered as quite crude estimates. Apart from that, [fig_ref] Figure 3 |: Temperature-dependent relaxation of hot electrons [/fig_ref] nicely confirms the predicted temperature dependence from equation [bib_ref] Chaos and quantum thermalization, Srednicki [/bib_ref].
We close with three remarks: first, refs 56,58 also implicitly confirm our prediction that the essential temporal relaxation (encapsulated by F(t) in equation (13)) is generically the same for different observables. Second, similar pump-probe experiments abound in the literature, but usually the pulse widths are too large for our purposes. Third, the temporal relaxation in Figs 1-3 has also been investigated numerically, but closed analytical results have not been available before [bib_ref] Relaxation and prethermalization in an isolated quantum system, Gring [/bib_ref] [bib_ref] Direct observation of electron thermalization and electronphonon coupling in photoexcited bismuth, Faure [/bib_ref].
Comparison with numerical results. [fig_ref] Figure 4 |: Relaxation of an integrable and a non-integrable fermionic model [/fig_ref] illustrates the very good agreement of our theory with Rigol's numerical findings from ref. 32, both for an integrable and an non-integrable example. A similar agreement is found for all other parameters and also for an analogous hardcore boson model examined in refs 31,32. On the other hand, a second observable considered in ref. [bib_ref] Quantum quenches and thermalization in one-dimensional fermionic systems, Rigol [/bib_ref] , deriving from the momentum distribution function, exhibits in all cases a significantly slower and also qualitatively different temporal relaxation. According to the discussion in the section 'Typical fast relaxation and prethermalization', it is quite plausible that the latter observable is indeed 'non-typical' in view of the fact that it represents a conserved quantity for fermions
[formula] with V ¼ t 0 ¼ V 0 ¼ 0 (ref. 32). [/formula]
In [fig_ref] Figure 5 |: Prethermalization in a one-dimensional electron gas [/fig_ref] , we compare our theory with the simulations of a different one-dimensional electron model from ref. [bib_ref] Photon-assisted tunneling versus tunneling of excited electrons in metal-insulator-metal junctions, Thon [/bib_ref]. In doing so, the pertinent temperature T has been estimated as follows: the textbook Sommerfeld expansion for N electrons in a one-
[formula] dimensional box yields E ¼ E 0 [1 þ (3p 2 /8)(k B T/E F ) 2 ] [/formula]
, where E is their total energy, E 0 ¼ (1/3)NE F the ground-state energy, E F ¼ (p' N/gL) 2 /2m the Fermi energy, L the box length, m the electron mass and g: ¼ 2s þ 1 ¼ 2 (s ¼ 1/2 for electrons). Assuming that the pulse acts solely on the small well implies N ¼ 16, LC15 nm (ref. [bib_ref] Photon-assisted tunneling versus tunneling of excited electrons in metal-insulator-metal junctions, Thon [/bib_ref] , and E À E 0 C0.045 eV (see in ref. [bib_ref] Photon-assisted tunneling versus tunneling of excited electrons in metal-insulator-metal junctions, Thon [/bib_ref]. Altogether, we thus obtain TC170 K.
The remnant 'fluctuations' of the numerical data in Figs 4 and 5 can be readily explained as finite particle number effects (see [fig_ref] Figure 4 |: Relaxation of an integrable and a non-integrable fermionic model [/fig_ref] in ref. [bib_ref] Quantum quenches and thermalization in one-dimensional fermionic systems, Rigol [/bib_ref] and [fig_ref] Figure 1 |: Prethermalization of ultracold atoms [/fig_ref] , and their temporal correlations are as predicted below equation. The seemingly rather strong fluctuations in [fig_ref] Figure 5 |: Prethermalization in a one-dimensional electron gas [/fig_ref] are a fallacy since the systematic changes themselves are very small.
Next, we turn to the numerical findings for a qubit in contact with a spin bath by the Trauzettel group from ref. [bib_ref] Equilibration in closed quantum systems: Application to spin qubits, Hetterich [/bib_ref]. The agreement with our theory in [fig_ref] Figure 6 |: Thermalization of a spin qubit coupled with a bath [/fig_ref] is as good as it possibly can be for such a rather small dimensionality of D ¼ 2 7 . Indeed, the remaining differences nicely confirm the predictions below equation (13), regarding both their typical order of magnitude (13) and (20) with T ¼ 5 nK. The pertinent effective temperature has also been roughly estimated in ref. 29 (see [fig_ref] Figure 2 |: Ultrafast relaxation of hot electrons [/fig_ref] and is still compatible with our present fit T ¼ 5 nK. As discussed at the end of the section 'Typical fast relaxation and prethermalization', the depicted prethermalization is followed by a much slower, global thermalization 29 , which is omitted in the present figure.
[formula] D A ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Trfr 2 ð0Þg=D p ¼ 1 ffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffi 2 À 6 =2 7 p [/formula]
' 0:01 and their temporal correlations (where we exploited that Tr{r 2 (0)} ¼ 2 À 6 for the particular initial condition r(0) adopted in [fig_ref] Figure 6 |: Thermalization of a spin qubit coupled with a bath [/fig_ref].
Our final example is Bartsch and Gemmer's random matrix model from ref. 17. Referring to the notation and definitions in the caption of [fig_ref] Figure 7 |: Thermalization in a random matrix model [/fig_ref] , one readily sees that the considered observable A is a conserved quantity for the unperturbed Hamiltonian (l ¼ 0). In agreement with our discussion in the section 'Typical fast relaxation and prethermalization', A is therefore still 'almost conserved' for small l and indeed exhibits a slow, exponential decay towards A h i r mc ¼ 0 (see [fig_ref] Figure 1 |: Prethermalization of ultracold atoms [/fig_ref]. Upon increasing l, one recovers the much faster, non-exponential decay of our present theory (see [fig_ref] Figure 1 |: Prethermalization of ultracold atoms [/fig_ref]. Unfortunately, the l-value 1.77 Â 10 À 3 from [fig_ref] Figure 1 |: Prethermalization of ultracold atoms [/fig_ref] 17 is still somewhat too small and the eigenvalues E 1 , ..., E 6,000 are not any more available (I asked the authors). Therefore, we repeated the numerics from ref. 17 on our own for l ¼ 7 Â 10 À 3 . The resulting agreement with equationin [fig_ref] Figure 7 |: Thermalization in a random matrix model [/fig_ref] is very good, and the temporal correlations of the deviations as well as their typical order of magnitude D A ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Trfr 2 ð0Þg=D p ¼ 2 ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffi 1=6; 000 p ' 0:03 are as predicted below equation.
We close with two remarks: first, there is no fit parameter in any of the above examples apart from hAi r(0) in [fig_ref] Figure 4 |: Relaxation of an integrable and a non-integrable fermionic model [/fig_ref] (13) and (20) with temperatures as indicated and convoluted with a Gaussian of 100 fs FWHM (see also main text). The conversion of a given fluence into a temperature change of the electron gas is not obvious. In particular, the estimates provided in ref. [bib_ref] Direct observation of electron thermalization and electronphonon coupling in photoexcited bismuth, Faure [/bib_ref] seem not very reliable to us: first of all, [fig_ref] Figure 6 |: Thermalization of a spin qubit coupled with a bath [/fig_ref] in ref. [bib_ref] Direct observation of electron thermalization and electronphonon coupling in photoexcited bismuth, Faure [/bib_ref] indicates a temperature of ca. 250 K at four different time points B200 fs before the pump pulse, while the actual temperature of the unperturbed system is known to be 130 K. Second, the temperature error bars in [fig_ref] Figure 6 |: Thermalization of a spin qubit coupled with a bath [/fig_ref] [fig_ref] Figure 3 |: Temperature-dependent relaxation of hot electrons [/fig_ref] should coincide, while their actual agreement is only moderately better than for the 'bare' curves in [fig_ref] Figure 3 |: Temperature-dependent relaxation of hot electrons [/fig_ref]. For all these reasons, we used the temperature as a fit parameter in the present figure.
[formula] h ¼ k B ¼ t ¼ V ¼ 1 and focusing on parameters t 0 ¼ V 0 , the model is integrable if t 0 ¼ V 0 ¼ 0 and non-integrable otherwise. [/formula]
A quantum quench generates an initial pure state out of equilibrium, whose energy corresponds to that of a canonical ensemble with temperature T ¼ 2. As detailed in ref. [bib_ref] Quantum quenches and thermalization in one-dimensional fermionic systems, Rigol [/bib_ref] , the considered observable dN k (t) is a dimensionless descendant of the density-density structure factor. The depicted data are from [fig_ref] Figure 1 |: Prethermalization of ultracold atoms [/fig_ref] ,j of ref. [bib_ref] Quantum quenches and thermalization in one-dimensional fermionic systems, Rigol [/bib_ref]. Lines: theoretical predictionsand [bib_ref] Chaos and quantum thermalization, Srednicki [/bib_ref] with T ¼ 2.and [bib_ref] Chaos and quantum thermalization, Srednicki [/bib_ref] , exploiting the estimate T ¼ 170 K from the main text, and neglecting the finite temporal width (20 fs) of the pulse. As in Figs 1-3, we are actually dealing with a prethermalization process within the smaller well. The subsequent global thermalization is much slower due to the high barrier between the wells. Considering that A h i r mc is the only remaining fit parameter in the theory from equations (13) and [bib_ref] Chaos and quantum thermalization, Srednicki [/bib_ref] , the agreement with the simulations is remarkably good. In particular, the two very different observables CT(t) and DP(t) are indeed governed by the same F(t), as predicted by equations (13) and [bib_ref] Chaos and quantum thermalization, Srednicki [/bib_ref]. model in [fig_ref] Figure 4 |: Relaxation of an integrable and a non-integrable fermionic model [/fig_ref] , one may question whether the considered system exhibits thermalization in the first place, as is tacitly assumed in equation. In Supplementary Note 2, we argue that equationindeed is expected to still remain valid in such cases if A h i r mc is replaced by the pertinent non-thermal long-time asymptotics (which, in turn, is estimated from the numerical data in [fig_ref] Figure 4 |: Relaxation of an integrable and a non-integrable fermionic model [/fig_ref].
# Discussion
Our main resultimplies thermalization in the sense that a generic non-equilibrium system with a macroscopically welldefined energy becomes practically indistinguishable from the corresponding microcanonical ensemble for the overwhelming majority of all sufficiently late times. Apart from the concrete initial and long-time expectation values (that is, hAi r(0) and A h i r mc in equation (13)), the temporal relaxation (that is, F(t) in equation (13)) depends only on the spectrum of the Hamiltonian within the pertinent interval of non-negligibly populated energy eigenstates, but not on any further details of the initial condition or the observable. This represents one of the rare instances of a general quantitative statement about systems far from equilibrium.
The theory agrees very well with a wide variety of experimental and numerical results from the literature (though none of them was originally conceived for the purpose of such a comparison). We are in fact not aware of any other quantitative analytical explanation of those data comparable to ours. Indeed, the usual paradigm to identify and then analytically quantify the main physical mechanisms seems almost hopeless here. In a sense, our present approach thus amounts to a different paradigm: there is no need of any further 'explanations', since the observed behaviour is expected with overwhelming likelihood from the very beginning, that is, unless there are special a priori reasons to the contrary.
Similarly as in refs 46,49-51, generic thermalization is found to happen extremely quickly (unless the system's energy or temperature is exceedingly low). Moreover, the temporal decay is typically non-exponential. A main prediction of our theory is that these features should in fact be very common (at least in the form of prethermalization), but often they are unmeasurably fast or they have simply not been looked for so far. Conversely, most of the usually considered observables and initial conditions are actually quite 'special', namely, exceptionally slow, 'almost conserved' quantities. A better understanding of those principally untypical but practically very common thermalization processes remains an open problem [bib_ref] Quantum systems equilibrate rapidly for most observables, Malabarba [/bib_ref] [bib_ref] Extremely quick thermalization in a macroscopic quantum system for a typical nonequilibrium..., Goldstein [/bib_ref].
# Methods
Basic matrices. According to the section 'Analytical results', the unitary U represents the basis transformation between the eigenvectors |ni (n ¼ 1, ..., D) of the Hamiltonian H and those of the observable A. Denoting the eigenvalues of A by l n and the eigenvectors by c v j i (n ¼ 1, ..., D), the matrix elements of U are thus U nn : ¼ n j c v h i . Accordingly, the matrix elements of r(0) in the basis of H are related to those in the basis of A via
[formula] r mn ð0Þ ¼ X D m;n¼1 U mm r mn U Ã nn ;ð21Þ [/formula]
where r mn :¼ c u rð0Þ j jc v h i . Similarly, the matrix elements of A satisfy
[formula] A mn ¼ X D x¼1 U mx l x U Ã nx ;ð22Þ [/formula]
and hence
[formula] r mn ð0ÞA nm ¼ X D m;n;x¼1 r mn l x U mm U Ã nn U nx U Ã mx :ð23Þ [/formula]
As announced below equation (3), we work (without loss of generality) in a reference frame (or reference basis of H) so that only H (and thus |ni) depends on U, while A and r(0) (and thus c v j i) are independent of U. Hence, r mn and l x on the right-hand side of equations (21)-(23) are independent of U.
Derivation of equation [bib_ref] Long-time behavior of macroscopic quantum systems: commentary accompanying the english translation of..., Goldstein [/bib_ref]. As a simple first exercise, let us average equation [bib_ref] Finite-size scaling analysis of the eigenstate thermalization hypothesis in a one-dimensional interacting..., Ikeda [/bib_ref] over all uniformly (Haar) distributed unitaries U, as specified in the section 'Analytical results'. Since the factors r mn l x on the right-hand side are independent of U, we are left with averages over the U matrix elements. Such averages have been evaluated repeatedly and often independently of each other in the literature, see, for example, refs 5,61-63, a key ingredient being symmetry arguments due to the invariance of the Haar measure under arbitrary unitary transformations. Particularly convenient for our present purposes is the formalism adopted by Brouwer and Beenakker, see ref. [bib_ref] Diagrammatic method of integration over the unitary group, with applications to quantum..., Brouwer [/bib_ref] , and further references therein. The general structure of such averages is provided by equation (2.2) in ref. [bib_ref] Diagrammatic method of integration over the unitary group, with applications to quantum..., Brouwer [/bib_ref] , reading
[formula] U a1 b1 . . . U am bm U Ã a1 b 1 . . . U Ã anb n h i U ¼ d mn X P;P 0 V P;P 0 Y n j¼1 d aja PðjÞ d bjb P 0 ðjÞ :ð24Þ [/formula]
Quoting verbatim from ref. [bib_ref] Diagrammatic method of integration over the unitary group, with applications to quantum..., Brouwer [/bib_ref] , 'the summation is over all permutations P and P 0 of the numbers 1, ..., n. The coefficients V P,P 0 depend only on the cycle structure of the permutation P À 1 P 0 . Recall that each permutation of 1, ..., n has a unique factorization in disjoint cyclic permutations ('cycles') of lengths c 1 , ...., c k (where n ¼ P k j¼1 c j ). The statement that V P,P 0 depends only on the cycle structure of P À 1 P 0 means that V P,P 0 depends only on the lengths c 1 , ..., c k of the cycles in the factorization of P À 1 P 0 . One may therefore write V c1 ; :::; ck instead of V P,P 0 .' The (13), (14) and [bib_ref] English translation by Tumulka, R. Proof of the ergodic theorem and the..., Von Neumann [/bib_ref]. Due to the above-mentioned initial condition and the quite small dimension D ¼ 2 7 , the approximations (18)- [bib_ref] Chaos and quantum thermalization, Srednicki [/bib_ref] are not very well satisfied by the actual energy eigenvalues E 1 , ... , E 128 (kindly provided by the authors of ref. [bib_ref] Equilibration in closed quantum systems: Application to spin qubits, Hetterich [/bib_ref]. Hence, we have evaluated F(t) in equation (13) directly via equations (14) and (8). (13), [bib_ref] Concentration of measure for quantum states with a fixed expectation value, Müller [/bib_ref] and [bib_ref] English translation by Tumulka, R. Proof of the ergodic theorem and the..., Von Neumann [/bib_ref]. Similarly as in [fig_ref] Figure 6 |: Thermalization of a spin qubit coupled with a bath [/fig_ref] , the numerically obtained energies E 1 , ..., E 6,000 were found to satisfy equations (18)-(20) not very well, hence we have directly evaluated equations [bib_ref] English translation by Tumulka, R. Proof of the ergodic theorem and the..., Von Neumann [/bib_ref] and [bib_ref] Concentration of measure for quantum states with a fixed expectation value, Müller [/bib_ref].
explicit numerical values of all V c1; :::; ck with nr5 are provided by the columns 'CUE' of Tables II and IV in ref. [bib_ref] Diagrammatic method of integration over the unitary group, with applications to quantum..., Brouwer [/bib_ref]. Further remarks: the labels m and n in equation [bib_ref] Finite-size scaling of eigenstate thermalization, Beugeling [/bib_ref] have nothing to do with those in equation [bib_ref] Finite-size scaling analysis of the eigenstate thermalization hypothesis in a one-dimensional interacting..., Ikeda [/bib_ref]. Equation (24) equals zero unless m ¼ n. Every label a j must have a 'partner', that is, its value must coincide with one of the a j 's and vice versa, since otherwise the product over the Kronecker delta's d aja PðjÞ in equation [bib_ref] Finite-size scaling of eigenstate thermalization, Beugeling [/bib_ref] would be zero for all P's. Note that some a j 's may assume the same value, but then an equal number of a j 's also must assume that value. Likewise, every b j needs a 'partner' among the b j 's and vice versa.
Adopting the abbreviation
[formula] X mn :¼ r mn ð0ÞA nm ½ Uð25Þ [/formula]
and the renamings a 1 : ¼ m, a 2 : ¼ n, b 1 : ¼ m, b 2 : ¼ x and b 3 : ¼ n, equation (23) yields
[formula] X a1a2 ¼ X b1 ;b2 ;b3 r b1 b3 l b2 U a1b1 U a2b2 U Ã a1b2 U Ã a2b3 h i U :ð26Þ [/formula]
The connection with equation (24) is established via the identifications a 1 : ¼ a 1 , a 2 : ¼ a 2 , b 1 : ¼ b 2 and b 2 : ¼ b 3 . Therefore, if b 1 ab 2 , then the only potential 'partner' of b 1 is b 2 , and only if their values coincide, that is, b 3 ¼ b 1 , the corresponding summands may be non-zero. The same conclusion can be drawn if b 1 ¼ b 2 . We thus can rewrite equation [bib_ref] Thermal equilibrium of a macroscopic quantum system in a pure state, Goldstein [/bib_ref] with equation (24) as
[formula] X a1a2 ¼ X b1;b2 r b1b1 l b2 X P;P 0 V P;P 0 Y 2 j¼1 d aja PðjÞ d bjb P 0 ðjÞ ;ð27Þwhere b 1 ¼ b 2 and b 2 ¼ b 1 . [/formula]
There are two permutations of the numbers 1 and 2, namely, the identity and one, which exchanges 1 and 2. Denoting them as P 1 and P 2 , respectively, and observing that b j ¼ b P2ðjÞ , equation [bib_ref] Exploring local quantum many-body relaxation by atoms in optical superlattices, Cramer [/bib_ref] can be rewritten as
[formula] X a1 a2 ¼ X 2 k¼1 Y 2 j¼1 d ajaP k ðjÞ X 2 l¼1 V Pk;Pl S lð28ÞS l :¼ X b1;b2 r b1b1 l b2 Y 2 j¼1 d bjb P 2 ðP l ðjÞÞð29Þ [/formula]
For l ¼ 1, the two Kronecker delta's in equation [bib_ref] Relaxation and prethermalization in an isolated quantum system, Gring [/bib_ref] both require that b 1 ¼ b 2 and hence
[formula] S 1 ¼ X b1 r b1 b1 l b1 ¼ Trfrð0ÞAg:ð30Þ [/formula]
The last equality can be verified by evaluating the trace in the eigenbasis of A, see above equation [bib_ref] Thermalization of interacting fermions and delocalization in Fock space, Neuenhahn [/bib_ref]. In the same way, one finds that
[formula] S 2 ¼ X b1 ;b2 r b1 b1 l b2 ¼ Trfrð0ÞgTrfAg ¼ DTrfr mc Ag:ð31Þ [/formula]
In the last equation, we exploited that Tr{r(0)} ¼ 1 and r mc : ¼ I/D, see below equation [bib_ref] Long-time behavior of macroscopic quantum systems: commentary accompanying the english translation of..., Goldstein [/bib_ref]. Observing that the two Kronecker delta's in equation (28) equal one if k ¼ 1 or if k ¼ 2 and a 1 ¼ a 2 , the overall result is
[formula] X a1a2 ¼ A h i rð0Þ V P1;P1 þ d a1 a2 V P2;P1 À Á þ D A h i r mc V P1 ;P2 þ d a1a2 V P2 ;P2 À Á ;ð32Þ [/formula]
where, as usual, hAi r(0) : ¼ Tr{r(0)A} and A h i r mc :¼ Tr r mc A f g. Finally, the coefficients V Pk;Pl are evaluated as explained below equation (24): if k ¼ l, then P À 1 l P k ¼ P 1 factorizes in two cycles of lengths c 1 ¼ c 2 ¼ 1, that is, V Pk;Pl ¼ V c1;c2 ¼ V 1;1 . Likewise, if kal, then P À 1 l P k ¼ P 2 consists of one cycle with c 1 ¼ 2, that is, V Pk;Pl ¼ V 2 . Referring to columns 'CUE' and rows 'n ¼ 2' of Tables II and IV in ref. 63 yields V 1,1 ¼ 1/(D 2 À 1) and
[formula] V 2 ¼ À 1/[D(D 2 À 1)]. [/formula]
Returning to the original labels m and n in equation (25), we thus can rewrite equation (32) as
[formula] X mn ¼ A h i rð0Þ D À d mn D D 2 À 1 ð Þ þ A h i r mc Dd mn À 1 D 2 À 1 :ð33Þ [/formula]
As a consequence, we can infer from equations (4) and (25) that A h i r av ¼ DX nn and with equation (33) that
[formula] A h i r av ¼ A h i rð0Þ 1 D þ 1 þ A h i r mc D D þ 1 :ð34Þ [/formula]
Hence, one readily recovers equation [bib_ref] Long-time behavior of macroscopic quantum systems: commentary accompanying the english translation of..., Goldstein [/bib_ref]. A relation remarkably similar to our present equation [bib_ref] Long-time behavior of macroscopic quantum systems: commentary accompanying the english translation of..., Goldstein [/bib_ref] , albeit in a quite different physical context, has been previously obtained also in ref. [bib_ref] An exactly solvable model for the integrability-chaos transition in rough quantum billiards, Olshanii [/bib_ref] (see equation [bib_ref] Entanglement and the foundations of statistical mechanics, Popescu [/bib_ref] therein).
Derivation of equation [bib_ref] Normal typicality and von Neumann's quantum ergodic theorem, Goldstein [/bib_ref]. Without any doubt, there are much faster ways to obtain equations [bib_ref] Onset of quantum chaos in one-dimensional bosonic and fermionic systems and its..., Santos [/bib_ref] or [bib_ref] Many-body localization phase transitions, Pal [/bib_ref]. The advantage of our present way is that it can be readily adopted without any conceptual differences (albeit the actual calculations become more lengthy) to more demanding cases like
[formula] x 2 ðtÞ Â Ã U ¼ A h i 2 rðtÞ h i U À A h i rðtÞ h i 2 U ;ð35Þ [/formula]
see equation (10).
To evaluate the last term in equation (35), we recast equation (6) with equations (7) and (9) into the form
[formula] A h i rðtÞ h i U ¼ F 0 ðtÞhAi rð0Þ þ F 0 ðtÞ A h i r mc þ R 1 ðtÞ ð 36Þ R 1 ðtÞ :¼ F 0 ðtÞ A h i r mc À A h i rð0Þ D 2 À 1 ð37Þ F 0 ðtÞ :¼ 1 À F 0 ðtÞ ð 38Þ F 0 ðtÞ :¼ 1 D 2 X D m;n¼1 e iðEn À EmÞt=' ¼ fðtÞ j j 2 ;ð39Þ [/formula]
where f(t) is defined in equation [bib_ref] English translation by Tumulka, R. Proof of the ergodic theorem and the..., Von Neumann [/bib_ref]. Similarly as in equation (15), one sees that F 0 ðtÞ; F 0 ðtÞ 2 ½0; 1 for all t. Denoting by l max and l min the largest and smallest among the eigenvalues l 1 , ..., l D of A, the range of A is defined as D A : ¼ l max À l min . Furthermore, we can and will add a constant to A so that l min ¼ À l max without any change in the final conclusions below. It readily follows that |l n |rD A /2 for all n and hence that
[formula] A k h i r D A =2 ð Þ kð40Þ [/formula]
for arbitrary density operators r and k 2 N. We thus can infer from equation (37) that
[formula] R 1 ðtÞ j j D A =ðD 2 À 1Þ:ð41Þ [/formula]
Likewise, one finds upon squaring equation (36) that
[formula] A h i rðtÞ h i 2 U ¼ F 0 ðtÞ A h i rð0Þ þ F 0 ðtÞ A h i r mc 2 þ R 2 ðtÞ ð 42Þ R 2 ðtÞ j j 3D 2 A =ðD 2 À 1Þ :ð43Þ [/formula]
Turning to the first term on the right-hand side of (35), one can infer, similarly as in equations (25) and (26), from (1) and (23)
[formula] r b1b5 l b2 r b3 b6 l b4 U a1b1 . . . U a4b4 U Ã a1b 1 . . . U Ã a4b 4 h i U ;ð45Þwith b 1 : ¼ b 2 , b 2 : ¼ b 5 , b 3 : ¼ b 4 and b 4 : ¼ b 6 . [/formula]
Similarly as below equation (26), it follows that only those summands may be non-zero, for which b 1 and b 3 have 'partners' among b 2 and b 4 , and vice versa. This condition can be satisfied in two ways: (i) b 5 ¼ b 1 and b 6 ¼ b 3 ; and (ii) b 5 ¼ b 3 , b 6 ¼ b 1 and b 1 ab 3 . The latter condition is due to the fact that the case b 1 ¼ b 3 is already covered by (i). Exploiting equation [bib_ref] Finite-size scaling of eigenstate thermalization, Beugeling [/bib_ref] and with the abbreviationã :¼ a 1 ; :::; a 4 ð Þand likewise forb,b and so on, we thus obtain
[formula] Xã ¼ X ðiÞ a þ XX b;b1 = ¼ b3 r b1 b3 l b2 r b3b1 l b4 X P;P 0 V P;P 0 Y 4 j¼1 d aja PðjÞ d bjb ðiiÞ P 0 ðjÞ ;ð48Þwhereb ðiÞ :¼ ðb 2 ; b 1 ; b 4 ; b 3 Þ andb ðiiÞ :¼ ðb 2 ; b 3 ; b 4 ; b 1 Þ. [/formula]
There are 4! ¼ 24 permutations P of the numbers 1, 2, 3 and 4. Adopting the shorthand notation [P(1)P(2)P(3)P(4)] to explicitly specify a given P, these 24 permutations are: P 1 ¼ ½1234; P 2 ¼ ½2134; P 3 ¼ ½3214; P 4 ¼ ½4231; P 5 ¼ ½1324; P 6 ¼ ½1432; P 7 ¼ ½1243; P 8 ¼ ½2143; P 9 ¼ ½3412; P 10 ¼ ½4321; P 11 ¼ ½1342; P 12 ¼ ½1423; P 13 ¼ ½3241; P 14 ¼ ½4213; P 15 ¼ ½2431; P 16 ¼ ½4132; P 17 ¼ ½2314; P 18 ¼ ½3124; P 19 ¼ ½2341; P 20 ¼ ½2413; P 21 ¼ ½3421; P 22 ¼ ½3142; P 23 ¼ ½4312; P 24 ¼ ½4123:
Observing that b ðiÞ j ¼ b P8 ðjÞ and b ðiiÞ j ¼ b P19ðjÞ for all j ¼ 1, ..., 4, it is quite straightforward but very arduous to explicitly carry out the sums over P 0 andb in equations [bib_ref] Time scales in the approach to equilibrium of macroscopic quantum systems, Goldstein [/bib_ref] and [bib_ref] Generic evaluation of relaxation time for quantum many body systems: analysis of..., Monnai [/bib_ref] , and the sum overã in equation (44), yielding
[formula] A h i 2 rðtÞ h i U ¼ X 24 k¼1 f k ðtÞTðP k Þ;ð49Þ [/formula]
where the functions f k (t) are given by ð50Þ and the coefficients T(P) are given by
TðPÞ ¼ D 2 A h i 2 r mc ðV P;P8 þ V P;P24 Trfr 2 ð0ÞgÞ þ D A 2 r mc ðV P;P10 Trfr 2 ð0Þg þ V P;P19 Þ þ D A h i r mc A h i rð0Þ ðV P;P2 þ V P;P7 þ V P;P20 þ V P;P22 Þ þ D A h i r mc Trfr 2 ð0ÞAgðV P;P12 þ V P;P14 þ V P;P16 þ V P;P18 Þ þ A h i 2 rð0Þ ðV P;P1 þ V P;P9 Þ þ Trf½rð0ÞA 2 gðV P;P3 þ V P;P6 Þ þ A 2 rð0Þ ðV P;P11 þ V P;P13 þ V P;P15 þ V P;P17 Þ þ Trfr 2 ð0ÞA 2 gðV P;P4 þ V P;P5 þ V P;P21 þ V P;P23 Þ :
To explicitly evaluate equations (49)-(51), we still need the coefficients V Pk;Pl for all k,lA{1, ..., 24}. They are obtained as explained below equation (24): defining j ¼ j(k,l) implicitly via P j ¼ P À 1 l P k , one finds by factorizing each P j into its disjoint cycles and exploiting Tables II and IV of ref. 63 that V Pk;Pl is given by V 1;1;1;1 ¼ D À 4 for j ¼ 1; V 2;1;1 ¼ À D À 5 for j ¼ 2; :::; 7; V 2;2 ¼ D À 6 for j ¼ 8; :::; 10; V 3;1 ¼ 2 D À 6 for j ¼ 11; :::; 18; V 4 ¼ À 5 D À 7 for j ¼ 19; :::; 24; ð52Þ up to correction factors of the form 1 þ OðD À 2 Þ on the right-hand side of each of those relations. One thus is left with finding P j ¼ P À 1 l P k for all 24 2 pairs (k,l). To mitigate this daunting task, we have restricted ourselves to those summands in equation [bib_ref] Quantum systems equilibrate rapidly for most observables, Malabarba [/bib_ref] , which are at least of the order D À 1 . Along these lines, one finally recovers with equations (35), [bib_ref] Quantum mechanical evolution towards equilibrium, Linden [/bib_ref] and (42) the result [bib_ref] Normal typicality and von Neumann's quantum ergodic theorem, Goldstein [/bib_ref].
Derivation of equation. While the essential steps in deriving equationhave been outlined already in the main text, we still have to provide the details of the statements below equation: our first observation is that R 1 (t) in equation [bib_ref] Absence of thermalization in nonintegrable systems, Gogolin [/bib_ref] amounts to the systematic (U independent) part of the omitted corrections in equation, and equation [bib_ref] Canonical thermalization, Reimann [/bib_ref] to the bound announced below equation.
By means of a straightforward (but again very tedious) generalization of the calculations from the preceding subsection, one finds that
[formula] xðtÞxðsÞ ½ U ¼ Cðt; sÞ D 2 A Tr r 2 ð0Þ f g D þ O D 2 A D 2 ;ð53Þ [/formula]
where C(t, s) has the following six properties: first, C(t, s) ¼ C(s, t) ¼ C( À t, À s) for all t, s. Second, |C(t, s)|r9 for all t,s. Third, C(t, 0) ¼ 0 for all t. Fourth, C(t, s)*0 for |t À s|-N, cf. equation [bib_ref] Typicality for generalized microcanonical ensembles, Reimann [/bib_ref]. Fifth, Cðt; sÞFðt À sÞhðA À A h i r mc Þ 2 i r mc for t, s-N. Sixth, given s, the behaviour of C(t, s) as a function of t is roughly comparable to that of F(t À s) for most t.
Though we did not explicitly evaluate the last term in equation (53), closer inspection of its general structure shows that it can be bounded in modulus by cD 2 A =D 2 for some c, which is independent of t, s, D, A, r(0) and H. Moreover, there is no indication of any fundamental structural differences in comparison with the leading and next-to-leading order terms, which we did evaluate. In other words, the last term in equation (53) is expected to satisfy properties analogous to those mentioned below equation [bib_ref] General relaxation time of the fidelity for isolated quantum thermodynamic systems, Monnai [/bib_ref]. Recalling that the purity Tr{r 2 (0)} satisfies the usual bounds 1 ! Tr r 2 ð0Þ f g!Tr r 2 mc È É ¼ 1=D, we thus recover the properties of x(t) announced below equation.
[fig] Figure 1 |: Prethermalization of ultracold atoms. The considered observable 'mean-squared contrast' quantifies the spatial correlation of the matter-wave interference pattern after coherently splitting a Bose gas into two quasi-condensates (see ref.29 for more details). Symbols: experimental data fromFig. 2a of ref.29. Line: theoretical prediction [/fig]
[fig] Figure 2 |: Ultrafast relaxation of hot electrons. A first laser pulse (at t ¼ 0) 'heats up' the electron gas in a thin ferromagnetic film, whose rethermalization is then probed by means of a second laser pulse. As detailed in ref.56, the considered observable 'differential transmission' quantifies the magneto-optical polarization rotation of the probe laser light. Symbols: experimental data fromFig. 2aof ref. 56. Dotted: theoretical prediction(13)and(20) with T ¼ 310 K and F(to0): ¼ 1. Solid: convolution of the dotted line with a Gaussian of 35 fs FWHM, accounting for the finite widths of the pump and probe laser pulses (see also main text). Similarly as inFig. 1, on larger timescales than covered by the present figure, the prethermalized electrons also exhibit non-negligible interactions with the lattice phonons and magnons, resulting in a much slower global relaxation of the compound electron-lattice system56 . Concerning the pertinent temperature T, a direct experimental estimate is not available for the set-up from ref. 56 (I contacted one of the authors), but it has been provided for a similar experiment by the same group in ref.65, except that the fluence (energy per spot area of the pump laser pulse) was 70 times larger than in ref.56. Taking all this into account, the estimate T ¼ 310 K adopted in the present figure seems very reasonable. [/fig]
[fig] Figure 3 |: Temperature-dependent relaxation of hot electrons. Symbols: similar pump-probe experiments as in Fig. 2, but now conducted on bismuth and for six different fluences (energy per spot area of the pump laser pulses). As detailed in ref. 58, the considered observable N exc quantifies (in a.u.) the number of excited electrons above the Fermi level. The depicted data are from Fig. 5b of ref. 58 for fluences (top-down) 0.12, 0.2, 0.36, 0.52, 0.68, and 0.84 mJ cm À 2 . Lines: theoretical prediction [/fig]
[fig] Figure 4 |: Relaxation of an integrable and a non-integrable fermionic model. The upper part of the plot refers to an integrable model, the lower to a nonintegrable one. Symbols: numerical results from ref. 32 for eight strongly correlated fermions on a one-dimensional lattice with 24 sites, described in terms of an extended Hubbard model with nearest-and next-nearest-neighbour hopping and interaction parameters t, t 0 , V and V 0 , respectively. Working in units with [/fig]
[fig] Figure 5 |: Prethermalization in a one-dimensional electron gas. Symbols: numerical results from ref. 59 for a one-dimensional model of the many-electron dynamics in an asymmetric double-well potential (emulating a metal-insulator-metal junction). Starting with 44 electrons in the ground state, a laser pulse-like electrical perturbation acts predominantly on the 16 electrons in the smaller, box-shaped well, and then their rethermalization is monitored via the charge transfer into the larger well (denoted in (a) as CT(t)), and via the change of the ground state population (denoted in (b) as DP(t)). Depicted are the numerical results fromFig. 8of ref. 59. For further details regarding the simulations, we refer to refs 59,67. Lines: theoretical predictions [/fig]
[fig] Figure 6 |: Thermalization of a spin qubit coupled with a bath. Solid: numerical results for the model with seven spin-1/2 degrees of freedom in an external magnetic field from ref. 60: a central spin (qubit) is randomly (and reasonably weakly) coupled with a bath of six spins. The initial state r(0) is the product of a totally mixed bath state and an eigenstate of the central spin component S x . Depicted are the data from Fig. 2 of ref. 60 for the central spin component S x . Dashed: theoretical prediction [/fig]
[fig] Figure 7 |: Thermalization in a random matrix model. Solid: numerical results for the random matrix model of the form H ¼ H 0 þ lV from ref. 17. Adopting dimensionless units with h ¼ 1, the D ¼ 6,000 eigenvalues of H 0 are chosen equidistant with level spacing 8.33 Â 10 À 5 (ref. 17). The matrix elements of A (observable) and V (perturbation) in the basis of H 0 satisfyA ik ¼ ( À 1) k d ik and V ki ¼ V Ã ik .Apart from the latter constraint, the real and imaginary parts of V ik are independent, normally distributed random numbers. The initial state is rð0Þ ¼ c j i c h j, where c j i is randomly sampled from the energy shell H under the constraint hAi r(0) C0.2 (ref. 17). Depicted are three representative numerical realizations for l ¼ 7 Â 10 À 3 akin toFig. 1bof ref. 17 (in dimesionless units). Dashed: theoretical prediction [/fig]
|
Patterns of care for older patients with stage IV non‐small cell lung cancer in the immunotherapy era
Background:Historically, older patients with advanced lung cancer have often received no systemic treatment. Immunotherapy has improved outcomes in clinical trials, but its dissemination and implementation at the population level is not well-understood. Methods: A retrospective cohort study of patients with stage IV non-small cell lung cancer (NSCLC) diagnosed age 66 or older from 2012 to 2015 was conducted using SEER-Medicare. Treatment patterns within one year of diagnosis were ascertained. Outcomes included delivery of (a) any systemic therapy; (b) any second-line infusional therapy, following first-line infusional therapy; and (c) any second-line immunotherapy, following first-line infusional therapy. Trends in care patterns associated with second-line immunotherapy approvals in 2015 were assessed using generalized additive models. Sociodemographic and clinical predictors of treatment were explored using logistic regression. Results: Among 10 303 patients, 5173 (50.2%) received first-line systemic therapy, with little change between the years 2012 (47.5%) and 2015 (50.3%). Among 3943 patients completing first-line infusional therapy, the proportion starting second-line infusional treatment remained stable from 2012 (30.5%) through 2014 (32.9%), before increasing in 2015 (42.4%) concurrent with second-line immunotherapy approvals. Factors associated with decreased utilization of any therapy included age, black race, Medicaid eligibility, residence in a high-poverty area, nonadenocarcinoma histology, and comorbidity; factors associated with increased utilization of any therapy included Asian race and Hispanic ethnicity. Among patients who received first-line infusional therapy, factors associated with decreased utilization of second-line infusional therapy included age, Medicaid eligibility, nonadenocarcinoma histology, and comorbidity; Asian race was associated with increased utilization of second-line infusional therapy. Conclusion: United States Food and Drug Administration (FDA) approvals of immunotherapy for the second-line treatment of advanced NSCLC in 2015 were associated with increased rates of any second-line treatment, but disparities based on social determinants of health persisted. K E Y W O R D S geriatrics, health services research, immunotherapy, lung cancer, patterns of care, SEER-Medicare 2020 | KEHL Et aL.
# | introduction
Lung cancer is the leading cause of cancer death in the United States; most patients have distant metastatic disease at diagnosis, and over 150 000 die annually.Although palliative-intent systemic anticancer therapy is available, estimates of the proportion of patients actually receiving such treatment on a population basis have varied from 26% to 81%.The high proportion of untreated patients likely relates to the fact that the historical backbone of systemic treatment has been cytotoxic chemotherapy. The median age at lung cancer diagnosis is 70 7 ; for elderly patients with multiple comorbidities, chemotherapy too often confers only a small benefit with substantial toxicity.Fortunately, novel therapies are advancing the treatment of advanced non-small cell lung cancer (NSCLC). Oral targeted therapy for driver mutations has improved outcomes compared with traditional chemotherapy.However, most patients do not have tumors with targetable mutations.In contrast, immunotherapy with immune checkpoint inhibitors (ICIs) has rapidly been incorporated into the standard of care for nearly all patients without targetable mutations since 2015, initially based on clinical trials demonstrating improved overall survival compared with second-line chemotherapy.Rates of high-grade adverse events are generally lower for immunotherapy than for cytotoxic chemotherapy for lung cancer.Immunotherapy may therefore be an option for patients who might otherwise forego systemic therapy. This dynamic-introduction of a widely available, entirely new class of therapy-could increase the proportion of patients in particular contexts who receive treatment at all, potentially constituting a mechanism of population-level impact that would not be evident in clinical trials. Under 5% of adults with cancer enroll on clinical trials, and elderly patients with multiple comorbidities are especially underrepresented.Real-world evidenceis therefore necessary to understand the impact of immunotherapy on treatment patterns and outcomes for patients with lung cancer on a population basis.
The objectives of this study, therefore, were to describe trends in first-and second-line systemic therapy delivery with the advent of the immunotherapy era among Medicare-insured older patients with stage IV NSCLC in the United States.
# | methods
## | study design
This was a retrospective cohort study of older patients with stage IV non-small cell lung cancer. The data source was SEER-Medicare, which contains linked cancer registry and administrative dataand is available from the National Cancer Institute (healt hcare deliv ery.cancer.gov/seerm edica re/). The analysis was declared exempt from review by our Institutional Review Board.
## | study subjects
Patients with a first primary lung cancer with distant metastasis at diagnosis were identified using cancer registry data. Diagnosis at age 66 or older and at least one year of Medicare parts A (inpatient), B (outpatient), and D (prescription drug) fee-for-service coverage before the registry diagnosis date was required, in order to estimate comorbidity antecedent to the diagnosis of cancer. Additional requirements included continuous fee-for-service Medicare A, B, and D coverage for at least one year after diagnosis or until death. Treatment pattern analyses were conducted within three groups: (a) all patients; (b) the subgroup that received any first-line infusional therapy and discontinued it within one year; and (c) the subgroup that received both first-and second-line infusional therapy within one year. In this analysis, the term "systemic therapy" refers to any medication used to treat cancer, and the term "immunotherapy" refers to immune checkpoint inhibitors. The term "infusional therapy" refers to any systemic therapy which would be billed under Medicare part B, including cytotoxic chemotherapy, immune checkpoint inhibitors, or other monoclonal antibodies such as bevacizumab. Oral cytotoxic chemotherapies that have an infusional equivalent, such as topotecan, would be included in this "infusional" category, although these oral cytotoxic agents do not play a major role in the treatment of NSCLC.The term "targeted therapy" refers to any oral tyrosine or serine-threonine kinase inhibitor, which would be billed under Medicare part D.
## | outcome variables
Treatment patterns were assessed over the 12 months following cancer diagnosis. This time window was chosen to facilitate assessments of secular trends in second-line treatment rates over time, by maintaining the diagnosis date as the index event. Alternative methods, such as defining the index event as the completion of first-line therapy or employing a time-to-event framework to enable assessment of treatments beginning more than a year after diagnosis, could have introduced confounding by clinical status at the completion of first-line therapy and made it difficult to assess secular trends; using those methods, patients still being followed later into the immunotherapy era would have been those with the best outcomes on first-line therapy.
The primary outcome variables of interest, therefore, were the rate of (a) any first-line systemic therapy within one year among all patients; (b) any second-line infusional therapy, among patients who received first-line infusional therapy and discontinued it within one year; and (c) and any second-line immunotherapy within one year among patients who received both first-and second-line infusional therapy. Treatment was defined based on claims for drugs used for the treatment of non-small cell lung cancer . Sensitivity analyses were performed to evaluate (a) the impact on the calculated treated proportion of including claims for delivery of cancer treatment without drug claims, among all patients; and (b) the impact on the calculated treatment proportions of assuming that any nonspecific drug claims (HCPCS codes J3490 or J9999, which may have been submitted by providers before specific immunotherapy drug codes were widely used) represented claims for immunotherapy.
## | line of therapy definition
In this analysis, a line of therapy was defined as any group of anti-cancer drugs, each of which was initiated within three weeks of the first drug in the group; and ending with the last date of administration of the last drug in the group prior to either the end of follow-up or initiation of a new drug not in the original group. Gaps in treatment alone were not considered to advance the line of therapy. Lines of therapy could contain one or more anticancer drugs. Carboplatin and cisplatin were considered equivalent for defining lines of therapy.
## | independent variables
Independent variables included date of diagnosis; age at cancer diagnosis, divided into categories (age 66-70, 71-75, 76-80, 81-85, or 86+); sex; race, as defined by SEER; Hispanic ethnicity; urban/rural status of county of residence 36 ; the ecological poverty rate among individuals aging 65-74 who resided in each patient's census tract, divided into quintiles; the ecological rate of college education in each patient's census tract, divided into quintiles; comorbidity as calculated using the NCI comorbidity index based on claims ranging from one year before diagnosis to three months before diagnosis; and any dual Medicaid enrollmentin the year prior to diagnosis. For the outcome of any second-line immunotherapy among patients who received both first-and secondline infusional therapy, date of diagnosis was excluded as an independent variable from the multivariable model to preserve model stability, given the strong collinearity between diagnosis date and immunotherapy treatment rate.
# | statistical analysis
Treatment pattern trends were analyzed using generalized additive models in which diagnosis date was the independent variable.Predictors of each treatment pattern outcome were explored using multivariable logistic regression. Analyses were performed using SAS, version 9.4, and R, version 3.5.1.
# | results
## | any systemic therapy
Cohort derivation is detailed in ; patient characteristics are provided in . In a sensitivity analysis, allowing nonspecific drug claims to define first-line therapy in the absence of any specific drug claims increased the proportion of treated patients only slightly, from 50.2% to 53.3%. When treatment patterns over the entire first year after diagnosis were analyzed, 944 patients (9.2% of the full cohort, or 18.2% of treated patients) received targeted therapy in the first line or thereafter. Of the remaining 4229 patients treated with infusional therapy only, just 286 (6.8%) remained on first-line therapy by one year after diagnosis; with censoring at the one-year timepoint, the median time on first-line treatment among patients receiving chemotherapy was only 2.0 months. The remaining 3943 patients constituted the cohort for subsequent second-line infusional therapy analyses. Of the 16 patients who received first-line immunotherapy, some also received targeted therapy, and the remainder were analyzed together with the patients who received first-line infusional therapy (data not shown to preserve confidentiality as per National Cancer Institute requirements). In a second sensitivity analysis, imposing an assumption that non-specific infusional systemic therapy claims (HCPCS codes J3490 or J9999) represented claims for immunotherapy had little impact on overall treatment pattern trends .
Patients who were older, Black, or had at least two comorbidities, nonadenocarcinoma histology, dual Medicaid coverage, or lived in high-poverty census tracts were less likely to receive any systemic therapy; patients living in rural areas were . were older, dually eligible for Medicaid, had nonadenocarcinoma histology, or had more comorbidities were less likely to proceed from first-line to second-line infusional therapy; patients of Asian/Pacific Islander/Other/unknown descent were more likely to do so ; . in 2015, 208 (45.9%) received second-line immunotherapy within one year. Proceeding to second-line immunotherapy was slightly more common among patients living in "metropolitan" (versus "large metropolitan" areas) and those who had squamous cell histology (Table 1; .
[formula] (N=907) (N=228) (N=6317) (N=3246) (N=2767) (N=4290) (N=642) (N=275) (N=3073) (N=1079) (N=5234) (N=2059) (N=2066) (N=2054) (N=2037) (N=2087) (N=2051) (N=2057) (N=2066) (N=2059) (N=2070) (N=3205) (N=7098) (N=565) (N=9738) (N=980) (N=797) (N=8526) (N=5021) (N=5282) (N=991)( [/formula]
## | second-line infusional therapy
## | second-line immunotherapy
[formula] (N=316) (N=98) (N=2284) (N=1082) (N=1114) (N=1747) (N=260) (N=117) (N=1189) (N=427) (N=1950) (N=708) (N=749) (N=805) (N=826) (N=855) (N=833) (N=777) (N=820) (N=775) (N=738) (N=945) (N=2998) (N=202) (N=3741) (N=328) (N=216) (N=3399) (N=2086) (N=1857) (N=140) (N=413)(N=839) [/formula]
# | discussion
In this analysis of treatment patterns among older patients with metastatic NSCLC diagnosed immediately before and after the first approvals of immunotherapy for NSCLC in the second-line setting, we found persistently low rates of any systemic treatment, with disparities by race and socioeconomic status, consistent with prior studies predating the immunotherapy era.The overall rate of systemic therapy delivery did not change substantially with the availability of second-line immunotherapy subsequent to its approval by the United States Food and Drug Administration (FDA) for the second-line treatment of squamous NSCLC in March 2015 and nonsquamous NSCLC in October 2015. However, among patients who received first-line infusional therapy, the use of immunotherapy for second-line treatment increased rapidly following FDA approval and was associated with a modest increase in the rate of any second-line therapy.
Overall, the uptake of immunotherapy among older adults on a population basis in the United States appears to have been rapid. Within one year of initial approval, most patients initiating second-line treatment were receiving immunotherapy. However, despite this rapid uptake, less than 10% of all patients diagnosed in 2015 received immunotherapy within a year of diagnosis. This was largely attributable to the high proportion of patients who received no systemic therapy at all, since few patients remained on first-line therapy for a full year . This finding has important implications for future research. Importantly, the indications for the use of immunotherapy expanded to include use in the first-line treatment of advanced NSCLC without targetable mutations in October 2016.Given the size of the previously untreated population, assessment of the effectiveness and value of first-line immunotherapy will need to measure the extent to which immunotherapy expands the population of treated patients. In other words, relevant effectiveness analysis will therefore compare not just immunotherapy versus traditional treatments, but immunotherapy versus no treatment at all. This study has several strengths. It is based on a cohort of older patients that is representative of the population of SEER registry sites. This facilitates unbiased examination of rates of any treatment, including first-line treatment, which is less feasible in cohorts drawn from oncology practices, 41,42 consisting of patients more likely to be treated. This distinction is especially important in light of the poor survival in our population-based cohort. Still, the results confirm a recent report from oncology practices that second-line treatment for NSCLC became more common in the immunotherapy era.In our analysis, it did not appear that the increase in second-line treatment within one year of diagnosis related to any substantial decrease in the proportion of patients remaining on first-line therapy for a full year, which was an uncommon outcome in this population across years of diagnosis . Limitations include the lag between clinical events and data availability in SEER-Medicare, such that it is not yet possible to assess the impact of first-line immunotherapy approvals 40,44,45 on population treatment patterns. Further research will be necessary to evaluate this impact when updated data become available. Finally, this analysis was restricted to patients with stage IV (de novo metastatic) disease; algorithms for defining recurrent disease using claims data 46 are challenging to apply to the question of rates of any treatment, since treatment claims are an important component of identifying recurrence.
In conclusion, a large proportion of older patients with metastatic NSCLC diagnosed from 2012-2015 did not receive systemic therapy, and socioeconomic disparities in these rates persisted. The advent of immunotherapy was associated with an increase in the proportion of patients receiving second-line treatment after discontinuing first-line chemotherapy. Assessment of the effectiveness and value of immunotherapy should include measurement of its impact on the proportion of patients receiving any treatment.
# Acknowledgments
This study used the linked SEER-Medicare database. The interpretation and reporting of these data are the sole responsibility of the authors. The authors acknowledge the efforts of the National Cancer Institute; the Office of Research, Development and Information, CMS; Information Management Services (IMS), Inc; and the Surveillance, Epidemiology, and End Results (SEER) Program tumor registries in the creation of the SEER-Medicare database.
## Conflicts of interest
Dr Kehl reports serving as a consultant to Aetion, Inc Dr Schrag reports receiving research funding from Pfizer and the American Association of Cancer Research, and serving as an editor and receiving personal fees from the Journal of the American Medical Association. Dr Hassett has no relevant conflicts of interest to disclose. |
A MT-TL1 variant identified by whole exome sequencing in an individual with intellectual disability, epilepsy, and spastic tetraparesis
The genetic etiology of intellectual disability remains elusive in almost half of all affected individuals. Within the Solve-RD consortium, systematic re-analysis of whole exome sequencing (WES) data from unresolved cases with (syndromic) intellectual disability (n = 1,472 probands) was performed. This re-analysis included variant calling of mitochondrial DNA (mtDNA) variants, although mtDNA is not specifically targeted in WES. We identified a functionally relevant mtDNA variant in MT-TL1 (NC_012920.1:m.3291T > C; NC_012920.1:n.62T > C), at a heteroplasmy level of 22% in whole blood, in a 23-year-old male with severe intellectual disability, epilepsy, episodic headaches with emesis, spastic tetraparesis, brain abnormalities, and feeding difficulties. Targeted validation in blood and urine supported pathogenicity, with heteroplasmy levels of 23% and 58% in index, and 4% and 17% in mother, respectively. Interestingly, not all phenotypic features observed in the index have been previously linked to this MT-TL1 variant, suggesting either broadening of the m.3291T > C-associated phenotype, or presence of a co-occurring disorder. Hence, our case highlights the importance of underappreciated mtDNA variants identifiable from WES data, especially for cases with atypical mitochondrial phenotypes and their relatives in the maternal line.
# Introduction
The introduction of whole exome sequencing (WES) in clinical settings has massively augmented diagnostic yield for intellectual disability (ID) and other neurodevelopmental disorders (NDD), and additionally identified many new disease-gene associations. Yet,~50-70% of individuals with ID/NDD remain undiagnosed [bib_ref] Meta-analysis and multidisciplinary consensus statement: exome sequencing is a first-tier clinical diagnostic..., Srivastava [/bib_ref]. The Solve-RD projectsystematically reanalyzes exomes and phenotypic data of 19,000 unsolved cases with rare disease from four European Reference Networks (ERNs) to elucidate the genetic etiology, including~5,000 cases from ERN-ITHACA (Intellectual Disability, TeleHealth and Congenital Anomalies; https://ern-ithaca.eu/). Exploration of mitochondrial DNA sequences extracted from WES data is part of this effort, as 27 of the 37 mitochondrial genes have a known diseasegene association (http://www.mitomap.org).
MT-TL1 encodes mitochondrial tRNALeu(UUR), involved in the synthesis of oxidative phosphorylation enzymes by adding leucine to the growing polypeptide chain of mtDNAencoded subunits during translation [bib_ref] Genetic, pathogenetic, and phenotypic implications of the mitochondrial A3243G tRNALeu(UUR) mutation, Finsterer [/bib_ref]. Pathogenic variants in MT-TL1 have been linked to several phenotypes associated with mitochondrial dysfunction, including mitochondrial myopathy, encephalopathy, lactic acidosis and stroke-like episodes (MELAS; MIM#540000) and myoclonic epilepsy associated with ragged-red fibers (MERRF; MIM#545000).
We report on a variant in MT-TL1 known to interfere with mitochondrial function, uncovered by systematic reanalysis of WES data, illustrating the underexposed potential of WES-based analysis of mtDNA in identifying variants with clinical consequences.
# Methods
## Patient inclusion
All individuals (or legal representatives) in the Solve-RD project provided consent, compliant with local ethical guidelines and the Declaration of Helsinki. For this case, the Radboudumc Ethics Board approved the study . For publication of photos, additional consent was obtained.
## Wes
Diagnostic trio-based exome sequencing (proband and parents) was performed as described previously [bib_ref] Diagnostic exome sequencing in persons with severe intellectual disability, De Ligt [/bib_ref] using DNA isolated from whole blood.
## Data sharing
Human Phenotype Ontology-coded clinical data were uploaded along with BAM files to the RD-Connect Genome-Phenome Analysis Platform (https://platform.rdconnect.eu/), and deposited at European Genome-Phenome Archive (EGAZ00001527897), as part of the Solve-RD infrastructure. The variant and phenotype were submitted to the Leiden Open Variation Database (individual number 00328346, phenotype number 0000246573, variant number 0000713909).
## Variant identification
Systematic re-analysis of WES data is described by Matalonga et al.. Details specific to this case include mapping, calling, and annotation of mtDNA using MToolBox pipeline (version 1.0) [bib_ref] MToolBox: a highly automated pipeline for heteroplasmy annotation and prioritization analysis of..., Calabrese [/bib_ref] , with mapping against rCRS (for mtDNA) and GRCh37/Hg19 (for genomic DNA) as reference sequences, allowing the detection of heteroplasmy levels and prioritization of variants. The following parameters were applied to identify possible diseaseassociated variants: (1) coverage ≥8-fold; (2) heteroplasmy fraction ≥1%; (3) GeneBank allele frequency (MITOMAP) <0.2%; (4) "Confirmed" or "Reported" disease association in MITOMAP; and (5) reported "Pathogenic" (ACMG, class 5) or "Likely pathogenic" (ACMG, class 4) in ClinVar.
## Heteroplasmy validation
Confirmation of mitochondrial heteroplasmy was performed on blood and urine of the proband, mother and sisters using routine diagnostic procedures (PGM Ion Torrent Technology).
# Results
## Clinical characteristics
We report on a 23-year-old male proband with a complex neurodevelopmental and neuromuscular phenotype, who remained undiagnosed despite extensive diagnostic evaluation in a tertiary center. Family history was unremarkable, with two healthy older sisters and non-consanguineous parents. After uncomplicated pregnancy and delivery at term (normal birth weight (3840 g) and length (52 cm); head circumference 34 cm, 0 SD; Apgar score 10), first concerns about development arose around 3 months of age. At age 15 months, there was severe developmental delay, consisting of hypotonia, delayed motor, social and communicative milestones, and secondary microcephaly (44 cm, −2.5 SD). Brain MRI (at 15 months; repeated at age 14 years) showed supratentorial pachygyria and frontoparietal polymicrogyria , with white matter abnormalities in the posterior limb of the internal capsule . Cerebellum and corpus callosum showed no deformities and EEG did not show epileptiform activity at age 15 months.
He developed spastic tetraparesis, orofacial dystonia, and dystonia of hands and feet. Epilepsy manifested at 7 years of age with frontal focal-onset seizures, with and without secondary tonic-clonic generalization. No myoclonus was observed. Deterioration occurred at age 17 with episodes of severe headaches, accompanied by nausea, vomiting, hematemesis, pallor and perspiration, coinciding with epileptiform activity on EEG in the left temporal lobe. Brain CT was interpreted as normal at that time . Different anti-epileptic drugs were prescribed, of which several resulted in adverse drug reactions including erythema multiforme (carbamazepine, oxcarbazepine), muscle weakness (clobazam, pregabalin), or obstructive sleep apnea syndrome (clobazam), resolving after cessation of the respective medication. Because of poor seizure control, 24-h continuous EEG monitoring was performed at 18 years and showed bilateral independent and simultaneous periodic discharges, which were more frequent during sleep than during wakefulness. He is now largely seizure free on low-dose carbamazepine. Other medical problems were severe progressive neuromuscular scoliosis , bilateral hip dysplasia A B C Neuroimaging displays pachygyria, polymicrogyria, white matter abnormalities, and loss of gray-white differentiation. Coronal MRI images (A) in the phase-sensitive inversion recovery sequence at age 14 years. The images in the upper row show exemplary regions with a microgyrated aspect (red arrow) as compared to the contralateral region (blue arrow). The images in the lower row show exemplary regions with pachygyration (red arrow). In these regions, the gyrus-sulcus pattern is lost as compared to the contralateral side (blue arrow). Polymicrogyria and pachygyria appear most prominent in the frontoparietal cortical areas. Axial T2-weighted MR images (B; left) and fluid attenuation inversion recovery (FLAIR) images (B; right) at the level of basal ganglia at age 14 years. The middle row shows a magnification of the basal ganglia derived from images in the upper row, with a schematic representation in the lower row. 1: caudate nucleus; 2: putamen; 3: globus pallidus; 4: thalamus; dotted line: white matter in between the basal ganglia representing the area of the internal capsule. In the posterior limb of the internal capsule, white hyperintensities are present. These can be recognized by their T2-weighted/FLAIR hyperintense aspect and suggest microstructural white matter degeneration. In addition, the globus pallidus on both sides shows a hypo-intense aspect on the T2-weighted images and FLAIR sequence. Axial CT scan images (C) in brain tissue setting (W/L: 90/40 HU) at age 17 years. Slides should be viewed from left to right to follow the caudocranial axis; the upper-left corner shows the most caudal slide, the lower-right corner shows the most cranial slide. The red circles indicate the hypodense configuration of the temporal lobe with loss of gray-white differentiation. Gray-white differentiation refers to the appearance of the interface between cerebral white matter and cerebral gray matter on brain CT imaging. Loss of gray-white differentiation often indicates the occurrence of cytotoxic edema. In turn, cytotoxic edema is typical for infarction and hypoxicischemic encephalopathic syndromes. A clear asymmetry between the left temporal lobe and the right temporal lobe can be observed. The hypodense configuration involves the superior, middle, and inferior temporal gyri (Color figure online).
with luxation, drooling with excessive tenacious stringy mucus, and severe feeding difficulties requiring gastric tube feeding and resulting in low body weight (age 18: length 165 cm, −2.6 SD; weight 39.5 kg, weight-forlength −3.1 SD; head circumference 54 cm, −1.9 SD). Ophthalmological assessment, visual and brainstem D C E A auditory evoked potentials, and electrocardiography were normal. Facial dysmorphisms at 19 years of age included a long hypotonic face, full eyebrows, long palpebral fissures, a prominent nose and nasal bridge, high palate, gingival hyperplasia, and abnormal tooth implant , becoming more prominent over time .
Karyotype, genomic micro-array (Agilent 180k oligoarray), targeted analysis of WDR62 and ADGRG1, extensive lysosomal screening, and trio-based WES did not yield a diagnosis. Metabolic workup revealed elevation of multiple amino acids, including glycine, serine, threonine (in plasma and urine), lysine, methionine, and alanine (in plasma), in a pattern not described for a known metabolic disorder. Lactate in blood (1.8 mmol/l, age 20 years) and cerebrospinal fluid (1.0 mmol/l, age 15 months) were within normal range. Pharmacogenetic analysis showed a rare homozygous variant in CYP3A4 (NM_017460.5:c.878T > C;*18; rs28371759) unlikely to explain the largely dose-independent adverse drug reactions, and two rare HLA-types (HLA-A*0103; HLA-B*0835) of unknown significance.
## Systematic re-analysis revealed a known pathogenic variant in mt-tl1
Data of the proband and parents were included in the Solve-RD project. Prioritization of nuclear DNA variants did not yield diagnostically relevant variants, despite analysis of (de novo) variants in known disease-genes, particularly those associated with recessive or dominant cortical dysplasia, and in genes not yet implicated in NDD/ID. Variant prioritization from mtDNA revealed a variant in MT-TL1 [bib_ref] The m.3291T>C mt-tRNA(Leu(UUR)) mutation is definitely pathogenic and causes multisystem mitochondrial disease, Yarham [/bib_ref] , NC_012920.1:m.3291T > C (NC_012920.1:n.62T > C), known to affect mitochondrial function, at 22% heteroplasmy. The variant was absent from maternal WES data.
Validation was performed on blood of the proband and his mother using routine diagnostic procedures, displaying a heteroplasmy level of 23% in the index, compared to 4% in his mother. Urine heteroplasmy levels were 58% and 17% in index and mother, respectively. Follow-up of the family revealed heteroplasmy levels of 4% (blood) and 9% (urine) in one of the sisters, but the variant was absent from blood and urine of the other sister. We re-evaluated brain CT imaging of the index, performed at age 17 years, which in retrospect showed early signs of stroke, including loss of gray-white differentiation in the left temporal lobe , co-localizing with epileptiform activity seen on EEG at the time of onset of episodic symptoms. Additionally, multidisciplinary evaluation and comparison of his phenotype to previously published individuals with the same variant (Table 1) revealed that the proband exhibits several symptoms seen in other individuals (epilepsy, episodic headaches with emesis, feeding difficulties, low body weight, and neuromuscular problems) [bib_ref] Variable phenotypes in a family with mitochondrial encephalomyopathy harboring a 3291T>C mutation..., Sunami [/bib_ref] [bib_ref] MERRF/MELAS overlap syndrome due to the m.3291T>C mutation, Liu [/bib_ref] [bib_ref] The expanding phenotype of MELAS caused by the m.3291T > C mutation..., Keilland [/bib_ref] [bib_ref] MERRF and Kearns-Sayre overlap syndrome due to the mitochondrial DNA m.3291T>C mutation, Emmanuele [/bib_ref] [bib_ref] Neuromuscular syndrome associated with the 3291T->C mutation of mitochondrial DNA: a second..., Uziel [/bib_ref] , but also features that were not described before (abnormalities of brain gyration, facial dysmorphisms, early age of onset of developmental delay).
# Discussion
Systematic re-analysis of existing WES data of unresolved cohorts can efficiently yield additional diagnoses [bib_ref] Reanalysis of whole exome sequencing data in patients with epilepsy and intellectual..., Li [/bib_ref]. Yet, re-analysis rarely includes evaluation of mtDNA although pathogenic mtDNA variants underly many rare diseases. This case illustrates the importance of including mtDNA in re-analysis, as the identification of the MT-TL1 variant is of medical relevance to the proband and his sisters of childbearing age.
Mitochondrial disorders associated with mitochondrial tRNA genes are characterized by both genotypic and phenotypic heterogeneity, with poor genotype-phenotype correlations [bib_ref] The m.3291T>C mt-tRNA(Leu(UUR)) mutation is definitely pathogenic and causes multisystem mitochondrial disease, Yarham [/bib_ref] [bib_ref] Mitochondrial tRNA mutations and disease, Yarham [/bib_ref] [bib_ref] mtDNA heteroplasmy level and copy number indicate disease burden in m.3243A>G mitochondrial..., Grady [/bib_ref]. The m.3291T > C variant described here is located in the T-loop of mitochondrial tRNALeu(UUR). This variant was shown to result in respiratory chain enzyme deficiency and its pathogenicity was proven by single muscle fiber mtDNA analysis, showing high heteroplasmy levels in cytochrome c oxidase deficient muscle fibers [bib_ref] The m.3291T>C mt-tRNA(Leu(UUR)) mutation is definitely pathogenic and causes multisystem mitochondrial disease, Yarham [/bib_ref]. Individuals carrying the m.3291T > C variant display a broad phenotypic spectrum varying from mild symptoms to severely debilitating disease , largely overlapping with features observed in the index. As no additional diagnostically relevant variants could be identified, it remains unclear whether the early age of onset of developmental delay, gyration defects, and dysmorphisms are attributable to a second rare genetic disorder, or expand the m.3291T > C phenotypic spectrum. However, despite gyration defects Clinical and radiology images show severe scoliosis, dysmorphisms, and dental crowding. A Clinical photographs without traction (left) and with traction (middle) showing asymmetry of the chest and rib protrusion on the left side. Radiograph of the vertebral column (right) with anteroposterior view in supine position shows a slight left convex curvature of the upper thoracic spine and severe right convex scoliosis of thoracolumbar spine (Cobb's angle ±75°) with axial rotation (asymmetric projection of spinous processes and pedicles) and asymmetry of the thoracic cavity. B Frontal and C profile facial photographs of the proband (age 19 years), showing a long face with hypotonic appearance, long palpebral fissures, a prominent nose, and small simple ears. The photograph of the mouth (D) shows crowded teeth with gingival hyperplasia (age 21 years). Facial photographs between age 5 months and age 23 years (E), showing a progression of facial dysmorphisms with advancing age. Only mild dysmorphic features are observed in early childhood, including ptosis and a long philtrum (age 5 months to 4 years). However, the proband develops a progressively pronounced long hypotonic face with open mouth (e.g., photographs at 19 versus 10 versus 2 years of age), and has crowded teeth at age 21 (D), whereas teeth appear less crowded at earlier ages (age 7, 13, and 17 years). being uncommon in individuals with MT-TL1 variants, polymicrogyria has been reported for m.3243A > G in two unrelated individuals with atypical MELAS phenotypes, additionally exhibiting other atypical symptoms also observed in the proband, including hypertonia, early onset developmental delay [bib_ref] A3243G mitochondrial mutation associated with polymicrogyria, Keng [/bib_ref] [bib_ref] Bilateral polymicrogyria and MELAS/ A3243G mutation. A very uncommon association, Vidal [/bib_ref] and facial dysmorphisms [bib_ref] A3243G mitochondrial mutation associated with polymicrogyria, Keng [/bib_ref]. The proband's facial dysmorphisms might be secondary to muscle tone abnormalities. Hence, we concluded that the MT-TL1 variant could at least in part, but possibly completely, explain the proband's phenotype.
In conclusion, we describe a male proband carrying a mtDNA variant confirmed to interfere with mitochondrial function, that was identified in systematic, large-scale reanalysis on a large cohort of individuals with unresolved (syndromic) ID through the collaborative Solve-RD project. Our observations suggest that re-analysis encompassing the mtDNA interpreted from WES data may successfully yield novel unanticipated diagnoses in unexplained cases of ERN-ITHACA with implications for reproductive choices of relatives in the maternal line.
## Compliance with ethical standards
Conflict of interest The authors declare no competing interests.
Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
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 license, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license 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 license, visit http://creativecommons. org/licenses/by/4.0/.
[table] Table 1: Phenotypic comparison of the case presented in this study and those previously published to assess the phenotypic spectrum associated with the MT-TL1: m.3291T > C variant (NC_012920.1:m.3291T > C; NC_012920.1:n.62T > C). Normal (serum 1.8 mmol/l; age 20 years; and cerebrospinal fluid 1.0 mmol/l, age 15 months) Not tested Heteroplasmy levels in the mother: 4% (blood), 17% (urine). The variant was observed in one of the sisters with heteroplasmy levels of 4% (blood) and 9% (urine), but undetectable in blood and urine of the other sister COX cytochrome c oxidase, ID intellectual disability, MELAS mitochondrial myopathy, encephalopathy, lactic acidosis, and stroke-like episodes, [/table]
|
Bilateral Mandibular Condylar Fractures with Associated External Auditory Canal Fractures and Otorrhagia
Case reportA 19-year-old female fell 20 feet from a roof and presented with a complete laceration of the lower lip and bilateral otorrhagia. A maxillofacial computed tomography (CT) exam(Fig. 1) revealed a nondisplaced fracture at the mandibular symphysis and bilateral temporomandibular joint (TMJ) fractures. The right TMJ fracture was at the condylar neck. The left TMJ fracture was within the intracapsular portion of the condyle. There was lateral displacement of the left TMJ. There were comminuted fractures of the anterior walls of both external auditory canals (EAC) with acute blood in the canals(Fig. 2). The EAC fracture fragments were more inferiorly and posteriorly displaced on the left(Fig. 3).RCR Radiology Case Reports | radiology.casereports.net 24 A rare case of bilateral mandibular condylar fractures associated with bilateral external auditory canal fractures and otorrhagia is reported. The more severe external auditory canal fracture was present on the side of high condylar fracture, and the less severe external auditory canal fracture was ipsilateral to the condylar neck fracture. A mechanism of injury is proposed to account for such findings. . Maxillofacial CT (0.625 mm axial acquisitions, isotropic reconstructions) of a young female patient after a 20-foot fall. Axial image shows a nondisplaced mandibular symphyseal fracture.
## 3a 3b
# 3c discussion
Otorrhagia is a potential complication of mandibular condylar fractures (1). Fractures of the ipsilateral petrous bone are likely the cause of otorrhagia. Penetration of sharp fracture fragments of the mandibular ramus into the auditory canal may also be another cause (2). There is a paucity of published literature describing the relationship between condylar fractures and EAC fractures. There have been only a few case reports, and only one of those is in the general radiology literature [bib_ref] External auditory canal fracture secondary to mandibular trauma, Chong [/bib_ref]. This case report contributes further understanding of the mechanism of such injury and posttraumatic otorrhagia.
The EAC and TMJ are intimately located. The EAC is divided into an outer cartilaginous one-third and an inner osseous two-thirds. The TMJ is located anterior to the inner osseous canal. The anterior and inferior walls of the osseous canal are composed of the tympanic portion of the temporal bone. Due to this intimate relationship, trauma to the TMJ is known to produce trauma to the EAC (4).
Fractures of the TMJ are often due to high impact injuries such as road traffic accidents, violence or falling. The vast majority of these fractures are either high condylar or subcondylar fractures. High condylar fractures are fractures of the intracapsular condyle (condylar head). Subcondylar fractures occur at a weak point in the ramus at 2-3 cm caudal to the condylar neck. The condylar neck is also another weak point.
The patient in this case report suffered a right condylar neck fracture and a left high condylar fracture. As shown on figures 2 and 3, the fragments of the anterior wall of the left EAC were displaced more posteriorly and inferiorly than those on the right . This suggests that the force applied to the left EAC was greater. One possible explanation is the shorter distance between a high condylar fracture and the ipsilateral temporal bone. With a condylar neck or subcondylar fracture, such distance is longer, allowing for greater dissipation of force. Angulation at the left TMJ in this patient may also contribute a shearing component to the fractures and cause lateral dislocation of the joint.
Bilateral mandibular condylar fractures are frequently associated with concomitant fractures of the mandibular body [bib_ref] Condylar fractures of the mandible. I. Classification and relation to age, occlusion,..., Lindahl [/bib_ref]. It is likely that condylar fractures are a result of indirect force while mandibular body fractures are due to direct impaction [bib_ref] Consideration of 180 cases of typical fractures of the mandibular condylar process, Maclennan [/bib_ref]. In this patient the exertion of force from falling was not fully absorbed in the area of its primary application at the mentum, with the remainder of the force transmitted to the condyles. Fractures of the condyles in these injuries are thought to be protective of brain injury as intracranial displacement of the condyles is avoided [bib_ref] Surgical versus nonsurgical treatment of unilateral dislocated low subcondylar fractures. Discussion, Oikarinen [/bib_ref]. The direction, degree, magnitude and precise point of application of the force and the state of dentition and the occlusal position of the mandibles determine the displacement of condylar fractures [bib_ref] Fractures of the mandibular condyle: a review of 466 cases. Literature review,..., Zachariades [/bib_ref]. Fractures of the condyles are more likely to occur with an open mouth.
Computed tomography is the diagnostic modality of choice for mandibular condylar fractures. It is important to recognize potential association between mandibular condylar trauma and temporal bone fracture. This case report highlights the need to scrutinize the temporal bones, particularly the petrous portion as well as the EAC on CT for fractures in a patient with condylar fractures. Such scrutiny is especially important on the side with a high condylar fracture as there is a greater impaction on the ipsilateral temporal bone. . Maxillofacial CT (0.625 mm axial acquisitions, isotropic reconstructions) of a young female patient after a 20-foot fall. Sequential parasagittal images show more posterior displacement of EAC fractures (curved arrows) on the left (C is lateral to D) than on the right (A is lateral to B). Note that the parasagittal images are from comparable sagittal planes as evidenced by similar sections through the semicircular canals.
## 3d
[fig] Figure 1B ,: C. Maxillofacial CT (0.625 mm axial acquisitions, isotropic reconstructions) of a young female patient after a 20-foot fall. Reconstructed coronal (B,C) images show an angulated right condylar neck fracture (B) and a comminuted left high condylar fracture (B, curved arrow). Arrow in C points to lateral dislocation of left TMJ. [/fig]
[fig] Figure 2A, Figure 2B: Maxillofacial CT (0.625 mm axial acquisitions, isotropic reconstructions) of a young female patient after a 20-foot fall. Axial images show comminuted fractures of the anterior walls of both EAC's (arrowheads). The fluids in the auditory canals likely represent acute blood. Maxillofacial CT (0.625 mm axial acquisitions, isotropic reconstructions) of a young female patient after a 20-foot fall. Axial images show comminuted fractures of the anterior walls of both EAC's (arrowheads). The fluids in the auditory canals likely represent acute blood. [/fig]
[fig] Figure 2C 2B, Figure 3: Maxillofacial CT (0.625 mm axial acquisitions, isotropic reconstructions) of a young female patient after a 20-foot fall. On the coronal view, the fragments are more inferiorly displaced on the left side than on the right side (curved arrows in C).2C Maxillofacial CT (0.625 mm axial acquisitions, isotropic reconstructions) of a young female patient after a 20-foot fall. Sequential parasagittal images show more posterior displacement of EAC fractures (curved arrows) on the left (C is lateral to D) than on the right (A is lateral to B). Note that the parasagittal images are from comparable sagittal planes as evidenced by similar sections through the semicircular canals. [/fig]
|
Application of inulin in bread: A review of technological properties and factors affecting its stability
Due to its dual function, inulin is an important prebiotic compound in the cereal industry, especially in bread production. In other words, improving technological features and creating health properties (such as reducing the risk of type 2 diabetes, heart disease, metabolic syndrome, and osteoporosis) have led to the widespread use of this compound. Inulin has many important technological functions in bread, including its ability to interact with water, create structure, and influence rheological properties, texture, and overall acceptability of the final product. Nevertheless, bread processing conditions can influence the structural integrity of inulin and thus affect its technological efficiency. Therefore, this review article aims to investigate the technological properties and factors affecting the stability of inulin during bread processing conditions. Generally, the addition of inulin could considerably improve the technical performance of bread. However, the stability of inulin depends on the formulation components, type of fermentation, and baking process.
## | introduc ti on
Bread is one of the essential components of the diet, which is mainly made from wheat flour [bib_ref] Inulin enriched wheat bread: Interaction of polymerization degree and fermentation type, Mohammadi [/bib_ref]. Gluten protein is responsible for the unique properties of wheat dough [bib_ref] Wheat bread: Potential approach to fortify its lysine content, Meybodi [/bib_ref] , which constitutes more than 80% of wheat proteins.
It consists of two proteins, gliadin and glutenin, which play a critical role in producing the viscosity and elasticity of the dough. Gluten can also affect some factors such as water absorption, dough mixing time, the storage capacity of gas (produced during processing), and ultimately the quality of bread (Natalia [bib_ref] Technological benefits of inulin-type fructans application in gluten-free products-A review, Drabińska [/bib_ref].
There are some challenges regarding the production and consumption of some bakery products, especially wheat bread. Although the use of whole-wheat flour improves the nutritional properties of the finished bread due to its high content of fiber and antioxidant compounds, there are several problems with this type of flour. Some of these vulnerabilities include nutritional issues such as reduced mineral bioavailability (due to phytic acid and insoluble fibers) and technological problems, including reduced specific volume, moisture, and water activity as increased hardness [bib_ref] Rheological properties of wheat flour dough and pan bread with wheat bran, Boita [/bib_ref]. Therefore, white wheat flour (containing 21% bran) is recommended to produce bulk wheat bread [bib_ref] Comparison of the impact of Lactobacillus casei and Lactobacillus rhamnosus on acrylamide..., Dastmalchi [/bib_ref]. In addition, wheat flour-based baked goods are also prohibited for people with celiac disease and other gluten-related disorders. Despite switching to a gluten-free diet such as gluten-free bread, it is impossible to benefit from the desired technological features of these products [bib_ref] Impact of different S. cerevisiae yeast strains on glutenfree dough and bread..., Horstmann [/bib_ref]. Furthermore, there is also a need to freeze dough as it is a convenient and cost-effective way to meet the nutritional needs of people who are pressed for time and do not have access to various food sources such as travelers, mountaineers, astronauts, and soldiers. However, this procedure is associated with adverse changes in the technological characteristics of wheat bread, including a reduction in specific volume, increased fermentation time, and a weakening of the gluten network structure [bib_ref] Effects of inulin on protein in frozen dough during frozen storage, Ke [/bib_ref]. Hence, it is challenging to fortify white wheat bread and simulate gluten structure in gluten-free bread.
In recent years, dietary fiber has attracted much attention because of its beneficial effects on health, including improving gastrointestinal health and preventing constipation, cardiovascular disease, and cancer [bib_ref] Chemical, physical, and baking properties of apple fiber compared with wheat and..., Chen [/bib_ref]. These compounds are divided into two categories depending on their solubility in water: "soluble dietary fiber" and "insoluble dietary fiber" [bib_ref] A study on fibre addition to gluten free bread: Its effects on..., Sciarini [/bib_ref]. It is reported that the addition of insoluble dietary fiber negatively affects the technological properties of the final product [bib_ref] Kinetic study of staling in breads with high-amylose resistant starch, Arp [/bib_ref].
It may also increase the postprandial glucose level of fortified food formulation by providing a physical barrier and preventing structure formation. Soluble fiber, on the other hand, can increase lipid and carbohydrate metabolism and lower cholesterol and blood glucose levels after eating [bib_ref] Inulin dietary fiber with functional and health attributes-A review, Kalyani Nair [/bib_ref].
Prebiotics are indigestible and soluble dietary fibers that are resistant to gastric acid and hydrolysis by mammalian enzymes. In the digestive tract, they could selectively stimulate certain probiotic bacteria. Consequently, prebiotics may reduce the risk of type 2 diabetes, metabolic syndrome, osteoporosis, inflammatory bowel disease, irritable bowel syndrome, obesity, and dysentery [bib_ref] Prebiotic fiber modulation of the gut microbiota improves risk factors for obesity..., Parnell [/bib_ref]. One of the most important prebiotic compounds with wide application in the bakery industry is inulin, which, in addition to functional and desirable effects on the technological properties of bread, has several beneficial effects on health [bib_ref] The effect of pH, temperature and heating time on inulin chemical stability, Glibowski [/bib_ref].
Considering the importance of wheat bread in human nutrition and the need for its enrichment and mimicking of the threedimensional network of gluten in some applications, the inclusion of inulin as soluble dietary fibers needs to be studied. Since the technological properties of bread, which are crucial for consumer acceptance, are highly influenced by the formulation, this study aims to review the effects of inulin on the technological properties of bread.
Furthermore, since inulin's health-promoting effects are attributed to its regular and specific consumption in intact form, the factors affecting its stability during bread processing will be investigated.
## | in ulin: chemi c al and he alth propertie s
Inulin is a water-soluble storage polysaccharide composed of fructose units linked by 1: 2 beta bonds with a degree of polymerization (DP) of 2-60 [bib_ref] Inulin: Properties, health benefits and food applications, Shoaib [/bib_ref]. 1 illustrates its benefits.
Although the main extraction source of inulin is chicory root [bib_ref] Effect of the addition of different fibres on wheat dough performance and..., Wang [/bib_ref] , it can be found in more than 36,000 plant species such as bananas, onions, artichokes, asparagus, leeks, chicory root, garlic, and pickled potatoes. There are three categories of inulin, including short chains (DP < 10), long chains (DP > 23), and natural (DP = 2-60) types. Short-chain inulin has higher solubility and sweetness than the long-chain type and is used as a sugar substitute and mouthfeel enhancer in low-calorie products. Long-chain inulin, on the other hand, is used as a filler and fat substitute because it is less soluble and more viscous (Natalia [bib_ref] Technological benefits of inulin-type fructans application in gluten-free products-A review, Drabińska [/bib_ref]. The inulin compound is colorless and tastes neutral. Although it has little effect on the taste of the finished product, it can significantly affect the organoleptic properties of this product in terms of DP [bib_ref] Inulin dietary fiber with functional and health attributes-A review, Kalyani Nair [/bib_ref].
The β-configuration of inulin prevents it from being hydrolyzed by enzymes in the upper gastrointestinal tract of humans, allowing it to enter the large intestine intact and be selectively fermented by beneficial intestinal bacteria. According to this property, inulin belongs to the category of prebiotics and low-calorie dietary fibers (compared to other carbohydrates) [bib_ref] Inulin enrichment of gluten free breads: Interaction between inulin and yeast, Morreale [/bib_ref]. This compound can also reduce the glycemic response, enhance gastrointestinal health, affect fat metabolism, and reduce the risk of osteoporosis and colon cancer [bib_ref] Inulin dietary fiber with functional and health attributes-A review, Kalyani Nair [/bib_ref]. The use of artichoke extract (as a source of inulin) in rat's drinking water has been shown to lower total cholesterol and low-density lipoprotein (LDL) and increase high-density lipoprotein (HDL) levels [bib_ref] The effect of artichoke supplement on lipid metabolism in rats subjected to..., Naci [/bib_ref]. Considering the functional benefits for celiac patients, the addition of 12% inulin-type fructans (ITFs) to gluten-free bread formulation (based on rice flour (50%) and potato starch (50%)) has also been documented to increase calcium absorption, reduce glycemic index (from 71 to 48), and glycemic load (from 12 to 8) [bib_ref] Effects of prebiotic inulin-type fructans on structure, quality, sensory acceptance and glycemic..., Capriles [/bib_ref]. The use of 100 g inulin in the diet of heterogeneous rats with human flora is also associated with increased sulfomucin production and decreased sialomucin (indicating a less risk of colon cancer) [bib_ref] Intestinal mucin distribution in the germ-free rat and in the heteroxenic rat..., Fontaine [/bib_ref].
Since the health benefits of inulin require adequate consumption, including inulin in bread may be a good strategy in a cerealbased diet to ensure adequate intake. However, the technology performance of this combination is also crucial. Considering the F I G U R E 1 Benefits of using inulin Increase the number of beneficial intestinal bacteria
## Aid in calcium absorption
Helps you feel fuller for longer Improve good cholesterol importance of the technological properties of bread for consumer acceptance, the following section examines the influences of inulin addition on these properties.
## | technolog ic al char ac teris tic s of b re ad containing in ulin
Technological characteristics are effective parameters in predicting the mixing behavior of the dough, fermentation characteristics, baking efficiency of food production, and textural characteristics of the final product. The textural, rheological, and sensory characteristics of bread are the most important technological characteristics that determine the final product's quality and the consumer's acceptance [bib_ref] The effect of inulin and fructooligosaccharide supplementation on the textural, rheological and..., Morris [/bib_ref]. In this regard, water absorption of bread formulation can significantly affect its technological characteristics by influencing the above parameters.
Consequently, an increase in water absorption can improve the specific volume properties, dough expansion [bib_ref] Rice varieties in relation to rice bread quality, Han [/bib_ref] , and the formation of large cavities in the bread crumb (the bread crumb is a white porous structure located under the crust, while the crust is the brown surface layer of the bread) [bib_ref] Inulin enriched wheat bread: Interaction of polymerization degree and fermentation type, Mohammadi [/bib_ref] , as well as the reduction in the hardness and staling rate of the bread [bib_ref] The effect of inulin and fructooligosaccharide supplementation on the textural, rheological and..., Morris [/bib_ref]. Furthermore, the amount of water absorption also determines the color of the bread crust, with low humidity promoting the Millard reaction and creating a darker color [bib_ref] Inulin enriched wheat bread: Interaction of polymerization degree and fermentation type, Mohammadi [/bib_ref].
It is worth mentioning that the incorporation of inulin into bread formulation could improve its technological properties due to its high capacity to interact with water [bib_ref] Influence of the soluble fibres inulin and oat β-glucan on quality of..., Hager [/bib_ref]. Moreover, it can prevent the formation and recrystallization of ice crystals and avoid damaging the proteins during the dough-freezing process.
Additionally, the presence of hydrogen groups in the inulin structure and the hydrophobic reactions between inulin and proteins can have a protective effect on the gluten network. For instance, it was reported that frozen dough containing 2.5% long-chain inulin had a greater protective effect on gluten protein compared with that containing 5% short-chain inulin. This phenomenon can be explained by the formation of denser gluten structures by the former type of inulin. Long-chain inulin is also used in gluten-free bread as a substitute for gluten (due to the formation of the gel structure) which could improve the technological properties of the final bread [bib_ref] Technological quality of bread from rice flour with spirulina, Selmo [/bib_ref]. Several parameters, including the presence or absence of gluten and the strength of the gluten network, influence the effects of inulin on the main technological properties of wheat bread, such as water absorption, special volume, texture, color, and staling, which will be discussed in detail in the following sections.
## | water absorption
Water is one of the main components in the formation and stability of dough. Water absorption is defined as the ratio of water volume to flour weight required to produce dough with standard consistency [bib_ref] The impact of salt reduction in bread: A review, Belz [/bib_ref]. The quality and shelf life of products are strongly influenced by the moisture distribution and the interaction of water with other compounds in the dough formulation [bib_ref] Effect of inulin with different degree of polymerization on plain wheat dough..., Luo [/bib_ref]. In this regard, as the baking temperature of the bread increases, the water acts as an insulator, preventing the evaporation of moisture from the bread crumb. Furthermore, water absorption of the dough and the resulting moisture can directly affect the texture, volume, color, and stale properties of bread [bib_ref] The effect of inulin and fructooligosaccharide supplementation on the textural, rheological and..., Morris [/bib_ref].
Since inulin hydrocolloids have abundant hydroxyl groups in their structure and form gel-like structures, they may play an important role in water absorption by maintaining and increasing it [bib_ref] Effect of inulin with different degree of polymerization on plain wheat dough..., Luo [/bib_ref].
Factors such as the amount and strength of protein, the type and content of starch, the presence of dietary fiber in the formulation, DP, and the level of inulin replacement affect the performance and behavior of inulin concerning water absorption by influencing gel formation [bib_ref] Rheological parameters of dough with inulin addition and its effect on bread..., Bojnanska [/bib_ref] [bib_ref] Non-gluten proteins as structure forming agents in gluten free bread, Ziobro [/bib_ref]. Additionally, adding gluten protein to the dough increases its water absorption, stability time, and strength [bib_ref] Physicochemical and structural properties of wheat gluten/rice starch dough-like model, Fan [/bib_ref]. presents the effect of inulin with different properties on water absorption of different pieces of bread investigated in previous studies. According to the previous research, the addition of inulin reduces the water absorption and moisture of the dough, as inulin creates a structure with low elasticity and high viscosity, by weakening the gluten network and at the same time forming areas of high water inclusion capacity [bib_ref] Effect of inulin on rheological, textural, and structural properties of brown wheat..., Ahmed [/bib_ref] [bib_ref] Inulin enriched wheat bread: Interaction of polymerization degree and fermentation type, Mohammadi [/bib_ref]. Therefore, less water is needed to achieve a dough with the desired consistency. It should be noted that short-chain inulin had a greater effect on reducing dough water absorption compared with long-chain inulin in constant water content. This can be a result of the fact that low DP inulin contains simple mono-and oligosaccharide sugars that compete with starch granules for water absorption, which delays the gelatinization of starch and reduces the dough water absorption. However, long-chain inulin was not significantly different from the control sample in terms of water absorption. This could be due to the lower solubility and ability to form a gel-like structure of the longer chain [bib_ref] Influence of inulin on physical characteristics and staling rate of gluten-free bread, Ziobro [/bib_ref]. Additionally, the powder form of inulin reduces the water absorption of the dough more than the gel form due to its low water solubility. However, gels have a higher rate of interaction with water [bib_ref] Physicochemical characterization and stability of inulin gels, Chiavaro [/bib_ref] Morris & Morris, 2012). As a consequence, it can be argued that inulin contributes crucially to the stability and structure of the final dough by influencing water absorption. were reduced with the substitution of 5% of FS and 5% of FI, but 2.5% of FXL was not significantly different from the control sample.
## Ta b l e 1 effect of inulin properties on water absorption in different breads
## | specific volume
The specific volume index determines the ability of the dough to retain and develop the carbon dioxide produced during the fermentation and baking process [bib_ref] Impact of emulsifiers on the quality and rheological properties of glutenfree breads..., Nunes [/bib_ref]. It is affected by factors such as water absorption rate, the intensity of the baking process, the performance of the fermentation process, and the aeration rate during kneading.
The addition of inulin, as one of the factors affecting the specific volume, is influenced by several factors, including the type of dough matrix (presence or absence of gluten and gluten network strength), type of fermentation process (Saccharomyces cerevisiae yeast and sourdough), amount, DP, and form of inulin (powder or dehydrating . Generally, the inclusion of inulin in the dough may alter the formation of disulfide bonds and retention of CO2 gases by delaying the gelatinization of starch and reducing water absorption [bib_ref] Influence of inulin on physical characteristics and staling rate of gluten-free bread, Ziobro [/bib_ref]. It is revealed that the addition of inulin to gluten-free bread improves several features such as viscosity, gas storage capacity during the baking process, and specific volume. These positive effects are attributed to its hydrophilic properties and the formation of a gel-like structure. However, inulin addition to gluten-containing bread reduces the specific volume compared with the control sample due to a reduction in the elasticity of the dough and weakening of the gluten network [bib_ref] Textural and staling characteristics of steamed bread prepared from soft flour added..., Kou [/bib_ref]. There have been inconsistent results in previous studies, which were discussed below.
Regarding inulin DP, it is reported that short-chain inulin cannot form a gel; natural inulin forms a gel in concentrations above 30% w/w. In contrast, long-chain inulin forms a gel in the range of 20%-40% w/w at room temperature due to its lower solubility in water. Consequently, it positively affects the increase in the specific volume [bib_ref] Physicochemical characterization and stability of inulin gels, Chiavaro [/bib_ref]. Likewise, [bib_ref] The effect of inulin and fructooligosaccharide supplementation on the textural, rheological and..., Morris [/bib_ref] revealed that compounds containing long-chain inulin in glutencontaining bread had a higher specific volume than short-chain types which is due to the formation of a gel-like structure and compensation for the weakening of the gluten network. However, short-chain inulin in gluten-free bread is easily hydrolyzed by yeasts due to the presence of simple monosugars and oligosaccharides, increasing gas production and thus increasing specific volume [bib_ref] Influence of inulin on physical characteristics and staling rate of gluten-free bread, Ziobro [/bib_ref]. As observed in a study by [bib_ref] Influence of inulin on physical characteristics and staling rate of gluten-free bread, Ziobro [/bib_ref] , by adding short-and medium-chain inulin (DP < 10 and DP ≥ 10) to gluten-free bread (based on corn starch and potato starch), the specific volume increased by 25%-33% and 24%-59%, respectively, compared with the control sample. In contrast, the sample containing long-chain inulin did not differ significantly from the control sample.
Considering the type of fermentation, it has been reported that fermentation with sourdough improves the specific volume resulting from the synergistic effect of bacterial lactic acid on the metabolic activity of Saccharomyces cerevisiae and the further release of carbon dioxide. In addition, inulin can partially compensate for the weakening of the gluten network caused by this type of fermentation by producing exopolysaccharides. [bib_ref] Inulin enriched wheat bread: Interaction of polymerization degree and fermentation type, Mohammadi [/bib_ref] explored that the specific volume of fermented wheat bread with mixed fermentation based on sourdough was significantly higher than samples fermented with Saccharomyces cerevisiae. Also, wheat bread containing long inulin fermented chains with sourdough showed a higher specific volume compared with short types of chains fermented with yeast. In this regard, the exopolysaccharides produced during the fermentation of sourdough and fermented long-chain inulin can synergistically compensate the decrease in the gluten content,
due to the addition of inulin level [bib_ref] Factors affecting gel formation of inulin, Kim [/bib_ref]. Gluten-free bread fermented with yeast with normal invertase activity (Y 1 ) has a higher specific volume than bread fermented with yeast with low invertase activity (Y 2 ) due to higher enzymatic activity and gas production [bib_ref] Inulin enrichment of gluten free breads: Interaction between inulin and yeast, Morreale [/bib_ref]. As for the effect of inulin addition form on the specific volume, the results of showed that inulin inclusion in wheat bread as a gel (containing 13.2% protein) led to a higher specific volume than the powdered state because of increased water absorption . Also, increasing inulin replacement as one of the factors affecting specific volume showed that with the increase in inulin substitution (from 5% to 10% w/w) in Chinese steamed bread (CSB), the specific volume decreased to 7.66%. This can be attributed to the decrease in dough elasticity due to the weakening of the gluten network [bib_ref] Effects of short-chain inulin on quality of Chinese steamed bread, Liu [/bib_ref].
## | texture
The texture is one of the most important parameters in bakery products, which significantly affects consumer acceptance. Especially for elderly people and those with swallowing and chewing disorders, the texture of food is important. Therefore, the right texture plays an important role in product superiority. presents the effect of inulin addition on the textural properties of different bread.
As the most important indicator of textural properties, hardness is directly related to chewiness [bib_ref] Influence of potato flour on dough rheological properties and quality of steamed..., Liu [/bib_ref]. It was reported that the hardness of wheat bread samples increased with the addition of inulin, which is related to the DP of inulin (Morris TA B L E 2 Effect of different inulin addition on textural properties of different breads. Arabic bread
## Type of bread
## Inulin properties
## Matrix properties
## Parameters
# Results
## References
## Chicory-derived inulin
Inulin addition at 1.25%, 2.5%, and 3.75%, respectively 20.09, 21.09, and 24.06 N for inulin addition at 1.25%, 2.5%, and 3.75%, respectively 0.01, 0.02, and 0.03 for inulin addition at 1.25%, 2.5%, and 3.75%, respectively
The hardness and cohesiveness increased with increasing the level of inulin replacement [bib_ref] Effect of inulin on rheological, textural, and structural properties of brown wheat..., Ahmed [/bib_ref] & [bib_ref] The effect of inulin and fructooligosaccharide supplementation on the textural, rheological and..., Morris [/bib_ref]. Overall, the hardness of wheat bread kernels increased with the addition of inulin because the elasticity of the dough decreased, the gluten network was weakened, and thus the gas storage capacity was reduced. The hardness degree is also affected by the type of fermentation. For example, wheat bread samples fermented by mixed fermentation based on sourdough had significantly higher hardness and chewiness than samples fermented by Saccharomyces cerevisiae. A possible explanation for this could be increased proteolytic activity in the mixed fermentation based on sourdough. Therefore, hydrolysis of various gluten residues results in a system with less flexibility and elasticity [bib_ref] Functional effects of phytate-degrading, probiotic lactic acid bacteria and yeast strains isolated..., Fekri [/bib_ref].
Cohesiveness is an indicator of the internal bands' strength of the product. Higher cohesion means a better ability to maintain the integrated structure of the bread, which tends to decrease with increasing inulin content in wheat bread due to the reduction in disulfide bonds and the weakening of the gluten network.
## | color
Color is an important quality characteristic of bread, which is measured with a colorimeter. This characteristic can be determined by the following parameters: brightness index (L*) varying between 0 (black) and 100 (white), as well as the blue-yellow index (b*) and green-red index (a*), which range from −127 to +127 [bib_ref] Effect of inulin and resistant starch on some of qualitative properties of..., Moghaddasi [/bib_ref].
Inulin, as one of the effective factors for bread color, increases the rate of the Maillard reaction, the formation of brown nitrogen polymers, and melanoidin during baking. Hence, inulin addition can improve the color of the bread crust (especially in gluten-free bread, which usually has a weak color compared to wheat bread) [bib_ref] Technological benefits of inulin-type fructans application in gluten-free products-A review, Drabińska [/bib_ref]. It is noticeable that the color of the bread crust mainly results from the Millard reaction products, while the bread crust color is mainly influenced by the components of the formulation [bib_ref] The effect of soya flour and flaxseed as a partial replacement for..., Conforti [/bib_ref]. In this context, [bib_ref] Inulin enriched wheat bread: Interaction of polymerization degree and fermentation type, Mohammadi [/bib_ref] showed that the crumb color of wheat bread enriched with inulin fermented by mixed fermentation based on sourdough had the highest specific volume and the lowest brightness. This is because the increase in enzymatic activity by the sourdough microorganisms leads to a more porous texture and greater light scattering, thus reducing brightness (L*). Considering the effect of inulin replacement level as one of the practical factors in creating crust color, [bib_ref] Effects of short-chain inulin on quality of Chinese steamed bread, Liu [/bib_ref] found that the replacement of 5% inulin increased the L* values of CSB crust compared with the control sample. On the other hand, inulin is whiter than flour. Also, Inulin can bind with starch, protein, and water to create a uniform structure that reflects light.
Therefore, this structure produces a whiter color when light strikes the uniform surface of the bread [bib_ref] Effects of short-chain inulin on quality of Chinese steamed bread, Liu [/bib_ref].
Considering the effect of the yeast type used during the fermentation process on the amount of color produced, [bib_ref] Inulin enrichment of gluten free breads: Interaction between inulin and yeast, Morreale [/bib_ref] identified that gluten-free bread fermented with yeast Y 1 showed greater a* and b* values compared with gluten-free bread fermented with yeast Y 2 because yeast Y 1 can increase free sugar production and Millard reaction rate compared to Y 2 [bib_ref] Inulin enrichment of gluten free breads: Interaction between inulin and yeast, Morreale [/bib_ref]. Similarly, [bib_ref] Inulin enriched wheat bread: Interaction of polymerization degree and fermentation type, Mohammadi [/bib_ref] explored that wheat bread samples containing short-chain inulin (DP < 10) fermented with sourdough had lower L* and higher a* and b* than samples containing long-chain inulin fermented with Saccharomyces cerevisiae yeast. Since crust color is mainly due to the Millard reaction, samples with short-chain inulin fermented with sourdough are more susceptible to the Millard reaction than samples with long-chain inulin fermented with Saccharomyces cerevisiae due to the more simple sugars and increased enzymatic activity. Overall, the bread color containing inulin is also affected by the DP of inulin. In this regard, [bib_ref] Effect of soluble dietary fibre addition on rheological and breadmaking properties of..., Peressini [/bib_ref] showed that the sample containing short-chain inulin had a higher Millard reaction rate and, consequently, a browner crust color than the sample containing long-chain inulin. This can be attributed to the presence of more reducing sugars in the former type [bib_ref] Effect of soluble dietary fibre addition on rheological and breadmaking properties of..., Peressini [/bib_ref]. It should be noted that the state of the added inulin has no significant effect on color. In this regard, [bib_ref] Influence of the soluble fibres inulin and oat β-glucan on quality of..., Hager [/bib_ref] reported that the color of the crust was the same for all levels (2.5%, 5%, and 7.5%) of added inulin in two forms (gel and powder) [bib_ref] Influence of the soluble fibres inulin and oat β-glucan on quality of..., Hager [/bib_ref].
## | staling
The staling process occurs when hydrophilic compounds of flour migrate from the crust to the crumb [bib_ref] Effect of soluble dietary fibre addition on rheological and breadmaking properties of..., Peressini [/bib_ref] as a result of the interaction between hydrophilic flour compounds (gluten-starch) and retrograded starches [bib_ref] Influence of inulin on physical characteristics and staling rate of gluten-free bread, Ziobro [/bib_ref]. The bread staling occurs through the accumulation of amylopectin in the starch structure. Baking, on the other hand, causes the starch to gelatinize, and the amylase to separate from the amylopectin during the process. During cooling, gelatinized starch changes from amorphous to crystalline (retrogradation). Thus, the linear amylose molecules and the linear part of amylopectin can be categorized as rigid crystal structures [bib_ref] Technological benefits of inulin-type fructans application in gluten-free products-A review, Drabińska [/bib_ref]. In general, the staling process is associated with a decrease in bread quality as well as an increase in hardness. This process leads to adverse changes in the appearance, taste, and texture of the product, which lowers customer acceptance and increases waste [bib_ref] Influence of inulin on physical characteristics and staling rate of gluten-free bread, Ziobro [/bib_ref]. However, white bread is one of the most important diet components but has the shortest commercial shelf life among processed food products due to its rapid staling [bib_ref] Influence of inulin on physical characteristics and staling rate of gluten-free bread, Ziobro [/bib_ref].
Inulin inoculation, as one of the effective factors impacting the staling rate, causes the partial distribution of water. It can bind to starch through hydrogen bonds and change the water distribution between starch and protein in gluten-containing bread [bib_ref] Textural and staling characteristics of steamed bread prepared from soft flour added..., Kou [/bib_ref]. Furthermore, inulin, as a hydrocolloid substance with high water-holding capacity, could reduce water loss during bread storage [bib_ref] Influence of the soluble fibres inulin and oat β-glucan on quality of..., Hager [/bib_ref]. In other words, inulin prevents amylose leaching during baking by forming a barrier around starch granules, which affects the staling rate in the first hours of storage [bib_ref] Bread staling, Maga [/bib_ref] [bib_ref] Influence of inulin on physical characteristics and staling rate of gluten-free bread, Ziobro [/bib_ref]. Factors affecting inulin performance on bread staling are shown in .
The effect of inulin on the staling rate of gluten-free and glutencontaining bread depends on several factors, including inulin characteristics such as DP, extraction source, and addition method [bib_ref] Bread staling, Maga [/bib_ref] [bib_ref] Influence of inulin on physical characteristics and staling rate of gluten-free bread, Ziobro [/bib_ref]. In this regard, the study results of [bib_ref] Technological benefits of inulin-type fructans application in gluten-free products-A review, Drabińska [/bib_ref] showed that the staling rate of gluten-free bread decreased by adding 5% inulin compared to the control bread sample [bib_ref] Technological benefits of inulin-type fructans application in gluten-free products-A review, Drabińska [/bib_ref]. By contrast, [bib_ref] Influence of the soluble fibres inulin and oat β-glucan on quality of..., Hager [/bib_ref] found an increase in staling rate with the addition of 5% inulin, probably due to differences in inulin extraction sources [bib_ref] Influence of the soluble fibres inulin and oat β-glucan on quality of..., Hager [/bib_ref]. Similarly, [bib_ref] Influence of inulin on physical characteristics and staling rate of gluten-free bread, Ziobro [/bib_ref] study documented that with the increasing DP inulin, the hardness and staling rate of gluten-free bread increased. This phenomenon can be attributed to the reduction in specific volume (due to lower gas production) by the use of long-chain inulin (DP > 23) [bib_ref] Influence of inulin on physical characteristics and staling rate of gluten-free bread, Ziobro [/bib_ref]. This outcome is contrary to that of [bib_ref] Bread staling studies. III. Effect of pentosans on dough, bread, and bread..., Kim [/bib_ref] , who observed that as DP increased, the staling rate decreased due to the formation of gel structure and less moisture loss.
As Bread storage time had a significant effect on inulin performance in changing the stale rate, as a study by [bib_ref] Textural and staling characteristics of steamed bread prepared from soft flour added..., Kou [/bib_ref] showed that inulin increased the stale rate of soft flour bread (12.9% moisture, 9.7% protein, 0.47% ash, and 22.2% gluten protein) in a short time (less than a day) after baking. The staling rate of bread decreased over a long period (more than a day) because inulin increases starch crystallization during short-term storage. Additionally, another possible reason could be its high ability to enclose water and decrease the migration of water from the crumb to the crust [bib_ref] Textural and staling characteristics of steamed bread prepared from soft flour added..., Kou [/bib_ref]. Likewise, [bib_ref] Effects of prebiotic inulin-type fructans on structure, quality, sensory acceptance and glycemic..., Capriles [/bib_ref] observed similar results by investigating that gluten-free bread fortified with inulin in the first hours (2, 24, and 48) after baking was firmer than the control bread. However, over long-term storage, inulin-containing bread had a softer texture caused by the redistribution of water between components with a lower staling rate than control bread [bib_ref] Effects of prebiotic inulin-type fructans on structure, quality, sensory acceptance and glycemic..., Capriles [/bib_ref]. Overall, it can be argued that the staling rate of gluten-free and gluten-containing bread decreased with the increasing amount of inulin addition, regardless of its type. This is because inulin contains a large number of hydrophilic groups that slow down the migration of water from the breadcrumbs to the crust.
## | fac tor s affec ting the s tab ilit y of in ulin during the b re ad pro ce ss ing s tag e
The lack of dietary fiber in most bread and the worldwide importance of wheat bread as a staple food make inulin fortification an excellent strategy for improving bread quality and human health.
Despite the resistance of β-configuration of inulin to the hydrolytic activity of upper gastrointestinal enzymes, its stability may be affected by the formulation components, pH and acidity of the fermentation process, and baking temperature during the processing step [bib_ref] Inulin enriched wheat bread: Interaction of polymerization degree and fermentation type, Mohammadi [/bib_ref]. Therefore, the most important factors affecting inulin stability were discussed in this study. illustrates the factors affecting the inulin stability during the bread preparation process.
## | formulation components
The main ingredients in bread are flour, water, salt, and yeast which are consumed after fermentation and baking [bib_ref] A review on the gluten-free diet: Technological and nutritional challenges, El Khoury [/bib_ref].
There is about 40% free water in the dough [bib_ref] The impact of salt reduction in bread: A review, Belz [/bib_ref]. Water absorption is one of the most important parameters determining the stability of inulin. It depends on the protein content, the presence or absence of gluten, the degree of starch damage (starch content), and the amount of inulin added and its properties. Also, water absorption is effective in determining the thermal sensitivity of inulin during the baking process. [bib_ref] The impact of salt reduction in bread: A review, Belz [/bib_ref] identified that increasing the amount of free water at temperatures above 135°C causes thermal degradation of long-chain polysaccharides [bib_ref] The impact of salt reduction in bread: A review, Belz [/bib_ref]. As the most important component of flour, protein effectively increases the quality of the final product and the stability of inulin by improving the water absorption power, the viscosity of the dough, and the ability to retain the fermented gas [bib_ref] Texture and other physicochemical properties of whole rice bread, Kadan [/bib_ref]. Water absorption and inulin retention are also influenced by the source and extent of starch damage, as well as the quantity and quality of protein and flour particle size (the smaller the particle size, the more water is absorbed and the firmer the dough). Approximately 80%
## F i g u r e 2 factors affecting the performance of inulin on bread staling
Fermentation process Formulation components
## Degree of inulin polymerization
Baking temperature
## Inulin replacement level
Water absorption and moisture of wheat bread is starch, which plays a crucial role in its structure [bib_ref] A review on the gluten-free diet: Technological and nutritional challenges, El Khoury [/bib_ref]. Starch is mainly linear, but amylopectin, which is branched (1-6), causes swelling or water absorption. Hence, starches with a high concentration of amylopectin will absorb more water compared with amylose. Furthermore, damaged starch absorbs more water than intact one; consequently, it can affect the inulin stability [bib_ref] A review on the gluten-free diet: Technological and nutritional challenges, El Khoury [/bib_ref]. Since the increase in pH affects the stability of inulin, the addition of salt together with the strengthening of the gluten network [bib_ref] The impacts of salt reduction strategies on technological characteristics of wheat bread:..., Pashaei [/bib_ref] can reduce the water activity and the activity of microorganisms in the dough pH, especially yeast and lactic acid bacteria, and increase the stability of inulin. In addition, salt reduction may increase yeast activity, which may lead to increased consumption of free sugars and decreased stability of inulin. Therefore, all of the above may affect the stability of inulin during processing.
## | fermentation process
One of the stages of bread preparation is the leavening of the dough, which can be performed by fermentation and mechanical or chemical agents. During bread preparation, fermentation is not meant to ensure longer shelf life, but to convert cereal flour into a usable and consumable product [bib_ref] Update in bread fermentation by lactic acid bacteria, Rollán [/bib_ref]. In other words, during the fermentation, yeasts in the dough use carbohydrates to produce carbon dioxide and alcohol and synthesize acids and aromatics that improve the taste, aroma, and shelf life of the bread [bib_ref] Inulin enrichment of gluten free breads: Interaction between inulin and yeast, Morreale [/bib_ref]. Generally, the fermentation process can be carried out by two methods: Saccharomyces cerevisiae yeast and fermentation with sourdough. Only Saccharomyces cerevisiae yeast is used in the yeast fermentation process, but sourdough is a mixture of water and flour fermented by lactic acid bacteria and yeasts [bib_ref] Inulin enriched wheat bread: Interaction of polymerization degree and fermentation type, Mohammadi [/bib_ref]. In comparison, yeast-based bread has a pH of about 5-6, while sourdough bread has about 4 because of the lactic acid bacteria it contains. Various fermentation processes can alter the stability of inulin by affecting the pH and thus the optimal conditions for enzyme activity.
Despite several benefits of fermentation, previous studies have shown the destructive effects of fermentation on inulin stability. [bib_ref] Inulin enriched wheat bread: Interaction of polymerization degree and fermentation type, Mohammadi [/bib_ref] documented that inulin stability was lower in sourdough fermentation than in yeast due to longer fermentation time and higher enzymatic activity [bib_ref] Inulin enriched wheat bread: Interaction of polymerization degree and fermentation type, Mohammadi [/bib_ref]. Also, because the yeast invertase enzyme is effective in degrading inulintype fructans, it was reported that gluten-free bread fermented with yeast with low invertase activity had higher stability of inulin than gluten-free bread fermented with yeast with normal yeast invertase activity. Moreover, the study results by [bib_ref] The effect of pH, temperature and heating time on inulin chemical stability, Glibowski [/bib_ref] reported that the stability of inulin in acidic products (pH ≤ 4) decreased at temperatures above 60°C, while in products with a pH ≥5, there was no inulin degradation even at 100°C [bib_ref] The effect of pH, temperature and heating time on inulin chemical stability, Glibowski [/bib_ref]. It has also been demonstrated that the prebiotic activity of inulin is stable for 24 h at low pH (3-6) [bib_ref] Effect of thermochemical treatment on the structure of inulin and its gelling..., Glibowski [/bib_ref]. However, when it was heat treated (85°C) for 30 min, its prebiotic activity decreased [bib_ref] Effect of processing conditions on the prebiotic activity of commercial prebiotics, Huebner [/bib_ref]. Since heating at low pH is associated with a decrease in prebiotic activity, inulin addition to acidic products is limited, especially when the process temperature exceeds 60°C.
## | baking process
Baking bread is a physical, biochemical, and microbiological process that involves putting the sticky, spongy, and inedible dough in the oven and creating a combination with a unique flavor or bread [bib_ref] Bread baking-A review, Mondal [/bib_ref]. The most apparent changes in this period are volume expansion, crust formation, inactivation of yeasts and enzymes, protein coagulation, partial gelatinization of starch, and moisture loss [bib_ref] The impact of salt reduction in bread: A review, Belz [/bib_ref]. Baking time and temperature affect the stability of inulin; at the same baking temperature, shortchain inulin is more vulnerable than long-chain inulin [bib_ref] Effect of thermochemical treatment on the structure of inulin and its gelling..., Glibowski [/bib_ref]. The study by [bib_ref] Inulin enriched wheat bread: Interaction of polymerization degree and fermentation type, Mohammadi [/bib_ref] Baking even increase after processing. As reported in the study by [bib_ref] Effect of dry heated inulin on selected intestinal bacteria, Bohm [/bib_ref] , dry heat (135-195°C) for 30 min leads to the destruction of 20%-100% of fructan chains and the production of new lowmolecular-weight products (Di-D-fructose dianhydrides). Moreover, its addition to the fecal culture medium significantly stimulated the growth of Bifidobacterium and Enterobacteriaceae and reduced considerably the growth of potential pathogens.
One of the reactions that occur on a large scale during bread baking is the Millard reaction, which is a nonenzymatic reaction between amino acids and reducing sugars at temperatures above 110°C. During this mechanism, increasing the surface temperature of the product causes the evaporation of unbonded water (free water), decreases the moisture content, reduces the water activity of the crust, and increases the Millard reaction rate [bib_ref] Effect of processing conditions on the prebiotic activity of commercial prebiotics, Huebner [/bib_ref].
Furthermore, factors affecting the decrease in pH (including the type of fermentation and the browning reaction associated with the loss of the amino group) can influence the Millard reaction rate. A possible explanation for the effect of inulin on the Millard reaction is reducing sugars, which accelerates the Millard reaction by decreasing the amount of free water. The Millard reaction may be favored when low-molecular-weight fructans are hydrolyzed during processing, thus increasing the amount of free sugar (especially fructose). Thereby, the Millard reaction is accelerated by inulin in formulations. Millard reaction can also reduce the prebiotic activity of carbohydrates. Several studies have demonstrated that when inulin is converted into short-chain oligosaccharides or monosaccharides, its prebiotic function diminishes and brown color is formed during the reaction [bib_ref] Effect of processing conditions on the prebiotic activity of commercial prebiotics, Huebner [/bib_ref].
## | con clus ion
Overall, adding inulin to bread improved its technological properties.
The purpose of this study was to examine the technological characteristics of bread enriched with inulin and to analyze factors such as the DP of inulin and the inulin type that affect the stability of inulin to improve its nutritional performance and health benefits. The effects of inulin replacement in bread formulation were analyzed such as textural and color, specific volume, water absorption, and staling pointed. In addition, several vital factors in the enhancement of these effects were studied, including the type of bread, the DP of inulin, and the type of fermentation. Inulin-type fructans reduce the staling rate and maintain the freshness of gluten-free and gluten-containing bread due to their hydrophilic properties. Furthermore, the effect of formulation components, fermentation, and baking process on the stability of inulin revealed that short-chain inulin had less stability than long-chain inulin. This could be attributed to the presence of simple sugars (which are more accessible to yeasts) as well as low pH. On the other side, due to the high content of reducing sugars in the structure of inulin, short-chain inulin increases the Millard reaction rate and forms a more attractive crust color. Overall, it can be concluded that inulin is almost stable under acidic conditions, but is highly susceptible to degradation by heating at low pH levels.
## Ack n owled g m ents
Fasa University of Medical Sciences supported this study, and so is gratefully acknowledged.
## Fu n d i n g i n fo r m ati o n
Fasa University of Medical Sciences supported this study, grant no.
400148.
## Co n fli c t o f i nte r e s t
The authors declare that they have no conflict of interest.
## Data ava i l a b i l i t y s tat e m e n t
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.
## Co n s e nt fo r pu b li c ati o n
Not applicable.
## E th i c s a pprova l
This study was approved by the Fasa University of Medical Sciences.
Approval ID: IR.FUMS.1400.115.
## O rci d
Neda Mollakhalili-Meybodi https://orcid. org/0000-0001-8473-7517
Amene Nematollahi https://orcid.org/0000-0001-8643-7745
## R e fe r e n c e s
|
Current data and strategy in glioblastoma multiforme
Glioblastoma multiforme (GBM) or astrocytoma grade IV on WHO classification is the most aggressive and the most frequent of all primary brain tumors. Glioblastoma is multiforme, resistant to therapeutic interventions illustrating the heterogeneity exhibited by this tumor in its every aspect, including clinical presentation, pathology, genetic signature.Current data and treatment strategy in GBM are presented focusing on basic science data and key clinical aspects like surgery, including personal experience; adjuvant modalities: radiotherapy, chemotherapy, but also for experimental approaches. Therapeutic attitude in recurrent GBM is also widely discussed.Key words: gliob la st oma mu lt iform e ( GBM), surg ical str ate gy, che mot herap y, exp erim ent al appr oa che s, recurren cy t reatme nt
# Introduction
Malignant brain tumors are among the cancers with the most dismal prognosis and are additionally one of the most expensive cancers to treat [bib_ref] In vivo tracking of neural progenitor cell migration to glioblastomas, Tang [/bib_ref]. The main culprit for adult CNS neoplasm is metastasis from other sites (especially lung, GI tract, breast cancer). Distinct from these, one encounters several types of primary brain malignancies, with tumors arising from glial cells (gliomas) being the most common. Glioblastoma multiforme (GBM) or grade 4 astrocytoma by the classification of World Health Organization is the most aggressive and frequent of all primary brain tumors [bib_ref] Recent advances in the molecular genetics of primary gliomas, Kitange [/bib_ref] ; the term "multiforme" illustrate the heterogeneity exhibited by this tumor in its every aspect, including clinical presentation, pathology, genetic signature and response to different therapies.
## Epidemiology
GBM accounts for 12-15% of all brain tumors and 50-60% of astrocytomas. The incidence of GBM is less than 10 to 100,000, but the median survival of a little over 1 year from diagnosis makes it a considerable public health issue. Interestingly, the incidence is fairly constant worldwide, leading to the logical inference that environmental, geographical and nutritional factors probably don't play a major role in this particular cancer, where genetics is more likely to tip the scale of etiology. The peak incidence is between 45 and 70 years, with a crest to 58 years; only 8.8% of children with central nervous system tumors had GBM [bib_ref] Glioblastoma multiforme in children, Dohrmann [/bib_ref] and congenital cases are extremely rare. A troubling observation, GBM incidence is increasing in all age groups, especially in adults, which cannot be explained by the aging population, better imaging techniques or earlier detection; no familial predisposition was found. It is slightly more common in men, with a male-to female ratio of 1.5:1 -the reasons for this gender distribution are yet unknown. Blacks are somewhat protected, as their incidence is lower than other ethnic groups like whites, latinos and Asians. A study has found GBM patients are generally born in winter, particularly in the month of February (40%); a perinatal viral origin (infection with an unknown oncovirus, integrating in the genome) has subsequently been put forward, but no hard evidence has yet been found to support it.. The association with an oncogenic virus remains controversial; anecdotically, SV40 or cytomegalovirus have been involved. Most GBM appear to be sporadic, although several genetic disorders are associated with increased incidence, including tuberous sclerosis, neurofibromatosis type 1 and type 2, von Hippel Lindau disease, Turcot and Li-Fraumeni syndromes. There is a proven association between GBM and exposure to ionizing radiation or polyvinyl chloride (a polymer commonly used in construction), but no links have been found between GBM and smoking, diet, cellular phones or electromagnetic fields .
## Historical annotations
1863 -Virchow, cited bywas the first one to recognize the glial origin of this tumor. 1888 -Bramwell is pointing out the surgical dilemma created by the highly infiltrative nature of GBM: "the tumor tissue is never limited by a capsule and it is impossible to say without microscopical examination where the tumor tissue ceases and the normal brain tissue begins". 1914 -Mallory [7] coined the term "glioblastoma multiforme". 1926 -Bailey and Cushing definitively changed the name from spongioblastoma multiforme to GBM, term also preferred by Zulch, Russel and Rubinstein. 1940 -Scherer and Kernohan recognized GBM sometimes develops by progression from a lower grade lesion; the idea was later on sustained by molecular studies showing a characteristic sequential accumulation of genetic alterations from diffuse astrocytoma (WHO grade II) to anaplastic (grade III) and to GBM. Most cases, however, are thought to arise de novo. Scherer [8] has also paid due attention to the migration of tumor cells away from the main tumor mass through the brain parenchyma, describing the "secondary structures of Scherer".
## Clinical considerations
Occasionally, GBM is asymptomatic until it reaches an enormous size. History is generally short, less than a few months and not uncommonly symptoms begin abruptly; in more than 50% of cases, this is explained by the rapid development of increased intracranial pressure, with compression and infiltration of the surrounding brain. Physical findings depend on the location, size and rate of growth of the tumor, as with any other CNS tumor: headache, partial or generalized seizures in 30-60% of patients, contralateral slowly progressive hemiparesis, sudden onset of hemiplegia with depressed level of consciousness -as a stroke in evolution [9], generated by intratumoral bleeding, directly correlated with the level of expression of vascular endothelial growth factor (VEGF) [10]; sensory disturbances, subtle personality changes, visual field defects. As survival gets longer, an increasing number of patients experience: cognitive problems, neurologic deficits resulting from radiation necrosis, communicating hydrocephalus, occasionally cranial neuropathies, polyradiculopathies from leptomeningeal spread.
## Genetics
Overexpression of several oncogenes and mutations leading to loss of function of key tumor suppressor genes concur to create one of the most aggressive cancers. Amplification of oncogenes such as the epidermal growth factor receptor (EGFR), as well as loss of tumor suppressors like PTEN on chromosome 10, p53 on chromosome 17, or p16/INK4A are some of the most common genetic alterations in GBM . Adult malignant astrocytoma was one of the first cancers shown to have frequent TP53 mutations . The p53 disfunction disrupts the downstream p14 ARF pathway, impedes the process of apoptosis and fosters further genomic instability. Most commonly, the mutations occur with loss of the remaining wild-type allele, such that the tumor cells express only aberrant p53 . This distribution of mutations is highly similar to those indicated in other human cancers and in which codons 175, 248, 273 are revealed as the most common targets of alteration, contributing substantially to the overall high incidence of alterations in p53's DNA binding domain. GBMs are often classified with respect to their clinical history: either as primary GBM, without indication of association with a lower malignancy precursor or as secondary GBM, having evolved from a lower grade astrocytoma. For the latter, which constitute <10% of all GBM, TP53 mutations are frequent and occur in nearly two-thirds of the cases . For the more common primary GBM, only 25-30% of the cases have p53 mutations. In total and without subclassification, approximately two thirds of all GBM have wild-type p53 status. The clinical division in primary and secondary GBM is also relevant from a basic science point of view, as different genetic events are incriminated in their development. For primary GBM, the main factors are considered to be EGFR amplification, loss of PTEN and INK4A [16]; all occurring almost simultaneously or in a brief period of genomic instability, correlating with a rapid clinical history, often shorter than 3 months. For secondary GBM, there is a distinct succession of genetic events and the course is longer: initially, activation of signal transduction pathways, usually through overexpression of PDGF, PDGFR, FGF2 (instead of EGFR); p53 mutation, CDK4 amplification, Rb (retinoblastoma gene) loss and eventually, loss of PTEN unleashes the Akt pathway, leading to downstream activation of mTOR (mammalian target of rapamycin) and other genes (NF-kB, an important transcription factor promoting proliferation; GSK3; anti-apoptotic genes such as BAD etc.) with devastating consequences . The main way of activating Ras in GBM is through EGFR (either overexpressed and stimulated by exogenous growth factors, or constitutively active [17] as is the case for certain mutant variants, like the truncated EGFRvIII) but binding of different growth factors to other tyrosine kinase receptors is another important upstream event. For the rare pediatric GBM, it is believed that EGFR signaling is revived up by overexpression of YB-1 (Y-Box binding protein 1). Things, unfortunately, are incredibly more complex than the deregulation of a few signaling pathways. For instance, differential recruitment of mRNAs to polysomes (thus optimizing protein synthesis) contributes to glioma formation, being an important player in the oncogenic Ras and Akt signaling . Such observations lead to cell-existential questions like: transcriptional control versus translational control, which is more essential to cell transformation? Does aberrant regulation of translation lead to cancer? Would it be a cause or consequence of cancer progression? Pathology GBM is preferentially located supratentorially -fewer than 50 cerebellar cases have been described in literature, in the cerebral hemispheres, more often in the frontal lobes than the temporal lobes or basal ganglia (although a combined fronto-temporal mass is particularly typical of GBM). GBM of the brainstem (malignant brain stem glioma) are quite infrequent and often affect children , while spinal cord is a rare site for this neoplasm -if it occurs, the cervical and thoracic segments are most likely to be affected; when concomitant with a brain mass, some possibly develop as a "drop pseudo-metastasis" from tumor cells that invaded the ventricular system. It usually presents as an irregular mass in the white matter, infiltrating the surrounding parenchyma by coursing along white matter tracts. Although tumor cells are considered to be already disseminated at time of diagnosis far in the surrounding parenchyma, it is generally a single mass, while true multifocal glioblastoma usually have distinct histological appearance and are most likely policlonal, usually presenting as simultaneous infra and supratentorial masses -incidence of 2,5% [20, 21]. Most GBMs are intraparenchymal with an epicenter in the white matter, some are largely superficial, in contact with the leptomeninges and dura, but without invading the subarachnoid space. Tipically unilateral, involving much of an entire lobe, not uncommonly it involves the corpus callosum and crosses the midline to produce the characteristic "butterfly" appearance of bilateral invasion.
Grossly, it appears topographically diffuse, a poorly delineated mass with no capsula, with prominent areas of old and recent hemorrhage (extensive areas of yellowishbrown to red discoloration) and necrosis (as much as 80% of the total tumor mass), cystic areas sometimes alternating with firm tissue. Microscopically, characteristic histopathological features include: cellular and nuclear polymorphism, nuclear atypia, high mitotic activity, vascular trombosis, microvascular proliferation and necrosis, with regions of pseudopalisading (created by viable tumor cells bordering areas of necrosis). Among the above, the presence of necrosis within the tumor is considered crucial, showing the aggressiveness of the cells (they outgrow the blood supply, regardless of their angiogenic potential) and it is a sine-qua-non diagnostic criterion to histopathologically "upgrade" an anaplastic astrocytoma to GBM. Another important diagnostic clue is the presence of secondary structures such as perineuronal and perivascular "satellitosis" -a result from the interaction of glioma cells with host brain structures, especially in certain regions, like the subpial zone of the cortex, in the subependymal region, and through the white matter tracts (intrafascicular spread, considered a "favorite" route of migration and invasion for the tumor cells). The cell composition is heterogenous (again -"multiforme"), especially for GBM resulting from progression of diffuse astrocytomas (grade II): fusiform, round, pleomorphic (small, undiferentiated, lipidized, granular, giant) astrocytes -may reflect the emergence of a new tumor phenotype through stepwise acquisition of genetic alterations. Histologic tumor variants which do not alter the prognosis of the tumor (with the exception of gliomatosis who has no surgical management and a more accelerated course), include: giant cell glioblastoma, gliosarcoma and gliomatosis cerebri. Extension within and along perivascular spaces are common, but invasion of the vessel lumen seems to occur infrequently, correlating with a very low incidence of haematogenous spread to extraneural tissues. Metastatic spread of GBM occurs in less than 5% of cases, late in the illness course and it was almost unheard of before the adjuvant therapy. Dissemination within the CSF pathways is so rare that it does not seem to justify the use of prophylactic irradiation of the spinal cord. In practice, GBM metastasis is considered an exceptional event. Local dissemination into subcutaneous tissues of the scalp, face and neck can occur if dura is not closed at the time of surgery. Penetration of dura is possible, although not frequent, while invasion of venous sinuses and bone is exceptional; peritoneal metastasis via ventriculoperitoneal shunt is also a possibility.
## Imaging
MRI with and without contrast is the most sensitive and specific study, particularly useful in evaluating tumor extension and subacute and chronic hemorrhage collections. On T 1 -weighted images, the mass has lowsignal intensity, presenting as a central hypodensity surrounded by a thick enhancing rim of tumor with thick, irregular walls, corresponding to the cellular and highly vascularized peripheral area of the neoplasm. Marked gadolinium enhancement indicates angiogenesis and vascular permeability. The tendency to infiltrate along the white matter tracts and involvement or crossing of the corpus callosum can also be observed. On T 2 -weighted images, it appears as a high-signal intensity mass, however the area is broader, less well defined and overlaps with the surrounding vasogenic edema; this imaging modality is best at revealing the perilesional edema. The CT scan appearance is variable and less informative and cannot replace the MRI as study of choice despite advantages of cost and time. It typically shows irregularly shaped lesions with peripheral ring-like zone of contrast enhancement around a dark, central area of necrosis, usually inhomogeneous. Surrounding edema has a hypodense or isodense appearance. Functional imaging, such as positron-emission tomography (PET scan), single-photon emission computed tomography or MR spectroscopy, although cost-prohibitive for routine clinical practice, may provide useful information. The level of regional consumption of radioactively labeled glucose (used in PET scanning) correlates with cellular metabolism and reduced survival. These techniques may help differentiate the tumor from other benign mass lesions, brain abscess, toxoplasmosis and help refine the differential diagnosis between treatment-related radiation necrosis versus tumor recurrence, so difficult to distinguish in MRI or otherwise (e.g. levels of choline and lipids are different on MR spectroscopy). They may also be helpful to define the margins of the tumor for surgical resection, planning for the radiation fields and to improve the diagnostic accuracy in case a small biopsy sample is taken.
## Differential diagnosis
Other primary brain tumors (meningioma, metastasis, astrocytoma, oligodendroglioma); features making GBM a more likely diagnosis include the irregular shape of the mass, the central necrosis, the extensive surrounding edema and mass effect. Metastasis should always be considered; it is less likely when the patient is realtively young and has no history of a primary tumor. Brain abscess should also be mentioned but it is typically identifiable as a thin, regular rim of enhancement around a central cavity. Other infections (toxoplasmosis, cysticercosis) may also have a close radiologic appearance. Primary CNS lymphoma may be, occasionally, "butterfly-shaped" involving the corpus callosum. Multiple sclerosis lesions -"concentric sclerosis of Balo", a borderline rare form of multiple sclerosis may be difficult to separate from GBM by clinical presentation and radiologic appearance.
## Treatment
Several factors concur to make GBM treatment notoriously difficult. First, the tumor cells themselves, despite their relatively rapid cycle, are quite resistant to conventional therapies. In addition, brain has a limited capacity to repair itself, any damage may be definitive and consequential. Last but not least, before the advent of temozolomide (TMZ), adequate penetration of the bloodbrain barrier (BBB) by chemotherapeutics could not be achieved without dose-limiting systemic side effects. The mainstay of therapy consists of surgery, radiation and chemotherapy. Objectives of surgery range from merely confirming the diagnosis (biopsy), to alleviating symptoms of mass effect and ICP (debulking or cytoreductive surgery, resecting as much as it is safe without worsening patient's neurologic deficits), to aggressive attempts to improve not only the quality of life, but also influence survival significantly. In addition to tumor-targeted therapy, one has to treat several associated phenomena.
Peritumoral edema may respond to a potent corticosteroid (Dexamethasone) given 4 to 10 mg every 4 to 6 h, diminishing mass effect and lowering intracranial pressure, with a decrease in headache and drowsiness.
Seizures treatment is required to only 40% of patients. An appropriate anticonvulsant, with minimal side effects and cytochrome P450 interference (enzyme inducers can increase the metabolism and clearance of some chemotherapeutic agents), should first be tried as monotherapy.
Prevention of thromboembolic disease is a major concern for patients with GBM, as the incidence has been reported to be as high as 35-40%. Prophylactic use of anticoagulation has not been recommended because of increased risk of intracranial hemorrhage; alternatives include appropriate mobilization and physical therapy, calf protection such as SCDs (sequential compressive devices) and radio-interventional placement of an inferior vena cava filter (Greenfield ® filters).
Occupational, speech therapy, emotional and psychological support are also important, especially as the emphasis shifts to palliative and supportive care (a point reached, unfortunately, in the evolution of a majority of GBM patients).
## Surgery
Bennett and Godlee are credited with the first successful removal of a glial tumor (1884), cited by Iacob. The extent of surgical resection depends on location and eloquence of the brain areas involved, but surgery is always an incomplete debulking, since GBM is a highly infiltrating tumor and cannot be resected completely. In a seminal study by Wilson [23], the percentage of tumor cells in the entire cell population was quantified as a function of distance from the "visible" tumor edge and the averages were found to be 6% at 0-2 cm away (hence, the margin considered for "radical" resection should not be less than 2 cm) and more troubling, 1.8% for 2-4 cm and 0.2% at more than 4 cm away (e.g., in the contralateral hemisphere). Whether aggressive, "radical" surgery prolongs survival is still debatable, but several studies suggest a close inverse correlation between survival and the amount of residual tumor observed on postoperative MRI scans . Partisans of radical resection maintain several advantages, such as: good relief of ICP, reversal of some neurologic deficits, lowering seizure incidence or even abolishing them, a definitive pathology diagnosis by reducing sampling error and the assumption that a "more cytoreductive" surgery may facilitate adjuvant treatment modalities and ultimately improve survival. Arguments against radical resection stem from the inherent invasiveness of GBM, which cannot be totally resected anyway; in addition, there might be a potential for facilitating tumor cells migration by the act of surgery and the possibility of surgical complications, new neurological deficits (thinking to "primum non nocere", "first, do no harm"). If pursued, radical resection may be improved by careful pre-operative planning, use of intraoperative MRI or at least 3D -image guidance for tumor delineation and electrophysiological mapping to help preserve eloquent areas; also, use of 5aminolevulinic acid (ALA) for influencing fluoresceineguided resections has recently been reported to increase survival (median of 17.7 months vs. 12.9 months) [25]. The routine use of robot arms has not yet been proven to influence survival.
## Radiation therapy
Early clinical trials revealed a modest, yet undeniable efficacy of external beam radiotherapy (RT) in treating GBM, based on the damage of ionizing radiation in the DNA helix by electrons and free radicals. RT is usually started within 4 to 6 weeks after surgery [26] and administered in a standard fractionated regimen over 6 to 7 weeks [27, 28]. The standard dose of external beam RT is 60 Gy in single daily fractions of 1.7-2 Gy, 5 times a week, applied to a limited field that includes the enhancing volume on CT scans with a 2-3 cm margin or 2 cm margin beyond Flair T 2 -weighted MR images. Whole brain RT does not improve survival when compared to the more precise and targeted three-dimensional conformal RT. Targeted variants of RT have been developed that are so focused, that they are closer to surgery -hence their generic name, radiosurgery. The Gamma knife has been introduced half a century ago by Lars Leksell and is now at its 4 th generation, the Leksell Perfexion (introduced in clinical practice in 2007), boasting 201 sources of 60 Co. These devices triangulate hundreds of gamma rays (of low individual energy) in a single spot, so "ground zero" receives a very high dose of radiation, with a sharp decline in exposure for the surrounding tissue. The Cyberknife and LINAC are other highly focused radiosurgery systems. This technique is generally used to treat small (<3 cm), radiographically well-defined lesions in surgically inaccessible or eloquent areas of the brain or in patients with serious coexisting medical illnesses, unsuitable for open resection. Growth of the tumor during the course of radiation is an extremely poor prognostic sign. A major reason for radioresistance is the significantly lower oxygenation of tumors compared to surrounding cortex . This may be quantified by measuring the level of a transcription factor, hypoxiainducible factor-1 alpha , a hypoxic sensor that is also an ominous indicator of tumor radioresistance. Stereotactic brachytherapy involves using stereotactic techniques to accurately place radioactive isotopes: 125 I, Californium within the tumor. Typically, brachytherapy delivers an additional 50-60 Gy of radiation, bringing the total dose of radiation up to 110-120 Gy. This is indicated in patients with unifocal, well-defined, supratentorial tumors less than 5 cm in diameter that do not involve the corpus callosum, brain stem or ependymal surfaces; currently as a salvage modality, in recurrent GBM after repeat resection of the tumor. Recently, instead of solid seeds, a variant has been developed where a temporary balloon catheter filled with liquid radiation is implantedthe GliaSite RTS (radiotherapy system). The device is an inflatable, silicone, balloon catheter that is inserted in the resection cavity and filled with an aqueous solution of organically-bound 125 I (Iotrex), which delivers 40-60 Gy at 0.5-1 cm from the balloon surface over a course of 3-7 days. Another adjuvant modality, thermotherapy, involves microwave antennas operating at 2450 or 915 MHz to deliver heat for one hour at 450C before and after 3-5 days of interstitial brachytherapy. Boron neutron capture therapy (BNCT) is an experimental form of RT where the damage occurs through the interaction between a beam of thermally slowed neutrons (created in a mini-nuclear reactor) with boron -10, which is injected to a patient and preferentially binds to tumor cells. Still investigational and costly, this is a treatment modality with great promise: Hatanaka [30] reported a 5-year survival rate of 50% with few complications, after combining intra-arterial polyhedral boron anion with focused thermal neutron radiation. The mean survival after optimal surgery and adjuvant RT (60 Gy) is about a year: 12.1 months in one of the latest studies . There are several limitations to RT that make adjuvant chemotherapy a must: the infiltrative nature of GBM, the risk of radiation necrosis and radiation-induced permanent neuronal damage (e.g. radiation encephalopathy), as well as the radio-resistance of some tumors, intrinsic or acquired.
## Chemotherapy
In attempts to further improve survival beyond that offered by RT, many chemotherapeutics have been tested for effectiveness in the treatment of GBM. Among these, alkylating agents have demonstrated some benefit; either chloroethylating drugs like carmustine (BCNU), lomustine (CCNU) or methylating agents like temozolomide (TMZ), are used in the majority of GBM clinical protocols [31]. The chloroethylating agents readily penetrate the bloodbrain barrier due to high hydrophobicity and they can also be administered orally . They act by forming O 6chloroethylguanine lesions, which lead to G-C interstrand crosslinks [33], after a mean time of only 10 hours (8-12h) . It is thought that as few as 2 to 5 such lesions (interstrand crosslinks) can trigger apoptosis in a tumor, as well as a normal, cell . However, an aggressive regimen with these agents causes considerable side effects [37], leading to dose reductions and corresponding decreases in therapeutic efficacy. Methylating agents such as TMZ show reduced toxicity toward normal cells and are much better tolerated. Oral administration of TMZ, either concomitant with radiotherapy, followed by adjuvant TMZ or as adjuvant TMZ alone after completion of radiation, is increasingly becoming standard of care for GBM patients, at least in those countries that can afford the high cost of TMZ therapy . The use of TMZ has been significantly increased as a result of a phase III trial that showed survival advantage to newly diagnosed GBM patients receiving TMZ with standard radiotherapy . Regarding its mechanism of action, O 6methylguanine (O 6 -mG) is perhaps the most biologically relevant lesion generated by TMZ; consequently, TMZ therapy has little effect in tumors where the added methyl group is removed due to the intact enzymatic activity of O 6 -methylguanine-DNA-methyltransferase (MGMT) [31, 39-41]. The importance of MGMT epigenetic silencing through promoter methylation was shown as a favorable prognostic factor for improved response to TMZ in a multiinstitutional study of GBM patients . However, additional analysis of the results of this study showed that 10% of the patients having tumors with nonmethylated MGMT survived more than 2 years, which is considered long-term survival for GBM patients. This observation indicates that MGMT methylation status is not the sole predictor of GBM response to TMZ and this drug of choice should not be withheld from any GBM patient, unless enrolled in investigational trials. Second-line cytotoxic agents, for patients who do not respond to the first-line drugs discussed above, include carboplatin, etoposide, oxaliplatin, and irinotecan. Sometimes, procarbazine and vincristine may be added to CCNU (lomustine) as the PCV regimen, which used to be a first line approach before the supremacy of TMZ. A few other chemotherapy approaches warrant a brief mention.
Gliadel is a small wafer that contains the chemotherapeutic drug carmustine (BCNU) and a biodegradable polyanhydride copolymer. Up to 8 Gliadel wafers are implanted in the resection cavity, designed to release BCNU slowly over a period of 2-3 weeks. Clinical trials results have shown that Gliadel prolonged survival in a statistically significant (albeit modest) manner in both newly diagnosed patients and patients with recurrent glioblastoma multiforme when used as adjunctive therapy to surgery and/or radiation therapy -13,9 months compared to placebo implants -11,6 months [42]. Nota bene, Gliadel wafers may have terrible side-effects: seizures, brain edema, wound healing problems, intracranial infections, delay in the administration of other drugs. Chronic administration of chloroquine -150-300 mg dose of chloroquine daily starting 1 day after surgerygreatly enhanced the response of GBM to antineoplastic treatment. Because the cytotoxicity of chloroquine on malignant cells is negligible, these favorable results appear mediated by its strong antimutagenic effect that precludes the appearance of resistant clones during radiotherapy and chemotherapy . Other experimental approaches to the chemotherapy of GBM include: antiangiogenic agents like anti-VEGF monoclonal antibodies, e.g. Bevacizumab (Avastin), anti-FGF antibodies, monoclonal antibodies targeting EGFR like Erlotinib and Gefitinib; inhibitors of other tyrosine kinase receptors (e.g. PDGFR); inhibitors of kinesin Eg5; mTOR inhibitors like: Everolimus, Sirolimus, Temsirolimus; farnesyl-transferase inhibitors like: Tipifarnib, Lonafarnib. A promising avenue of individualized research is an antitumoral vaccine made from specific proteins isolated from glioblastoma cells following resection, in order to generate a specific immune response to that particular tumor [44].
## Authors' experience
We analyzed data from 118 patients with a final histopathological diagnosis of GBM, treated in the Neurosurgery Clinic of the University Hospital, Bucharest, in the last 6 years . Patients received either surgery or biopsy and adjuvant therapy in the form of radiotherapy, chemotherapy, following standard protocols. Overall, the median survival was 49 weeks. There was a significant difference in survival between the 73 patients who have undergone resection (57 weeks) and the 45 patients who were only biopsied (38 weeks) (p<0.01). Obviously, this difference is at least partly due to selection bias, as patients with multifocal and large bilateral tumors ("butterfly GBM") who presented in an advanced consumptive state (poor Karnofsky scores) or who were not good candidates for extensive open surgery due to advanced age and comorbid disorders, were more often offered biopsy. presents prognostic factors and survival in our cohort as a function of age, confirming that, contrary to many other malignancies, younger is better for GBM patients.
## Management of recurrent gbm
Semantically, there are two distinct clinical entities: 1) Tumor recurrence: after complete surgical removal of the visible mass, documented radiologically (MRI), the tumor re-appears after a variable free interval, usually within 2 cm of the original site (correlate with [23]), although 10% of patients may develop new lesions at distant sites.
2) Tumor progression: after a subtotal excision, the mass never disappears completely from imaging studies and after a while there is radiological documentation of an increase in tumor size.
There are many factors favoring recurrence: subtotal excision (either erroneously or intentional, in order to protect eloquent areas); tumor multicentricity; histopathological features of aggressiveness (extensive angiogenesis, high mitotic index etc.), genetic features (deregulated genome with many mutations accumulated); clinical factors (e.g. Karnofsky score, age, comorbidities, other factors that may influence administration of adjuvant therapy) and perhaps most important, the inherent infiltrative nature of GBM -at the time of diagnosis, there are already malignant cells disseminated at a distance from the bulk of the tumor, mocking the scalpel. Recurrence usually means the tumor is becoming more aggressive, genetically and clinically and it has acquired resistance to the adjuvant therapy. For the patients and their families, it is important to do something to offer a longer life expectancy, maintaining at the same time a good quality of life for a reasonable period, without supplementary neurological deficits. For the society, the most embarrassing problem is the costs of different procedures: surgery, chemotherapy, radiotherapy. For the neurosurgical community, reoperation for recurrent malignant gliomas is still a difficult dilemma. So what is to be done for these patients? Unfortunately, additional radiation has limited control on further tumor growth and would potentiate neurologic toxicity. It has been shown that prior chemotherapy had significant correlations with time to tumor progression and survival, but further cycles of chemotherapy are problematic in these patients, due to their biological status and the heightened risk of cumulative myelotoxicity (as the recurrence is usually fast, contrary to some other cancers). Radiosurgery may be an option, for tumors of appropriate size, pattern and location. Open surgery is also to be considered, it could be life saving to patients with lesions that produce increased intracranial pressure and mass effect. From a life quality standpoint, to selected patients with recurrent GBM, repeat surgery could be beneficial and a reasonable therapeutic alternative. If the tumor is better circumscribed on recurrence, extensive removal is indicated, as radical as first-time surgery. On the contrary, patients over 60 years old, with bad preoperative function, low performance status and a short symptom free period after the first surgery, have a bad prognosis. Significant peritumoral edema, symptomatic mass effect, location near eloquent areas are other negative prognostic factors for surgery. Timing is also important. If tumor recurrence occurred within 6 months after a radical resection and radiotherapy, progression or recurrence after second surgery are much quicker; in case a year or more has passed after the initial presentation, it would be worthwhile to excise the recurrent tumor.
# Conclusions
The prognosis of GBM is still dismal. Without treatment, the median survival from the diagnosis is 3 months (death is usually due to cerebral edema and increased intracranial pressure). With maximal treatment, median survival is only 14 months, despite major improvements in neuroimaging, neurosurgery, radiation treatment techniques and the advent of temozolomide -"bullets but no magic"- . The survival at 2 years is 16% and at 3-years is 5%. Only approximately one in 5000 GBM patients survives for decades. That being said, patients with GBM are not universally incurable, with an ever increasing albeit small fraction of patients who fight and survive the disease. Progress in improved outcomes for GBM patients will require the identification and development of therapeutics with high specificity for brain tumor cells and that have the ability to access the entire intracranial compartment.
## Age group
|
Adenoviral protein E4orf4 interacts with the polarity protein Par3 to induce nuclear rupture and tumor cell death
The tumor cell-selective killing activity of the adenovirus type 2 early region 4 ORF4 (E4orf4) protein is poorly defined at the molecular level. Here, we show that the tumoricidal effect of E4orf4 is typified by changes in nuclear dynamics that depend on its interaction with the polarity protein Par3 and actomyosin contractility. Mechanistically, E4orf4 induced a high incidence of nuclear bleb formation and repetitive nuclear ruptures, which promoted nuclear efflux of E4orf4 and loss of nuclear integrity. This process was regulated by nucleocytoskeletal connections, Par3 clustering proximal to nuclear lamina folds, and retrograde movement of actin bundles that correlated with nuclear ruptures. Significantly, Par3 also regulated the incidence of spontaneous nuclear ruptures facilitated by the downmodulation of lamins. This work uncovered a novel role for Par3 in controlling the actin-dependent forces acting on the nuclear envelope to remodel nuclear shape, which might be a defining feature of tumor cells that is harnessed by E4orf4.
# Introduction
The discovery of interactions between viral proteins and actin assembly machineries provides crucial information about how actin structure organization is regulated and controls cell behaviors. The adenovirus (Ad) type 2 early region 4 ORF4 (E4orf4) exerts a cell-killing activity that correlates with a dramatic reorganization of actin structures and impairs cell structural integrity. Oncogenic transformation potentiates E4orf4-induced cell death, suggesting that E4orf4 targets tumor-defining features. These features should be broadly conserved in vivo, as E4orf4 can inhibit the development of various aggressive tumors in Drosophila without causing significant damage to healthy tissues. However, the tumor cell-defining features that are targeted by E4orf4 remain elusive.
In cancer cell lines or cells transformed with viral oncogenes, the down-modulation of myosin II, which inhibits E4orf4-induced actin remodeling, also interferes with cell death induction, therefore arguing for a role for cell contractility. Mechanistically, E4orf4 may affect actomyosin contractility through interactions with various signaling pathways that show mutual regulatory mechanisms, including Src kinases, PP2A-and PP1-regulatory subunits, Rho GTPases, and Hippo intermediates. Cellular mechanics are manipulated by E4orf4, but the clear mechanism involved is yet to be defined, as is its relevance for E4orf4-induced tumor cell killing.
Epithelial cancers characteristically deregulate cell mechanics. Cancer cells often show reduced stiffness and generate more contractile forces as a result of both cellular intrinsic oncogenic alterations and microenvironment changessubnetwork organization can contribute to mechanical deregulation and malignant cell behaviors. For instance, polarity protein signaling network rewiring can promote new interactions with the actin assembly machineries that increase cell invasiveness. Additionally, changes in nuclear envelope (NE) protein expression, which modify nuclear shape and rigidity, influence cell metastatic potential. The mechanical forces generated by contractile actin filaments are transferred to the nucleus through physical connections provided by the LINC (linker of nucleoskeleton and cytoskeleton) complex. This dynamic scaffold enables the nucleus to sense and respond to mechanical forces by modifying nuclear organization and gene expression. How abnormal mechanical forces in tumor cells transfer signals to the nucleus is unclear.
In this study, we sought to interrogate the relevance of actomyosin contractility for E4orf4's tumor cell-selective action along with the mechanisms involved. We provide compelling evidence that E4orf4 subverts Par3 polarity protein signaling to regulate its own spatial dynamics and perturb nuclear mechanics in a tumor cell-selective manner. The data suggest that Par3 may act by the transfer of perinuclear actomyosin forces to modulate nuclear and cell shape changes. We infer that Par3 defines an important pathway to control nucleocytoskeletal coupling in tumor cells.
# Results
E4orf4's tumoricidal activity is associated with actomyosin-regulated changes in nuclear shape To determine the relevance of actomyosin changes for E4orf4's tumoricidal effect, toxicity in nontumorigenic epithelial cells and tumorigenic cell lines from various origins was initially characterized. There is currently no clear biochemical readout for E4orf4-induced cell killing, which is p53 and caspase independent in most tumorigenic cell lines tested. E4orf4 also modifies mitochondrial dynamics and could, therefore, impact cell metabolism before cell death. Thus, cytotoxicity would be better addressed by measuring cell proliferation and membrane integrity.
Clonogenic activity in breast cell lines, including nontumorigenic MCF10A cells and metastatic triple-negative MDA-MB-231 cells, was first determined after transduction of cells with recombinant Ads encoding E4orf4 or GFP as a control. MCF10A cells exhibited an approximately two-to fourfold higher survival rate relative to MDA-MB-231 cells after transduction at identical plaque-forming units (PFUs) per cell of Ad-E4orf4, which was associated with similar or higher E4orf4 levels in MCF10A cells, supporting a cancer cell-selective killing effect.
To perform a robust analysis in cell lines from various origins, recombinant Ad-E4orf4-mCherry or Ad-RFP was used to monitor the transduction efficiency microscopically. Trypan blue dye exclusion assays were performed at various times to measure cell proliferation and the percentage of dead cells within a population. Remarkably, E4orf4 had no detectable impact on cell proliferation or viability over a 3-d period in two additional nontumorigenic cell lines, HaCaT and RPE-1 cells, which were efficiently transduced . HaCaT cells are spontaneously immortalized from adult skin keratinocytes and retain full differentiation capacity, while RPE-1 cells are near-diploid human retinal pigment epithelial cells immortalized with human telomerase reverse transcriptase (hTERT;. In marked contrast, E4orf4 reduced cell proliferation in several types of cancer cell lines, including triple-negative breast cancer cells with epithelial or mesenchymal properties , human bone osteosarcoma epithelial cells (U2OS), prostate carcinoma epithelial cells (22Rv1), and cervix carcinoma epithelial cells (HeLa;B, Proliferation). Cancer cell lines also displayed significant increases in cell death rates at all time points, even those displaying lower transduction efficiencies . S1 A, Cell death). These results indicate that tumorigenic cells from various origins are sensitized to E4orf4-induced cell killing.
Next, E4orf4-induced morphological phenotypes and their dependence on actomyosin contractility were analyzed. Nontumorigenic epithelial cells expressing E4orf4-mCherry displayed little changes in actin cytoskeleton organization and no significant increase in cell blebbing or nuclear condensation compared with control cells expressing RFP (Figs. 2 A and S2 B; MCF10A, HaCaT, and RPE-1). In contrast, all cancer cell populations exhibited increased rates of cell blebbing and nuclear condensation in response to E4orf4and, B and C), in line with our previous findings. Both phenotypes were severely blocked by blebbistatin, which downmodulates actomyosin contractility through the inhibition of myosin II. Furthermore, blebbistatin increased the nuclear-to-cytoplasmic ratio of E4orf4 in breast cancer cells (MDA-MB-468) without perturbing E4orf4 localization in nontumorigenic MCF10A cells. These results imply that the regulation of E4orf4's tumoricidal activity is dependent on actomyosin contractility.
It was further observed that E4orf4-responsive cells accumulated robust actin bundles proximal to their nucleus that correlated with pronounced nuclear deformation; compare x-z views). Quantification of nuclear circularity in lamin A/C-labeled cells clearly revealed a tumor cell-selective effect. E4orf4-induced nuclear dysmorphia, visible before the onset of cell blebbing, was characterized by striking disorganization of the nuclear lamina (Figs. 2 D and S1 C). Nuclear dysmorphia was also reduced by cell treatment with blebbistatin. These results argue that actin remodeling may cause nuclear compression. In line with this, E4orf4-expressing cells often accumulated lamin A/C at NE sites that were codistributed with apical actin fibers, forming lamina folds that are typical of mechanical compression. These results indicate that E4orf4 tumoricidal activity correlates with actomyosin-driven changes in nuclear shape.
E4orf4 interacts with Par complex proteins and promotes epithelial cell polarity We then sought to identify E4orf4 partner proteins that could regulate actomyosin dynamics. To this end, affinity purification coupled to mass spectrometry (AP-MS) was performed using breast cancer MDA-MB-231 cells that display high transduction efficiency and cell blebbing rates. Cells were transduced with a control Ad (Ad-LacZ) or an Ad encoding E4orf4 (Ad-Flag-E4orf4), and E4orf4 immune complexes were purified and processed for liquid chromatography coupled to tandem mass spectrometry (LC-MS/MS). We combined the data from three different experiments in order to perform a significance analysis of interactome (SAINT) using a stringent threshold of 0.98. Among five high-confidence interactions, which include previously identified interactors (PP2A subunits;, we selected Par3 (PARD3;for functional analyses, based on the role of Par polarity proteins in the spatial control of actin dynamics. While Par3 was previously reported as an E4orf4 partner, the relevance of this interaction has not been addressed.
Coimmunoprecipitation (IP) studies revealed that E4orf4 interacts with the three isoforms of Par3 in various tumorigenic cell lines, along with other Par complex proteins (Par6 and PKCɩ) that are recruited by the Par3 signaling scaffold, B and C, arrows; and Figs. S2 A and 4 B). Downmodulation of Par3 in 293T-shPar3 cells, which stably express a Par3-specific shRNA (293T-shRNAPar3;, reduced coprecipitation of Par6 and PKCɩ, while the complex was reconstituted by introducing GFP-Par3, but not GFP . Interactions between E4orf4 and Par complex proteins were also detected in nontumorigenic MCF10A cells. However, the ensuing impact on cell morphodynamics may depend on the cellular context, which could shape interactions between polarity proteins and the actin assembly machineries.
Because of Par3's role in epithelial cell polarity, we examined the impact of E4orf4 in MCF10A cells that are proficient in assembling cadherin-containing adherent junctions but defective in the assembly of ZO-1-containing tight junctions. MCF10A cells were transduced with Ad-LacZ or Ad-E4orf4 at low confluency and grown after confluency in 2D monolayers. Cellular junction assembly was monitored after 3 and 6 d by confocal microscopy, after E-cadherin and ZO-1 immunostaining. Remarkably, E4orf4-expressing cells exhibited a more rapid expansion of E-cadherin-labeled adherent junctions at day 3, as revealed by the significant increase in cortical E-cadherin staining, F and G). Moreover, we observed the presence of E4orf4-expressing cells in which ZO-1 staining showed a more contiguous apical staining that contrasted with the fragmented and cytoplasmic staining of ZO-1 in control cellsH, y-z views; magenta arrows). These results suggest that E4orf4 promotes assembly of cell-cell junctions in nontumorigenic epithelial cells rather than cell blebbing and death, as widely observed in cancer cells.
E4orf4's interaction with Par3 enables the corruption of actomyosin dynamics in tumorigenic cells In the context of cancer cells, Par3 has both tumor-suppressive and tumor-promoting functions, of which the latter would involve its scaffolding function. A crucial role for Par3 now emerges in the segregation of actomyosin structures. Thus, we hypothesized that subversion of Par3's scaffolding function by E4orf4 contributes to its tumoricidal effect on cell mechanics.
To test this hypothesis, the structural requirements for E4orf4-Par3 interaction were first characterized. We found that Par3 binding was mediated by a conserved motif across Ad serotypes, proximal to a K88 residue previously shown to impact the E4orf4-Par3 interactionA;. Substitution of a single negatively charged residue (D90A) reduced binding, while modification of three residues (DYV to GGG, to minimize structural alteration) blocked Par3 binding without impairing interaction with PP2A. Conversely, an E4orf4 construct that is defective in PP2A binding (F84A) was able to interact with Par3, indicating that these interactions are independent . To identify the E4orf4-binding domain on Par3, we performed IP studies in 293T-shPar3 cells after transfection of E4orf4 and GFP-Par3 constructs that contain deletions of key functional domains, including the oligomerization domain (ΔNter), Par6-binding domain (ΔPDZ), a protein kinase C-binding domain (ΔPKC-BD), or the C-terminal domain (ΔCter;, B-D). Only GFP-Par3 (ΔCter) showed a significant binding loss after reciprocal IPs, E and F). These results agreed with findings inshowing that E4orf4 can associate with PKCɩ and Par6 via Par3.
We then sought to visualize the in situ subcellular interaction of E4orf4 and Par3 by performing a Duolink proximity ligation assay (PLA) in HeLa cells. Numerous fluorescent foci were detected in HeLa cells expressing WT E4orf4, which were enriched proximal to important nuclear deformations, C and D, WT; and. Foci numbers were decreased by ∼85% in cells expressing E4orf4 (GGG), corroborating the importance of a loss of Par3 binding. Phosphorylated PKC substrates were also detected at sites of nuclear deformation that would indicate major sites of E4orf4-Par3 interaction. Intriguingly, loss of Par3 binding led to an approximately two-to threefold increase in the nuclear-to-cytoplasmic ratio of E4orf4, Data are the means ± SEM (n ≥ 3); *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001; ns, not significant. (C) Representative IB of E4orf4-mCherry levels assessed 48 h after transduction. (D) The percentage of cells expressing E4orf4-mCherry relative to the total number of infected cells, as assessed by fluorescence microscopy 48 h after transduction. See alsoA. recapitulating the effect seen during myosin II inhibition. This result suggests that E4orf4 that is unable to interact with Par3 is retained in the cell nucleus, as it is during the inhibition of cell contractility with blebbistatin. We next compared the ability of E4orf4 (GGG) to trigger the actomyosin-regulated phenotypes that match E4orf4's tumoricidal activity using U2OS and HeLa cells. These cells display a large projection area and clearly delineated nuclear/cytoplasmic (N/C) boundaries that facilitate quantification of morphological changes. Remarkably, E4orf4 (GGG) showed a severe loss of function, being unable to efficiently induce cell blebbing and nuclear condensation (Figs. 4 F and S3 C). Importantly, cancer cells expressing E4orf4 (GGG) maintained a relatively circular nucleus with few lamina folds, contrary to cells expressing E4orf4 (WT) showing a reduction of nuclear circularity (Figs. 4 G and S3 C). To further interrogate the role of Par3, the phenotypes of E4orf4-expressing cells were analyzed after Par3 depletion using siRNAs or lentivirus carrying Par3-specific shRNAs. We found that Par3 depletion blocked E4orf4-induced cell blebbing and nuclear condensation by >50% in both U2OS and HeLa cells. Furthermore, Par3 depletion also prevented E4orf4-induced nuclear deformations and lamina folds, as observed with blebbistatin treatment . The tumoricidal phenotypes, which were also impaired in 293T-shPar3 cells, were restored by the reintroduction of rat GFP-Par3, corroborating a Par3-specific effect. As expected, Par3-GFP constructs with functional domain deletions were unable to restore the tumoricidal phenotypes, F-J). These results support the requirement of Par3 for a scaffoldingrelated function.
Since similar phenotypes could be achieved by either depleting Par3 or inhibiting myosin II, we analyzed the impact of Par3 on E4orf4-induced actomyosin contractility. Immunoblotting of myosin light chain phosphorylation (Thr18/Ser19) revealed that E4orf4-induced myosin II activation was not significantly inhibited by Par3 depletionD, PP-MLC2;. This result indicates that Par3 may act in parallel to promote actin structure reorganization.
Par3 engages an E4orf4 positive feedback loop, which causes E4orf4's nuclear exit by nuclear bleb rupture To determine how Par3 and actomyosin contractility regulate E4orf4 spatial dynamics leading to tumoricidal phenotypes, the reversible action of blebbistatin on E4orf4 spatial dynamics was harnessed. This permitted the synchronization of cells expressing mainly nuclear E4orf4 while blocking nuclear shape changesA, schematics). Blebbistatin removal caused the nuclear-to-cytoplasmic ratio of E4orf4 to be restored within 1 hB, blebbistatin + 60' release). The associated changes to nuclear shape were imaged by real-time spinning disk confocal microscopy.
Live-cell imaging revealed that cancer cells from diverse origins expressing E4orf4 commonly showed a high frequency of nuclear blebs Several nuclear blebs collapsed in the cytoplasm, provoking a sudden nuclear efflux of E4orf4, D and G, see "Rupture" on heatmap-colored images; Video 1, E4orf4+siCTL compared to Video 2, RFP only; and Videos 3 and 4, arrows). In contrast, cells expressing a nuclear-targeted RFP or GFP protein only showed a much lower frequency of nuclear bleb formation, which was not associated with the nuclear marker effluxH and Video 2). E4orf4-induced nuclear bleb formation was also observed at a high frequency in cells that were not treated with blebbistatin, ruling out a drug-sensitizing effect (∼46.8%;C, vehicle). Importantly, recurrent nuclear bleb formation was markedly decreased by depletion of Par3 (∼38% to 68% inhibition;E, siPar3-3 and 3-9; and Fig. 6 I, shPar3B), while the proportion of nuclear blebs that resorbed was increased (∼37.5% in siPar3 versus ∼15.5% in siCTL;, D and F; and Video 1, E4orf4+-siPar3). Such effects are consistent with a reduction of cytoplasmic E4orf4 upon Par3 depletion and suggest that E4orf4 exits the nucleus through nuclear bleb ruptures.
Nuclear bleb collapse could reflect transient NE ruptures due to focal forces generated by perinuclear actin structures. Recent findings revealed that cancer cells are prone to undergo transient NE rupture and repair when they are submitted to physical constraints, a feature that could sensitize them to E4orf4 action . To directly determine nuclear integrity, E4orf4-mCherry was coexpressed with a GFP protein fused to an NLS. The GFP-NLS reporter exhibits rapid efflux bursts into the cytoplasm upon NE rupture and is then gradually reimported into the nucleus as NE integrity is restored. Control cells in interphase displayed a low incidence of NE rupture/repair (HeLa cells: ∼10% over a 10 h-period,B; U2OS cells: ∼4.8% over a period of 4 h,. In contrast, ∼44.5% to 61.3% of E4orf4expressing U2OS and HeLa cells, respectively, showed NE ruptures associated with a sharp drop of GFP-NLS nuclear intensity as it spilled into the cytoplasm, A-C and E; and Videos 5 and 6, rupture and repair). Strikingly, repetitive cycles of NE rupture/repair were associated with a loss of NE integrity in ∼15% to 25% of E4orf4-expressing cells, B and C, asterisk; In cells depleted of Par3, nuclear ruptures and loss of NE integrity were reduced by ∼50% and ∼73%, respectively. E4orf4-induced NE rupture is potentially regulated by nucleocytoskeletal connections, such as NE rupture caused by contractile actin bundles. To evaluate this possibility, cells were transduced with lentiviruses carrying shRNA sequences to deplete SUN1/2 proteins and disrupt the LINC complex that connects perinuclear actin bundles to the NE. SUN1/2 proteins were efficiently depleted in U2OS cells using a mixture of two distinct shRNA sequences for each protein, as reported previously . When E4orf4-mRFP and the GFP-NLS reporter were imaged in these cells, a significant reduction in nuclear blebs and NE ruptures were observed (Figs. 6 I and 7 E; ∼75%-85% inhibition). Under these conditions, E4orf4induced nuclear dysmorphia was not significantly present as compared with RFP-expressing cellsF; circularity index). Furthermore, E4orf4-induced cell blebbing and nuclear condensation were dramatically impaired (inhibited by ∼55% to 67%; Figs. 7 G and S4 C). Transfection of a dominant-negative GFP-KASH2 to disrupt the binding of endogenous nesprins to SUN proteins also interfered with E4orf4 phenotypes in HeLa cells (∼27% inhibition;, A-C;. These results strongly suggest that E4orf4's tumoricidal effect is linked to a subversion of nucleocytoskeletal coupling. Moreover, they argue that a positive feedback loop ascribed to NE rupture and nuclear E4orf4 efflux would exacerbate the actomyosin forces transmitted to the nucleus and jeopardize nuclear integrity.
Localization of Par3 proximal to nuclear lamina grooves is associated with a directional movement of actin filaments and NE rupture Results suggest that Par3 could function to link actomyosin contractility and nuclear shape changes in response to E4orf4. In line with these findings, Par3-GFP was recruited to perinuclear actin structures that are associated with nuclear dysmorphia. Since exogenous Par3-GFP tended to spontaneously cluster, we immunostained fixed cells for endogenous Par3 in order to examine the impact of E4orf4 by high-resolution confocal imaging. The signal specificity was confirmed in cells transfected with Par3-specific siRNAs. Airyscan confocal imaging revealed that E4orf4 modulated the localization of Par3 in tumorigenic U2OS cells but had little effect in HaCaT, a nontumorigenic cell line expressing similar Par3 levels (The Human Protein Atlas). In U2OS cells, E4orf4 induced Par3 clustering that segregated around the nucleus, contrasting with a relatively diffuse localization of Par3 in unpolarized, exponentially growing HaCaT cells;. The distinct localization of Par3 was clear upon quantification of Par3-lamin A/C dual-labeled fixed cells on single-slice high-resolution confocal images, highlighting an overlap between NE and Par3 only in E4orf4-expressing U2OS cells. Most remarkably, Par3 clusters showed a clear alignment with NE folds at the nuclear surface, suggesting that Par3 may delineate sites of nuclear mechanical compression.
To directly investigate the relationships among Par3, actin structure dynamics, and nuclear shape changes, U2OS cells coexpressing E4orf4-mCherry and LifeAct-GFP were imaged to track the behavior of actin structures within cells in relation to NE rupture (Videos 7, 8, and 9). In E4orf4-cells transfected with siCTL, actin structures mainly exhibited a directional movement toward and over the nucleus, which coincided with nuclear bleb formation and nuclear E4orf4 efflux due to NE ruptureA, siCTL arrows; and Videos 7 and 9). This directional and retrograde movement of actin structures was disrupted by Par3 depletion, as evidenced by kymograph measurementsA, siPar3, dotted line;, B and C; and Video 8). Quantification of actin structure velocity revealed that in control E4orf4-cells, actin filaments moved toward the nucleus with an average speed of 0.129 µm min −1 (shown by the peak of the curve), while in Par3-depleted cells, they barely moved toward the nucleus, with a fivefold reduction in average speed (0.026 µm min −1 ;. Importantly, we found a linear correlation between mean average velocity values of actin filaments in single cells and their respective time to nucleus rupture (R 2 = 0.8945;. Notably, among the Par3-depleted cells, the few that showed NE rupture also displayed higher actin filament velocities and positive movement values toward the nucleus (white circles,. Together, the results suggest that Par3 clustering at the NE can direct a retrograde movement of actin structures that contributes to increasing nuclear mechanical stress.
Par3 regulates spontaneous NE rupture facilitated by downmodulation of lamins Finally, the impact of Par3 on NE rupture mediated by actinbased nucleus confinement in the absence of E4orf4 was determined. To this end, lamina proteins were partially depleted to recapitulate lamina deficiencies and induce spontaneous NE rupture, as reported before. HeLa cells were transfected with a mixture of lamin A/Cand lamin B1-specific siRNAs that achieved moderate lamin depletion and increased the occurrence of spontaneous NE rupture by approximately twofold as assessed by live-cell imaging of the RFP-NLS marker, A-C; and Video 10;. Strikingly, when Par3 was also depleted in these cells, the occurrence of spontaneous NE ruptures was reduced back to near the basal levels observed in control cells (siCTL), even though lamins were similarly depleted, B and C). These results suggest that depletion of Par3 alleviates the actin-dependent constraints on the nucleus that cause NE rupture. We conclude that Par3 is a bona fide regulator of the actin-dependent forces controlling nuclear mechanics in tumor cells.
used. (F and H) Merged plane views from confocal image stacks of MCF10A cells transduced with lacZ or E4orf4 showing staining of E-cadherin or ZO-1; F-actin was labeled with phalloidin; N.A., not applicable. The dotted lines in H indicate the position of the y axis presented as y-z views on the right panels; magenta arrowheads indicate ZO-1 clusters induced by E4orf4 (n ≥ 3). Bars: 20 µm. (G) Quantification of the cortical/cytoplasmic ratio of E-cadherin for datasets represented in F; means ± 95% CI; ****, P < 0.0001.
## Dziengelewski et al.
Journal of Cell Biology
# Discussion
Taken together, our data suggest that E4orf4's tumoricidal action involves an early loss of nuclear integrity. This loss is predominantly due to corruption of a new Par3 scaffolding function at the NE, which may control the spatial dynamics of actin structures and thereby force transmission to the nucleus. As a result of the E4orf4-Par3 interaction, clustering of Par3 could provide spatial information to polarize actin structures and promote their retrograde translocation toward the nucleus, the efficiency of which would depend on nucleocytoskeletal connections. We suggest that Par3 enables E4orf4 to engage a positive feedback loop that exploits nucleocytoskeletal coupling in tumorigenic cells. These cells are more prone to nuclear ruptures, due to their reduced nuclear stiffness and increased actomyosin contractility. This vicious cycle would build upon NE ruptures causing an efflux of E4orf4 and other molecules that would further amplify actomyosin contractility and NE ruptures, ultimately overcoming the cellular capacity to repair the NE. Prolonged NE integrity loss during interphase would perturb several essential functions and jeopardize genomic integrity. We propose that disruption of NE mechanical regulation represents a prime lesion responsible for the extreme alterations of tumor cell structural organization. We found that disrupting myosin II, Par3, or LINC complex functions can recapitulate the same inhibitory phenotypes on E4orf4-induced morphodynamic changes, suggesting that they act on the same pathway for efficient force transmission to the nucleus. Since Par3 also regulated spontaneous NE rupture induced by lamin depletion, we infer here that a Par3 scaffolding is widely implicated. The exact mechanism whereby Par3 would control the spatial linkage of contractile actin filaments at the nuclear surface is an outstanding issue with potential implications for tumor cell metastatic potential. Metastasizing cancer cells that encounter tight interstitial spaces are particularly prone to transient NE rupture. This process is viewed as a new mechanism dependent on confinement that promotes genomic instability and tumor progression. Therefore, such a mechanism could be an important part of the tumor-promoting functions of Par3.
In conclusion, our data indicate that oncogenic rewiring of the Par complex protein network could define novel protein partners for Par3 that regulate NE dynamics, genome integrity, and metastatic potential. In that context, E4orf4 offers a unique tool to uncover new vulnerabilities in malignant cells.
# Materials and methods
## Expression vectors, mutagenesis, ads, and baculoviruses
The following expression vectors were described previously: Flag-E4orf4(WT)-mRFP, Flag-E4orf4(F84A)-mRFP, HA-E4orf4(WT); pEGFP-KASH2 and pEGFP-KASH2ext; and GFP-Par3 (rat Par3 cDNA in pEGFP-C1; Hidalgo-Carcedo et al., 2011). The GFP-Par3 deletion mutants were designed by PCR using rat GFP-Par3 as a template and the following primers:
Par3_DNterm_fwd: 59Phos-ATGAAGCTCGTACAAGTCCCCAA-39; Par3_DNterm_rev: 59-GGTGGCGATATCGCTAGCAA-39; rPar3_ delPDZ_fwd: 59Phos-AGGCGGATCAGCAGATGTAACGAGTTG-39; rPar3_delPDZ_rev: 59-TACCATATCATCGAGGGAAAAGTTGGGC-39; rPar3_delPKC_fwd: 59Phos-GAGAAGAGGGACAAGGCAGA GAAGAAAAAG-39; rPar3_delPKC_rev: 59-GCGTTCTCTGTC ATCCAACTCTGTTTCTAT-39; Par3_DCterm_fwd: 59Phos-GTCGACGG TACCGCG-39; Par3_DCterm_rev: 59TGGGTCTCCATTCATTTGTC-39.
GFP-NLS (NLS from SV40 large T antigen: PKKKRKKV) was described before. E4orf4-D90A and DYV-GGG mutants were designed by PCR using the QuickChange Site-Directed Mutagenesis kit (Stratagene). Flag-E4orf4(WT)-mRFP in pcDNA 3 or the HA-E4orf4(WT) was used as templates with the following primers:
Orf4 All constructs were confirmed using DNA sequencing. Ad-LacZ and Ad-GFP were described previously. Ad-Flag-E4orf4(WT)-mCherry, Ad-Flag-E4orf4(F84A)-mRFP, and Ad-Flag-E4orf4(GGG)-mRFP were produced by Welgen. Adenovirus type C strain 5 containing the different E4orf4 variants was produced by subcloning into recombinant adenoviral vector lacking the E1 and E3 regions. The E1 deletion was replaced with an expression cassette consisting of a cytomegalovirus promoter with a bacterial tetracycline resistance operon, the gene of interest, and a poly A signal. Recombinant Ads were amplified in the HEK293VR cell line expressing the tetracycline repressor (a gift from Philip E. Branton, McGill University, Montreal, Québec, Canada). All recombinant Ads were confirmed by sequencing of the inserted sequences. Virus titers were determined using the AdenoX Rapid titer kit (Clontech Laboratories; #631028). Baculovirus encoding RFP-NLS or GFP-NLS were purchased from Invitrogen/Thermo Fisher Scientific (CellLight Nucleus-RFP, BacMam 2.0, C10603; CellLight Nucleus-GFP, Bac-Mam 2.0, C10602).
## Lentivirus-shsun1/2 production
The following individual SUN1-and SUN2-specific shRNAs were engineered using two oligonucleotides (IDT) that were annealed, phosphorylated, and ligated into AgeI and EcoRI restriction sites of pLKO.1 vector (Addgene; plasmid #8453): shSUN1-1 (59-CCG GCAGATACACTGCATCATCTTTCTCGAGAAAGATGATGCAGTG TATCTGTTTTTTG-39 and 59-AATTCAAAAAACAGATACACT GCATCATCTTTCTCGAGAAAGATGATGCAGTGTATCTG-39), shSUN1-2 (59-CCGGGAACTAGAACAGACCAAGCAACTCGAG TTGCTTGGTCTGTTCTAGTTCTTTTTTG-39 and 59-AATTCA AAAAAGAACTAGAACAGACCAAGCAACTCGAGTTGCTTG GTCTGTTCTAGTTC-39), shSUN2-1 (59-CCGGGCCTATTCAGACGTT TCACTTCTCGAGAAGTGAAACGTCTGAATAGGCTTTTTTG-39 and depletion in U2OS cells using mixtures of two sequences targeting SUN1 or SUN2 (shSUN), as estimated by loading increasing amounts of control cell extracts (shCTL or siCTL, 1, 1/2, and 1/4); GAPDH levels, loading control. (E) The percentage of shRNA-treated U2OS cells expressing RFP or E4orf4-mCherry together with GFP-NLS as a marker of NE rupture, which showed transient NE rupture or loss of NE integrity (no repair over >30 min), as scored relative to GFP-NLSexpressing cells; data represent means (n ≥ 92 cells; n = 2) or means ± SEM (n ≥ 85 cells; n = 3). (F) Quantification of the nuclear circularity index for datasets represented inC . (G) The percentage of U2OS cells showing cell blebbing and nuclear condensation, as scored relative to RFP-or E4orf4-mCherry-expressing cells for datasets represented in, or means ± SEM [n ≥ 363 cells; n = 3]). *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001; ns, not signifciant.
## Dziengelewski et al.
Journal of Cell Biology For double knockdown of SUN1 and SUN2, pLKO.1 vector containing shSUN2 was amplified by PCR using primers 59-TTTAAATAGAATTCGAGACTAGCCTCGAGCGGC-39 and 59-TGCCATTTGTCTCGAGGTCGA-39 in order to produce a sequence with the U6 promoter followed by shRNA. After digestion by EcoRI, this sequence was ligated into EcoRI restriction sites of pLKO.1 vector containing shSUN1 to generate shSUN1-1/2-2 and shSUN1-2/ 2-1. Lentiviral particles were generated by transfecting 293T cells with 12 µg pLKO.1 vector, 6 µg psPAX2 packaging plasmid (Addgene; plasmid #12260), and 2 µg pMD2.G envelope plasmid (Addgene; plasmid #12259), as described previously.
## Antibodies and chemicals
The following antibodies and drugs were used: Anti-Par3 (#07-330) was from Millipore; anti-GFP 3E6 (A11120) for IP was from Molecular Probes; anti-Flag M2 (F3165) and anti-vinculin (V9131) were from Sigma-Aldrich; anti-RFP (600-401-379) was from Rockland; anti-E-cadherin (610181) and anti-PKCι (610175) were from BD Transduction Laboratories; anti-GAPDH ZG003 (39-8600) and anti-ZO-1 1A12 (33-9100) were from Thermo Fisher Scientific; anti-GFP B-2 for Western blot (sc-9996), antilamin A/C 636 (sc-7292) for immunofluorescence (IF) anti-myosin light chain 2 (sc-15370), and anti-Pard6B H-64 (sc-67392) were from Santa Cruz; anti-lamin B1 (ab16048) and anti-lamin A/C for Western blot (ab133256) were from Abcam; anti-HA.11 (MMS-101R) was from Covance; and anti-phospho-myosin light chain 2 (Thr18/Ser19; 3674) and anti-phospho(Ser) PKC substrate (2261), and hTERT RPE-1 cells (ATCC) were cultured in high-glucose DMEM (GIBCO) with 10% FBS; MCF10A cells (ATCC) were cultured in DMEM F12 with 5% horse serum, 0.5 µg/ml hydrocortisone, 10 µg/ml insulin, 20 ng/ml epithelial growth factor, and 100 ng/ml cholera toxin (Sigma-Aldrich) and maintained as described previously. Cells were grown in a humidified atmosphere in 5% CO 2 or in 0% CO 2 when cultivated in Leibovitz's L-15 medium. MCF10A, HaCaT, hTERT RPE1, 22Rv1, HeLa-CCL2, U2OS, and MDA-MB-231 cell lines were verified using short tandem repeat profiling (ATCC human short tandem repeat profiling cell authentication service). All cell lines were tested negatively for mycoplasma contamination.
293T-shCTL and 293T-shPar3 stable lines were generated using the MISSION shRNA kit (Sigma-Aldrich). 293T cells (ATCC) were transduced with 0.3 colony-forming units of lentivirus, and stable clones were selected with 3 µg/ml puromycin, isolated, and assayed for their level of Par3 knockdown by Western blot analyses. The MISSION shRNA transduction particles were as follows: shPar3A, reference. TRCN0000118134; sequence, 59-CCGGGCCATCGACAAATCTTATGATCTCGAGATCAT AAGATTTGTCGATGGCTTTTTG-39.
## Transfection and sirna experiments
Plasmid transfection was performed using Lipofectamine 2000 or Lipofectamine 3000 (Thermo Fisher Scientific; HeLa and U2OS cells) or the calcium-phosphate method (293T cells;. For siRNA transfection, cells were plated on 35-mm Petri dishes and cultured for >16 h before transfection. siRNA duplexes were transfected overnight at 50-60 nM using the RNAiMax transfection reagent (In-Vitrogen). Cells were washed, replated at a 1/7 dilution in 35-mm Petri dishes, and cultured for at least 48 h before performing subsequent analyses. The siRNA duplexes were purchased from Thermo Fisher Scientific (standard A4 grade); sequences of the sense strands are as follows: siPar3-3, 59-GAU GGCGACCUUCGAAAUAUU-39; siPar3-9, 59-GGGCAAAUCCCA AGAGGAAUU-39; siLamin A/C, 59-UGUUCUUCUGGAAGUCCA GUU-39; and siLamin B1, 59-CGCGCUUGGUAGAGGUGGAUU-39. Control siRNA (siCTL): AllStars Negative Control was purchased from Qiagen, target sequence: 59-CAG GGTATCGACGATTACAAA-39.
## Cell transduction with viral vectors
Adenovirus and baculovirus infections were performed in cell suspension in culture medium supplemented with 8 µg/ml polybrene (Sigma-Aldrich). HeLa (2 × 10 5 /2 ml), U2OS (1.75 × 10 5 /2 ml), 22Rv1 (4 × 10 5 /2 ml), MDA-MB-231 (2.75 × 10 5 /2 ml), sections is shown as a reference (in gray), and the arrow designates the time of NE rupture in a siCTL-treated cell. (C) Quantification of actin structure velocity toward the nucleus in E4orf4-expressing cells, as measured from the probability distribution from all the visible actin structure velocity distributions; siCTL, C, n = 9 cells; siPar3, s, n = 8 cells. The plain lines provide the Gaussian fit of the normal distribution. Actin structures moving toward the nucleus were defined as having a positive speed, whereas those moving away as having a negative speed. (D) Linear plot of the time of nuclear rupture as a function of the speed of actin structures toward the nucleus, showing a linear relationship (R 2 = 0.8945); each data point corresponds to the average actin speed ± SD and rupture time ± 2.5 min (e.g., half the acquisition rate) for a single cell, and the shaded area provides the 95% CI of the linear regression. *, P < 0.05.
## Dziengelewski et al.
Journal of Cell Biology MDA-MB-468 (1.8 × 10 5 /2 ml), HaCaT (2 × 10 5 /2 ml), and hTERT RPE-1 (1.3 × 10 5 /2 ml) cells were transferred to a 5-ml tube and gently mixed with Ad and/or baculovirus particles. CellLight reagents were added at four particles per cell . HeLa, U2OS, MDA-MB-231, MDA-MB-468, HaCaT and hTERT RPE-1 cells were infected with Ad-RFP (15 PFU/cell; completed with 60 PFU/cell Ad-LacZ) or Ad-E4orf4-mCherry (WT or GGG; 75 PFU/ cell); 22Rv1 was infected with Ad-RFP (2 PFU/cell; completed with 18 PFU/cell Ad-LacZ) or Ad-E4orf4-mCherry (20 PFU/cell). The cell-virus suspensions were incubated at 37°C with light agitation for 15-30 min, seeded on 35-mm Petri dishes and incubated at 37°C. Culture medium was changed 16-20 h after infection. For shRNA knockdown, U2OS (1.5 × 10 5 / 2 ml) cells were infected with lentiviral particles in the presence of 8 µg/ml polybrene. 48 h after infection, cells were replated, and selection of lentivirus-infected cells was performed by adding puromycin (1 µg/ml) for 48 h at 37°C. The resulting cells were harvested and infected with Ad-E4orf4 or Ad-RFP (96 h after infection with lentivirus-shRNA). The MISSION Par3-specific shRNA transduction particles were as follows: shPar3A, reference TRCN0000118134; sequence, 59-CCGGGCCATCGACAAATCTTA TGATCTCGAGATCATAAGATTTGTCGATGGCTTTTTG; shPar3B, 59-CCGGCCATCGACAAATCTTATGATACTCGAGTATCATAAGAT TTGTCGATGGTTTTTG-39.
## Proliferation and cell death assay
To measure long-term survival by colony-forming assay tests, MCF10A (3.5 × 10 5 ) and MDA-MB-231 (2.8 × 10 5 ) cells were infected with Ad-GFP as control or Ad-Flag-E4orf4 at varying PFUs per cell (20, 40-50, 75-100 PFU/cell) in 35-mm Petri dishes. The culture medium was changed 16 h after infection, and duplicate cultures were harvested 24 h after infection for Western blot analyses. 40 h after infection, the cells were trypsinized, counted, and replated in each of three 60-mm Petri dishes at 2 × 10 4 , 2 × 10 2 , and 1 × 10 2 for each condition (technical replicates). The cells were grown at 37°C for 2 wk, and the resulting colonies were fixed in methanol/acetic acid (3:1), stained in a solution of 0.25% Trypan blue in 5% acetic acid, and scored manually. Cell proliferation and death were assessed from estimations of cell numbers according to a Trypan blue dye exclusion assay. At 24, 48, and 72 h after infection, the culture medium containing dying cells was recovered in a 5 ml-tube together with trypsinized adherent cells. The cell suspension was gently mixed with Trypan blue (0.4% in PBS 1X) at a 1:1 ratio and then rapidly counted in a hemacytometer. Each cell count was performed three to five times using independent Trypan blue cell suspension mixtures from the same sample for statistical reproducibility. White, refringent cells were considered viable, whereas dark, blue cells that incorporated the dye were considered nonviable.
## Immunoprecipitation and western blot experiments
Cells were suspended in their culture medium, pelleted, and washed with PBS. The cell pellet was lysed in 500 µl of 1% digitonin lysis buffer (0.1 M MES-NaOH, pH 6.5, 1 mM magnesium acetate, 0.5 mM EGTA, 200 µM sodium orthovanadate, and 1% digitonin with protease inhibitors [Roche; 1X Complete];, B-D; and. Alternatively, cells were lysed by three freeze-thaw cycles in 500 µl lysis buffer B (20 mM Tris-HCl, pH 7.6, 150 mM NaCl, 1 mM EDTA, 0.1% octylphenoxy poly(ethyleneoxy)ethanol (IGEPAL), 1 mM Na 3 VO 4 , 10 mM NaF, 40 mM β-glycerophosphate, 1X Complete [Roche], and 1 mM DTT;, B and E). Lysates were centrifuged at 10,000 g for 15 min at 4°C in 1.5-ml tubes, and the supernatant was transferred to a new tube. Lysates were then precleared with 20 µl Dynabeads protein G (Life Technologies) for 30 min at 4°C. 20 µl Dynabeads protein G were incubated with 5 µg anti-Flag M2 (Sigma-Aldrich) or 1 µg anti-GFP (Molecular Probes) antibody in 500 µl lysis buffer for 30 min at 4°C. The lysate was then incubated with Dynabeads protein G for 1 h at 4°C with gentle agitation. Immune complexes were collected using a magnetic stand (EMD Millipore), washed three times in lysis buffer, and transferred to a fresh tube. Equal amounts of IPs were resolved by SDS-PAGE and analyzed by Western blot. Protein levels were determined after lysing cells in SDS sample buffer (62.5 mM Tris-HCl, pH 6.8, 2.3% SDS, 10% glycerol, 5% B-mercaptoethanol, 0.05% bromophenol blue, and 1 mM phenylmethylsulfonyl fluoride) and protein concentrations were measured using the DC protein assay reagent (Bio-Rad). Immunoblotting was performed as described previously.
## Ap-ms experiments
For AP-MS experiments, MDA-MB-231 cells plated on 150-mm Petri dishes (eight dishes per condition), were infected with Ad-Flag-E4orf4-mCherry or control Ad-lacZ recombinant Ads (75 PFUs/cell). Cells were collected and lysed in digitonin buffer 16 h after infection. IP was performed using anti-Flag M2 antibody, and proteins were eluted using the Flag peptide (100 µg/ml), overnight at 4°C. The eluate was collected and concentrated on a Vivaspin 500 Protein Concentrator 10-kD Column (GE Healthcare Life Sciences) and then diluted into the appropriate amount of 3X SDS sample buffer. The eluate and the Dynabeads in SDSsample buffer were loaded on a 6-10% gradient polyacrylamide gel and allowed to migrate until the migration front reached 10 cm. The gel was stained with Coomassie blue (50% methanol vol/vol, 10% acetic acid vol/vol, and 0.1% Coomassie blue wt/vol) and washed in destaining solution (10% methanol vol/vol and 10% acetic acid vol/vol) overnight at RT. Bands of interest were extracted from small pieces of gels that had been cut to cover all protein samples and placed in 96-well plates and then processed for MS/MS analyses. Each AP-MS experiment was performed in biological triplicate (n = 3) to perform SAINT analyses, as reported previously, using Flag controls to filtrate background contaminants. Mass spectrometry was performed by the Proteomics platform of the CR-CHU de Québec Université Laval on a TripleTOF 5600 (AB Sciex) mass spectrometer equipped with an ion source (NanoSpray III; AB Sciex) and coupled to a nanopump and by the Proteomics And Bioanalytical Mass Spectrometry of the Institute for Research in Immunology and Cancer (Montreal University, Quebec, Canada) on a quadrupole/TOF Q-STAR (AB Sciex) mass spectrometer. MS/MS samples were analyzed using Mascot (Matrix Science) that was set up to search the UR100_12_05_-Homo_sapiens_9606 database (128,347 entries), with a fragment ion mass tolerance of 0.50 D and a parent ion tolerance of 2. was used to validate MS/MS based peptide and protein identifications. Peptide identifications were accepted if established at greater than 95% probability to achieve a maximum of 1.2% false discovery rate. Protein identifications were accepted if they contained at least two identified peptides. Peptide and protein probabilities were assigned by the Protein Prophet algorithm.
Immunofluorescence, PLA, and microscopic analyses Cells plated on fibronectin-coated glass coverslips were fixed with 4% paraformaldehyde (Biobasics) for 20 min at RT. DNA was stained with cellpermeable Hoechst. Cells were permeabilized with 0.2% Triton X-100 in PBS containing 1 mM MgCl 2 for 15 min at RT. F-actin was stained with a 1/100 dilution of Alexa-Fluor phalloidin in PBS for 20 min at RT. Immunostaining of proteins was performed as described previously, and coverslips were mounted on glass slides using the ProLong Glass Antifade Mountant (Invitrogen; P36980). PLA detection was performed according to the manufacturer's instructions using PBS containing 5% milk and 1 mM MgCl 2 as a blocking solution. Antibodies (anti-Par3 diluted at 1/100 and anti-HA diluted at 1/500) were incubated in PBS supplemented with 1% milk and 1 mM MgCl 2 . Volocity software was used to count the number of PLA foci per cell after background subtraction using the "Find objects" tool in the Measurements thumbnail. Detection was according to intensity using an intensity threshold set to optimally detect the foci, minimal object size = 0.3 µm 3 , and "Separate touching objects" (object size guide = 0.3 µm 3 ).
Cell blebbing and nuclear condensation were quantified by visual inspection of fixed specimens using an AxioObserver Z1 system with a 60× oil 1.25 NA objective and an Axiocam MRm charged-coupled device camera controlled by Zen software (Carl Zeiss). Cell blebbing was determined by phalloidin staining, and nuclear condensation was scored based on a decreased nuclear surface with condensed chromatin, as visualized by Hoechst staining. N/C intensity ratio and nuclear circularity index were measured on confocal images acquired with a PerkinElmer Ul-traVIEW spinning disk confocal (60× oil 1.4 NA) equipped with an Hamamatsu C9100-50 Electron Multiplier EMCCD camera at −50°C driven the Volocity software version 6.3 (Quorum Technologies), or with an Orca Flash 4.0LT+ sCMOS camera (Hamamatsu Photonics K.K.) driven by the NIS Elements software version 5.02.00 (Nikon.ca). The N/C ratio was estimated from the mean fluorescence intensity of the nuclear surface relative to the cytoplasmic surface, as delineated on the median confocal plane using Volocity software version 6.3 or Fiji software (Im-ageJ 2.0.0-rc-69/1.52n). The nuclear circularity index was estimated from maximum intensity projection images of lamin A/C-stained cells using the "shape descriptor" tool of Fiji software (ImageJ 2.0.0-rc-69/1.52n). Image processing was done using Fiji software and was limited to background subtraction, brightness and contrast adjustment, and deconvolution using Volocity software, when indicated.
High-resolution confocal imaging of Par3 and lamin A/C staining on fixed cells was performed using a 63×/1.4 C-Plan Apo oil-immersion lens on a Zeiss LSM 800 microscope in Airyscan mode. Image acquisition and processing of Airyscan datasets was performed with the Zen Blue 2.6 software. For Par3 cluster analysis, the FIJI software (ImageJ 2.0.0-rc-69/1.52n) was used to apply an Intermode threshold, and a region of interest was then manually drawn to encompass each cell. The area of all clusters was quantified by using the "analyze particles" tool. No size or circularity limitation was included. Par3 and lamin A/C intensity profiles were analyzed using the plot profile tool of Fiji software (ImageJ 2.0.0-rc-69/1.52n).
Live-cell imaging Confocal microscopy of live cells was performed with a Perki-nElmer UltraVIEW spinning disk confocal (40× oil 1.3 NA, or 40× 0.75 NA, or 20× 0.50 NA), equipped with a humidified, 5% CO 2 thermoregulated chamber. Cells were plated on fibronectincoated glass dishes (Corning; Fluorodish) the day before transfection or infection. For E4orf4-induced nuclear bleb formation, NE rupture, and actin structure dynamics, cells were imaged after release from a 2-h treatment with blebbistatin (50 µM); imaging was performed at 7 h after transfection of HeLa cells with Flag-E4orf4-mRFP or 20 h after infection of U2OS or MDA-MB-468 cells with Ad-RFP, Ad-GFP, and Ad-Flag-E4orf4-mCherry, with or without CellLight Nucleus-GFP, BacMam 2.0, or Ad-LifeAct-GFP. Where required, cells were transfected with siRNAs or infected with lentivirus shRNAs at least 48 h before transfection of infection with E4orf4 constructs. Images (multiple fields/sample) were acquired at 5-, 1.5-, or 1-min intervals for 3-12 h. Analysis of actin structure dynamics was performed on maximum intensity projections of image stacks from timelapse movies using MATLAB and ImageJ. A median filter was applied to the image stacks (1-pixel radius), and a line (5 pixels wide) was drawn perpendicular to actin network and across the nucleus over the width of the cell. A kymograph was extracted along the drawn line for both the actin and E4orf4 fluorescence channels. The actin kymograph was then used to measure the speed at which actin fibers were translating within the cells by calculating the slope of the visible fiber components. The E4orf4 channel was then used as a reference of the nucleus position to establish the relative speed of actin structure. The time point 0 min corresponds to the beginning of cell recording, 24 h after recombinant virus transduction (64 h after siRNA transfection). All visible actin filaments, but not the cell cortex, were considered in the analysis. For analysis of spontaneous NE rupture induced by lamina protein depletion, HeLa cells plated at low confluency were transfected with Par3-specific siRNAs (50 nM), and 48 h later, the cell cultures were treated with a mixture of lamin A/C-and lamin-B1-specific siRNAs used at 15 nM to achieve moderate knockdowns and avoid toxicity (∼20% lamina protein depletion). Concomitantly, cells were infected with a recombinant baculovirus encoding RFP-NLS overnight (Cell-Light Nucleus-RFP, BacMam 2.0). Cell imaging was started 16 h after lamin depletion, and images were recorded at 5-min intervals for 12 h.
## Statistical analyses
For experiments examining the difference between two conditions, the mean values of individual experiments were Journal of Cell Biology analyzed using the Mann-Whitney test, a nonparametric test suitable to compare two small unpaired groups of ordinal or numerical variables. For quantitative analyses examining the proportions of cells displaying phenotypes of membrane blebbing, nuclear condensation, nuclear bleb formation, NE rupture, and cell death, data were analyzed using the Fisher's exact test, which is used for small sample sizes with nominal/categorical variables. To analyze the difference in cell proliferation, statistical analysis was performed using the two-way ANOVA test. To analyze the difference in E4orf4 N/C ratio and nuclear circularity index, individual measurements for each cell were pooled together for each condition and analyzed using the Kruskal-Wallis test with Dunn's multiple comparisons, which is a nonparametric test suitable to compare several groups of numerical variables. For actin structure velocities, statistical significance was analyzed using a two-tailed Student's t test.
Statistical calculations were performed using Prism 8.1.1 (GraphPad Software) statistical software. The significance of P values is shown as *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. In many cases of one or two independent experiments, data were collected in different nontumorigenic and tumorigenic cell models or from replicates of two or more independent siRNAs/shRNAs, adding robustness to the results. Phenotypic data were collected through the analysis of many fields to properly represent the entire sample, and technical counts for independent experiments were obtained by two independent investigators or repeated two or three times.
## Data availability
Proteomic data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repositorywith the dataset identifier PXD014329. All data that support the findings of this study are available from the corresponding author upon reasonable request.
Online supplemental materialshows representative images of transduction efficiencies and analysis of cell blebbing and determination of nuclear condensation phenotypes in cells from various origins expressing RFP alone or E4orf4-mCherry, showing the actomyosindependent tumor cell-selective effects of E4orf4.shows structure-function analyses, supporting a requirement for Par3 binding and scaffolding function for E4orf4-induced cell blebbing and nuclear condensation in 293T cells.A provides representative images of the PLA assay controls related to, C and D;B shows the recruitment of atypical protein kinase C (aPKC) substrates at nuclear deformation sites.C shows loss-of-function phenotypes of the E4orf4-GGG mutantA shows the efficiency of Par3 depletion in HeLa cells., B and C show E4orf4 loss-offunction phenotypes upon depleting Par3 or SUN1/2., A-C show the effects of dominant-negative KASH2 constructs on E4orf4 phenotypes in HeLa cells.D shows Par3-GFP localization relative to F-actin in an E4orf4-expressing HeLa cell.E shows Par3 depletion efficiency in single cells. Video 1 shows the distinctive dynamics of nuclear blebs in representative E4orf4-mRFP-expressing HeLa cells transfected with siCTL or siPar3-3. Video 2 shows the absence of nuclear blebbing in control HeLa cells expressing mRFP-NLS. Video 3 shows a representative U2OS cell expressing E4orf4-mCherry undergoing nuclear bleb rupture and E4orf4 efflux. Video 4 shows a representative E4orf4-mCherry-expressing MDA-MB-468 cell undergoing repetitive cycles of nuclear bleb formations and ruptures along with E4orf4 signal effluxes. Videos 5 and 6 show representative HeLa and U2OS cells expressing E4orf4-mCherry undergoing NE rupture and repair. Videos 7, 8, and 9 show representative dynamics of actin filaments relative to nuclear rupture in U2OS cells expressing E4orf4-mCherry and GFP-LifeAct, showing that Par3 depletion inhibits the recruitment of actin filaments at the nuclear surface. Video 10 shows transient NE rupture after lamin depletion in HeLa cells.. Par3 depletion, as SUN1/2 depletion, impairs E4orf4 tumoricidal phenotypes. (A) Representative IB of Par3 depletion in E4orf4-expressing HeLa cells (siPar3-3 and 3-9), as estimated by loading increasing amounts of control extracts (siCTL, 1, 1/2, and 1/4); GAPDH: loading control. (B and C) Maximum intensity projections of confocal image stacks of control E4orf4-mCherry-expressing HeLa or U2OS cells (siCTL and shCTL), as compared with cells treated with Par3-specific siRNAs (B), or transduced with lentivirus carrying Par3-specific shRNAs (B) or a mixture of SUN1/2-specific shRNAs (C), showing representative phenotypes. Please note that Par3 depletion, as SUN1/2 depletion, inhibited cell blebbing and nuclear condensation/dysmorphia; F-actin staining, green; lamin A/C staining, magenta. Bars: 20 µm; n ≥ 3. |
Estimating the cost of referral and willingness to pay for referral to higher-level health facilities: a case series study from an integrated community case management programme in Uganda
Background: Integrated community case management (iCCM) relies on community health workers (CHWs) managing children with malaria, pneumonia, diarrhoea, and referring children when management is not possible. This study sought to establish the cost per sick child referred to seek care from a higher-level health facility by a CHW and to estimate caregivers' willingness to pay (WTP) for referral. Methods: Caregivers of 203 randomly selected children referred to higher-level health facilities by CHWs were interviewed in four Midwestern Uganda districts. Questionnaires and document reviews were used to capture direct, indirect and opportunity costs incurred by caregivers, CHWs and health facilities managing referred children. WTP for referral was assessed through the 'bidding game' approach followed by an open-ended question on maximum WTP. Descriptive analysis was conducted for factors associated with referral completion and WTP using logistic and linear regression methods, respectively. The cost per case referred to higher-level health facilities was computed from a societal perspective. Results: Reasons for referral included having fever with a negative malaria test (46.8 %), danger signs (29.6 %) and drug shortage (37.4 %). Among the referred, less than half completed referral (45.8 %). Referral completion was 2.8 times higher among children with danger signs (p = 0.004) relative to those without danger signs, and 0.27 times lower among children who received pre-referral treatment (p < 0.001). The average cost per case referred was US$ 4.89 and US$7.35 per case completing referral. For each unit cost per case referred, caregiver out of pocket expenditure contributed 33.7 %, caregivers' and CHWs' opportunity costs contributed 29.2 % and 5.1 % respectively and health facility costs contributed 39.6 %. The mean (SD) out of pocket expenditure was US$1.65 (3.25). The mean WTP for referral was US$8.25 (14.70) and was positively associated with having received pre-referral treatment, completing referral and increasing caregiver education level. Conclusion: The mean WTP for referral was higher than the average out of pocket expenditure. This, along with suboptimal referral completion, points to barriers in access to higher-level facilities as the primary cause of low referral. Community mobilisation for uptake of referral is necessary if the policy of referring children to the nearest health facility is to be effective.
# Background
Malaria, pneumonia and diarrhoea are among the leading killer diseases among children aged five years and below. Efforts in realizing the Millennium Development Goals are tailored towards tackling the three diseases at community level through integrated community case management (iCCM) [bib_ref] Tackling pneumonia and diarrhoea: the deadliest diseases for the world's poorest children, Gupta [/bib_ref] [bib_ref] Community case management of childhood illnesses: policy and implementation in Countdown to..., De Sousa [/bib_ref]. iCCM involves diagnosis and treatment of malaria, pneumonia and diarrhoea by community health workers (CHWs) in addition to referral of severely sick children, children whose conditions cannot be treated by CHWs, and newborns. However, slow implementation of the approach has been noted in the 75 'countdown' countries which together account for more than 95 % of all maternal, newborn, and childhood mortality [bib_ref] Community case management of childhood illnesses: policy and implementation in Countdown to..., De Sousa [/bib_ref].
One key component of integrated community approaches is a well functional referral system which has the ability to send referred cases to higher-level facilities [bib_ref] Adopting integrated management of childhood illness module at local level in Bangladesh:..., Khan [/bib_ref]. However, not all children referred to obtain care from higher-level facilities access referral treatment, indicating a gap in health services access [bib_ref] Understanding caretakers' dilemma in deciding whether or not to adhere with referral..., Simba [/bib_ref] [bib_ref] Coping with paediatric referral-Ugandan parents' experience, Peterson [/bib_ref]. Previous studies have highlighted the challenges faced by caregivers of children referred, including problematic transport options, lack of money, competing responsibilities, perceived poor quality of care at the health facility and improvement in the child's condition following prereferral treatment [bib_ref] Understanding caretakers' dilemma in deciding whether or not to adhere with referral..., Simba [/bib_ref] [bib_ref] Coping with paediatric referral-Ugandan parents' experience, Peterson [/bib_ref] [bib_ref] Community referral in home management of malaria in western Uganda: a case..., Kallander [/bib_ref]. There is general agreement that the cost of referral must decrease to make referral completion an achievable goal [bib_ref] Coping with paediatric referral-Ugandan parents' experience, Peterson [/bib_ref]. Moreover, the limited ability of first level health facilities to handle severely ill children in sub-Saharan Africa has been documented [bib_ref] Management of severely ill children at first-level health facilities in sub-Saharan Africa..., Simoes [/bib_ref] [bib_ref] Are hospitals prepared to support newborn survival? -An evaluation of eight first-referral..., Opondo [/bib_ref]. There is also a general concern about the high out of pocket expenditure for health care among households in resource-poor settings, particularly health care sought from higher-level facilities [bib_ref] Understanding the impact of eliminating user fees: utilization and catastrophic health expenditures..., Xu [/bib_ref] [bib_ref] Out-ofpocket payments for under-five health care in rural southern Tanzania, Manzi [/bib_ref]. Furthermore, people are often willing to pay (WTP) for western medicine if the cause of disease is perceived to be biomedical [bib_ref] The paradox of the cost and affordability of traditional and government health..., Muela [/bib_ref]. However WTP does not translate into actual expenditure; as the ability to pay is influenced by other factors. One of the factors that further limit low-income earner's ability to pay for biomedical treatment, despite WTP, is the lack of flexibility in how and when fees should be paid, as more flexible payment options would allow time to save or accrue the funds needed [bib_ref] The paradox of the cost and affordability of traditional and government health..., Muela [/bib_ref]. The Uganda iCCM guidelines dictate that children with danger signs seen by a CHW should be referred to the nearest heath centre following pre-referral treatment for suspected severe malaria, pneumonia, dehydration or other danger signs in young children and newborns. Caregivers of children are likely to incur recurrent costs if the first referral centre is not able to handle complicated cases. The mean out of pocket expenditure for a completed referral was estimated to be US$11 in a study conducted in Uganda in 2003, a value that is way too costly for an average caregiver from rural Uganda [bib_ref] Coping with paediatric referral-Ugandan parents' experience, Peterson [/bib_ref].
The evaluation of health care programmes and of the economic burden of disease is most meaningful when examined from a societal viewpoint, considering all costs, regardless of to whom they accrue, i.e. both health system and household costs, however studies with such analysis design are still scanty in low-income countries [bib_ref] Household costs for treatment of severe pneumonia in Pakistan, Sadruddin [/bib_ref] [bib_ref] The economic burden of inpatient paediatric care in Kenya: household and provider..., Ayieko [/bib_ref]. The contingent valuation method, which elicits people's WTP, can be used to measure the value attached to a health service delivery process, even in low-income countries [bib_ref] When do the "dollars" make sense? Toward a conceptual framework for contingent..., O'brien [/bib_ref] [bib_ref] Willingness-to-pay for a rapid malaria diagnostic test and artemisinin-based combination therapy from..., Hansen [/bib_ref]. The stated WTP for referral can therefore estimate a caregivers' valuation of referral and can be used as a basis for financial incentives fostering referral completion. This study used a societal perspective to examine average cost per child referred to a higher-level health facility as well as WTP for referral among the caregivers of the referred children. Such an approach is useful in identifying the brunt of costs borne by child caregivers in relation to service providers and can be used for planning interventions for a functional and integrated community referral system devoid of financial and economic barriers to referral completion.
# Methods
## Setting
The study was conducted in early 2013 in four of the nine mid-western districts of Uganda where iCCM had been implemented since 2010. Based on records from the Uganda health facility inventory, the area was registered to have a total of 276 health centres; 192 government owned, 51 private not for profit and 33 private health centres (HCs), serving a catchment area population of 2.2 million people (18 % aged < 5 years). The nomenclature of the health services in Uganda is based on the services they provide and the catchment area they are intended to serve. In Uganda's health system hierarchy, the lowest level health centre operating at the village level (HC I) is the CHW who works from home. Among the HCs with a physical structure, the lowest level is a HC II located at the parish level that offers outpatient services through a comprehensive nurse, followed by HC III offering in-patient services through clinical officers at the sub-county level. HC IVs located at the county level have doctors who offer inpatient services while hospitals located at the district level offer specialist services. All health centres offer basic health promotive and curative services and staff from higher level HCs supervise their peers at the lower level HCs. Patients with complex disease conditions are referred to either the regional or national referral hospitals offering more comprehensive specialist services. However, alternative health service providers such as private clinics, drug shops and traditional healers, all without designated catchment areas are common in the study area. On average most of the HCs with a physical structure in the study area were level II or III and most villages had at least two CHW trained on delivering iCCM.
According to Uganda's village health team (VHT) strategy, two CHWs (locally referred to as VHT members) in each village should be trained on iCCM in addition to regular health education and promotion activities. They are provided with job aids and are entrusted with the role of diagnosing and treating children between 2-59 months with uncomplicated disease whilst referring those with danger signs, such as convulsions, unconsciousness, chest in-drawing, excessive vomiting and chronic conditions. Additionally, CHWs are expected to offer safe referrals for children they cannot treat, such as sick newborns and children who do not fall within 2-59 months age bracket. In the study area, CHWs diagnosed malaria and 'fast breathing' pneumonia with the help of a simple rapid diagnostic test and respiratory timer respectively. They treated malaria with artementher-lumefantrine combination therapy, pneumonia with amoxicillin and diarrhoea with a combination of ORS and zinc. They were also required to offer pre-referral rectal artesunate to children with signs of severe malaria, artementher-lumefantrine to children with fevers lasting more than 7 days, ORS to children with diarrhoea, and amoxicillin to children with signs of severe pneumonia before referral to the nearest health facility.
## Study design and procedures
Data were drawn from a sample of caregivers of sick children referred to seek care from higher-level healthcare facilities by CHWs. The sample size was limited to children who had been referred within two weeks of the interview. The desired sample size was calculated in relation to the expected mean cost of referral. However, at the time of the study, there were no studies on referral from which mean societal costs could be estimated. There were two existing studies conducted in Uganda addressing higherlevel facility referrals which provided only median estimates of out of pocket expenditures incurred by child caregivers [bib_ref] Coping with paediatric referral-Ugandan parents' experience, Peterson [/bib_ref] [bib_ref] Community referral in home management of malaria in western Uganda: a case..., Kallander [/bib_ref]. In the absence of suitable estimates for societal costs of referral, the sample size was estimated using the best alternative available evidence. Therefore, the sample size was estimated using the mean out of pocket expenditure as a proxy for total costs. The mean out of pocket expenditure was calculated from the dataset of the Kallander et al. 2006 study [bib_ref] Community referral in home management of malaria in western Uganda: a case..., Kallander [/bib_ref]. Using the sampsi command in STATA 12 (StataCorp LP, College Station, TX), assuming a mean of 8950 UGX, a standard deviation of 16,860 UGX [bib_ref] Community referral in home management of malaria in western Uganda: a case..., Kallander [/bib_ref] , a hypothesised mean referral cost of 15,000 UGX in the sample (given current inflation and cost of living), and adjusting for cluster design effects and a loss to follow up of 5 %, a total of 174 referred children was required to detect the mean cost per completed referral equivalent to the hypothesised mean. With the village as the primary sampling unit and assuming uniform distribution, four districts were selected out of nine districts. Twelve sub-counties were selected from a list of all the sub-counties in the four selected districts. Under the assumption that a CHW would have referred at least one child within two weeks of interview, 50 villages were sampled from the list of sub-counties using probability proportional to population size of the sub-county. Referred children were identified through an interview with CHWs from the selected villages. Caregivers of referred children were traced with the help of the CHW and were given a facilitator-administered questionnaire by a trained research assistant. Most parts of the questionnaire were adapted from previous research on referral in integrated management of childhood illnesses and home based management of fever [bib_ref] Coping with paediatric referral-Ugandan parents' experience, Peterson [/bib_ref] [bib_ref] Community referral in home management of malaria in western Uganda: a case..., Kallander [/bib_ref] [bib_ref] Management of severely ill children at first-level health facilities in sub-Saharan Africa..., Simoes [/bib_ref]. One part of the questionnaire assessed socio-demographic characteristics of the referred child and his/her caregiver. Another part of the questionnaire assessed the referral process with special emphasis on the circumstances under which the child was referred, and the steps taken by the caregiver thereafter. Referral completion was operationally defined as receiving care from a recognised iCCM referral health centre supervising the CHWs. Caregivers obtaining care from non-iCCM designated health facilities were classified as incomplete referrals warranting further investigation. The questionnaire had a set of open-ended questions aimed at establishing both direct costs, such as monetary expenses incurred by the caregiver during the referral processes, as well as nonfinancial costs, such time spent on caring for the sick child. In order to establish the importance caregivers attached to following referral advice from CHWs to seek health care from higher-level facilities, contingent valuation methods were used to elicit caregivers WTP for referral. Caregivers were asked about their preferred referral health facilities and the reasons as to why they preferred these health facilities. The 'bidding game' followed by an open-ended question on maximum WTP was chosen as the method of elicitation. This was because the culture of bargaining for services (including health services) is popular in Uganda making it easy for participants to understand the 'bidding game' [bib_ref] Willingness-to-pay for a rapid malaria diagnostic test and artemisinin-based combination therapy from..., Hansen [/bib_ref]. Additionally, the method has been shown to be reliable in eliciting WTP [bib_ref] Valuing the benefits of a health intervention using three different approaches to..., Onwujekwe [/bib_ref]. While assessing the WTP for referral using the 'bidding game' , participants were presented with a scenario where they were supposed to assume they had a sick child and had gone to seek care from the CHW. However, on this particular occasion the CHW was unable to treat their sick child because he or she had noticed that the child had a danger sign, or he or she did not have enough skills to handle the child's symptoms, or he/she had no drugs to treat the child. The CHW then requests the caregiver to seek further treatment from his or her own previously stated preferred health facility. The respondent was then asked if he/she would take his/her child to the preferred referral health facility if it would cost him/ her a total of US$0.5 in costs related to the process of seeking and obtaining full treatment. The starting bidding value of US$0.5 was estimated from interviews with a small sample of participants. If the participant answered yes to the bid, the price was raised to US$1.0 and then US$2.0 etc. on condition that the participant accepted the previous level of WTP. On rejecting a higher bid, the respondent was asked to mention the maximum amount that he/she would be willing to pay for referral . This was done in order to give the respondents an opportunity to give a more specific value of WTP.
The study adopted a societal viewpointwhile estimating the cost of referral. Therefore, a costing matrix capturing all costs (financial and economic) related to management of children referred, regardless of to whom they accrued, was designed. CHW associated costs were captured through a questionnaire that assessed time spent managing children who required referral and resources utilised. The opportunity costs to the volunteering CHWs and the child caregivers for this time were estimated through market price value approaches. In these approaches, the time of volunteers is valued by a relevant market wage rate. Since Uganda is a low-income country that does not have a perfect competitive labour market due to high levels of unemployment, it was important to make assumptions about CHW and caregiver earnings. In the presence of an imperfect labour market, the literature often recommends choosing among minimum wage approaches, average labourer wage and time foregone as reported by the participants [bib_ref] Indirect cost in economic evaluation: the opportunity cost of unpaid inputs, Posnett [/bib_ref] [bib_ref] The amount and value of work time of community medicine distributors in..., Agyei-Baffour [/bib_ref]. The value of time spent by CHWs and caregivers while working with and caring for referred children, respectively, was therefore estimated in terms of foregone monthly income using an average income for rural dwellers as reported in the Uganda National household survey. This is because there is no minimum wage policy in Uganda and the average income foregone could not be estimated by some of the participants in this study due to limitations in the questionnaire. There were limited health facility costing studies conducted in Uganda, the best available evidence was from a study by Medical Sciences for Health (MSH) in which the minimum health care package at health centres in Uganda was costed [bib_ref] The cost of scaling up primary health-care services? comparisons from studies in..., Collins [/bib_ref]. The MSH study used the cost and revenue plus (CORE Plus) analysis tool to model direct and indirect costs associated with delivery of a health services by service protocol. Indirect costs were defined as Bidding game scenario costs that cannot be traced to one particular service. The researchers computed the cost per service delivered by taking the standard unit cost (staff, drugs and medical supplies and lab tests), and loading the indirect and operating costs proportionally across the services. The MSH study however excluded capital costs and depreciation costs since they are not normally part of the recurrent budgets of the health facilities. Therefore, direct and indirect health facility related costs were identified using the estimates available from the MSH study, [fig_ref] Table 1: Summary of type and source of costs explored per level % [/fig_ref] shows a summary of the exhaustive costs included in the matrix as well as the data sources.
# Statistical analysis
The overall objectives of the analysis were to estimate the cost of referral from a societal perspective, and to estimate the median caregiver WTP for referral. Direct and indirect costs associated with referral completion, were inputted into spreadsheets to estimate the average health facility type specific cost per child referred. Direct costs incurred by caregivers and CHW during referral were analysed for distribution. Non-financial costs incurred by both caregivers and volunteering CHWs in form of opportunity costs were computed basing on assumptions described below. According to the study dataset, the average stated monthly income for the caregivers who were predominantly farmers (72 %) was US$129. Most CHWs were also farmers (86 %) with an average monthly income of US$ 95.45. However the average monthly household income for rural areas (US$125.65) from the last Uganda National Household Income Surveywas used for calculating opportunity costs as this was deemed a more representative value. People in rural areas were assumed to work for six days a week for approximately 8 h a day. WTP values were logarithmically transformed and a linear regression model was used to identify the socio-demographic and economic factors associated with stated WTP for referral while adjusting for clustering effects at the village level. Logistic regression was used to identify which children were more likely to complete referral. Socioeconomic status was constructed through principle component analysis of ownership of household items, household construction materials, sanitation infrastructure and means of transportation; variables which are recommended by Uganda National Bureau of Statistics. The socioeconomic status variables that weighted heaviest in the analysis were house construction materials and type of toilet facility. All analyses were done in Microsoft excel and STATA version 12.
# Ethical statement
Written informed consent was obtained from child caregivers and CHWs before study participation. Institutional consent was obtained from Makerere University School of Public Health Institutional review Board and the Uganda National Council of Science and Technology (HS958).
# Results
Data were drawn from a sample of 203 children referred by CHWs to seek further care from the nearest health facility. The reasons for referral included danger signs, drug shortage and complaints of fever in children with a negative rapid diagnostic test for malaria [fig_ref] Table 2: Reasons for referral and reasons for not completing referral [/fig_ref]. Of the 197 traceable children, about half (54.2 %) did not complete referral. Overall, referral completion was significantly higher among children with danger signs relative to those without such signs (adjusted OR = 2.8; 95 % CI 1.4-5.4). Referral completion was also significantly lower among children who received pre-referral treatment compared to those who did not (adjusted OR = 0.27; 95 % CI 0.1-0.5). On the whole, 5.1 % of children were referred to other CHWs, 3.0 % were referred to the private sector and 91.8 % were referred to the public sector. Among the 107 caregivers who did not complete referral for their sick children [fig_ref] Table 2: Reasons for referral and reasons for not completing referral [/fig_ref] , stated referral reasons were no drugs available (37.3 %), fever with a negative malaria test (49.1 %) and danger-signs (21.8 %). Of these, 28 % stated that the child had improved at home, 25 % reported that the health facility was closed, especially in the evenings or on the weekend, 21 % were hindered by the long distances to the health facility, 12 % anticipated a drug shortage at the health facility and 7 % felt that the staff at the health facility could not be trusted.
The mean (SD) and median (range) WTP for referral among caregivers whose children had been referred by a CHW to higher-level facilities were $8.25 (14.70) and $3.92 (0.39-157.17) respectively. Caregivers of children who completed referral had a mean WTP of $9.56 , median $5.89 (0.39-157.17) compared to a mean of $7.13 (12.13) and median $3.92 (0.39-117.87) among those who did not complete referral. The cost was positively associated with provision of pre-referral treatment, referral completion and education level of the caregiver [fig_ref] Table 4: Adjusted predictors of willingness to pay for referral by caregivers [/fig_ref]. Preferred referral sites for stated WTP included public facilities (52 %), private health facilities (45 %) and mission hospitals (3 %). Private facilities were preferred mainly because of having good customer care while public facilities were preferred for having inpatient services and various specialist services.
# Discussion
The integrated community case management (iCCM) strategy relies on referral as a backup for a continuum of care for children who cannot be treated by CHWs. However, less than half of the children referred completed referral in this study. Low referral completion among febrile children with negative rapid diagnostic test has previously been reported in a study from Sierra Leone [bib_ref] Low referral completion of rapid diagnostic test-negative patients in community-based treatment of..., Thomson [/bib_ref]. This study population is similar to that of Sierra Leone in that the fever with a negative rapid diagnostic test for malaria was the predominant cause of referral. A majority of caregivers whose children did not complete referral resorted to giving treatment at home, or went to drug shops and private clinics. This underscores the failure of the current referral policy that relies on referring children to the nearest public health facility when circumstances do not permit community case management.
Among the referred children, those who received prereferral treatment were less likely to complete referral, compared to children who did not receive pre-referral treatment. This is consistent with a study from Tanzania [bib_ref] Understanding caretakers' dilemma in deciding whether or not to adhere with referral..., Simba [/bib_ref] that discussed the dilemmas faced by caregivers following temporary improvement in a child's condition at home. This indicates the need to establish clear referral procedures, including household follow up of children referred by CHWs, which is important in establishing strong referral systems. It is also necessary to ensure that referral messages are communicated effectively by CHWs. Although iCCM stipulates that referral is between CHWs and higher-level health facilities, in a few instances there was CHW to CHW referral, CHW to private clinic referral, and health facility to private clinic referral. This should be scrutinised in the light of potential delays in access to appropriate care that might arise. CHW to CHW referral should only be encouraged if availability of drugs can be verified to the caregiver prior to referral. As much as the role of the private sector in health system strengthening cannot be underestimated [bib_ref] Evaluating health systems strengthening interventions in low-income and middle-income countries: are we..., Adam [/bib_ref] [bib_ref] Working with the private sector for child health, Waters [/bib_ref] , caution should be taken as the mismanagement of patients by drug shops and clinics is widely described in the literature [bib_ref] Private sector drug shops in integrated community case management of malaria, pneumonia,..., Awor [/bib_ref]. WTP methods have been used in low-income country context to establish demand for health services as well as to justify subsidies on commodities [bib_ref] Willingness-to-pay for a rapid malaria diagnostic test and artemisinin-based combination therapy from..., Hansen [/bib_ref] [bib_ref] Willingness and ability to pay for artemisinin-based combination therapy in rural Tanzania, Saulo [/bib_ref] [bib_ref] Willingness to pay for zinc treatment of childhood diarrhoea in a rural..., Akhter [/bib_ref]. Median WTP for referral was higher than average out of pocket expenditure. Median WTP for referral was also higher than average out of pocket expenditure combined with opportunity costs incurred by caregivers. WTP was positively associated with provision of pre-referral treatment, referral completion and education level of the caregiver. The presence of a high median WTP for referral compared to both out of pocket expenditure and opportunity costs combined implies that it is not caregiver attitude towards that is primarily affecting referral completion, but health system factors such as long distances to health facilities. This is backed up by the finding that caregivers of children who received pre-referral treatment were likely to have higher WTP for referral even though they were less likely to complete referral.
Overall total out of pocket expenditure constituted a significant proportion of the total cost among children referred to higher-level health facilities. Previous studies on community referral and referral between health facilities in African contexts have expressed the need to lower the cost of referral [bib_ref] Coping with paediatric referral-Ugandan parents' experience, Peterson [/bib_ref] [bib_ref] Community referral in home management of malaria in western Uganda: a case..., Kallander [/bib_ref] and to improve quality of care at referral facilities [bib_ref] Management of severely ill children at first-level health facilities in sub-Saharan Africa..., Simoes [/bib_ref] [bib_ref] Quality of hospital care for seriously ill children in less-developed countries, Nolan [/bib_ref] [bib_ref] Assessment of inpatient paediatric care in first referral level hospitals in 13..., English [/bib_ref] [bib_ref] Perceived quality of care for common childhood illnesses: facility versus community based..., Nanyonjo [/bib_ref] in order to make referral an achievable goal. Although the overall out of pocket expenditures incurred by caregivers in this study were generally lower than those in previous studies [bib_ref] Community referral in home management of malaria in western Uganda: a case..., Kallander [/bib_ref] child caregivers still absorbed a considerable amount of the referral cost. The lower out of pocket expenditure might be an indication of improved financial and geographical access in iCCM areas since CHWs counsel caregivers to take children to the nearest health facility. The cost per child completing referral increased with the level of health facility. However this variability in cost by health centre level needs to further be scrutinised as they may indicate variability in quality of care provided at the various levels of health facilities [bib_ref] The cost of scaling up primary health-care services? comparisons from studies in..., Collins [/bib_ref].
Study limitations include the dearth of studies presenting the societal costs of referral from which the sample size could be estimated however the best available proxy (out of pocket expenditure) was used to calculate the sample size. There was a possibility of recall bias, which we tried to minimise by using a two weeks recall period. Another limitation is that the study relied on the average national income for rural dwellers to calculate assumptions about earnings of CHWs. However this is justifiable as other methods such as reported income foregone can prove difficult to validate in rural settings [bib_ref] The amount and value of work time of community medicine distributors in..., Agyei-Baffour [/bib_ref]. The health facility costs were extrapolated from a study that excluded capital costs and depreciation costs. We also made assumptions about the time worked by CHWs based on the official working time of a typical Ugandan, which may not be true for all CHWs. The 'bidding game' has been criticised for introducing a starting point bias and the openended technique has been questioned, as patients who are naive about a health care programme may not be able to attach a valid value to it [bib_ref] Willingness to pay: a valid and reliable measure of health state preference?, O'brien [/bib_ref] [bib_ref] The contingent valuation method in health care, Klose [/bib_ref]. The starting price in this study was established through interviews with a small sample of participants. It was subsequently made as low as possible, was the same for all participants and was followed by an open-ended question on maximum WTP. The cost that a person claims to be willing to pay may deviate from what they would actually be willing to pay should they be confronted with the actual situation. However the WTP interview was conducted among only referred children who were given a realistic choice of their preferred health centre of referral.
# Conclusion
There was a suboptimal referral completion rate among children referred by CHWs to higher-level facilities. While caregivers' out of pocket expenditure contributes significantly to the cost of referral, the mean WTP for referral was higher than the average out of pocket expenditure spent by most caregivers. This suggests that if appropriate continuum of care depends on referral to the nearest health facility as the sole support plan for children who cannot be managed by CHWs, then strategies are needed that can increase compliance with referral. Removal of other access barriers, such as transport costs and medical fees, may be necessary.
[fig] Figure 2: A flow diagram of the referral pathway for children referred to higher-level health facilities by community health workers (N = 203). Place of first referral.Place of second referral. Referral completion and pathway for non compliers [/fig]
[table] Table 1: Summary of type and source of costs explored per level %) were referred to a HC II, 81 (44.7 %) were referred to a HC III and 27 (14.9 %) were referred to a HC IV or hospital. Of the children referred to these higher-level facilities, 8 (4.45 %) were referred onwards to another health care facility. From a societal perspective, the average cost per child referred was $4.89 per child referred and USD $7.35 per child completing referral. The average cost per case referred increased with the level of health facility; from US$4.34 to US$6.68 for HC II and HC IV, respectively. Similarly the average cost per case completing referral increased from US$5.89 for HC IIs to US$11.32 for HC IVs(Table 3). Among the children who completed referral, caregiver out of pocket expenditure contributed 34.5 % to the unit cost per referral completed, caregivers' and CHWs' opportunity costs contributed 37.1 % and 3.4 % respectively and health facility costs contributed 26.3 % (Table 3). The overall median out of pocket expenditure was US$ 0 (mean US$1.65, range US$0 to US$12.45). The median out of pocket expenditure for caregivers whose children [/table]
[table] Table 2: Reasons for referral and reasons for not completing referral [/table]
[table] Table 3: Average cost per referred case overall and by health facility level [/table]
[table] Table 4: Adjusted predictors of willingness to pay for referral by caregivers [/table]
|
Epsilon wave disappearance by catheter ablation for ventricular arrhythmia from the left ventricular outflow tract
# Introduction
Arrhythmogenic right ventricular cardiomyopathy (ARVC) is pathologically characterized by fibrofatty replacements. [bib_ref] Arrhythmogenic right ventricular cardiomyopathy, Basso [/bib_ref] It is reported that the subtricuspid/peritricuspid area and the left ventricular inferolateral wall may be the only affected regions in the early stages of ARVC. [bib_ref] Arrhythmogenic right ventricular cardiomyopathy: evaluation of the current diagnostic criteria and differential..., Corrado [/bib_ref] This lipomatous or lipofibromatous composition causes delayed excitation. This late conduction through diseased myocardium is described as an epsilon wave, which is included in the revised Task Force diagnostic criteria. [bib_ref] Diagnosis of arrhythmogenic right ventricular cardiomyopathy/dysplasia: proposed modification of the task force..., Marcus [/bib_ref] Herein, we report a case of ARVC with epsilon wave, in which radiofrequency catheter ablation (RFCA) on left ventricular outflow tract (LVOT) eliminated the epsilon wave.
## Case report
A 26-year-old Japanese man had a mild dilatation and dysfunction of the right ventricle (RV) on a medical checkup and was subsequently referred to our hospital. Cardiac magnetic resonance (CMR) imaging showed regional RV akinesis in the anterior wall, a mild increase in RV end-diastolic volume /body surface area ratio of 107.7 mL/m 2 , and a reduced RV ejection fraction (RVEF) of 41.6%, but without late gadolinium enhancement [fig_ref] Figure 1: Cardiac magnetic resonance imaging showing a mild increase in the right ventricular... [/fig_ref]. Electrocardiography (ECG) showed an epsilon wave and inverted T waves in the V 1 and V 2 leads, as well as an incomplete right bundle branch block (RBBB). Although the 12-lead ECG showed frequent ventricular extrasystoles, the Holter monitoring did not show 500 ventricular extrasystoles per 24 hours. No relatives were previously diagnosed with ARVC or had sudden cardiac death. Endomyocardial biopsy and genetic testing were not performed. Chest computed tomography with contrast showed no abnormalities in the coronary arteries and lung field, nor dilated pulmonary artery. In the Revised Task Force criteria, [bib_ref] Diagnosis of arrhythmogenic right ventricular cardiomyopathy/dysplasia: proposed modification of the task force..., Marcus [/bib_ref] 1 major and 2 minor criteria were satisfied, and the diagnosis of ARVC was confirmed. As the patient did not show left ventricular dysfunction or any history of ventricular tachyarrhythmia, we instructed him to avoid high-intensity endurance exercises and followed him annually without medical or implantable cardioverter-defibrillator therapy. [bib_ref] HRS expert consensus statement on evaluation, risk stratification, and management of arrhythmogenic..., Towbin [/bib_ref] Four years later, the patient complained of palpitation, and the Holter ECG showed an increased burden of premature ventricular complexes (PVC) (10%-16%). Repeat CMR imaging showed a further decrease in the RVEF (29.4%), whereas the RV end-diastolic volume / body surface area ratio was unchanged and late gadolinium enhancement was not observed. As the pharmacological treatment with metoprolol did not reduce the PVC frequency, we performed RFCA.
The 12-lead ECG with PVC revealed a left bundle branch block morphology with a northwest axis deviation and QS
## Key teaching points
Early activation in premature ventricular complexes (PVCs) and a low-amplitude electrogram during sinus rhythm were not revealed in the right ventricular outflow tract.
Radiofrequency catheter ablation targeting frequent PVCs from the left ventricular outflow tract eliminated the epsilon wave.
Epicardially extending myocardial strands affected by the remodeling of arrhythmogenic right ventricular cardiomyopathy may be the cause of the epsilon wave. morphology in the V 1 lead, with notching on the downward deflection, suggesting its origin from the right coronary cusp-left coronary cusp (RCC-LCC) commissure. [bib_ref] Electrocardiographic and electrophysiologic features of ventricular arrhythmias originating from the right/left coronary..., Bala [/bib_ref] Anatomically, the RCC-LCC commissure abutted the midseptal aspect of the RV outflow tract (RVOT); we initially performed mapping in the RVOT, which did not reveal early activation in PVC or low-amplitude electrogram during sinus rhythm. Substrate mapping of the entire RV was not performed. Next, we mapped the aortic root and observed a discrete potential activated 60 ms earlier than the QRS onset in the RCC-LCC commissure, red point). The RF application in this region only decreased the PVC frequency, but did not eliminate the PVC. Further mapping in the LVOT beneath the aortic cusp showed a discrete potential 83 ms earlier than the QRS onset, blue point). Additionally, a delayed activating potential was observed during sinus rhythm, white arrow), although without a late potential. The RF application at this point eliminated the PVC completely. Furthermore, the delayed activating potential almost disappeared. Interestingly, the 12-lead ECG showed a clear alteration of the notching at the end of the QRS complex in leads III, aVf, and V 1 -V 3. Of note, the QRS morphology during sinus rhythm showed complete RBBB because of catheter contact during the mapping in the RVOT. An ECG after RFCA without complete RBBB revealed the epsilon wave disappearance, right panel). CMR imaging after 4 months showed no improvement in RVEF and RV mild dilation. Repeat ECG recordings consistently showed the absence of epsilon waves. The Holter ECG after RFCA showed only a few PVCs (0.2%), and the patient was free from palpitation for ˃6 months, with no epsilon wave reappearance.
# Discussion
To the best of our knowledge, this is the first report showing that ablative therapy from left-side outflow tract modified the epsilon waves.
Epsilon wave is prominent in the right precordial leads and manifests as a low-amplitude potential between the end of the QRS complex and the beginning of the T wave. [bib_ref] Epsilon wave: a review of historical aspects, Perez-Riera [/bib_ref] Epsilon waves are believed to represent delayed activation of the RV free wall, and their presence reflects more diffuse RV involvement. The current case differs in 2 points. First, we observed delayed activation not in the RVOT, but in the LVOT opposite to the septal aspect of the RVOT. Second, epsilon wave was observed in a patient with structurally moderate RV remodeling. With the fact that the delayed potential was observed in the anteroseptal LVOT and the epsilon waves disappeared by ablation on this area, we believe that delayed activated myocardium was present in the epicardial side of the septal LVOT, and that ablative intervention electrically isolated these myocardia or modified the conduction pattern in this area, leading to the disappearance of the epsilon waves. Intracardiac electrograms before and after ablation. Of note, a sharp delayed activating potential (black arrow) during sinus rhythm before ablation almost disappeared after radiofrequency application. C: Twelve-lead electrocardiography during ablation at the LVOT (blue point in. The radiofrequency application eliminated the PVC and changed the notching of QRS waves during sinus rhythm (red arrows).
ARVC is a progressive disease, and patients do not always show all diagnostic manifestations at the initial assessment. Kirubakaran and colleagues 7 reported the early-stage ARVC patients without structural remodeling and that these patients have limited scarring in the epicardial RV outflow. Epsilon waves are known as a maker of the depolarization disorder of ARVC. [bib_ref] Epsilon wave: a review of historical aspects, Perez-Riera [/bib_ref] However, it is of note that ARVC also shows various ECG changes related to the depolarization disorder: terminal activation delay and RBBB, in addition to Epsilon waves. [bib_ref] Electroanatomic correlates of depolarization abnormalities in arrhythmogenic right ventricular dysplasia/cardiomyopathy, Tanawuttiwat [/bib_ref] Actually, in our case, low-amplitude signals in leads V 1 -V 2 were observed at the same timing of the terminal of the QRS complex, and one might define the ECG change as atypical RBBB patterns. [bib_ref] High interobserver variability in the assessment of epsilon waves: implications for diagnosis..., Platonov [/bib_ref] Tanawuttiwat and colleagues 8 analyzed the association between these depolarization disorders and the anatomical location with scarring and delayed activation, in which they showed the latest activated area was located in the basal RVOT and free wall dominantly in the epicardial side. Recent review article by Anderson and colleagues 10 described that myocardial strands extend distally beyond the anatomic ventriculoarterial junction toward the nontubular junction. Importantly, these myocardial strands are opposite to the LVOT in the anteroseptal direction, and therefore we speculate that these epicardially extending myocardial strands were affected by the remodeling of ARVC, showing the epsilon waves in our case.
There have been several reports describing the modification of the QRS complex in ARVC patients. Nogami and colleagues [bib_ref] Changes in the isolated delayed component as an endpoint of catheter ablation..., Nogami [/bib_ref] reported that ablation eliminated isolated delayed components in the intracardiac electrograms, which corresponded to the modification of the signal averaged electrocardiogram. Recently, Lee and colleagues 12 reported a case in which extensive low-voltage area and scar over the RVOT and body were demonstrated, and epicardial ablation suppressed the epsilon wave. On the contrary, Caldwell and colleagues 13 stated that the extensive ablation for RVOT produced local delayed activation and manifested epsilon waves. In contrast to these above cases, it is interesting that our case showed that ablation from anteroseptal LVOT modified the epsilon waves in an ARVC patient with structurally moderate RV remodeling.
[fig] Figure 1: Cardiac magnetic resonance imaging showing a mild increase in the right ventricular end-diastolic volume without late gadolinium enhancement (A and B). Short-axis views in the end-diastole (C) and end-systole (D) showing regional right ventricle akinesis in the anterior wall. [/fig]
[fig] Figure 2 A: : Twelve-lead electrocardiography (ECG) before (left) and after (right) radiofrequency catheter ablation (RFCA). B: Magnified ECG in the V 1 lead before (top) and after (bottom) RFCA. [/fig]
[fig] Figure 3 A: : Activation map for the premature ventricular complex (PVC) in the aortic root and left ventricular outflow tract (LVOT) and intracardiac electrograms. B: [/fig]
|
Applying a Yin–Yang Perspective to the Theory of Paradox: A Review of Chinese Management
This paper examines the paradox in management through the yin-yang lens and how to apply the yin-yang perspective in Chinese indigenous management research through a literature review. Compared with several western philosophical perspectives, yin-yang places more emphasis on the interaction of interdependent contradictions and better describes the complex dynamics of both contradictions. Our review finds that the scholars propose yinyang as a supplement to and an optimization of western management. The review focuses on connecting the yin-yang perspective with the paradox in management, cross-cultural management, and practice of organizational management.
# Introduction
Western modern management theory has influenced the formation of mainstream Chinese management science after it was introduced to China. However, many scholars have found that directly applying western management systems does not fit Chinese enterprises and organizations well. [bib_ref] C' theory: a Chinese philosophical approach to management and decision-making, Cheng [/bib_ref] [bib_ref] The C-theory: on Chinese philosophical approach to decision making, leadership and management, Cheng [/bib_ref] [bib_ref] Toward an integrative framework of indigenous research: the geocentric implications of Yin-Yang..., Li [/bib_ref] [bib_ref] The unique value of yin-yang balancing: a critical response, Li [/bib_ref] [bib_ref] Paths exploration for Chinese indigenous management research, Xie [/bib_ref] For example, western scholars have introduced cooperative governance thinking and discussed building a cooperative relationship between government and non-profit organizations in public service. However, in view of the unique characteristics of China's non-profits, there is a lack of research on the dynamic changes in the opposition and unity between the Chinese government and non-profits. In addition, in cross-cultural organizations in China, the relationship of confrontation and coordination between China and foreign nations is constantly emerging in organizational issues, and the division of power and responsibility is often in dispute. [bib_ref] Integrating Taoist Yin-Yang thinking with Western nomology: a moderating model of trust..., Du [/bib_ref] [bib_ref] The traditional Chinese philosophies in inter-cultural leadership: the case of Chinese expatriate..., Lin [/bib_ref] An important reason for this is the contradiction and divergence between western management philosophy and traditional Chinese philosophy. [bib_ref] Yin Yang: a new perspective on culture, Fang [/bib_ref] [bib_ref] Integration of Chinese and Western philosophies, Li [/bib_ref] Therefore, more detailed and indigenous research, theory, and work are required, rather than the simplistic reproduction of western models based on "assumptions, perspectives and biases that are simply just not applicable or appropriate in China". [bib_ref] Trends in Chinese management and business: change, confucianism, leadership, knowledge & innovation, Rowley [/bib_ref] To fully explain some behaviours and performances of Chinese indigenous organizations, the use of Chinese traditional philosophical thinking is required to modify, supplement, enrich, or even replace traditional western concepts or theories in some cases. [bib_ref] C' theory: a Chinese philosophical approach to management and decision-making, Cheng [/bib_ref] China's indigenous organizational management research needs to adopt the perspective of localization, use local management philosophy and thought, and integrate it with western culture and management thought so as to adopt its strengths and compensate for its weaknesses. [bib_ref] The dilemma and access of Chinese indigenous management research, Lu [/bib_ref] With contradictions in demands emerging in organizational management, an increasing number of western scholars have analysed the organizational environment of globalization, and dynamic and diversified competition from the perspective of paradox in management. [bib_ref] Toward a theory of paradox: a dynamic equilibrium model of organizing, Smith [/bib_ref] The western paradox in management is similar to, but very different from, the Chinese traditional yin-yang. [bib_ref] Exploring paradox: toward a more comprehensive guide, Lewis [/bib_ref] [bib_ref] Changing Chinese values: keeping up with paradoxes, Faure [/bib_ref] [bib_ref] Global implications of the indigenous epistemological system from the East: how to..., Li [/bib_ref] [bib_ref] Management paradox and Yin Yang balancing, Su [/bib_ref] By contrast, eastern philosophies stress the need to avoid simplistic distinctions. [bib_ref] Exploring paradox: toward a more comprehensive guide, Lewis [/bib_ref] Thus, we propose a paradox in management through the yin-yang lens and reveal how to apply the yin-yang perspective to Chinese indigenous management research through a literature review. Can yin-yang philosophy provide a new conceptual frame of reference or perspective to supplement the western management system as well as theoretical support for the indigenous research of organizational management, and, if it does, how?
## Paradox in management
Paradox has many conceptual meanings in western academic circles, which are commonly used in the fields of traditional philosophy and psychology. [bib_ref] Exploring paradox: toward a more comprehensive guide, Lewis [/bib_ref] In traditional western philosophy, philosophers regard human existence as paradoxes, such as life and death, and good and bad. Aristotle, Hegel, and other traditional philosophers described paradox as irrational and unsolvable problems or double constraints. [bib_ref] Adding complexity to theories of paradox, tensions, and dualities of innovation and..., Smith [/bib_ref] Psychologists have emphasized the cognitive nature of paradox to study its impact on creativity and mental health. [bib_ref] Exploring paradox: toward a more comprehensive guide, Lewis [/bib_ref] With more conflicting demands, opposing perspectives, or seemingly illogical findings emerging in western organization management, an increasing number of western scholars have analysed the global, dynamic, and competitive organizational environments from the perspective of paradox in management.Ford and Backoff define paradox in management as "some "thing" that is constructed by individuals when oppositional tendencies are brought into recognizable proximity through reflection or interaction". When trying to understand the ambiguous and dynamic world, reality is often simplified as a polarization of either/or, which hides complex mutual relations.A paradox can express all kinds of contradictory but interwoven elements. believes that organizations need to balance innovation and existing capabilities effectively. If the organization does not innovate, there will be no competition prospect, yet excessive innovation will reduce profits. Only by finding a balance between the two can an organization be stable and efficient. [bib_ref] Exploration and exploitation in organizational learning, March [/bib_ref] Similarly, Lewis believes that the root of paradox lies in the simultaneity of the two poles, though the cognitive nature of the two poles often conceals the coexistence of conflicts. In contrast to either/or choices, paradoxical tensions signify two sides of the same coin, which is a whole that exists at the same time. [bib_ref] Exploring paradox: toward a more comprehensive guide, Lewis [/bib_ref] When western management scholars analyse the paradox in management, they tend to reveal the differences and opposites of the dominant bipolar nature or manifestation of contradiction.
## Yin-yang: a chinese perspective
Yin-yang is a Chinese philosophy and lifestyle that embraces paradox, dynamics, and change.It involves three tenets: dialectical, holistic, and dynamic. In Chinese philosophy, the union of yin and yang is a state of equilibrium. [bib_ref] Bian (change): a perceptual discourses of I Ching, Chen [/bib_ref] The yin-yang model captures the idea that there are no opposites-there is no absolute borderline between black (yin) and white (yang); a dot of yin exists in yang, and a dot of yang also exists in yin. [bib_ref] Yin Yang: a new perspective on culture, Fang [/bib_ref] Displaying the dynamic interaction between two opposite sides, the yin-yang model [fig_ref] Figure 1: Model of yin-yang [/fig_ref] embraces the coexistence of opposite elements in the same environment. [bib_ref] The unique value of yin-yang balancing: a critical response, Li [/bib_ref] Yin and yang form a dialectical balance that reflects the natural and organic core of existence, in which forces are complementary. [bib_ref] Toward an integrative framework of indigenous research: the geocentric implications of Yin-Yang..., Li [/bib_ref] In the yin-yang model, yin represents feminine forces, such as the moon, water, darkness, passivity, intuition, softness, contraction, and yielding in the universe, while yang represents masculine forces, such as the sun, fire, light, activity, rational thought, hardness, expansion, and assertiveness. [bib_ref] Chinese communication characteristics: a yin-yang perspective, Fang [/bib_ref] Thus, although the S line divides the circle into two sides, the dark and white areas are interdependent and interpenetrative. [bib_ref] Intercultural communication influence of confucius institutes: a Yin-Yang perspective, An [/bib_ref] The small eye in the white and dark areas demonstrates the interwoven and dynamic nature of the yin and yang forces, respectively. [bib_ref] Intercultural communication influence of confucius institutes: a Yin-Yang perspective, An [/bib_ref] Scholars in the field of traditional Chinese medicine often interpret yin-yang as a philosophical theory. However, strictly speaking, yin-yang is not a speculative philosophy of western rationalism. In this study, the concept of yin-yang does not constitute a theory in academic research-there is no formulation of yin-yang theory. In modern academic research, yin and yang are essentially a concept and mode of thinking, so scholars like to call it yin-yang thinking or frame 3-5 or yin-yang perspective. [bib_ref] Yin Yang: a new perspective on culture, Fang [/bib_ref] [bib_ref] Chinese communication characteristics: a yin-yang perspective, Fang [/bib_ref] Based on this, yin-yang thinking is defined as a dynamic, dialectical, comprehensive, holistic, and harmonious way of thinking. [bib_ref] Bian (change): a perceptual discourses of I Ching, Chen [/bib_ref] It has three main characteristics. First, it emphasizes the relationship of opposition and complementarity, that is, yin and yang may be two sides of a contradiction but the two aspects of contradiction or opposition are not absolute; rather, they are interrelated, and they feel and push each other. [bib_ref] Adding complexity to theories of paradox, tensions, and dualities of innovation and..., Smith [/bib_ref] Second, competition exists in rival forces. [bib_ref] Bian (change): a perceptual discourses of I Ching, Chen [/bib_ref] The opposite movement of yinyang is mutual dependence, which needs to see the alternate development of the two sides from the perspective of change and development. The third is the pursuit of harmony and balance. As a whole, the universe is antagonistic to yin-yang, which leads to conflicts. However, the ultimate ideal state is the great harmony between the two sides of the contradiction,forming a stable and sustainable development. The yin-yang perspective is the thinking and analysis of dynamic duality based on yin-yang thinking. It emphasizes that all things are each composed of one yin and one yang, which are not only contradictory but also complementary. The yin-yang perspective seems the best choice to coordinate and integrate the opposing views to regard all disputes or differences as two sides of yin-yang, so as to avoid prejudice in any aspect. [bib_ref] Toward an integrative framework of indigenous research: the geocentric implications of Yin-Yang..., Li [/bib_ref] In recent years, the wellspring of Chinese philosophical thoughts, the yin-yang philosophy has been widely discussed in Chinese and international academics, particularly those related to medical science, information science, social science disciplines and cross-disciplinary studies. For example, Langevin et al 26 develop and evaluate a method for quantitative evaluation of yin and yang scores in human subjects for the purposes of traditional Chinese medicine research; Zhang and Zhang [bib_ref] YinYang bipolar logic and bipolar fuzzy logic, Zhang [/bib_ref] propose bipolar fuzzy sets in the field of information science to reveal a yin-yang fusion of two interactive fuzzy subsystems. This paper mainly focuses on the related aspects of yin-yang in the field of management, and tries to explore the essential relationship between paradox in management and yin-yang from the philosophical root.
## Paradox and yin-yang
There have been some scholars comparing the western paradox framework and Chinese yin-yang to some extent (eg, Chin et al. [bib_ref] Confucian business model canvas in the Asia Pacific: a Yin-Yang harmony cognition..., Chin [/bib_ref] Li et al.Redding;Li; [bib_ref] Global implications of the indigenous epistemological system from the East: how to..., Li [/bib_ref] Luo and Zheng 31 ). However, we try to systematically put forward the common characteristics and essential differences between paradox and yin-yang. The concept of the binary paradox is similar to the "unity of opposites" emphasized by yin-yang thinking in traditional Chinese philosophy. However, unlike yin-yang, the paradox in management lacks the dynamic characteristics of "competition in rival forces" and "mutual transformation". [fig_ref] Figure 2: Western philosophical perspectives and yin-yang in the study of management paradox [/fig_ref] is a comparison of several western philosophical perspectives and yin-yang thinking in the study of paradox in management.
In the study of paradox in management, discussions on some western traditional philosophical perspective are similar to those on the yin-yang perspective, but there are also some essential differences. From Aristotle's logical point of view, the two sides of the contradiction are completely independent, without development and movement. This is an isolated and static either/or way of thinking. [bib_ref] Toward an integrative framework of indigenous research: the geocentric implications of Yin-Yang..., Li [/bib_ref] Hegel criticized Aristotle's metaphysics and put forward the dialectics of internal contradiction, which states that everything is self-contradictory.The use of Hegel's dialectics to study paradox in management emphasizes the unity of opposites, which is the logical thinking of both/ and. This is consistent with the basic characteristics of opposition and unity of yin-yang thinking. However, Hegel put forward the concept of aufheben of the two opposite sides, believing that the contradiction between new and old contradictions is the inevitable law of the development of things, in which process contradictions will eventually be eliminated; 32 this is different from the transformation of yin-yang thinking for two opposite sides. From the perspective of yin-yang, there may be two situations in which the two opposite sides are "competition in rival forces", similar to Hegel' s aufheben. When the contradiction between the two opposite sides
## Dovepress
## 1593
intensifies to a certain extent, there will be a revolutionary change in which one side attempts to eliminate the other side,that is, "Qi hui nai wang" mentioned in the book of I-Ching. Second, while the two opposite sides were more through negotiation in the process of "competition in rival forces" in the struggle, not all but a part of them was transformed, resulting in "yang flourishing and yin declining" or "yin flourishing and yang declining", thus moving towards the harmonious ideal state of "middle" and "great harmony". [bib_ref] A Yin-Yang theory of human communication, Chen [/bib_ref] Bohr [bib_ref] The quantum postulate and the recent development of atomic theory, Bohr [/bib_ref] proposed the application of the "complementarity principle" to explain quantum mechanics and used the yin-yang diagram to express the characteristics of the complementary principle. His complementary principle is similar to yin-yang thinking, which is that the opposite sides repel and complement each other. Both emphasize the opposition and unity of contradiction, and think that the two sides are contrasting and complementary. However, similar to the logic of most western dualism, the principle of complementarity is still based on either/or in epistemology, emphasizing that the opposite poles are exclusive and cannot be integrated. [bib_ref] The quantum postulate and the recent development of atomic theory, Bohr [/bib_ref] This is different from the idea that the opposite sides of yin-yang thinking emphasize mutual penetration and can form a harmonious and unified whole. 17 Generally speaking, western dialectics, represented by Aristotle and Hegel, is a contradictory dialectics focusing on the dichotomy of matter and self and the development of rational logic. Chinese traditional dialectics, represented by Lao Tzu, Confucius and the book of I Ching, is a harmonious dialectics focusing on the unity of man and nature and the development of perceptual intuition. The main difference between the two is that the contradictory dialectics emphasizes the binary opposition of phenomenon and essence, subject and object, material and spirit, with logic as the core; the harmonious dialectics emphasizes the life movement process of human and nature and the harmonious process of opposition and mutual transformation of yin-yang, with perception as the basic. Traditional Chinese philosophy places more emphasis on the interaction of interdependent contradictions. [bib_ref] Bian (change): a perceptual discourses of I Ching, Chen [/bib_ref] In the first sentence of Tao Te Ching, Lao Tzu mentioned that "Tao that can be spoken of is not the true one". It seems that the inexplicable eternal Tao referred to here is consistent with the "irrational complex ambiguity" mentioned in the western paradox. [bib_ref] Exploring paradox: toward a more comprehensive guide, Lewis [/bib_ref] But "Dao begets one, one is two, two is three, and three is all things" in yin-yang philosophy further interprets the dynamic interoperability of both sides of the contradiction, which is usually displayed as either/or in the western paradox in management. Traditional Chinese philosophy with yin-yang thinking as the main research perspective can better describe the complex and changeable duality as well as the complex dynamics of both contradictions than the western paradox theory. [bib_ref] Exploring paradox: toward a more comprehensive guide, Lewis [/bib_ref] In organizational management, the key is how to effectively deal with these paradoxes and unify these paradoxes and contradictions in the balance of dynamic development. This dynamic balance implies "the unity of opposites", "the ceaseless change of production and reproduction", or "middle" and "great harmony" in yin-yang philosophy. In summary, the literature review finds that yin-yang provides a new perspective and a more appropriate indigenous method for the western paradox.
## Yin-yang perspective in organizational management
As a classical traditional Chinese philosophy, yin-yang thinking or perspective has been actively explored and applied to research and practice by contemporary scholars. However, the application of yin-yang in the field of management is still in the exploratory stage. The introduction of the Western modern management theory has gradually helped form the mainstream Chinese management system. However, there are contradictions and differences between traditional Chinese philosophy and western management philosophy. Applying the western management system directly may not fit Chinese organizations well. [bib_ref] C' theory: a Chinese philosophical approach to management and decision-making, Cheng [/bib_ref] Therefore, to fully explain some Chinese organizations' behaviour, we should use the yin-yang perspective to modify, supplement, enrich, or even replace western traditional concepts or theories. The yin-yang perspective can be used as a macro research paradigm in the east to supplement or replace the western either/or logic. [bib_ref] Toward an integrative framework of indigenous research: the geocentric implications of Yin-Yang..., Li [/bib_ref] [bib_ref] Yin Yang: a new perspective on culture, Fang [/bib_ref] Therefore, applying the yin-yang perspective is more in line with the characteristics of Chinese indigenous research. Yin-yang is very important for improving the western one-sided, static, linear either/or logic and the absolute binary separation theory. [bib_ref] Toward an integrative framework of indigenous research: the geocentric implications of Yin-Yang..., Li [/bib_ref] [bib_ref] The unique value of yin-yang balancing: a critical response, Li [/bib_ref] [bib_ref] Paths exploration for Chinese indigenous management research, Xie [/bib_ref] As yin-yang is especially useful for theorizing in Chinese organizations, an increasing number of Chinese scholars have used this philosophy in research (see . On the one hand, domestic scholars in interdisciplinary fields such as management and philosophy try to use traditional Chinese philosophy to adjust and correct the "acclimatization" of western management theory system to Chinese enterprises. [bib_ref] Paths exploration for Chinese indigenous management research, Xie [/bib_ref] [bib_ref] Management paradox and Yin Yang balancing, Su [/bib_ref] On the other hand, many overseas Chinese scholars believe that traditional oriental philosophy, especially the yin-yang perspective, is particularly suitable for Chinese transnational organizations. [bib_ref] A Yin-Yang model of organizational change: the case of CBG, Jing [/bib_ref] [bib_ref] Balancing formality and informality in business exchanges as a duality: a comparative..., Lin [/bib_ref] proposed that yin-yang thinking can regard the opposition between globalization and localization as a dynamic dualism to avoid any possible onesided bias. Li 17 proposes a duality map tool of yin-yang and believes that dynamic dialectical yin-yang thinking is more suitable for paradox management than the static dialectical logic of either/or and both/and. In the binary diagram of organizational paradox management, the relationship between social value and organizational profit, exploration and utilization, competition, and cooperation can provide better solutions from the perspective of the balance of yin-yang. [bib_ref] Global implications of the indigenous epistemological system from the East: how to..., Li [/bib_ref]
## Connecting the paradox in management and yin-yang
In studying the paradox in management, western scholars found that the dichotomy of traditional western philosophy is too simple and directly shows the management paradox as a conflict between two poles, when the use of Chinese yin-yang thinking or perspective can better explain and analyse its complexity and dynamic multilateralism at a deeper level. [bib_ref] Exploring paradox: toward a more comprehensive guide, Lewis [/bib_ref] believes that compared with Hegel's dialectics, the differences in yin-yang thinking lie in integrity, conditional mutual transformation, and the permanence of contradictions. In traditional western philosophy, Aristotle's either/or and Hegel's both/and are often used as the core of dualistic logic, while the dualistic framework of traditional Chinese philosophy takes either/ and of the balance of yin-yang as the core of logic, emphasizing the interaction and harmonious unity of the two opposites. However, by criticizing Perter P. Li's either/ and logic, [bib_ref] Toward an integrative framework of indigenous research: the geocentric implications of Yin-Yang..., Li [/bib_ref] Li 37 believed the both/and logic is an indispensable element of yin-yang, which cannot adequately describe the diversity of epistemological expressions and methodological prescriptions derived from yin-yang thinking. Li believed the structure of Hegel's dialectical sublation in paradox is almost same as that of the change process of yin-yang. [bib_ref] Can Yin-Yang guide Chinese indigenous management research?, Li [/bib_ref] Although the essence of yin-yang is similar to the paradox in the west, [bib_ref] Chinese strategic thinking on competitive conflict: insights from Yin-Yang harmony cognition, Chin [/bib_ref] there are significant differences in determining priorities and implementing decisions in the field of strategic management.Cheng 1 put forward the C theory, pointing out that many western scholars only emphasized the overall culture and style of Chinese management system and psychological characteristics but did not explore the deep ideological and philosophical basis of Chinese management system and style. C stands for "creativity, change, culture, Chinese tradition and Confucianism and their creative". [bib_ref] The C-theory: on Chinese philosophical approach to decision making, leadership and management, Cheng [/bib_ref] integrates humanized management and rational management to form the C principle to guide the practical development and implementation of Chinese management.
## Cultural research of organizational management and cross-cultural management based on the yin-yang perspective
In cross-cultural management and cultural research of organizational management, some scholars use the yinyang perspective to highlight the dynamic, complementary, and holistic perspectives. Fang 8 was the first to propose the use of the dialectical dynamic method to analyse culture, which is quite different from the mainstream bipolar paradigm analysis in cross-cultural management. Through a comparative analysis of eight paradoxical values in modern Chinese society, namely "guanxi and professionalism, face and self-expression, thrift and materialism, family orientation and individualism, aversion to and respect for law, hierarchy and competence, long-term and short-term orientation, traditional creeds and modern approaches", Faure and Fang 16 point out that contemporary Chinese culture has undergone major cultural changes, but in terms of the thinking process, modern Chinese society remains anchored to the classical yinyang approach. Fang 9 proposed to conceptualize culture as a value orientation with internal contradiction through "situation, context, and time" from the perspective of yinyang. Therefore, there is a paradoxical value in any culture, which is a process of generation, existence, strengthening and complementation, forming the integrity, dynamic, and dialectic of culture. Chen and Miller 39 emphasized that western management culture should not "lead" but "meet" eastern management culture. They believed that Chinese indigenous organizations must learn from both eastern and western cultures and create better things through the integration of the two. Sometimes it is necessary to integrate opposites, which must also become part of the dual cultural orientation with "social good and self-interest, trust-based and legal relationships, teamwork and individual achievement, risk taking and caution, business and society, locally sensitive and foreign". [bib_ref] West meets East: toward an ambicultural approach to management, Chen [/bib_ref] Pauluzzo et al 40 described a cross-case analysis on yinyang balancing [bib_ref] Global implications of the indigenous epistemological system from the East: how to..., Li [/bib_ref] and on the yin-yang philosophy of culture with situation, context, and time 9 to find that crosscultural learning can help manage cultural dilemmas and lead to greater effectiveness in specific settings. Chen [bib_ref] Zhong (centrality), self-competence, and social/communication competence: a Chinese perspective, Chen [/bib_ref] and An et al [bib_ref] Intercultural communication influence of confucius institutes: a Yin-Yang perspective, An [/bib_ref] proposed from the perspective of organizational communication that the dialectical interaction between yin-yang is the basis of the formation of human communication theory, which provides a new research perspective for cross-cultural communication management. Chen 42 proposed a communitarianism model based on the dialectics of yin-yang about cultural differences to guide the balance and integration of individualism and collectivism in cross-cultural management. In the latest literature, Chin et al 28 propose a new framework for a Confucianism business model based on an integrative view of paradox and culture angle, which is connected to the impact of culture in constructing the frame of business models in the Asia-Pacific area. In that model, social legitimacy, institutional enablers, and institutional disablers were put forward in a yin-yang harmony cognition. [bib_ref] Confucian business model canvas in the Asia Pacific: a Yin-Yang harmony cognition..., Chin [/bib_ref] Applying the Yin-Yang Perspective to Guide the Practice of Organizational Management
Applying the yin-yang perspective to guide the analysis of organizational management practice, Li et al 43 carried out a case study based on the five characteristics of yin and yang embodied in the management of a Chinese enterprise -"inaction, emptiness, hardness and softness, moderation and harmony"-and proposed that embedding the thought of yin-yang into the enterprise growth model is important for improving the core competitiveness and benefits of enterprises. In China's local management practice, enterprises applying the philosophy of yin-yang will form a "humanized" management paradigm that combines rules with human nature. 2, [bib_ref] Yin yang and company growth: a case study of a coal company..., Li [/bib_ref] Jing and Van de ven, [bib_ref] A Yin-Yang model of organizational change: the case of CBG, Jing [/bib_ref] through a case study of an enterprise in China, proposed three key aspects related to the concept of yin-yang and organizational change: situation, process, and action. From the yin-yang perspective, they highlighted the role of "potential" and action strategies and the dialectics of "inaction". Lee and Reade 44 tested 97 Chinese employees in three Japanese subsidiaries in China to investigate the leadership and followership dynamic in foreign firms in China. They found yin-yang offers an approach to the leadership and cosmopolitan followership which have a positive effect on employee commitment. Chin et al [bib_ref] Chinese strategic thinking on competitive conflict: insights from Yin-Yang harmony cognition, Chin [/bib_ref] put forward an explanation for the cognitive framework of yin-yang, and expounds how to use it as a meta theoretical tool to describe the contradictory and complementary dynamics of Chinese strategic thinking. They proposed eight paradoxical situations facing Chinese organisations based the yin-yang change and raise four strategic propositions regarding how competing Chinese higher education institutes handle the conflicting yet complementary dynamics in China. Horak and Long 45 analysed the relationship between power and trust in the enterprise organization from the yin-yang perspective, and believed that the relatively natural, mutual integration, and mutual dependence shown by the analysis of yin-yang was more appropriate than that by traditional western management theory in terms of power and trust. Lin et al 7 explained the leadership style of Chinese expatriate managers from the perspective of traditional Chinese philosophy. The leadership style of Chinese expatriate managers is influenced by some factors in Chinese traditional philosophy and foreign cultural context, which is consistent with the view of yinyang balance. 7
## Implications of the study and directions for future research
The positive growth of China's economy during the period of COVID-19 can be displayed and promoted as a successful development experience in management. Therefore, we believe that the Chinese experience in organizational management can also be concerned and used for reference by non-Chinese audiences. Sorting out the yinyang perspective, a significant characteristic of Chinese management is a contribution to the literature research of management disciplines. We do not emphasize that yinyang is superior than western management theory. We propose that yin-yang perspective can supplement the deficiency of western management theory from dynamic and dialectical aspects. The literature research of management disciplines should not only contain the spirit of western scientific management but also absorb the wisdom of Chinese philosophical management. Chinese indigenous organizations, even some other non-Chinese organizations, can learn from both eastern and western cultures and create better things through the integration of the two. According to the review, applying the yin-yang perspective to the theory of paradox in management is well suited for research on Chinese indigenous characteristics. We examine the paradox in management through the yinyang lens and proposes an application of the yin-yang perspective examined in Chinese indigenous management research. Based on a literature review, the study emphasizes that yin-yang should be part of management theory to guide the practice of organizational management. We believe that our study makes a significant contribution to the literature because it discusses how to improve the practice of organizational management in a cross-cultural background, in the backdrop of an increasing number of Chinese multinationals in world markets and of foreign multinationals in Chinese markets. This study also proposes some directions and implications for future research on the yin-yang perspective by applying it to China's nonprofit organizations, the practice of organizational management in the cross-cultural background, and the awareness of the limitations and deficiencies of the yin-yang perspective.
## Theory and theory development
Compared with previous studies, we find that the Chinese concept of yin-yang emphasizes more on the interaction of interdependent contradictions, and further explains the dynamic interaction between the two sides of the contradiction, which is not deeply involved in the western management paradox. The traditional Chinese philosophy with yin-yang perspective as the main research perspective can better describe the complex and changeable duality, and better describe the complex dynamics of both sides of the contradiction than the western paradox of management. We think that the traditional Chinese philosophy of yinyang provides a new perspective for the western management paradox, which is the optimization and supplement of the western management paradox. Future research on Chinese indigenous management should be a combination of Chinese and western, and while the essential structure is still the modern western management system, it should be supplemented by and improved upon in terms of effect, purpose, and function through the yin-yang perspective.
Whether Chinese culture such as yin-yang philosophy is independent of or complementary to modern western culture will provide ideological and intellectual basis for China's business management.
## Applying yin-yang perspective to non-profit organizations in china
Scholars have proposed traditional theories for non-profit organizations, such as market failure, government failure, and contract failure, using Aristotle's thinking of either/or to assume that non-profits and government are antagonistic. In modern times, many scholars consider that nonprofits and governments can work synergistically, similar to Hegel's both/and dialectics, in ways that meet public needs and benefit both parties.However, the illustration of dualism or exclusive opposites in western literature does not best interpret the relationship between non-profits and the Chinese government. Most existing non-profits in China are directly differentiated from the administrative departments of the government, as a result of the national administrative reform, and takes a "path of generation within the system".From the perspective of "path dependence", these non-profits are bound to be more or less "official", or strongly or weakly dependent on the government. Based on this, some Chinese scholars state that the unique characteristics of China's non-profits is the "dual nature of government and people", which means that the behaviour of the organization is subject to the dual control of administrative mechanism and autonomous mechanism. [bib_ref] Association development and organizational system reconstruction, Wang [/bib_ref] China's non-profits are characterized by the duality of government and people, which makes the relationship between their own development and the government's macro supervision contradictory but often able to effectively deal with the paradox. Therefore, it is important to apply a yin-yang perspective to investigate how and to what extent the relations between the non-profits and the government are based on the former's unique characteristics.
## Practice of organizational management in the cross-cultural background
From the literature on yin-yang in organizational management, in addition to Chinese local organizations, applying the yin-yang perspective is especially suitable for crosscultural organizations with Chinese backgrounds. This includes Chinese organizations abroad and foreign organizations in China. On the one hand, Chinese organizations abroad must face contradictions, coordination, and cooperation in leadership and management between the foreign and Chinese cultures. Many of China's overseas organizations also need to consider factors such as the Chinese government's macro supervision. On the other hand, foreign organizations in China are also facing confrontation, cooperation, and adjustment between Chinese cultures and their own. The market-oriented view that western organizations hold needs to consider local policies, regulations, and relations with the government. The underlying logic of value creation in some Chinese business models led by the government may not be dominated by commercial market logic, but by the ultimate vision to fulfil both social and commercial missions. [bib_ref] Confucian business model canvas in the Asia Pacific: a Yin-Yang harmony cognition..., Chin [/bib_ref] The yin-yang perspective, which embodies opposition, competition, and pursuit of harmony and balance, can better deal with these complex and dynamic relationships.
## Limitation and deficiency of the yin-yang perspective
Li [bib_ref] The danger of Chinese exceptionalism, Li [/bib_ref] and Li et al 29 alert indigenous Chinese management researchers to the danger of Chinese exceptionalism and overconfidence in the belief that yin-yang is "superior to all" other cognitive frames or logical systems in dealing with organizational paradoxes. It is also necessary to integrate western management ideas, take their strengths and compensate for their weaknesses, and adopt a way of combining Chinese and western philosophy. [bib_ref] C' theory: a Chinese philosophical approach to management and decision-making, Cheng [/bib_ref] [bib_ref] Integration of Chinese and Western philosophies, Li [/bib_ref] [bib_ref] Trends in Chinese management and business: change, confucianism, leadership, knowledge & innovation, Rowley [/bib_ref] The traditional thinking of yin-yang has some limitations, such as unclear logical analysis and causal relationships, too general and abstract concepts of yin-yang, and lack of operational management practices and tools. [bib_ref] Management paradox and Yin Yang balancing, Su [/bib_ref] When applying the yin-yang perspective to the analysis of management, we should pay attention not to apply it mechanically and directly, and should not be overconfident that yin-yang is "superior" to all other cognitive frameworks or logical systems in managing paradoxes.The perspective of yin-yang can supplement the deficiency of western management theory from the dynamic and dialectical aspects, and the formal logic of either/or in the west can also make the indigenous research more rigorous and detailed, strengthen the elaboration of abstract and general logic of yin-yang and causality, reduce the use of local research of traditional Chinese philosophy, and simplify the complexity.
# Conclusion
Through a comparison between Chinese and western dualistic dialectics, the yin-yang perspective seems more suitable for paradox management than the static dialectical logic of either/or and both/and. The yin-yang philosophy treats debates or controversies as dynamic dualities to avoid any one-sided biases. However, the application of traditional Chinese philosophy in the field of management is still in the exploratory stage. Scholars have mainly focused on comparing Chinese and western management theories, involving the yin-yang perspective in the cultural research of organizational management and cross-cultural management. In recent years, some scholars have begun to apply the yin-yang perspective to guide the practice of Chinese indigenous organizational management. However, yin-yang has some limitations, such as unclear logical analysis and causal relationships that cannot be applied independently without the western management system. The yin-yang perspective needs to integrate western culture and management. Therefore, in theory, future research on indigenous organizational management should study how to better integrate Chinese and western management systems, that is, by applying western management systems from the yin-yang perspective. In practice, a comparative analysis from the yin-yang perspective based on the characteristics of Chinese organizations in a cross-cultural background would be useful.
[fig] Figure 1: Model of yin-yang. https://doi.org/10.2147/PRBM.S330489 DovePress Psychology Research and Behavior Management 2021:14 1592 [/fig]
[fig] Psychology: Research and Behavior Management 2021:14 https://doi.org/10.2147/PRBM.S330489 [/fig]
[fig] Figure 2: Western philosophical perspectives and yin-yang in the study of management paradox. https://doi.org/10.2147/PRBM.S330489 DovePress Psychology Research and Behavior Management 2021:14 [/fig]
[fig] ••: It should not only contain the spirit of western scientific management but also absorb the wisdom of Chinese philosophical management. • Yin-yang perspective can supplement the deficiency of western management theory from dynamic and dialectical aspects. • Yin-yang thinking needs to be integrated with western culture and management thought. The binary paradox in organizational management is similar to the yin-yang thinking. • Yin-yang perspective is better than the paradox of western management to describe the complex dynamics of both sides of the contradiction. • Yin-yang provides a new perspective and an appropriate way to solve the paradox. Involving yin-yang perspective in the cultural research of organizational management and cross-cultural management Fang 2005, 2012; Faure and Fang 2008; Chen 2008, 2016; An et al 2018; Chen 2018; Pauluzzo et al 2018 • There is a whole, dynamic, and dialectical paradoxical value in any culture. • The dialectical interaction between yin and yang is the basis of the formation of human communication theory • Contemporary Chinese culture has undergone great changes, but the traditional thinking of yin-yang still runs through modern Chinese society. Emphasizing that yin-yang should be included in management theory, and should guide the practice of organizational management Li et al 2011; Du et al 2011; Lu et al 2013; Jing and Van de Ven. 2014; Horak and Long 2018; Lin et al 2018; Lee and Reade 2018; Chin et al 2020 • In China's local management practice, the enterprises using yin-yang thinking will form a "humanized" management paradigm that combines rules with humanity. • The concept of change plays an important role in guiding organizational change in Chinese local enterprises. • Yin-yang is used to guide the balance and integration of individualism and collectivism in cross-cultural management. • As a local study of management, the idea is not just to explore the philosophy of yin-yang but to develop new knowledge of management from the perspective of yin-yang philosophy. https://doi.org/10.2147/PRBM.S330489 DovePress Psychology Research and Behavior Management 2021:14 [/fig]
[table] 1595 Table 1: Psychology Research and Behavior Management 2021:14 https://doi.org/10.2147/PRBM.S330489 Organizational Management Literature from Yin-Yang Perspective: A Summary [/table]
|
A Case of Recurrent Vogt-Koyanagi-Harada Disease Successfully Treated with Adalimumab in Young Female Adult Patient
## A case of recurrent vogt-koyanagi-harada disease successfully treated with adalimumab in young female adult patient
Dear Editor, Vogt-Koyanagi-Harada (VKH) disease is characterized by diffuse, bilateral, granulomatous panuveitis that can be associated with auditory, neurological, and cutaneous manifestations. Early and aggressive systemic corticosteroids are the primary initial therapy; however, despite proper treatment using corticosteroids and immunosuppressants, a number of patients experience recurrent attacks or steroid-associated complications. Adalimumab, an antitumor necrosis factor α monoclonal antibody, is indicated for several inflammatory conditions that may be associated with intraocular inflammation. In this case report, we describe a 26-year-old female with recurrent VKH disease that could not be controlled with a steroid and immunosuppressant. Adalimumab treatment alone was successful in achieving remission without long-term recurrence.
A 26-year-old Korean female presented to our clinic with one week of bilateral blurred vision that had initiated eight months prior. The patient's uncorrected visual acuity was 20 / 22 (0.9) in the right eye and 20 / 30 (0.67) in the left eye. There were bilateral 1-2+ anterior chamber (AC) cells without keratic precipitates. Fundus exam showed a sunset-glow appearance with diffuse Dalen-Fuchs nodules, suggesting VKH disease. Optical coherence tomography showed mild choroidal thickening without subretinal fluid. Systemic workup was non-revealing. After maintenance treatment with oral prednisone (initiated with 30 mg/day and tapered weekly to 20 and 10 mg/ day) and methotrexate 15 mg/wk for 1 month, inflammation resolved, and the patient recovered vision in both eyes. During the follow-up period, the patient suffered frequent and intense relapses of anterior uveitis in both eyes when the dose of prednisone was reduced below 10 mg/day despite maintenance of methotrexate 15 mg/wk. A few years after initiation of systemic corticosteroid and multiple failed attempts of slow tapering, the patient complained about adverse effects such as weight gain, lethargy, depression, and gastrointestinal problems. Despite addition of mycophenolate mofetil 1,000 mg twice a day, the patient had persistent anterior chamber inflammation with posterior synechiaeand progressing chorioretinal atrophy threatening the posterior pole. In the following months, the patient developed tinnitus and vitiligo, indicating the diagnosis of complete VKH.
At that time, we started treatment with adalimumab (Humira; Abbvie, Chicago, IL, USA) 40 mg subcutaneously every 2 weeks, with resolution of anterior chamber inflammation 1 week after the first administration. During adalimumab treatment, itching eczematous lesions were noted over the whole body and were consistent with seborrheic dermatitis; these lesions resolved after 2 weeks of topical steroid cream. Sequential cessations of the oral steroid, mycophenolate mofetil, and methotrexate were possible. The patient has been on adalimumab monotherapy 40 mg biweekly for 15 months and 40 mg monthly for 6 months with best-corrected visual acuity 20 / 20 (1.0) in both eyes without signs of active uveitis or systemic symptoms. The fundus abnormality stopped progressing after the treatment with adalimumab. No other adverse events have been noted during adalimumab treatment.
To the best of our knowledge, this is the first case of recurrent VKH syndrome in Korea to be successfully treated with adalimumab. Adalimumab has been reported to be successful in controlling ocular inflammation with non-infectious uveitis. Couto et al.showed adalimumab to be safe and effective in controlling intraocular inflammation in 14 patients with recurrent VKH disease. Jeroudi et al.reported successful use of adalimumab in a case of refractory pediatric VKH, with rapid resolution of inflammation and favorable visual outcome.
High-dose systemic corticosteroids have been the mainstay of initial treatment for VKH disease. But prolonged use of systemic corticosteroids can lead to complications such as Cushing syndrome, hyperglycemia, adrenal suppression, and psychological problems. Although our patient showed improvement with a high-dose systemic corticosteroid, she had significant difficulty in tapering the steroid and suffered corticosteroid-associated complications. Immunosuppressive agents can be added as adjunctive therapy in refractory VKH or in patients who cannot tolerate long-term corticosteroids. Despite these treatment options, refractory cases have been reported; in the present case, adalimumab treatment was an adequate alternative. In adalimumab treatment in general, complications such as malignancy, inflammatory neurologic disease, opportunistic infections, reactivation of latent tuberculosis, and congestive heart failure may occur. However, in this case, no significant adverse events were observed except mild skin eruptions that were easily controlled with dermatologic treatment.
In conclusion, this case suggests that treatment with adalimumab could be an effective and safe therapeutic op-tion for controlling inflammation and vision preservation in patients with recurrent VKH disease. |
Determinants of quality of shared sanitation facilities in informal settlements: case study of Kisumu, Kenya
Background: Shared facilities are not recognised as improved sanitation due to challenges of maintenance as they easily can be avenues for the spread of diseases. Thus there is need to evaluate the quality of shared facilities, especially in informal settlements, where they are commonly used. A shared facility can be equated to a common good whose management depends on the users. If users do not work collectively towards keeping the facility clean, it is likely that the quality may depreciate due to lack of maintenance. This study examined the quality of shared sanitation facilities and used the common pool resource (CPR) management principles to examine the determinants of shared sanitation quality in the informal settlements of Kisumu, Kenya. Methods: Using a multiple case study design, the study employed both quantitative and qualitative methods. In both phases, users of shared sanitation facilities were interviewed, while shared sanitation facilities were inspected. Shared sanitation quality was a score which was the dependent variable in a regression analysis. Interviews during the qualitative stage were aimed at understanding management practices of shared sanitation users. Qualitative data was analysed thematically by following the CPR principles. Results: Shared facilities, most of which were dirty, were shared by an average of eight households, and their quality decreased with an increase in the number of households sharing. The effect of numbers on quality is explained by behaviour reflected in the CPR principles, as it was easier to define boundaries of shared facilities when there were fewer users who cooperated towards improving their shared sanitation facility. Other factors, such as defined management systems, cooperation, collective decision making, and social norms, also played a role in influencing the behaviour of users towards keeping shared facilities clean and functional. Conclusion: Apart from hardware factors, quality of shared sanitation is largely due to group behaviour of users. The CPR principles form a crucial lens through which the dynamics of shared sanitation facilities in informal settlements can be understood. Development and policy efforts should incorporate group behaviour as they determine the quality of shared sanitation facilities.
# Background
The sanitation target of the sixth Sustainable Development Goal(SDG) is to achieve access to adequate and equitable sanitation for all by 2030. Due to increasing urbanisation and informality, however, providing adequate sanitation in informal settlements is increasingly becoming a challenge [bib_ref] Neighbour-shared versus communal latrines in urban slums: A cross-sectional study in Orissa,..., Heijnen [/bib_ref]. Inadequate household sanitation facilities in informal settlements force residents to share the few available facilities, a practice that some authors have proposed as being the most practical alternative [bib_ref] Communal sanitation alternatives for slums: A case study of Kibera, Schouten [/bib_ref] [bib_ref] Shared sanitation facilities: A reality or mirage?, Kabange [/bib_ref]. In the classification of sanitation facilities, however, the Joint Monitoring Program (JMP) of the World Health Organization (WHO) and the United Nations Children's Fund (UNICEF) does not classify shared sanitation facilities as 'improved' facilities due to concerns related to, among others, cleanliness and maintenance.
In addition to cleanliness and maintenance, studies have also highlighted the importance of aspects such as hygienic status of sanitation facilities, state of the superstructure, presence of smell, presence of flies, and the state of the slab (especially in the case of pit latrines) in defining the quality of shared sanitation [bib_ref] Water, sanitation, and primary school attendance: A multi-level assessment of determinants of..., Dreibelbis [/bib_ref] [bib_ref] Assessing the impact of a school-based water treatment, hygiene and sanitation programme..., Freeman [/bib_ref] [bib_ref] Assessment of latrine use and quality and association with risk of trachoma..., Montgomery [/bib_ref] [bib_ref] Water, sanitation, hygiene and rural poverty: Issues of sector monitoring and the..., Garriga [/bib_ref] [bib_ref] Performance of pit latrines in urban poor areas: A case of Kampala, Nakagiri [/bib_ref] [bib_ref] Why are some latrines cleaner than others? Determining the factors of habitual..., Sonego [/bib_ref]. What is evident from these studies is that quality of sanitation facilities is determined by maintenance practices such as cleaning or lack thereof.
Unclean shared facilities may be due to a number of factors, including inadequate management practices of users. This inadequacy may lead to a scenario where users benefit from using a shared sanitation facility, but put little or no effort into its management. This scenario is similar to "the tragedy of the commons", depicted by :1244), where no one wants to make personal sacrifices for the good of all users. A common good or resource is one that can be utilised by all, but that is not owned by any one user. Every user, therefore, maximises benefits from the good/resource, but the costs are shared by all [bib_ref] Design principles and common pool resource management: An institutional approach to evaluating..., Quinn [/bib_ref]. For such goods, it is difficult to exclude any of the users, yet overexploitation takes away the ability of other users to use the same resource (subtractability) and eventually leads to depletion [bib_ref] Hardin and the commons: A critical appreciation and a revisionist view, Ostrom [/bib_ref] [bib_ref] Taking a common-pool resources approach to space sustainability: A framework and potential..., Weeden [/bib_ref]. Applying this theory to sanitation, it may be difficult to exclude users who benefit from shared facilities, but, overexploitation, such as misuse and lack of cleaning, reduces the ability of other users to use the facility.
To minimise the challenges of common goods, Elinor Ostrom recommends the common pool resource (CPR) management principles, which are elements/conditions that encourage users to work towards a common end of ensuring the sustainability of common resources. They are:
## Boundaries congruence with local conditions appropriation and provision
Collective choice arrangements Monitoring Graduated sanctions Conflict resolution mechanisms Recognition by external government authorities The organisation of these activities in multiple layers of nested enterprises [bib_ref] A review of design principles for communitybased natural resource, Cox [/bib_ref] [bib_ref] A behavioral perspective on the governance of common resources, Janssen [/bib_ref].
These principles may not be applicable in all contexts but they work well in self-governing institutions that require coordination and collective action from users [bib_ref] Studying the commons, governing common-pool resource outcomes: Some concluding thoughts, Agrawal [/bib_ref]. It is thus important to understand the local context within each system [bib_ref] The promise of common pool resource theory and the reality of commons..., Saunders [/bib_ref] [bib_ref] Generalizing the core design principles for the efficacy of groups, Wilson [/bib_ref] [bib_ref] Institutions and the environment, Ostrom [/bib_ref] [bib_ref] Design principles in long-enduring institutions of Japanese irrigation common-pool resources, Sarker [/bib_ref]. In the context of shared sanitation in informal settlements:
A shared sanitation facility can be equated to a scarce resource. Management of the facility is done by the users (appropriators). Quality and continued use of the shared facility depends on the users' management practices.
In urban areas, a household's benefits from sanitation depend largely on the actions of others [bib_ref] Realizing the right to sanitation in deprived urban communities: Meeting the challenges..., Mcgranahan [/bib_ref] [bib_ref] Slum inhabitants' perceptions and decision-making processes related to an innovative sanitation service:..., O'keefe [/bib_ref] and the CPR management principles are a possible avenue to an in-depth understanding of group actions influencing shared sanitation quality. The aim of this study was thus twofold: To examine the quality of shared sanitation facilities in informal settlements, and to use the CPR principles to investigate the determinants of shared sanitation quality.
## Study area -kisumu city
Kisumu is the main city in the western region of Kenya, with a population of approximately 420 000 people. Approximately sixty percent of the city's population lives in informal settlements. These settlements are faced with challenges such as lack of sanitation facilities. Most residents in the settlements are tenants who commonly live in compounds. A compound is a group of several tenant households, living in individual housing units which are all under one landlord. More often than not, these housing units are constructed next to each other and they share a common yard. Compound households also share amenities such as water and sanitation facilities [bib_ref] An enumeration and mapping of informal settlements in Kisumu, Kenya, implemented by..., Karanja [/bib_ref].
Statistics on access to sanitation in the settlements is scanty. An earlier study by Okurut and Charles [bib_ref] Household demand for sanitation improvements in low-income informal settlements: A case of..., Okurut [/bib_ref] revealed that 65% of the population in the settlements have access to 'improved' sanitation (as defined by the JMP). Nonetheless, the study pointed out that most of these facilities did not count as providing sustainable access to basic sanitation judging from indicators such as safety, privacy, dignity, and cleanliness. Common sanitation facilities in the settlements are traditional pit latrines with a few septic tanks [bib_ref] Sanitation policy and spatial planning in urban East Africa: Diverging sanitation spaces..., Letema [/bib_ref]. When these sanitation facilities are available, they are often shared within a compound [bib_ref] An enumeration and mapping of informal settlements in Kisumu, Kenya, implemented by..., Karanja [/bib_ref] [bib_ref] Socio-economic dynamics in slums and implications for sanitation sustainability in Kisumu, Simiyu [/bib_ref] , and as mentioned, are inadequate in cleanliness, privacy, and safety [bib_ref] Socio-economic dynamics in slums and implications for sanitation sustainability in Kisumu, Simiyu [/bib_ref]. Half of the compounds in the settlements lack sanitation facilities, and cases of flying toilets (the practice of defecating in a plastic bag and flinging it away) have been reported [bib_ref] An enumeration and mapping of informal settlements in Kisumu, Kenya, implemented by..., Karanja [/bib_ref]. The lack of sanitation is worsened by geographical conditions in the settlements, as high water tables, loose soils and flash flooding during the rainy season lead to the collapse of pit latrines. The practice of flying toilets/open defecation indicates a lack of sanitation facilities, although it also may be an indication of the dysfunctional and inadequate sanitation facilities which drive residents to open defecation.
# Methods
This study adopted a case study approach. A case study aims at a comprehensive exploration/understanding of the case(s) and its interaction within specific real-world contexts:126,:16, [bib_ref] Qualitative research designs and data gathering techniques, Nieuwenhuis [/bib_ref] :75-76,:66-68, [bib_ref] California: Sage Publishers, Flyvbjerg [/bib_ref] , thus providing answers to the how and why questions:289). A case in this study was a shared sanitation facility, and since a number of sanitation facilities (cases) were to be studied, the study qualified as a multiple case study design :139,:311). The study was limited to sanitation facilities that were shared by at least two households, within Kisumu's informal settlements (the context).
With a case study design, more than one method of data collection is recommended in order to provide a comprehensive exploration:, and for this reason, this study used both quantitative and qualitative methods. The study adopted an explanatory sequential mixed methods design which often begins with a quantitative phase, some initial data analysis, and then a qualitative phase in the same study area. The purpose of the qualitative phase is to further explain the results of the quantitative phase:38,:224, [bib_ref] Best practices in mixed methods for quality of life research, Klassen [/bib_ref] [bib_ref] Using mixed-methods sequential explanatory design: From theory to practice, Ivankova [/bib_ref] [bib_ref] A general typology of research designs featuring mixed ethods, Teddlie [/bib_ref] ,:552).
## Quantitative stage
An initial cross-sectional study was conducted during the dry season between January and March 2014. In order to calculate the sample size, the alpha level was set at 95% and the power at 90% to increase the representativeness of the sample to the population. Based on preliminary findings [bib_ref] Socio-economic dynamics in slums and implications for sanitation sustainability in Kisumu, Simiyu [/bib_ref] , the difference between compounds with sanitation facilities and those without sanitation facilities was 27.8. Similarly, the standard deviation, between those with and without sanitation facilities, was 0.. Two clusters were selected from each settlement.
Since the number of compounds in each cluster was not known, transect walks with community leaders were taken in each cluster in order to estimate the number of compounds. This estimate was then divided by the required sample size from each cluster to determine the sampling interval, which in most cases was three compounds. Compounds were selected if they had a sanitation facility that was shared by households within the compound. Selection of such compounds began from one end of each cluster towards the other end. Data was collected by research assistants who worked in a group of two. In each compound, a household was selected randomly. After identifying the household, assistants established if the adult household head or their spouse was available. The purpose and requirements of the research were then explained to the respondent. If they were willing to participate, they gave their oral consent, after which the interview began.
The data collection tool used was a structured interview guide:212,:344) which had closed-ended questions which the interviewer posed to the respondent. The interviewer completed the tool based on the responses given by the respondent. Respondents were asked questions relating to (among others) the type of residence, the location of the toilet and users of the toilets. After the interview, the shared sanitation facility that was used by members of the compound was inspected using an inspection tool that captured details of construction materials, location of the toilet, a rating of the cleanliness of the facility (from very dirty to very clean), and various components that define the quality of sanitation facilities as highlighted by various studies. These components of quality were hygiene factors, privacy factors and slab factors [fig_ref] Table 1: Quality of shared sanitation facility score sheet [/fig_ref].
For quality assurance and to ensure the validity of data, before beginning the survey the research assistants were trained on objectives of the research, administration of tools, handling respondents, and ethics of data collection. They were also taken through each question in the data collection tools to ensure that they understood not only the meaning of the question but also how to present the questions to the respondents. Such training ensured that all the questions were asked in a standard format. After the training, the tools were pre-tested and any issues that were not clear were rectified. Since the assistants worked in a group of two, one assistant interviewed the respondent and after the interview, the other assistant inspected the shared sanitation facilities. Both assistants and the main researcher (who was also involved in data collection) reviewed the inspection to ensure that they agreed on all aspects of the shared facility.
At the end of the survey, data had been collected from 85 compounds, with all respondents who had been selected consenting to participate. These data were transferred to Stata (v 13) for initial analysis. Some aspects of maintenance of shared sanitation had been raised during the quantitative data collection stage, showing the need to further investigate determinants of shared sanitation quality. Some of these aspects included reasons why some of the shared sanitation facilities were dirty, and how the clean facilities were kept clean. Since such aspects were beyond the scope of the quantitative survey and the data collection tool, a qualitative study was then designed using the CPR perspective.
## Qualitative stage
Driven by the inadequacies of the previous quantitative stage such as little details in answering the 'why' questions, this stage was carried out in December 2014. The design was informed by the characteristics of an explanatory sequential mixed methods design, in which a qualitative phase follows a quantitative phase to further explain the results obtained during the quantitative stage. Data were thus collected from the same settlements and clusters that had been selected during the cross-sectional study, but from different compounds. Just like the quantitative stage, the research assistants worked in a group of two. Compounds and household respondents were selected in a similar fashion as the quantitative stage, and respondents gave their consent for participating in the study. After combing through the cluster, more compounds were selected from neighbouring clusters in order to get more depth, variation and achieve saturation.
In each household, a guided and audio-recorded faceto-face interview was conducted with the adult household head within the compound, after which the shared facilities were inspected using the same inspection tool that had been used in the quantitative stage. One research assistant interviewed the respondent, while the other assistant (and the main researcher) recorded the interview, observed the respondent for any non-verbal communication, made notes, asked for clarification (if needed) and afterwards inspected the shared sanitation facilities within the compound.
The data collection tool was a semi-structured interview guide that had open-ended questions. Such a tool, unlike the structured interview guide that was used during the quantitative stage, allows for probing and clarification of answers, allows the researcher to guide the respondent so that they do not deviate from the main topic, and consequently, can reveal other relevant aspects that might have been missed in the tool [bib_ref] Qualitative research designs and data gathering techniques, Nieuwenhuis [/bib_ref] :87-88). In addition to questions related to the residence type and number of households as in the quantitative stage, the tool had questions on the management of shared sanitation facilities. These management questions were designed using the CPR principles that had been revised to make them applicable to the local context and to sanitation, hence:
a. Boundary definition of users and of the shared sanitation facility. b. Presence (or absence) of management rules/ structures. c. Contribution by individuals to the common good of the shared facility (e.g. cleaning). The interview guide had approximately twenty-three open-ended questions that covered each of these management themes. Interviews lasted at most an hour, depending on the answers given by the respondents. Selection and interviewing continued until the point of 'saturation' when new information was not forthcoming. Saturation in this study was defined by the principle that the sample size ought to be 'large enough' to provide a thick description and support convincing conclusions, but not too large to hinder a thorough analysis:421,425,:162). Selection therefore continued until a total of 40 respondents had been interviewed and the 40 toilets within their compounds also inspected.
To ensure the quality of the data collected, the same research assistants were involved in this second stage of qualitative data. They were again trained on the new data collection tool and its administration. The tools were pre-tested to ensure that the researchers and respondents understood the questions and that the questions were asked in the same format. A pilot study was initially carried out to assess the applicability of the common pool resource management principles to shared sanitation and to design the interview questions. Data collection teams always had a male and female to cater for circumstances when a respondent needed to be interviewed by someone of the same gender.
Overall, this study was strengthened by the use of quantitative and qualitative methods of data collection. Whereas the quantitative data collection methods identified the problem (quality of shared sanitation facilities), the qualitative methods provided an opportunity for a finer explanation of the issues identified during the quantitative stage (reasons explaining the quality of shared sanitation facilities). Having an initial quantitative stage and a follow up qualitative stage in the same settlements increased the sample size, decreased bias and increased the representativeness of the sample to the population.
## Data management and analysis
Quantitative data from all the inspected sanitation facilities were entered in EpiInfo and checked for any errors before transferring to Stata (v 13) for analysis.
Just like previous studies that have calculated the quality of sanitation as a score of the various attributes [bib_ref] Water, sanitation, and primary school attendance: A multi-level assessment of determinants of..., Dreibelbis [/bib_ref] [bib_ref] Assessment of latrine use and quality and association with risk of trachoma..., Montgomery [/bib_ref] , the quality of shared sanitation facilities was calculated as a score, summed from each of the three main factors (hygiene, privacy, and state of the slab). For hygiene and slab factors, if the answer to any of the questions was no, the facility scored 1, otherwise, it scored 0. However, it was the reverse for the availability of a hand-washing facility: 1 if yes, and 0 if no. For privacy-related factors, the score was 1 if the answer to any of the questions was yes, and 0 if otherwise.
To examine the determinants of quality, a standard multiple linear regression was performed with the total quality score as the dependent variable. The independent variables were settlements, the location of the toilet, superstructure and slab materials, toilet users, and number of households sharing a toilet. Two hypotheses were being tested: that poor-quality construction materials of the superstructure and the slab lead to lower quality of shared sanitation facilities; and that more households sharing a sanitation facility result in lower quality of shared facilities.
For the qualitative phase, initial analysis of data began while conducting field work to identify and refine any emergent issues that may have been missed and needed follow up in subsequent interviews. After data collection, all recordings were replayed by the main researcher in order to get an overall understanding of each respondent's story. The interviews were transcribed verbatim in Microsoft Word, and the main researcher then re-read the transcripts. The transcripts were then transferred to ATLAS.ti software. In ATLAS.ti, analysis followed a thematic content analysis approach. The transcripts were first coded based on frequently appearing words or issues (for instance, locking latrines). The codes were then merged into families which were the CPR themes that had been identified a priori (such as defined boundaries of a compound). The themes were then summarised in a matrix, (referred to as the Primary Documents table in ATLAS.ti, and presented as [fig_ref] Table 2: Descriptive summary of findings [/fig_ref] which presented the frequencies of these themes and codes within the shared sanitation facilities. Such a matrix revealed some cases that were 'out of the norm' , commonly referred to as deviant cases. Such cases often prompted the researcher to revisit the transcripts, compare the coding, and relate the cases to the theories in order to obtain a deeper understanding. This process led to finer explanation on possible reasons for the quality of shared sanitation facilities experienced during the quantitative stage. The convergence of the quantitative and qualitative findings was then reconciled at the point of interpretation of the data (analytic or interpretative integration) [bib_ref] Triangulation and integration: processes, claims and implications, Moran-Ellis [/bib_ref] by linking the CPR theory to shared sanitation quality in order to provide a richer discussion.
# Results
# Quantitative results
Tables 1 and 2 present a summary of the study findings. Apart from describing the aspects defining quality in the inspection tool, [fig_ref] Table 1: Quality of shared sanitation facility score sheet [/fig_ref] also presents the total count of facilities that exhibited these quality aspects.
All inspected sanitation facilities were pit latrines shared by averagely eight households [fig_ref] Table 2: Descriptive summary of findings [/fig_ref]. Most of these facilities (64%) were dirty (either slightly dirty or very dirty). Compounds with clean toilets had an average of seven households sharing a toilet, while dirty sanitation facilities had a mean of nine households.
Regression analysis results summarised in [fig_ref] Table 3: Summary of regression results of determinants of quality of shared sanitation facilities [/fig_ref] indicated an inverse relationship between quality and number of household users (p = 0.04; CI-0.22--0.001). Sanitation facilities constructed with a brick superstructure had two scores of better quality compared to sanitation facilities with iron sheets/mud/wood superstructure (p < 0.01; CI 0.91-3.11). Thus the hypotheses that shared sanitation facilities with more people have lower quality, and that poor construction materials lead to lower quality of shared sanitation facilities were accepted. However, for the slab material, there was no evidence to reject the null hypothesis (p = 0.13, CI-0.39-2.99), leading to the conclusion that there was no difference in quality between the slab construction materials. In addition, as shown in [fig_ref] Table 3: Summary of regression results of determinants of quality of shared sanitation facilities [/fig_ref] , regression results indicated no quality difference based on type of residence (tenants only or tenants with caretaker). The R-squared value of the final regression model suggested that the variables in the model explained only 26% of the shared sanitation quality. It was assumed that more factors, i.e. management practices, would further explain the quality of shared sanitation facilities.
## Qualitative results/applicability of cpr principles
Majority of the respondents were female tenants, with a few male tenants and female landladies. Most of these respondents were middle aged, and had children or lived in compounds where their neighbours had children. The CPR principles were exhibited in various ways:
## Boundary definition
Boundaries were demarcated in various ways: Toilets were situated within fenced and/or gated compounds and they were locked with padlocks. In compounds where toilets were locked, each household had a copy of the keys or one key was shared by at least two households. In other cases, keys were situated at strategic positions (e.g. on a wall) where they were accessible to households within the compound. Users acknowledged that toilets were locked to keep intruders, who often left the toilets dirty, away.
"We keep them [toilets] locked because other passersby and people from neighbouring compounds would want to use them." [A Female tenant]
Cases of users losing their keys, which eventually led to the toilets not being locked, were also reported. Such toilets were an easy target for illegitimate users, especially if they were not within fenced or gated compounds. The breaking of padlocks (in order to use the toilets) and stealing of materials used for the construction of the facilities, were also reported.
Often times, dirty facilities were left open for all to use, including members of other compounds. Users from compounds with such dirty and 'open for all' toilets did not feel the need to block outsiders from using their facilities.
"How and why should one prevent outsiders from using such a toilet? It is already too dirty." [A male tenant]
## Cleaning arrangements and rules of use
There were varied cleaning structures or patterns. Defined cleaning structures were commonly in the form of a duty rota, and each household had a specific day(s) when they cleaned toilets. It was not a written rota per se, but rather households followed an order (e.g. arrangement of houses within the compound) that ensured that all users participated in cleaning the toilet. Such structures worked best in compounds with fewer households who had good relations among themselves. Women were mostly responsible for cleaning the toilets.
"I clean on Monday and Wednesday, and the others also have two days of the week when they clean."
[formula] [A female tenant] [/formula]
In some compounds with live-in landlords, the landlords cleaned the toilets without involving tenants.
In other compounds, even with defined cleaning arrangements and rules of use, some users did not perform their cleaning duties as expected. At other times, when the person responsible had cleaned the toilet, other compound members soiled them, which led to other users not carrying out their cleaning responsibilities. A female tenant explained that it was common for other users to soil the toilets after the person responsible had cleaned them. When asked what she did in such a situation, she said:
"If someone else soils it [the toilet], I will ask them to clean it; else I just leave it dirty."
In compounds where cleaning rules and/or management structures were not well defined, toilets were often left dirty and would only be cleaned by any member who volunteered to clean.
"Nobody cares about this toilet … whoever is willing to clean it will do it …" Such toilets remained dirty, often for a number of days, before someone volunteered to clean. When sanitation facilities were left dirty, most respondents mentioned using their neighbour's facilities. A few admitted to using flying toilets or defecating in the open.
Often times, the common cleaning practice was to simply pour dirty soapy water, which had been used to clean clothes, over the toilet slab. Women often volunteered and took on the responsibility of cleaning sanitation facilities in order to protect their children from using unhygienic facilities. One female tenant explained why she cleaned the toilet in their compound every day without relying on anyone:
"I clean the toilet … because I have children … I do not want them to use a dirty toilet."
In most compounds, it was commonly felt that cleaning toilets was a woman's responsibility, and thus men were sometimes exempted from cleaning/management activities. However, in other compounds, all users (both male and female) were required to clean the toilets It was also noted that in some compounds, there had previously been a cleaning and/or management structure that was abandoned when users did not adhere to the rules. When rules were not adhered to, the toilet was either left dirty or would be cleaned by anyone who volunteered.
"We previously had a rota…. but members started complaining…they refused to clean the toilet.
## ….eventually, no one cleans it." [a female tenant]
## Collective decision-making
Meetings were held in some compounds, and all members were required to attend. During such meetings, sanitation issues affecting members were raised and discussed. Such meetings often led to collective decisions and the formulation of rules for the management of sanitation facilities.
## "we, the tenants, held a meeting and all agreed to it [toilet cleaning plan] [a female tenant]
These meetings were easier to coordinate in compounds with fewer households, or in compounds with a leader, such as a landlord or the caretaker (A person appointed by the landlord to be in charge of all tenants in a compound). The leader ensured that all users participated in decisions and that they carried out their duties as agreed upon in the meeting. On the other hand, it was difficult to arrive at a consensus in compounds with many members, even when there was a leader, often due to differences in opinion, uncooperative members, or the unavailability of all members during decision-making meetings.
## Monitoring
Monitoring was done to check on illegitimate users as well as on the condition of the toilets. In tenant-only compounds, one of the tenants would sometimes act as the leader, and in compounds with a live-in landlord, the landlord took up the responsibility. Other compounds had caretakers who took up this role. Monitoring was done in various ways. For example, when residents sat outside their houses during the day, they would easily identify any illegitimate users. In some tenant-only compounds, tenants themselves acted as monitors, a practice which was successful when they had good relations among themselves. A female tenant, who lived in a compound of three households, mentioned that they did not have any one person responsible for monitoring their toilet, but that they all did it together. When asked how they do it, she explained:
"We are all responsible, for example, if someone from a neighbouring compound comes to me asking to use the toilet, my next-door neighbour will not allow them."
Compounds with defined boundaries such as fences and gates needed less monitoring, as it was not easy for outsiders to sneak in and use the toilets; unlike compounds without boundaries.
## Conflict and its resolution
Cases of conflict were reported in instances when users soiled sanitation facilities after they had been cleaned. Conflicts were common among women, especially if children dirtied the toilets and their guardians did not clean them. These conflicts were at times physical fights, disagreements, exchange of words, or quarrels among compound members. At times conflict was experienced in a subtle way, for example, "People sulk at each other, others talk ill of those who dirty the toilet." [A female tenant] Conflicts were resolved in various ways, including discussions with the concerned parties individually or collectively; or involving a third partyoften the leader. It was also noted that without cleaning or management rules, users in compounds with dirty toilets often experienced little or no conflict since no one was in charge of cleaning the toilets.
## Sanctions
In most compounds, sanctions were administered by other compound members or the leader. Reported sanctions included buying new padlocks when keys were lost, or being forbidden to use the toilet if anyone lost the keys. Sometimes a landlord gave un-cooperating tenants a warning when they refused to abide by cleaning rules. If such tenants continued being uncooperative, they were asked to vacate the compound. A male landlord explained that he was very vigilant in ensuring that the tenants in his compound kept the toilet clean. When asked what he did if there were any tenants who did not follow the set cleaning rules, he explained:
"I give the [uncooperative] tenant three warnings, after which I ask them to vacate the compound."
[formula] [A male Landlord] [/formula]
In extreme cases, one landlord explained that when tenants stubbornly refused to abide by the toilet rules in the compound such as not cleaning the toilet after they soil it, they were reported to the local chiefs. [fig_ref] Table 4: Applicability of CPR principles in shared sanitation facilities in Kisumu's informal settlements... [/fig_ref] summarizes the application of the CPR principles in the clean and dirty sanitation facilities.
# Discussion
Defining the quality of shared sanitation facilities is critical if they are to be considered as improved sanitation by the JMP. In this study, shared sanitation quality was measured as a total entity that included the roof, superstructure, as well as hygienic conditions. Most of these indicators were used by other studies [bib_ref] Water, sanitation, and primary school attendance: A multi-level assessment of determinants of..., Dreibelbis [/bib_ref] [bib_ref] Assessing the impact of a school-based water treatment, hygiene and sanitation programme..., Freeman [/bib_ref] [bib_ref] Assessment of latrine use and quality and association with risk of trachoma..., Montgomery [/bib_ref] [bib_ref] Water, sanitation, hygiene and rural poverty: Issues of sector monitoring and the..., Garriga [/bib_ref] [bib_ref] Performance of pit latrines in urban poor areas: A case of Kampala, Nakagiri [/bib_ref] [bib_ref] Why are some latrines cleaner than others? Determining the factors of habitual..., Sonego [/bib_ref] [bib_ref] Seasonal variations and shared latrine cleaning practices in the slums of Kampala..., Kwiringira [/bib_ref] ] that aimed at assessing the hygienic conditions of sanitation facilities. The advantage of using all these indicators as used in this study is that the measure of quality is all-inclusive, and not only focused on hygienic aspects. For example, anyone using a shared sanitation facility would be more comfortable to use one that is not only clean but also offers privacy and shelter from weather conditions such as rainfall.
Results of this study reveal that household shared sanitation facilities were pit latrines, some of which had iron sheet as the superstructure material. Superstructures made out of iron sheets were likely to have a wooden slab and were often not hygienically clean compared to facilities with superstructure made from bricks which had better sanitation quality. Unlike iron sheets, bricks are more durable, they are likely to have fewer crevices, and thus offer better privacy. Moreover, due to a high water table that leads to the collapse of pit latrines in the study area, it is unlikely that one would construct a sanitation facility with a superstructure made of bricks and use poor quality slab material because the toilet easily collapses. Similar findings were reported in Uganda, where latrines with plastered brick superstructures were structurally sound and showed little signs of collapse during the rainy season [bib_ref] Performance of pit latrines in urban poor areas: A case of Kampala, Nakagiri [/bib_ref].
Quantitative results further showed that the facilities performed better in privacy aspects compared to hygiene and slab aspects. A closer look at the scores of hygiene and slab aspects (such as faecal matter and fluids on the slab which can be a public health risk) points to the fact that the facilities were not adequately maintained. Similar results are reported in Western Kenya [bib_ref] Water, sanitation, and primary school attendance: A multi-level assessment of determinants of..., Dreibelbis [/bib_ref] and Tanzania [bib_ref] Assessment of latrine use and quality and association with risk of trachoma..., Montgomery [/bib_ref] where although most facilities had complete superstructures, they were not hygienically clean as evidenced in their hygiene and slab aspects. Better performance of privacy aspects reveals that more attention is usually given to the provision of the 'structure' (hardware), and the poor performance in hygienic conditions points to less attention being given to the behaviour of users. Sustainability in sanitation is not only about the provision of hardware aspects. A great extent of shared sanitation quality is explained by 'soft' factors that are behavioural [bib_ref] Descending the sanitation ladder in urban Uganda: Evidence from Kampala slums, Kwiringira [/bib_ref] and related to maintenance/ management, and thus attention also needs to be directed to users' behaviour and practices.
To further examine the role of sanitation practices, findings revealed that the quality of shared sanitation decreased with an increase in the number of users, similar to findings from informal settlements in Uganda [bib_ref] Descending the sanitation ladder in urban Uganda: Evidence from Kampala slums, Kwiringira [/bib_ref] [bib_ref] Prevalence and determinants of the cleanliness of shared toilets in Kampala slums, Tumwebaze [/bib_ref] [bib_ref] Determinants of households' cleaning intention for shared toilets: Case of 50 slums..., Tumwebaze [/bib_ref]. The relationship between increasing number of users and decreasing quality, especially in informal settlements, is explained by users' behaviour. Studies have related the cleanliness of shared sanitation to individual behaviour [bib_ref] Why are some latrines cleaner than others? Determining the factors of habitual..., Sonego [/bib_ref] [bib_ref] Seasonal variations and shared latrine cleaning practices in the slums of Kampala..., Kwiringira [/bib_ref] [bib_ref] Determinants of households' cleaning intention for shared toilets: Case of 50 slums..., Tumwebaze [/bib_ref] [bib_ref] Shared toilet users' collective cleaning and determinant factors in Kampala slums, Tumwebaze [/bib_ref] such as having the intention to clean sanitation facilities. However, qualitative findings from this study suggest that all users may not share the same intent or have the same attitude. In addition, one individual's effort may not be as productive as a group's effort, and it is, therefore, important to investigate group dynamics and behaviour, especially with shared sanitation. Group dynamics, provide a more holistic picture as opposed to an individual's actions, which may or may not have as significant an effect on quality as the actions of a group would, hence the application of the CPR principles.
One of the CPR principles is boundary definition, and in the context of shared sanitation in informal settlements, it included strategies like locking shared toilets and having the toilets located in fenced compounds to keep intruders away. Such practices have also been reported in informal settlements in Nairobi [bib_ref] Tenancy and sanitation provision in informal settlements in Nairobi: Revisiting the public..., Wegelin-Schuringa [/bib_ref] , Uganda [bib_ref] Seasonal variations and shared latrine cleaning practices in the slums of Kampala..., Kwiringira [/bib_ref] [bib_ref] Descending the sanitation ladder in urban Uganda: Evidence from Kampala slums, Kwiringira [/bib_ref] and India [bib_ref] Sites of entitlement: Claim, negotiation and struggle in Mumbai, Mcfarlane [/bib_ref]. Defining boundaries is crucial because it identifies legitimate users, and it becomes easier to coordinate efforts among the legitimate users. Having a defined user group is also important in defining and implementing management structures and practices like cleaning, collective decision making, and monitoring. A defined user group and management structure do not, however, guarantee that facilities will be in proper hygienic conditions. Results of the current study, as well as studies from Nairobi [bib_ref] Tenancy and sanitation provision in informal settlements in Nairobi: Revisiting the public..., Wegelin-Schuringa [/bib_ref] and Bangladesh [bib_ref] Sanitation and personal hygiene: What does it mean to poor and vulnerable..., Reddy [/bib_ref] , show that some management system such as cleaning rotas break down after some time, implying that there is more that explains the quality of shared sanitation other than the users, defined boundaries and defined management structures.
In addition to users, defined boundaries, and defined management structures; cooperation from and among users is equally vital. Cooperation results when users are in communication, and it ensures that sanitation facilities are kept clean, the defined structures are implemented, monitoring and sanctioning are implemented, and conflicts are resolved. One way of attaining cooperation in groups is through collective decision making. Collective decision making ensures that decisions that are arrived at are favourable to everyone. Qualitative results, for example, indicated that residents who lived in compounds with dirty facilities rarely made decisions collectively. When users are in agreement, it is possible and easier to organise for collective action to ensure that shared facilities are of acceptable quality.
McGranahan [bib_ref] Realizing the right to sanitation in deprived urban communities: Meeting the challenges..., Mcgranahan [/bib_ref] also highlights the importance of collective action in sanitation by noting that when there is collective action, it becomes possible to solve local sanitation challenges.
It is also indicative from this study that management practices are interrelated, each working with another to influence the quality of sanitation facilities. For example, with a defined user group, proper management structures, and collective decision making; monitoring, implementing sanctions, and conflict resolution are easier. These practices were rarely carried out in compounds with dirty sanitation facilities in this study. Therefore, it is possible to effect some of the CPR management practices but still have dirty sanitation facilities.
Such a scenario where some management principles are implemented but shared facilities are not hygienically clean can partly be related to the number of users. Relating back to the issue of numbers, this and other studies [bib_ref] Descending the sanitation ladder in urban Uganda: Evidence from Kampala slums, Kwiringira [/bib_ref] [bib_ref] Prevalence and determinants of the cleanliness of shared toilets in Kampala slums, Tumwebaze [/bib_ref] [bib_ref] Determinants of households' cleaning intention for shared toilets: Case of 50 slums..., Tumwebaze [/bib_ref] show decreasing sanitation quality with increasing number of users. A large number of users is difficult to coordinate, make decisions collectively, and implement rules. It is difficult for people in large groups to trust all other participants, hence making it easier to free ride on the actions of others [bib_ref] A behavioral perspective on the governance of common resources, Janssen [/bib_ref]. Consequently, sanctions may not be easily implemented and effected, and conflict, which is mostly due to the neglect of responsibilities may arise, as reported in India [bib_ref] Sanitation and personal hygiene: What does it mean to poor and vulnerable..., Reddy [/bib_ref] and Ghana [bib_ref] Households' perception of community toilets in low income communities in Kumasi, Appiah [/bib_ref]. On the other hand, it may be possible to have a smaller number of users and also have dirty sanitation facilities. Such a situation may arise when for example, the few users do not have defined boundaries and management structures and do not collectively make decisions.
On the issue of numbers and in line with the discussion on classifying shared sanitation as improved or not, some authors have proposed minimum or maximum numbers allowable for sharing. Gunther et al.'sstudy recommended that facilities be shared by a maximum of four people. Kabange and Nkansah [bib_ref] Shared sanitation facilities: A reality or mirage?, Kabange [/bib_ref] suggest sharing with 2-3 families. One limitation of this study, though, is that it was not possible to establish a threshold number. Although it is clear that shared sanitation quality decreases with increasing number of users, it should be noted that the size of a household is crucial in defining a threshold. For example, compounds may have a number of housing units, which may be occupied by one individual in each unit. Individuals in such a compound may be able to work collectively and keep their sanitation facilities clean. Alternatively, a compound with a similar number of housing units that are occupied by families with an average of five members may have dirty shared facilities. Again, as earlier discussed, it is also possible to have fewer households sharing sanitation facilities but still have unhygienic and dirty facilities. Therefore, focusing on numbers alone is not enough to make recommendations; other determinants beyond the numbers are also important.
Aside from the already discussed CPR principles, another determinant of shared sanitation quality is the importance of good relations among users of sanitation facilities. With good relations, users can, for instance, take on other users' roles in cleaning sanitation facilities, or they are likely to resolve issues amicably thus resulting in less conflict. Good relations between landlords and tenants may also lead to productive management practices, even without the use of sanctions or monitors. For example, some individual tenants participated in the management of shared sanitation because it was a norm that every user should take part in management. Nonparticipation in behaviour such as cleaning can be viewed as 'abnormal' and is part of the reason why landlords were ready to evict non-cooperative tenants. Hence, it is possible that even when sanitation facilities are shared by many households and it may be expected that their sanitation facilities be dirty, there may be social rules and norms that guide users in ensuring that these facilities are kept clean. The CPR literature also proposes that where there are no rules, social ties may reduce conflict and facilitate the development of rules or social norms, which lead to the growth of beneficial behaviour that encourages cooperation [bib_ref] Collective action and the evolution of social norms, Ostrom [/bib_ref] [bib_ref] Interactions between organizations and networks in common-pool resource governance, Agrawal [/bib_ref] [bib_ref] The role of social networks in natural resource governance: What relational patterns..., Bodin [/bib_ref].
Further noted in this study were the aspects of gender and the role played by individual efforts that contributed to the good quality of shared sanitation. Results showed that women were more likely to clean sanitation facilities, rather than men. In addition, they volunteered to clean sanitation facilities because they had young children who would be exposed to the risk of disease. Women have often been responsible for sanitation, including cleaning, also reported among users of shared facilities in Uganda [bib_ref] Gender variations in access, choice to use and cleaning of shared latrines;..., Kwiringira [/bib_ref] and India [bib_ref] Sanitation and personal hygiene: What does it mean to poor and vulnerable..., Reddy [/bib_ref]. However, when users depend on actions of specific individuals and do not put in their own effort, it is possible that such beneficial actions of specific individuals may stop when circumstances change e.g. when children grow up, or when users relocate to other areas. Management of the shared sanitation facility may then end up being no one's responsibility. If only specific individuals participate in a group's common good, the practice may not be sustained long enough to ensure continued use of shared sanitation facilities. Eventually, as noted by Tumwebaze [bib_ref] Prevalence and determinants of the cleanliness of shared toilets in Kampala slums, Tumwebaze [/bib_ref] poor-quality sanitation facilities may not be used and users may resort to practices such as open defecation.
# Conclusion
This study has highlighted that the quality of shared sanitation facilities is not only influenced by hardware aspects but also software aspects. Hardware aspects include construction materials, while software aspects include the behaviour and practices of users. Software factors, which were investigated using common-pool resource management principles, show the importance of group dynamics and practices because they determine the quality of shared sanitation facilities. Shared sanitation facilities should be located where illegitimate users will not have access, and the legitimate users should have a management system that they agree upon collectively. With cooperation, collective action is possible, as the group works together to ensure that sanitation facilities are in good condition. Such an environment also enables the development of social norms that guide other users towards responsible behaviour with regard to shared sanitation facilities. The CPR principles provided useful insights into the complex dynamics of shared sanitation management. Emergent from this study is that, in relation to sanitation in informal settlements, focusing only on numbers may suggest fewer number of users per facility, and consequently more sanitation facilities in informal settlements, both of which may not be feasible. Attention should also be directed at practices that ensure cooperation among users for their common good. Otherwise, 'access' to sanitation does not always mean 'use' of sanitation facilities.
## Policy implications and recommendations for further studies
Policy makers and stakeholders, such as public health departments, should ensure that efforts are not only directed at increasing access but also at ensuring that shared facilities are in useable hygienic conditions. These efforts should involve stakeholders, such as landlords, tenants, and local leaders. As this study suggests, shared sanitation facilities can be kept hygienically clean if there is collective effort from users. Therefore, for policy development, the focus should not only be on number of users sharing a facility but also on the behaviour and practices of the users. With such consideration, it means that if managed adequately, household shared sanitation in Kisumu's informal settlements may be considered as improved sanitation. Development efforts should in addition to the provision of sanitation, also include aspects of safe and hygienic use of shared sanitation facilities and proper disposal of human excreta. Hygienic use of sanitation facilities will ensure that there is sustained use of shared sanitation facilities. Follow-up studies could be carried out to determine the number of users who can share sanitation facilities while ensuring that there is cooperation and coordination amongst them towards a common goal.
# Limitations
The study was carried out during the dry season, and it is possible that the results may be different during the rainy season. Being a case study design, the findings of this study are applicable within the context of Kisumu's informal settlements. This study may then be a basis for comparison with studies (perhaps with larger sample size) from other informal settlements.
[table] Table 1: Quality of shared sanitation facility score sheet [/table]
[table] Table 3: Summary of regression results of determinants of quality of shared sanitation facilities [/table]
[table] Table 2: Descriptive summary of findings [/table]
[table] Table 4: Applicability of CPR principles in shared sanitation facilities in Kisumu's informal settlements a [/table]
|
Simultaneous assessment of cell morphology and adhesion using aluminum nanoslit-based plasmonic biosensing chips
A variety of physiological and pathological processes rely on cell adhesion, which is most often tracked by changes in cellular morphology. We previously reported a novel gold nanoslit-based biosensor that is capable of real-time and label-free monitoring of cell morphological changes and cell viability. However, the preparation of gold biosensors is inefficient, complicated and costly. Recently, nanostructure-based aluminum (Al) sensors have been introduced for biosensing applications. the Al-based sensor has a longer decay length and is capable of analyzing large-sized mass such as cells.Here, we developed two types of double-layer Al nanoslit-based plasmonic biosensors, which were nanofabricated and used to evaluate the correlation between metastatic potency and adhesion of lung cancer and melanoma cell lines. Cell adhesion was determined by Fano resonance signals that were induced by binding of the cells to the nanoslit. The peak and dip of the Fano resonance spectrum respectively reflected long-and shortrange cellular changes, allowing us to simultaneously detect and distinguish between focal adhesion and cell spreading. Also, the Al nanoslit-based biosensor chips were used to evaluate the inhibitory effects of drugs on cancer cell spreading. We are the first to report the use of double layer Al nanoslitbased biosensors for detection of cell behavior, and such devices may become powerful tools for antimetastasis drug screening in the future.Cell adhesion is involved in a variety of physiological and pathological responses, such as cell differentiation, immune response, inflammation, embryogenesis, and tumor metastasis 1-4 . The process of cell adhesion can be divided into three steps, which include cell-substrate contact, cell spreading and cytoskeleton reorganization 5,6 . During the process of cell adhesion, the main observable change is the morphological transition of cells from spherical to flat. The surface membrane proteins involved in adhesion are called cell adhesion molecules (CAMs) and are classified as integrins, immunoglobulins, lymphocyte homing receptors, cadherins and selectins 7 . These CAMs are responsible for the specificity of intracellular binding as well as binding of cells to the extracellular matrix (ECM). As a result, selective binding allows a cell to adhere to a particular tissue, where it may perform essential cellular biological functions in the living organism. After CAMs bind to the ECM, tyrosine kinases such as Src and focal adhesion kinase (FAK) are recruited and activated, stimulating focal adhesion. Activated FAK then triggers the reorganization of α-actin and other structural cytoskeletal proteins, which induces the flattening of cell morphology and completion of cell adhesion 3,6,8 . The process of cell adhesion plays an essential role in cancer cell metastasis 4 , making CAMs potential targets for cancer therapy 9 . However, the influence of adhesion rate on the metastatic potency of cancer cells has not been fully elucidated. Thus, accurate tests of the correlation between metastatic efficacy and cancer cell adhesion may facilitate the development of new strategies for cancer treatment.Several methods can be used to assess cell adhesion. In conventional adhesion assays, colorimetric 10 or fluorometric 11 detection methods have been used to quantify cell adhesion; however, the experimental approaches are time-consuming, insensitive and complicated. Cell morphological changes can also be detected using optical 12 Published: xx xx xxxx opeN Scientific RepoRts | (2019) 9:7204 | https://doi.org/10.1038/s41598-019-43442-wwww.nature.com/scientificreports www.nature.com/scientificreports/ and electrical 13 approaches. Prism-based surface plasmon resonance (SPR) optical sensing systems and electrical biosensors have both proven useful for analyzing cell adhesion14,15. Both methods are sensitive, but optical assays are less invasive to the cells because electrical stimulation can potentially activate FAK and influence cell adhesion16.Localized SPR, generated by metallic nanostructures (e.g., nanoholes, nanorods, and nanoslits) 17 , and prism-based SPR, also known as conventional SPR, can both be used to perform real-time, label-free kinetic monitoring of biomolecules. Conventional SPR detection requires a glass prism and precise control of incident light angle 14 . Moreover, chip-based high-throughput and miniaturized systems are expensive and difficult to set up. In a conventional SPR system, cell morphology changes are detected via changes in the reflective intensity and angle of incident light through the prism. Therefore, the prism blocks the optical path needed for phase-contrast imaging. Because clear phase-contrast images of the test cells are not attainable with conventional SPR, it is not possible to precisely correlate cellular morphology responses and SPR signals from prism-based SPR sensors.To overcome this limitation, we designed and developed a cell adhesion assessment system (CAAS) using an inverted microscope to gather phase-contrast images of the cells and carry out simultaneous surface plasmonic sensing. We used nanostructured SPR chips that allow cell attachment and measurement of transmitted light through the chip. The resonance changes in the transmission spectra were quantified and correlated to observed cell behaviors.In the last two decades, an increasing number of studies have used localized SPR for biomolecule detection or cell behavior analysis 18 . Localized SPR can be generated by metal nanoparticles or periodic metal nanoslit arrays. Although gold is the most commonly used metal in localized SPR-based sensors, nanostructure-based aluminum (Al) sensors, such as nanoconcave arrays 19 , nanoholes 20 and triangular nanoparticles 21 , have been developed because Al is a cost-effective plasmonic material that is useful for short-wavelength surface plasmons. Previously we developed a non-labeling optical method based on localized SPR on gold nanoslit (AuNS) array film to study the influence of anti-cancer drugs or fluidic shear stress on adhesion, detachment, and mortality of adherent cells 22,23 . However, the preparation of such AuNS biosensors is inefficient, complicated and costly. Recently, a sensitive double-layer capped Al nanoslit (CPALNS) biosensor has been used to detect biomolecules 24 . The double-layer nanoslit biosensor showed better sensitivity than monolayer nanoslit sensors 24 . In addition to the previous CPALNS sensor, we also fabricated a novel design with a complementary structure to the CPALNS biosensor, a double-layer grooved Al nanoslit (GOALNS) biosensor. The respective structures at the nanoslit sites of the CPALNS and the GOALNS biosensors are ridges and grooves. Here, we tested the capabilities of the two novel localized SPR-based biosensors to detect cell adhesion properties of a panel of cancer cells.For nanoslit-based biosensors, the intensity change and the spectral shift of the Fano resonance dip/peak are used for molecule sensing 24,25 . In our study, we used resonance wavelength and intensity interrogation methods to evaluate cell adhesion kinetics. Specifically, the cell adhesion rate was determined by the time constant calculated according to the spectral intensity integration (dA), differential spectral peak intensity (dI), dip shift and peak shift of the Fano resonance signals. We then compared the adhesion rates among different types of cancer cells. It has been reported that inhibiting the activity of FAK in human melanoma 26 or human non-small-cell lung cancer cells 27 results in decreased metastatic capacity. In order to test for the correlations between metastatic potency and the adhesion rate among the cancer cells, metastatic human lung cancer and human melanoma cells were evaluated using the CPALNS and GOALNS biosensor chips. In addition, a cell adhesion inhibitor was applied to test whether the cell adhesion sensor chip can detect the effects of a drug on cell adhesion.
# Results and discussion
optical setup of the CAAs and Al nanoslit-based plasmonic biosensors. The configurations of the CAAS and the Al nanoslit-based biosensor chips are shown in Figs 1 and 2. By using the CAAS and the localized SPR-based Al biosensors, we expected to be able to simultaneously observe cell phase-contrast images and perform surface plasmonic sensing during cell adhesion testing. The CPALNS biosensor was shown to be more sensitive for protein detection than an AuNS sensor [bib_ref] Highly Sensitive Aluminum-Based Biosensors using Tailorable Fano Resonances in Capped Nanostructures, Lee [/bib_ref]. Furthermore, a monolayer AuNS sensor has been used to detect cancer cells in human blood [bib_ref] Label-free detection of rare cell in human blood using gold nano slit..., Mousavi [/bib_ref]. The monolayer AuNS sensor shown in [fig_ref] Figure 1: The optical setup and the nanoslit-based sensors for cell adhesion assessment [/fig_ref] was made by a thermal-annealing template-stripping method [bib_ref] Enhancing surface plasmon detection using template-stripped gold nanoslit arrays on plastic films, Lee [/bib_ref] , which is complicated and time-consuming. Moreover, compared to the AuNS sensor, the preparation time and effort for the CPALNS and the GOALNS biosensor were greatly reduced by using hot-embossing nanoimprint and injection molding (please see Materials and Methods). Therefore, these biosensors are more suitable for mass production as a commercial product. In addition, using Al has an additional benefit over using Au.
It has been recently reported that long-range SPR with a longer decay length (500-1000 nm) is more sensitive than the conventional SPR (100-200 nm) and suitable for analysis of large-sized mass such as cells [bib_ref] Selection of invasive and metastatic subpopulations from a human lung adenocarcinoma cell..., Chu [/bib_ref] [bib_ref] Bacterial pathogen surface plasmon resonance biosensor advanced by long range surface plasmons..., Wang [/bib_ref]. Compared to the gold-based SPR sensor, the Al-based SPR sensor has a longer decay length. For the surface plasmon wave propagating on a flat metal surface, the decay length l d (where the amplitude drops to 1/e) is determined primarily by the resonance wavelength λ and can be expressed as follows [bib_ref] Surface-Plasmons on Smooth and Rough Surfaces and on Gratings, Raether [/bib_ref] : where ε m and ε d are the relative permittivities of metal and the adjacent dielectric material, the wavelength dependence permittivity of Al and Au are obtained from previous studies 33, [bib_ref] Algorithm for the determination of intrinsic optical constants of metal films: application..., Rakic [/bib_ref]. In [fig_ref] Figure 2: The fabrication of aluminum nanoslit-based biosensor chips for cell adhesion studies [/fig_ref] , the calculated decay length at the wavelength of 470 nm for Al film is three folds longer than Au film. These studies suggested that Al nanoslit-based biosensors are more sensitive and suitable than the gold sensor for sensing a large mass analyte, such as cells.
[formula] ι λ π ε ε ε = + Im 2 ,(1) [/formula]
www.nature.com/scientificreports www.nature.com/scientificreports/ Design of the plasmonic biosensor chips for cell sensing. The CPALNS4c chip was designed to be used for cell sensing in a microfluidic system. A continuous-flow media supply system was connected to the CPALNS4c chip through the polymethylmethacrylate (PMMA) adaptors [fig_ref] Figure 2: The fabrication of aluminum nanoslit-based biosensor chips for cell adhesion studies [/fig_ref] , thereby enabling long-term observation periods. As shown in [fig_ref] Figure 2: The fabrication of aluminum nanoslit-based biosensor chips for cell adhesion studies [/fig_ref] , the GOALNS25c chip was designed to have an open-well format. The well-to-well distance is 9 mm, which is compatible with that of 96-well microplates. Additionally, the cover lid was designed to prevent reagent cross-contamination between wells. Thus, the chip may be used with automated liquid handling systems for screening of drugs that modulate cell adhesion. These features for chip-based and high throughput label-free detection make the Al plasmonic biosensor chips better than conventional SPR-based biosensors. The inverted microscope can detect cell morphological and spectral changes and is equipped with a motorized stage, camera, spectrometer, transparent heater and a computer control system. (b) (Modified from the microscope instruction manual) The optical path in the system is shown. The spectrum signal was detected from the side port of the microscope. The cell image was obtained from a camera mounted on the trinocular observation tube. The schematic diagram shows the (c) CPALNS and (d) GOALNS biosensors that were used for the cell adhesion sensing in current study. The grey layer indicates the deposition of Al on the nanofabricated PC film (blue) and plate (green).
www.nature.com/scientificreports www.nature.com/scientificreports/ optical properties of the nanoslit-based plasmonic biosensors. Transmission spectra of the CPALNS4c chip [fig_ref] Figure 3: The optical properties of aluminum nanoslit-based biosensors [/fig_ref] and the GOALNS25c chip [fig_ref] Figure 3: The optical properties of aluminum nanoslit-based biosensors [/fig_ref] were measured using our CAAS. In the water-filled chamber, the intensity spectrum of the CPALNS4c chip showed a Fano resonance peak and dip at 615 nm and 645 nm, respectively [fig_ref] Figure 3: The optical properties of aluminum nanoslit-based biosensors [/fig_ref]. When the chambers were filled with air, we observed a peak at 468 nm [fig_ref] Figure 3: The optical properties of aluminum nanoslit-based biosensors [/fig_ref] , which is close to the expected wavelength of 470 nm [bib_ref] Highly Sensitive Aluminum-Based Biosensors using Tailorable Fano Resonances in Capped Nanostructures, Lee [/bib_ref]. For the GOALNS25c chip, specific and obvious dips were observed in the intensity spectrum and transmission spectrum when the chip was in contact with water. Although the transmission spectra represent the feature of the resonance of nanoslit sensors, we used the intensity spectra to analyze the kinetics of cell adhesion. The use of intensity spectra for the analysis simplified the process and the spectral difference could be observed while the artifact from the light source was subtracted. www.nature.com/scientificreports www.nature.com/scientificreports/ The Fano resonance spectrum of the Al nanoslit-based biosensor is comprised of the 3-mode coupling resonance of Cavity resonance, Wood's anomaly and SPR [bib_ref] Highly Sensitive Aluminum-Based Biosensors using Tailorable Fano Resonances in Capped Nanostructures, Lee [/bib_ref]. In the previous study, Fano resonances could be easily modulated in CPALNS sensors by changing the ridge height of nanoslits and the deposited metal film thickness. Depending on the ridge height and the metal thickness, the transmission spectrum could range from a Wood's anomaly-dominant resonance (peak) to an asymmetric Fano profile (peak and dip) or an SPR-dominant resonance (dip). Moreover, the differential wavelength shifts of the localized-SPR peak and dip are determined by the period of the nanoslit sensor [bib_ref] Highly Sensitive Aluminum-Based Biosensors using Tailorable Fano Resonances in Capped Nanostructures, Lee [/bib_ref]. In this study, the transmission spectrum indicates that the Fano resonance of the CPALNS biosensor is an asymmetric Fano profile (peak at 610 nm, dip at 644 nm) [fig_ref] Figure 3: The optical properties of aluminum nanoslit-based biosensors [/fig_ref] , while the GOALNS biosensor shows an SPR-dominant (dip at 638 nm) resonance [fig_ref] Figure 3: The optical properties of aluminum nanoslit-based biosensors [/fig_ref].
The changes of the Fano resonance induced by the cell adhesion in the biosensor chips were further scrutinized. In the CPALNS4c chip, the Fano resonance exhibited a spectral redshift and intensity increase corresponding to the process of cell adhesion [fig_ref] Figure 3: The optical properties of aluminum nanoslit-based biosensors [/fig_ref]. The overall intensity changes together with the spectral shift were used to calculate dA, similar to the calculated parameter used for biomolecule sensing in previous studies [bib_ref] Dynamic monitoring of mechano-sensing of cells by gold nanoslit surface plasmon resonance..., Wu [/bib_ref] [bib_ref] Cell viability monitoring using Fano resonance in gold nanoslit array, Wu [/bib_ref] [bib_ref] Multiplex detection of urinary miRNA biomarkers by transmission surface plasmon resonance, Yeung [/bib_ref]. The composite parameter dA was then correlated with the cell adhesion process. In the GOALNS25c chip, the intensity spectra shown in [fig_ref] Figure 3: The optical properties of aluminum nanoslit-based biosensors [/fig_ref] revealed a significant redshift of the resonance during cell adhesion. Accordingly, the intensity of the resonance peak also increased with time during cell adhesion. As shown in [fig_ref] Figure 3: The optical properties of aluminum nanoslit-based biosensors [/fig_ref] ,f, the shifting www.nature.com/scientificreports www.nature.com/scientificreports/ in resonance dips stopped earlier than the intensity increasing of the resonance peaks in both Al biosensor chips (red line versus blue line). This result implied that the resonance dip and the peak of the Al nanoslit biosensor chips might sense different aspects of cellular behavior.
The sensors reported in this work could monitor molecular event in two different distance ranges. In our earlier study, the average calculated decay lengths were 338 nm for the resonance dip and 892 nm for the resonance peak in the CPALNS biosensor [bib_ref] Highly Sensitive Aluminum-Based Biosensors using Tailorable Fano Resonances in Capped Nanostructures, Lee [/bib_ref]. SPR signals generated by cell adhesion for two different types of sensors with different characteristic penetration depths were previously reported by Méjard et al. [bib_ref] Systematic study of the surface plasmon resonance signals generated by cells for..., Mejard [/bib_ref] Conventional SPR utilizes a single layer of thin gold film for sensing and is characterized by single short penetration dimensions. Alternatively, long-range SPR, which uses an additional fluoropolymer polydecafluoroxaheptadiene (Cytop) layer on the thin gold film, is characterized by long penetration dimensions but still offering single range of penetration. The plasmonic resonance penetration depths of conventional SPR and long-range SPR are 100-200 nm and 500-1000 nm, respectively [bib_ref] Systematic study of the surface plasmon resonance signals generated by cells for..., Mejard [/bib_ref]. Notably, prism-based SPR sensors can only detect a single depth at one time. Our CPALNS biosensor is preferable over prism-based SPR sensors because it is capable of simultaneous sensing with dual penetration (or detection) depth. Similarly, using GOALNS chip, the dynamics of cell adhesion detected by the resonance dip are different from those by the resonance peak. Thus, the GOALNS biosensor can also simultaneously sense molecular events in dual penetration depths. As a comparison, resonant waveguide grating (RWG) biosensor is also capable of cell adhesion assessment 37 , however, the sensor can only detect a single dimension changes. Our biosensors are preferable to prism-based SPR and RWG sensors because they are capable of simultaneous dual detection depth sensing.
The main changes during cell adhesion are related to cell morphological reshaping in a two-dimensional plane and in the longitudinal direction. Since our biosensors produced different signal responses corresponding to long-range and short-range behaviors during the cell adhesion test, the data may provide useful information regarding both the cellular flattening and the extension in the longitudinal direction.
Assessing cell adhesion by fano resonance and cell coverage changes using the CPALNS4c biosensor chip. Both phase-contrast images and the Fano resonance spectra can be simultaneously obtained in our CAAS. We first analyzed the cellular morphology changes in a lung cancer cell line, A549. [fig_ref] Figure 4: The Fano resonances and cell morphological changes during A549 cell adhesion in... [/fig_ref] shows that the round cells attached to the CPALNS biosensor surface became flattened over time. We identified the Fano resonance effective detection area (the red circles) by scanning the field-of-view of the phase-contrast microscope with a pinhole (opening diameter 15 μm, data not shown). The cell morphologies and Fano resonance inside the red circle, which contained approximately ten cells, were used for cell adhesion studies. The time constant of A549 adhesion process was determined by curve fitting the measured cell coverage changes [fig_ref] Figure 4: The Fano resonances and cell morphological changes during A549 cell adhesion in... [/fig_ref]. Meanwhile, the Fano resonance spectrum showed a gradual redshift over time, which stopped at about 60 min after cell seeding [fig_ref] Figure 4: The Fano resonances and cell morphological changes during A549 cell adhesion in... [/fig_ref]. The dynamic change in the cell-sensor contact area was quantified as dA of the Fano resonance spectral shift, a parameter that was also used in our previous studies [bib_ref] Cell viability monitoring using Fano resonance in gold nanoslit array, Wu [/bib_ref] [bib_ref] Multiplex detection of urinary miRNA biomarkers by transmission surface plasmon resonance, Yeung [/bib_ref]. The time constant for the cell adhesion obtained by the cell coverage calculation was similar to the constant obtained from the dA change calculation (39 min and 37 min, respectively) [fig_ref] Figure 4: The Fano resonances and cell morphological changes during A549 cell adhesion in... [/fig_ref]. Yashunsky et al. reported a system to simultaneously collect bright-field cell images and SPR signals in a cell adhesion assay [bib_ref] Real-time monitoring of epithelial cell-cell and cell-substrate interactions by infrared surface plasmon..., Yashunsky [/bib_ref]. However, the cell images are not clear, and it is difficult to identify the edges of individual cells, especially flattened cells. The phase-contrast cell images in our study provide a significant improvement in the image resolution compared to the bright-field images. Therefore, the cell adhesion may be accurately analyzed using from our CAAS. Moreover, within the first 10 min of the assay, the dA analysis showed a significant signal change, while the cell coverage did not. This difference in dA change indicates that small changes in the initial cell-surface contact can be distinguished by plasmonic signal sensing but not by analysis of cell morphology. This suggest the importance to incorporate both imaging and SPR methods for detailed cell adhesion analysis.
We further assessed the adhesion process of A549 using the resonance dip and peak induced by the cells on the CPALNS4c chips. As shown in [fig_ref] Figure 3: The optical properties of aluminum nanoslit-based biosensors [/fig_ref] ,b, the resonance dip shifts faster than the peak in A549 cell adhesion tests. Additionally, the time constant calculated by the shifts of resonance dip is shorter than both the constants by the peak and dA. These results support the notion that the cell adhesion processes detected by the Fano resonance dip and peak are different. In [fig_ref] Figure 4: The Fano resonances and cell morphological changes during A549 cell adhesion in... [/fig_ref] , no significant difference is shown between the constants calculated by the shifts of resonance peak and by dA, indicating that the cell adhesion responses analyzed by these two parameters tend to be identical. Therefore, we conclude that the resonance dip shift and peak shift/dA can be used to monitor short-and long-range cellular changes, respectively, in our cell adhesion test.
## Assessment of cell adhesion by fano resonance changes using the goalns25c biosensor chip.
As mentioned previously, the spectral shape of the GOALNS biosensor is SPR-dominant, which means that the dip may be used for monitoring short-range cellular changes. Previously we have detected biomolecular changes by spectra intensity changes in Fano resonance using a AuNS sensor [bib_ref] Enhancing surface plasmon detection using ultrasmall nanoslits and a multispectral integration method, Lee [/bib_ref]. In this work, the adhesion-induced Fano resonance dip shift and intensity changes of dI in the GOALNS25c chip were used to assess adhesion of a normal kidney epithelial cell line, Madin-Darby Canine Kidney (MDCK) cells. Detailed analysis methods are described in the Supplementary Materials and [fig_ref] Figure 4: The Fano resonances and cell morphological changes during A549 cell adhesion in... [/fig_ref]. As shown in [fig_ref] Figure 5: Assessment of MDCK cell adhesion by Fano resonance changes measured by the... [/fig_ref] , during the cell adhesion test, MDCK cells stopped spreading at 90 min after cell seeding. Moreover, morphological differences were not readily observed at 90 and 120 minutes during the test. The image result matches the cell adhesion measurement with the GOALNS25c chip in [fig_ref] Figure 5: Assessment of MDCK cell adhesion by Fano resonance changes measured by the... [/fig_ref] , in which the spectra at 90 and 120 mins are almost identical. The temporal changes in dI [fig_ref] Figure 5: Assessment of MDCK cell adhesion by Fano resonance changes measured by the... [/fig_ref] and the resonance dip shift [fig_ref] Figure 5: Assessment of MDCK cell adhesion by Fano resonance changes measured by the... [/fig_ref] were plotted and fitted to calculate the time constants of MDCK cell adhesion kinetics using a cell adhesion model reported earlier [bib_ref] Dependence of cancer cell adhesion kinetics on integrin ligand surface density measured..., Orgovan [/bib_ref].
A highly metastatic melanoma cell line, A375 cells, was used to test the performance of GOALNS25c chip. As shown in [fig_ref] Figure 6: The cell adhesion rates of normal and cancer cells [/fig_ref] -c, the time constants of A375 cell adhesion can be obtained from the Fano resonance dip shift, the peak shift and the dI changes along time. There was no difference between the time constants calculated by The time constants of cell adhesion were calculated from the dip shift, peak shift and dA kinetic curves, n = 5. Statistical significance was evaluated by a paired t-test. There is no significant difference between the peak shift and dA groups.
www.nature.com/scientificreports www.nature.com/scientificreports/ the peak shift and by the dI [fig_ref] Figure 6: The cell adhesion rates of normal and cancer cells [/fig_ref] , suggesting that both parameters can be used to monitor long-range cellular changes by the GOALNS biosensor. However, the variation in the time constant analyzed by the dI was smaller than that from the resonance peak shift calculation. Therefore, all subsequent cell adhesion processes were evaluated by recording and analyzing the resonance dip and the dI with the GOALNS25c chip.
When the process of cell adhesion progresses to late stages, the Fano resonance sensing response stopped, and no more spectral changes occurred. Therefore, the maximum value of the dI and the maximum shift of resonance dip can be used as a final readout for cell adhesion. By plotting the time constant versus dip shift maximum or dI maximum, the adhesion characteristics of MDCK cells could be classified into six categories [fig_ref] Figure 5: Assessment of MDCK cell adhesion by Fano resonance changes measured by the... [/fig_ref] based on the extension of spread and the adhesion rate. The extension of spread is categorized into two states: normal and low (marked with "−" in [fig_ref] Figure 5: Assessment of MDCK cell adhesion by Fano resonance changes measured by the... [/fig_ref] while the adhesion rate is categorized into three states: fast (marked as "+" in [fig_ref] Figure 5: Assessment of MDCK cell adhesion by Fano resonance changes measured by the... [/fig_ref] , normal and slow. [fig_ref] Figure 6: The cell adhesion rates of normal and cancer cells [/fig_ref] shows that the time constant analyzed by the dI and the resonance dip changes of MDCK cells were 31.4 min and 5.9 min, respectively. As expected, the cell adhesion rate was faster when calculated from the resonance dip shift. The result corresponds well to the fact that cell adhesion starts with rapid interaction in region that is in proximity to the surface. Moreover, the extent of spreading and the total cell numbers also may affect the dI maximum and the dip shift maximum. Shown in [fig_ref] Figure 6: The cell adhesion rates of normal and cancer cells [/fig_ref] , the coefficient of variation for the dip shift maximum was significantly smaller than that for the dI maximum, indicating that the dip shift maximum is better for determining the final extent of cell spread in CAAS using the GOALNS25c chip.
Comparing adhesion of cancer cells using the GOALNS25c biosensor chip. In order to explore the correlation between metastatic potency and cell adhesion, three lung cancer cell lines (CL1-0, CL1-5 and A549 cells) and two melanoma cell lines (A375 and SKMEL-24 cells) were assessed using the GOALNS25c chip.
As shown in [fig_ref] Figure 6: The cell adhesion rates of normal and cancer cells [/fig_ref] , time constant calculated by the dip shift was shorter than that by the dI in all tested cancer cell lines. Interestingly, for both types of calculation, the higher metastatic CL1-5 cells showed shorter time constants than the lower metastatic CL1-0 cells. This result indicates that the CL1-5 cells have higher cell adhesion www.nature.com/scientificreports www.nature.com/scientificreports/ rate than the CL1-0 cells. However, the cell adhesion rate of CL1-0 and A549 cells were similar according to our measurements.
In the adhesion test using GOALNS sensor, the long-and short-range cellular response in the longitudinal direction can be respectively analyzed by the dI changes and the dip shifts. The dynamic change in the longitudinal direction might help us to evaluate the cell adhesion processes such as the dynamic cellular properties or thickness change.
In our result, the time constant ratio was similar for A549 cells and CL1-5 cells, but different from that for CL1-0 cells (A549: 1.56, CL1-5:1.46, CL1-0: 2.26). According to a previous report, metastatic potency based on invasiveness assay is the highest for CL1-5 cells, moderate for A549 cells and the lowest for CL1-0 cells [bib_ref] Tumor-associated macrophages: the double-edged sword in cancer progression, Chen [/bib_ref]. The value of the ratio may correlate with the metastatic potency of lung cancer cells in our test.
For the melanoma cells, it has been reported that the metastatic potency of A375 cells is similar to that of SKMEL-24 cells. In our results, the SKMEL-24 showed shorter dI time constant than the A375 cells but the time constants by the dip shift are similar. These results shows that short-range detection indicates similar adhesion rate but long-range detection indicates differently. Additionally, the ratio of the time constant by the two types of calculation was also different in A375 and SKMEL-24 cells (A375: 2.44 and SKMEL-24: 1.3). Notably, the correlation between time constant ratio of dI to dip shift and metastatic potency of melanoma cells is not as clear as for the lung cancer cells and needs further investigation. Nevertheless, the dip shift time constant can be a potential parameter for determination of metastatic potency in cancer cells. www.nature.com/scientificreports www.nature.com/scientificreports/ The time constant ratio may indicates final cell thickness of the adherent cells. The thickness of adherent MDCK cells was reported to be 8-10 μm and A549 cells were reported to be approximately 2 μm [bib_ref] Apical beta 1 integrin in polarized MDCK cells mediates tubulocyst formation in..., Zuk [/bib_ref] [bib_ref] Real-time sensing of cell morphology by infrared waveguide spectroscopy, Yashunsky [/bib_ref] [bib_ref] Ptychography-a label free, high-contrast imaging technique for live cells using quantitative phase..., Marrison [/bib_ref]. The time constant ratio of MDCK cells and A549 cells was 5.19 and 1.34, respectively, suggesting that smaller time constant ratio may correspond to thinner adherent cells. It is reasonable that thicker cells would require more time to form a thicker monolayer. Thus, the time constant ratio of dI to dip shift might be used for estimating the thickness of the adherent cells.
## The effects of inhibitors on adhesion measurements by the goalns25c biosensor chip. we
next tested whether small molecules that alter cell adhesion activity could potentially be identified by our biosensor chips. We used the time constant calculated by resonance dip shift to compare the adhesion rate of different kinds of cancer cells in the GOALNS25c chip. Additionally, dot plots of the time constant versus the dip shift maximum further elucidated the characteristics of cell adhesion under various stimuli. Using these analyses, we tested the effect of an adhesion inhibitor on cancer cell adhesion rates using the GOALNS25c chip. Focal adhesion formation is known to be controlled by FAK 9 , and the FAK inhibitor, FAKi-14, has been reported to suppress the cell adhesion in human pancreatic cancer [bib_ref] A novel small molecule inhibitor of FAK decreases growth of human pancreatic..., Hochwald [/bib_ref]. Thus, we treated four types of cancer cells with FAKi-14 and assessed the effect of the drug on the cell adhesion.
As shown in [fig_ref] Figure 7: The effects of a cell adhesion inhibitor on the adhesion of melanoma... [/fig_ref] , two cancer cell lines, A375 and CL1-5 were insensitive to the inhibitor, while the other two lines, SKMEL-24 and CL1-0 were responsive to the drug. In [fig_ref] Figure 7: The effects of a cell adhesion inhibitor on the adhesion of melanoma... [/fig_ref] , it shows that the responses of the inhibitor-treated cells could not be distinguished from those of the non-treated cells. On the other hand, one or two subgroups appear in the left-lower corner of the plot for CL1-0 and SKMEL-24 cells [fig_ref] Figure 7: The effects of a cell adhesion inhibitor on the adhesion of melanoma... [/fig_ref]. The appearance of these subgroups suggests that these cells were significantly affected by FAKi-14 treatment and showed a poor extension of spreading. The spreading of the FAKi-treated cells were interrupted and reached steady state earlier than the normal cell. Moreover, the incomplete and interrupted adhesion compared to the normal cells result in an increased adhesion rate (i.e. lower time constant). Additionally, a dose-dependency was observed in FAKi-14-treated SKMEL-24 cells, wherein both the dip shift time constant and the dip shift maximum decreased with increasing concentration of the inhibitor [fig_ref] Figure 7: The effects of a cell adhesion inhibitor on the adhesion of melanoma... [/fig_ref]. While the cell adhesion rates of A375 and CL1-5 cells were not affected by drug concentrations as high as 10 μM [fig_ref] Figure 7: The effects of a cell adhesion inhibitor on the adhesion of melanoma... [/fig_ref] , the cell adhesion rate of SKMEL-24 and CL1-0 were significantly increased by the drug at a concentration of 10 μM. The suppression of the cell spreading by the inhibitor was also observed in the images of SKMEL-24 cells [fig_ref] Figure 7: The effects of a cell adhesion inhibitor on the adhesion of melanoma... [/fig_ref] and CL1-0 cells [fig_ref] Figure 7: The effects of a cell adhesion inhibitor on the adhesion of melanoma... [/fig_ref].
To sum up, we can use two parameters (dip shift time constant and dip shift maximum) to evaluate the inhibitory effects of a compound on cell adhesion using GOALNS biosensor. As shown in [fig_ref] Figure 7: The effects of a cell adhesion inhibitor on the adhesion of melanoma... [/fig_ref] ,d, the dose-dependency curves derived from the dip shift time constant and the dip shift maximum show a similar trend. Despite the higher variation in the dip shift time constant, we consider both parameters can be used to assess cell adhesion response. It is interesting that the two melanoma cell lines exhibit similar metastatic potencies but differ in their sensitivities to FAKi-14. In these lines, it is possible that the FAK-activated adhesion pathway differs in terms of kinase activity or expression levels. Further experiments may be performed using extracellular matrix proteins, such as fibronectin, collagen and laminin to specifically activate the FAK adhesion pathway during the cell adhesion test. Nevertheless, in this report, we clearly show that the GOALNS25 chip is able to detect the effects of inhibitors on cell adhesion.
# Conclusion
In this study, we built a new system for simultaneous collection of cell morphology and Fano resonance signals to monitor cell adhesion. In this system, sensitive double-layer Al nanoslit biosensors were used to evaluate the adhesion of normal and cancer cells. The experimental spectra were collected, and the adhesion time constants were computed using dA, dI, peak shift and dip shift of the Fano resonance signals, which were induced by the binding of cells to the nanoslit. By using the Al nanoslit-based biosensor chip, we were able to sense both short-and long-range changes during the cell adhesion process. The time constants calculated from dip shifts and dI of the Fano resonance allowed us to compare the cell adhesion rates among normal and cancer cells. Additionally, we could examine certain features of cell adhesion using the dI-and dip shift time constants, dI maximum and dip shift maximum. Since the detection depth is different for dI and resonance dip, the time constant ratio of dI to dip shift was found to have some potential for estimating the thickness of adherent cells and the metastatic potencies of lung cancer cells. Assessment of cell adhesion by the resonance dip change is more reliable than the dI change because the resonance intensity change can be affected by damage to the nanoslit structure and cell scattering effects in the test. We successfully determined the adhesion rate of cancer cells as well as their response to an adhesion inhibitor with the Al nanoslit-based biosensor chip. Moreover, we are able to rapidly produce a large number of nanoslit-based biosensors at low-cost. Therefore, the Al nanoslit-based plasmonic biosensing chips show great potential for drug screening applications in the future.
# Materials and methods
the optical setup of cell adhesion assessment system. The setup for the CAAS is shown in [fig_ref] Figure 1: The optical setup and the nanoslit-based sensors for cell adhesion assessment [/fig_ref].
The system is based on an inverted phase-contrast microscope (IX71, Olympus) with an integrated motorized stage, proportional integral derivative (PID; TTM-J40-R-AB, JETEC Electronics Co.)-controlled transparent heater, spectrometer (V2000, Ocean Optics) and digital camera (EOS-D60, Canon). The temperature of the biosensor chip was measured by a K-type thermocouple (TPK-02A, TECPEL) clipped between the chip and the indium tin oxide glass (ITO glass, Part No. 300739, Merck) transparent heater. Images of cells were obtained with a digital camera. The light path configuration shown in [fig_ref] Figure 1: The optical setup and the nanoslit-based sensors for cell adhesion assessment [/fig_ref] allowed us to simultaneously collect spectrum signals and phase-contrast images.
www.nature.com/scientificreports www.nature.com/scientificreports/ Fabrication of spR biosensors and cell adhesion sensing chips. Two types of novel nanostructure sensors, CPALNS and GOALNS, were fabricated and used for cell adhesion assessment. The nanostructural schematics of the biosensors are shown in [fig_ref] Figure 1: The optical setup and the nanoslit-based sensors for cell adhesion assessment [/fig_ref] The scanning electron microscope images of the capped and the grooved aluminum nanoslits are shown in [fig_ref] Figure 7: The effects of a cell adhesion inhibitor on the adhesion of melanoma... [/fig_ref]. The nanostructure induces light in transverse and longitudinal [fig_ref] Figure 2: The fabrication of aluminum nanoslit-based biosensor chips for cell adhesion studies [/fig_ref] Fabrication of the nanoslit plastic substrate was described previously [bib_ref] Low-Cost and Rapid Fabrication of Metallic Nanostructures for Sensitive Biosensors Using Hot-Embossing..., Lee [/bib_ref]. Briefly, a 130-nm-thick diluted ZEP-520 resist (ZEP-520, Zeon Corp, Tokyo, Japan) was spin-coated onto a 4-inch silicon substrate. Periodic nano-groove arrays were fabricated in the resist using an electron beam drawing system. The period and width of the nano-groove arrays were 470 nm and 60 nm, respectively. The resist patterns were then coated with gold using a sputter and electroformed with Ni and Co to produce a ridged nanoslit metal mold (RNMD). The RNMD was then used in the production of nano-groove array plates by compression-injection molding. A picture of the nano-groove array plastic plate is shown in [fig_ref] Figure 2: The fabrication of aluminum nanoslit-based biosensor chips for cell adhesion studies [/fig_ref]. A grooved nanoslit metal mold (GNMD) was obtained by direct electroforming with the RNMD. The GNMD was used to produce the nano-ridge arrays by hot-embossing nanoimprinting. A picture of the nano-ridge array plastic film is shown in [fig_ref] Figure 2: The fabrication of aluminum nanoslit-based biosensor chips for cell adhesion studies [/fig_ref]. After depositing Al on the nanoslit array plastics with a thermal evaporator, we obtained the CPALNS and the GOALNS biosensors. A four-chamber cell adhesion assessment chip, called the CPALNS4c chip, was designed and fabricated using PMMA, double-sided tape and the CPALNS biosensor [fig_ref] Figure 2: The fabrication of aluminum nanoslit-based biosensor chips for cell adhesion studies [/fig_ref]. A 25-chamber cell adhesion assessment chip, called the GOALNS25c chip, was assembled by PMMA, double-sided tape and the GOALNS biosensor [fig_ref] Figure 2: The fabrication of aluminum nanoslit-based biosensor chips for cell adhesion studies [/fig_ref]. The volumes of the cell culture chambers in the CPALNS4c and the GOALNS25c chips were 54.7 mm 3 (6 × 6 × 1.52 mm) and 43.2 mm 3 (4.1 × 4.1 × 2.57 mm), respectively. All of the acrylic sheets and the double-sided tapes were fabricated by laser ablation using a CO 2 laser scriber (V-2000, LTT group, Taiwan), according to design patterns drawn in AutoCAD (Autodesk). Details of the micro-channel fabrication procedures using a CO 2 laser were provided in a previous work [bib_ref] Simultaneous chemical and electrical stimulation on lung cancer cells using a multichannel-dual-electric-field..., Hou [/bib_ref] [bib_ref] Direct-write laser micromachining and universal surface modification of PMMA for device development, Cheng [/bib_ref]. All components of the chip were disinfected by 30 min UV irradiation before assembly. The completed sensor chip assembly was then treated with oxygen plasma for 120 s at 200 W. Complete culture medium was incubated in a 5% CO 2 incubator at 37 °C overnight prior to injection into the culture chamber. Before the cell adhesion test, all chambers were washed once with distilled water and air dried for 20 min in a laminar flow cell culture hood.
Cell adhesion assessment. In order to enhance cell adhesion of the cancer cells, the biochip surface was coated with complete culture medium before cell injection. Approximately 90 μl (in the CPALNS4c chip) and 45 μl (in the GOALNS25c chip) of cells (5 × 10 5 cells/mL) were injected or loaded into the chamber of the cell adhesion sensing chip. The chip was then incubated on a computer-controlled motorized stage. The optical wavelength change was recorded by a spectrophotometer every two minutes during the cell adhesion assessment; the total experimental period was 2-4 h depending on the cancer cell type. Cell adhesion kinetic curves were expressed as the total integrated intensity changes and the specific spectra shift over time. where dA is the SPR response, I 0 (λ) is the intensity of the referenced spectrum (time = 0), I (λ) is the intensity of the spectrum at the indicated time point and I s (λ) is the intensity of a constant wavelength (resonance substrate mode), which served as an internal control. The spectral integration region was from 600 to 650 nm, in which Fano resonance occurred at the Al/medium interface. The cell adhesion response curve was plotted as dA over time, and the curves were fitted with an equation (one-phase association) in Prism 6 software (GraphPad). The time constant was obtained from the fitted equation. The dI was calculated as: where the dI is the cell adhesion response, I 0 (λ) is the intensity of the spectral band from 620 to 645 nm at time 0, I i (λ) is the intensity of the spectral band from 620 to 645 nm at an indicated time point, I max is the intensity peak of differential intensity spectral. The resonance dip shift, peak shift and the intensity peak of the differential spectra were obtained by fitting the data with a Gaussian distribution. The cell adhesion kinetic curves were plotted as dip shift, peak shift and dI over time and fit with the logistic equation for cell spreading to calculate time constants [bib_ref] Dependence of cancer cell adhesion kinetics on integrin ligand surface density measured..., Orgovan [/bib_ref].
# Data analysis.
[fig] Figure 1: The optical setup and the nanoslit-based sensors for cell adhesion assessment. (a) A photograph of the setup shows the CAAS. [/fig]
[fig] Figure 2: The fabrication of aluminum nanoslit-based biosensor chips for cell adhesion studies. (a) The design of the CPALNS4c chip. The CPALNS4c chip is composed of 8 PMMA adaptors, three layers of PMMA, four layers of double-sided tape and the CPALNS biosensor. (b) The photographs show the hot-embossing nanoimprinting polycarbonate film with nano-ridge arrays on the surface, and (c) the completed assembly of the CPALNS4c chip. (d) The design of the GOALNS25c chip. The GOALNS25c chip is composed of a PMMA cover, two layers of PMMA, double-sided tape and the GOALNS biosensor. (e) The photographs show the injection molding polycarbonate plate with nano-groove arrays on the surface, and (f) the complete assembly of the GOALNS25c chip. [/fig]
[fig] Figure 3: The optical properties of aluminum nanoslit-based biosensors. The optical properties of the doublelayer (a-c) capped and (d-f) grooved Al nanoslit biosensors in the respective CPALNS4c and GOALNS25c chips. (a,d) The intensity spectra and (b,e) the transmission spectra of the Al biosensor chips under the water-filled or air-filled conditions. The intensity spectra shift of the Fano resonance induced by (c) A549 and (f) MDCK cell attachment and spreading at 0, 60 and 120 mins after cell seeding in the CPALNS4c and GOALNS25c chips, respectively. [/fig]
[fig] Figure 4: The Fano resonances and cell morphological changes during A549 cell adhesion in the CPALNS4c chip. (a) The images show cell morphological changes over time during cell adhesion. The area inside the red circle is the effective area for resonance detection. (b) The cell coverage percentage was plotted against time, and the time constant of cell adhesion kinetics was calculated by curve fitting. (c) The intensity spectra showed a redshift of Fano resonance spectra over time during cell adhesion. The dA is shown for a unique segment of the spectral band (600-650 nm). (d) The dynamics of cell adhesion were plotted as dA over time, and the time constant of the cell adhesion reaction was calculated by curve fitting with one-phase associated equation. (e) [/fig]
[fig] Figure 5: Assessment of MDCK cell adhesion by Fano resonance changes measured by the GOALNS25c chip. The cell adhesion kinetic curves were plotted using (a) dI and (c) dip wavelengths. Dots represent data, solid lines are the fitted curves, n = 24. The time constants (by dI and dip shift) were calculated by equation fitting. The profile of cell adhesion was plotted as (b) the dI time constant versus the dI maximum, and (d) the dip shift time constant versus the dip shift maximum. Cell types were classified into six categories based on the extension of spread and the adhesion rate. The dots in (b,d) represent the MDCK cell populations with normal spread and adhesion rate. Scientific RepoRts | (2019) 9:7204 | https://doi.org/10.1038/s41598-019-43442-w [/fig]
[fig] Figure 6: The cell adhesion rates of normal and cancer cells. (a) The time constants of MDCK cell adhesion were obtained from curve fitting of the dI and the dip shift. (b) Coefficients of variation for the dI maximum and the dip shift maximum from cell adhesion assessment with the GOALNS25c chip, n = 24. (c) The time constants of cancer cells were calculated by dI and dip shift curve fitting. There were significant differences between dIand dip shift-calculated time constants in all cancer cell groups, n = 7-10 for each group; lung cancer cells: CL1-0, CL1-5 and A549; melanoma cells: A375 and SKMEL-24. All data are presented as mean ± S.D. (2019) 9:7204 | https://doi.org/10.1038/s41598-019-43442-w [/fig]
[fig] Figure 7: The effects of a cell adhesion inhibitor on the adhesion of melanoma and lung cancer cells. (a,c) Adhesion of cancer cells in response to different concentrations of cell adhesion inhibitor, FAKi-14. The plots show dip shift time constant versus dip shift maximum. (b,d) A dose-dependent effect of the cell adhesion inhibitor was observed for dip shift time constant and dip shift maximum. Data are presented as mean ± S.D. The images show (e) SKMEL-24 and (f) CL1-0 cell morphological changes over time during cell adhesion with or without cell adhesion inhibitor present. (2019) 9:7204 | https://doi.org/10.1038/s41598-019 [/fig]
|
Prospective multicentre cross-sectional audit among older Australians accessing health and residential aged care services: protocol for a national advance care directive prevalence study
# Introduction background
Advance care planning (ACP) is a coordinated communication and planning process that aims to clarify and share a person's values and preferences for medical treatment so as to guide healthcare decision-making should the person subsequently lose the capacity to make or communicate such decisions in the future. The ultimate goal of ACP is to ensure care received is consistent with the person's known preferences. [bib_ref] Definition and recommendations for advance care planning: an international consensus supported by..., Rietjens [/bib_ref] ACP has been shown to improve outcomes for patients, their families, healthcare staff and the healthcare system. [bib_ref] Overview of systematic reviews of advance care planning: summary of evidence and..., Jimenez [/bib_ref] [bib_ref] Efficacy of advance care planning: a systematic review and meta-analysis, Houben [/bib_ref] [bib_ref] The effects of advance care planning on end-of-life care: a systematic review, Brinkman-Stoppelenburg [/bib_ref] [bib_ref] A review of the implementation and research strategies of advance care planning..., Flo [/bib_ref] Ideally, ACP discussions should result in documentation recording a person's values and preferences and/or appointment of a Open access substitute decision-maker (SDM). Documentation is an important component of ACP. It has been shown to improve outcomes for people and provides information and support for SDMs, caregivers and clinicians who may be required to make treatment decisions on behalf of a person whose decision-making capacity is impaired. The names, scope and legal requirements of ACP documentation vary considerably within Australia and internationally. [bib_ref] Advance care planning in Australia: what does the law say?, Carter [/bib_ref] [bib_ref] Review of advance care planning laws across Australia: short report, Fountain [/bib_ref] [bib_ref] Advance directives as an instrument in an ageing Europe, Goffin [/bib_ref] [bib_ref] Comparisons of guardianship laws and surrogate decision-making practices in China, Japan, Thailand..., Tsoh [/bib_ref] [bib_ref] Advance care planning: whose agenda is it anyway, Russell [/bib_ref] [bib_ref] The misleading simplicity of advance directives, Hertogh [/bib_ref] In Australia, ACP documentation includes advance care directives (ACD), a term encompassing documents recognised by state-based legislation (statutory ACD: preferences for care or statutory ACD: appointment of SDM) or common law (non-statutory ACDs) which are completed and signed by a competent adult.While the specific execution requirements for these three types of ACDs vary, all must be completed by a person with decision-making capacity and they only come into effect once that capacity is lost. ACP documentation may also be completed on behalf of the person by a health professional or someone else (SDM/family). Additionally, medical treatment orders that describe action to be taken in an emergency, such as 'Goals of Care', 'Resuscitation Plans' and 'Physician Orders for Life-Sustaining Treatment', may also include reference to a person's preferences for care. [bib_ref] Use of the physician orders for life-sustaining treatment program in the clinical..., Hickman [/bib_ref] [bib_ref] Improving end-of-life communication and decision making: the development of a conceptual framework..., Sinuff [/bib_ref] These are completed and signed by a doctor.
For the potential benefits of ACDs and other documented outcomes of ACP to be realised, it is critical that documentation is accessible at the point of care should the person be unable to participate in decision-making. If healthcare providers do not have access to the person's documented preferences when treatment decisions are required, the person may receive care that is inconsistent with their preferences.In Australian health services, a range of systems are used to store and retrieve ACDs, including service-specific electronic health record systems, scanned medical records, hard-copy files or a combination.Most health services also now have access to 'My Health Record', a new centralised national e-health record that offers online storage of ACDs. While some services may have mechanisms to elicit existing ACDs on admission, many will rely on the person to provide his or her documentation to relevant providers.Although support for ACP and ACDs continues to grow in Australian policy and legislative frameworks, there is a lack of national data on the prevalence, content and quality of ACDs at the point of care in Australian health services. Without this information, governments have limited ability to monitor the effectiveness and impact of ACP policy and programmes, and organisations lack evidence to inform service-level programmes and initiatives to increase ACD uptake.
In 2017, an Australian-first pilot feasibility study was conducted by Advance Care Planning Australia (ACPA) to assess the prevalence of ACDs among older adults accessing Australian health and residential aged care services. [bib_ref] Prevalence and correlates of advance care directives among older Australians accessing health..., Detering [/bib_ref] Of 2285 health records audited across 51 sites, 30% contained at least one ACD. The ACD prevalence was significantly higher in residential aged care facilities (48%) compared with hospitals (16%) and general practices (3%) and varied considerably across participating sites. [bib_ref] Prevalence and correlates of advance care directives among older Australians accessing health..., Detering [/bib_ref] A further 20% of people had 'other ACP documentation'. This likely included documents that were not formal ACDs (eg, personal letters) or documentation completed on behalf of the person (eg, by a health professional, a family member or the SDM). However, the exact nature or authorship of this 'other ACP documentation' was unclear. [bib_ref] Prevalence and correlates of advance care directives among older Australians accessing health..., Detering [/bib_ref] During 2018-2020, ACPA has been funded by the Australian Government to deliver the National ACD Prevalence Study. This study will build on the findings and methodology of the pilot feasibility study to collect a national data set on the prevalence and characteristics of ACDs and other documented outcomes of ACP among older people accessing general practices, hospitals and residential aged care facilities across Australia. The study will provide data to assist organisations and governments understand how well ACP is being implemented in Australia and identify areas for improvement. It will also contribute to the methodology of undertaking research into the prevalence of ACP documentation.
## Aims and hypotheses
The aims of the National ACD Prevalence Study are to: 1. Determine the prevalence of ACDs and other documented outcomes of ACP among people aged ≥65 years accessing Australian general practices, hospitals and residential aged care facilities. 2. Assess the content and quality of identified ACDs. 3. Explore consistency between medical orders and the person's documented preferences for care. 4. Explore the characteristics of individuals and study sites where ACDs are present in the health record. Based on previous Australian studies, we hypothesise that the overall prevalence of ACDs will be low, and there will be more non-statutory ACDs and 'statutory ACD: appointment of SDM', than 'statutory ACD: preferences for care'. We expect to find various other ACP documents completed by health professionals, SDMs and/or family. We anticipate that prevalence will be highest in residential aged care facilities and lowest in general practice and expect a wide range of ACD prevalence rates within each sector. We hypothesise that there will be variation in the content and quality of ACDs, and that many ACDs will not meet signing and witnessing requirements specified in jurisdictional legislation or quality criteria outlined in the Australian National Framework for ACDs.We hypothesise that where there is an ACD, medical orders will be consistent with the person's preferences. We expect a range of individual and site factors to be associated with the presence of an ACD, including demographic and clinical characteristics and site-reported organisational support for ACP.
## Box 1 site eligibility criteria
General criteria: general practices, hospitals and residential aged care facilities are eligible to apply if they: 1. Are an accredited organisation according to sector requirements. 2. Have the approval and endorsement of their executive team. 3. Can nominate one staff member as the Study Lead (responsible for coordination of the study at the site and will be the key contact person for the research team). 4. Have internet, email and telephone access. 5. Have access to devices for online data collection (eg, computer, laptop or iPad). 6. Have policies in place about privacy and confidentiality. 7. Agree that the information provided in their application will be used to generate a Research Collaboration Agreement if successful and are willing to sign this preferably within 4 weeks of being notified of successful application. 8. Meet all sector-specific eligibility criteria. Sector-specific eligibility criteria Hospitals and residential aged care facilities are eligible to apply if they: 1. Expect that their site will have at least 50 patients/residents on the day(s) of the study who are aged 65 years or older and have been admitted for at least 48 hours.* 2. Have a records management system with the ability to extract a list of all admissions of people aged 65 years or older who have been admitted for more than 48 hours on the day(s) of the study. 3. Can nominate up to two additional staff members to collect the data for the study. 4. Are willing to support the Study Lead and data collector(s) to undertake mandatory online training in data collection procedures. 5. Have the capacity to review a minimum of 30 health records. Hospitals will also be required to obtain additional ethics approval and/ or a site-specific assessment at their site, ideally within 6-8 weeks of notification of successful application. ACPA will provide support in obtaining necessary approvals as required. * A minimum of 30 health records will be randomly selected from eligible records on the day(s) of the study. General practices are eligible to apply if they: 1. Expect that at least 30 patients aged 65 years or older will attend their practice on the day(s) of the study. 2. Have a records management system with the ability to extract a list of all people aged 65 years or older attending the practice on the day(s) of the study. 3. Can nominate up to two additional staff members to collect the data for the study OR agree to data collector(s) being provided by ACPA.
ACPA, Advance Care Planning Australia.
MEthodS And AnAlySIS overview of modifications to the original pilot study protocol The 2017 pilot feasibility study provided key learnings that have informed the current protocol for the National ACD Prevalence Study. Based on these learnings, a number of modifications were made to the study design, site recruitment processes and information collection, training and support for data collectors, data items and data collection tools. These modifications and associated rationale are outlined in relevant sections below.
## Study design
The National ACD Prevalence Study is a prospective multicentre cross-sectional study consisting of two parts 1 : sitelevel data, collected during the expression of interest process 2 and an audit of health records of eligible people accessing those services, conducted by trained staff from participating sites. The pilot study included a self-report survey of people whose records were included in the audit. [bib_ref] Protocol for a national prevalence study of advance care planning documentation and..., Ruseckaite [/bib_ref] The purpose of this survey was to collect information from the person's perspective, regarding their views about and experience with ACP. Feedback received indicated that the survey substantially increased the time and resourcing required to complete data collection, leading to significant additional burdens on study sites. In response, the survey has been removed from the current protocol. We anticipate that this will enable the participation of a wider range of smaller and regional/rural sites that would not have had the resource capacity to conduct both a health record audit and participant survey.
A first round of data collection for the National ACD Prevalence Study was completed in 2018-2019. One further round of data collection is anticipated in 2020. Further rounds of data collection are likely but will depend on funding availability, which is yet to be confirmed.
Sites will be eligible to participate in more than one round of data collection. However, as the study is cross-sectional rather than longitudinal, it is expected that the proportion of sites who participate in more than one data collection round will be low.
Part 1: study sites Study sites will include Australian general practices, public and private hospitals and residential aged care facilities, recruited from all eight Australian states and territories (jurisdictions). A minimum of 24 sites will be included in each round of data collection. recruitment Sites will be recruited through an expression of interest process, coordinated by ACPA and promoted via state and territory departments of health and stakeholder networks using newsletters and websites. Depending on responses to the expression of interest process for each round of data collection, additional sites may be approached by the project team to promote representativeness across sectors (residential aged care facilities, hospital and general practice) and jurisdictions. All prospective sites will be required to complete an online expression of interest application (see site-level data collection below). Based on feedback from the pilot study, detailed application guidelines including a set of general and sector-specific eligibility criteria (box 1) were developed.
Site-level data collection Section 1 of the expression of interest application assesses site eligibility. Sites must meet all criteria before they can proceed. Sections 2 and 3 of the application collects sitelevel data [fig_ref] Table 1: Site-level data variables Demographic information Jurisdiction, location, funding [/fig_ref]. Data items are informed by the pilot study, a literature review of system factors potentially important for ACP/ACD implementation and uptake, and guidance from the project advisory group. Study sites are also required to nominate a Study Lead and one to three data collectors in the expression of interest application.
Part 2: health record audit Audit participants will comprise people aged 65 years or older who are admitted to participating hospitals or residential aged care facilities for at least 48 hours prior to audit, or attending general practices on the nominated day(s) of the study. The requirement for admission for at least 48 hours prior to audit is to ensure adequate time for relevant documentation that have been provided or retrieved from the person, the SDM or another service. Sites will be required to audit the records of a minimum of 30 and a maximum of 50 eligible people. These lower and upper limits are to ensure that sites audit similar numbers of records, thus allowing for meaningful comparison between sites. In the pilot study, sites were required to audit 50 records. However, feedback during recruitment in the pilot study suggested that this precluded the participation of smaller sites that were unlikely to have at least 50 patients/residents meeting eligibility criteria, particularly in rural or regional areas. Thus, the minimum of 30 record audits was applied in the current study to increase opportunities for smaller services to participate, facilitating greater representativeness among participating sites.
Sites are required to nominate how many records they intend to audit (minimum of 30 and maximum of 50 records) prior to their study day(s).
## Record selection
In hospitals and residential aged care facilities, health records will be randomly selected from a list of all eligible people using a simple randomisation procedure, designed to protect against selection bias. On the first day of the study, the site Study Lead will contact their organisation's Health Information Management team (or similar) to obtain a list of current people who meet eligibility criteria. Each eligible person will be assigned a number chronologically, creating an 'Eligible Records List'. These chronological numbers will be used for randomisation. No identifiable information will be provided.
The Study Lead will then inform ACPA of the total number of eligible records, and the number (30-50 records) they intend to audit. Randomisation will be conducted by an ACPA researcher using a random number generator ('Research Randomizer', www. randomizer. org). Records will be assigned to group 1 (include) or group 2 (do not include) within an 'Allocation List', which will be returned to the Study Lead. The Study Lead will match the 'Allocation List' to their 'Eligible Records List' to determine which files to audit. Group 1 will also contain a supplementary list of 10 records which are to be used (consecutively) as needed if any of the initial list are unavailable (eg, patient discharged).
Feedback from Study Leads involved in the pilot study indicated that the randomisation procedure was difficult to understand and implement in practice. To address these issues, Study Leads will receive specific training in the randomisation process, and be invited to complete a trial randomisation procedure prior to the nominated study date. Detailed instructions are also provided in study manuals and online education.
For practicality purposes, consecutive eligible records will be audited in general practices until the required number has been achieved.
data collection: training, procedures and support Participating sites will identify up to three data collectors (may include Study Lead) to undertake the audit. Study Leads and data collectors are expected to complete compulsory online training (approximately 60 min) and will have access to a data collection manual [fig_ref] Table 2: National advance care Directive prevalence study [/fig_ref]. ACP legislation in Australia is determined by each jurisdiction, [bib_ref] Review of advance care planning laws across Australia: short report, Fountain [/bib_ref] and therefore jurisdiction-specific manuals will be provided. Researchers based at ACPA will be available via telephone for consultation throughout the study.
Data collection will occur during defined periods. Sites are required to nominate 1-3 consecutive days for data collection. Data collectors will obtain selected paper and/ or electronic records (including My Health Record, if applicable) and attempt to locate relevant documentation within 15 min of opening the record. This time frame was selected in recognition that for ACP documentation and/ or medical orders to be useful, they need to be located quickly. Data variables extracted will include demographic and clinical information, and details regarding the type, content and characteristics of ACP documentation and/ or medical orders [fig_ref] Table 3: Participant data variables [/fig_ref].
It is expected that data collection will be completed by staff from participating sites. It is not practical to use external data collectors for a national study of this size. Key learnings from the 2017 pilot study have been used to improve the accuracy of collection. Staff will undertake compulsory training in study methodology and data Example of a jurisdiction-specific flowchart provided to data collectors to help them classify documentation identified in the audit. collection. Importantly, by supporting staff within organisations to complete data collection, it is anticipated that staff will increase their knowledge and ability to undertake future audits, generating opportunities for implementing ACP initiatives within their services.
ACPA may provide data collector(s) to undertake audit in general practices, and in hospitals and aged care facilities in remote or regional areas that would otherwise be unable to participate due to limited resources. These auditors will not be members of the research team, will complete all required training and will meet privacy and confidentiality requirements of the organisation where they collect data.
All data will be entered directly into a password-protected online database specifically built for this project, and hosted by REDCap ( www. project-redcap. org). The database will be available via web-based user interface using a personal computer, laptop or tablet. Each data collector will receive a unique login name and password. Each site will have an identification (ID) number, which will be known only to the site and the researchers conducting the analyses. The database will have in-built validation and range checks to reduce data errors.
The pilot study identified important issues with some data variables. In particular, data collectors reported difficulty classifying ACP documentation. Furthermore, based on free text responses it is likely that some documents were incorrectly classified. Thus in the present study, further information is provided regarding the nature of different documents they may encounter and characteristics of each to facilitate consistent classification. While the pilot study showed ACD prevalence of 30%, a further 20% of people had 'other ACP documentation'. [bib_ref] Protocol for a national prevalence study of advance care planning documentation and..., Ruseckaite [/bib_ref] The nature of these documents is unclear. ACDs are documents completed by competent people and come into effect if the person subsequently loses capacity to make his or her own decisions. However, ACP documentation is sometimes completed by people other than the person concerned, which may also be of benefit for people without an ACD, who no longer have capacity. This is currently an underexplored area of ACP practice. Consequently, in this study, the process of facilitating correct classification of documentation has changed. We have developed a process and flowchart based on who completed documentation (figure 1) and have provided detailed instructions and examples in education and study manuals. Feedback from the pilot indicated that data collectors had difficulty classifying some medical conditions . To address this, data collectors will be instructed to record only current/active medical conditions, and a range of examples will be provided for each of the categories. The pilot also had a substantial amount of missing data (30%) for the Eastern Cooperative Oncology Group (ECOG) status variable. The revised audit tool includes an additional question asking data collectors to estimate the person's level of functional disability based on available information in the person's record if the actual ECOG rating is not available. These items may be combined during analysis into an overall 'estimated' level of functional disability. In the pilot, there was also evidence of data entry errors in the database. For this study, the database logic has been reviewed, more detailed instructions will be provided within the database and study manuals and additional in-built validation checks will be added. We will also conduct comprehensive user testing of the database and improve education provided to data collectors regarding database usage.
## Open access
data de-identification A study ID number will be assigned to each person on the audit list, and will be used for data entry. Participating sites will be able to match study IDs with specific people. Identifiable information will not be recorded or disclosed to the research team.
## Outcomes
The primary outcome will be the presence of at least one ACD (statutory ACD: preferences for care, statutory ACD: SDM and/or non-statutory ACD) that is located within 15 min of accessing the record. Secondary outcomes will include prevalence of other documented outcomes of ACP (documents completed by a health practitioner or someone else, such as family or the SDM), assessment of content and quality of ACDs and concordance between the person's preferences for care as documented in their ACD and medical treatment orders. Currently in Australia, there is no standard measure for quality and validity of ACDs. Therefore, documents will be assessed based on requirements specified in jurisdictional legislation, and quality criteria outlined in the Australian National Framework for ACDs,including whether the document contains the name, date of birth and address of the person, the date of completion, whether the document is signed by the person and/or witnesses and whether any instructions have been provided for the SDM (for statutory ACD: SDM only). To assess concordance between the person's preferences for care and treatment instructions documented in medical orders, the consistency between treatment and/or other preferences specified in the person's ACD will be compared with treatment limitations outlined in their medical treatment order.
## Reliability testing
To estimate any potential bias in rating the primary and secondary outcome variables, two data collectors independently rated the same 40 health records and a reliability analysis was undertaken. Both a percentage agreement and a kappa statistic were calculated for primary (total as well as for each of the three types of ACD) and secondary outcome variables (documentation completed by health professionals and someone else). For each of the primary outcome measures, percentage of agreement between the first and second data collector was 100% and kappa statistic level of agreement was very high [fig_ref] Table 4: Reliability analysis [/fig_ref]. For secondary outcome variables, percentage of agreement ranged between 97.1% and 100% and kappa statistic level of agreement was good or very good.
## Sample size estimation and justification
Sample size calculations are necessary in prevalence studies to ensure that estimates are obtained with adequate precision. [bib_ref] Sample size estimation in prevalence studies, Arya [/bib_ref] [bib_ref] Sample size estimation in epidemiologic studies, Hajian-Tilaki [/bib_ref] [bib_ref] Sample size calculation in medical studies, Pourhoseingholi [/bib_ref] The minimum number of records required for this audit was calculated as 505. This calculation assumes an expected average ACD prevalence of 0.3 (based on pilot study), [bib_ref] Prevalence and correlates of advance care directives among older Australians accessing health..., Detering [/bib_ref] confidence level of 95% and desired precision of ±4%.
While a simple estimate of prevalence shows that a minimum of 505 records are required, to allow for three health sectors to be represented across all eight jurisdictions, 24 sites are required. A minimum sample of 30 records from each of the 24 sites was chosen to minimise the data collection burden, while providing an adequate sample size for site-level results to be reported with a Open access precision of ±3.5%. Therefore, the minimum total sample size required will be 720 health records.
# Statistical analysis
Descriptive statistics will be calculated for the total sample and by major grouping site or participant variables. Overall prevalence of ACDs will be calculated for the total sample and separately for healthcare sectors and jurisdictions. Prevalence rates of the types of ACDs, and 'other documentation' will also be reported. Comparisons will be made using t-tests for the continuous type variables and χ 2 contingency table analysis for the categorical type variables.
Similar to the pilot study, descriptive statistics will be used to describe the characteristics and content of ACDs and other documented outcomes of ACP (eg, by a health professional or family/SDM), and the specific preferences for care specified in ACDs in comparison the content of medical orders. Due to the nesting of individual records within sites, generalised linear mixed model regression will be performed to determine the predictive value of demographic, clinical and site-level variables on the presence of ACD in the person's record. For all analyses, data will be weighted for relevant population characteristics (eg, age, gender, jurisdiction) as necessary, using the latest data from the Australian Bureau of Statistics. Other data sources for weighting such as hospital and aged care demographics will be accessed as required. The level of significance will be set at 0.05.
If sites participate in more than one data collection round, and there is an analysis that compares prevalence rates over time, then the model will include time, in order to account for repeated measures. If study sample size is insufficient to allow for such a model, data from these returning sites will be limited to the first round of data collection in which they participated.
## Project governance
This study is led by ACPA, and will be overseen by a project advisory group, who will meet approximately four times per year via teleconference. The advisory group is chaired by ACPA, and includes academics and clinicians with interdisciplinary expertise relevant to this study including medicine, nursing, allied health, law and policy. The advisory group also includes representation from hospital, general practice and aged care sectors. The advisory group will co-opt a statistician to provide methodological and statistical advice.
The advisory group will review and endorse the project methodology, and provide advice regarding site recruitment across sectors, risk management, intellectual property and ethical concerns. They will provide advice, drawing on their interdisciplinary expertise, on the collection and interpretation of data and on contextual and jurisdictional aspects of the study.
## Confidentiality and privacy
To avoid any potential breach of confidentiality, only staff of the participating organisations directly involved in collection of data will have access to health records. ACPA auditors will be required to adhere to privacy principles of the services they attend. Data will be handled, stored and disposed off according to the Australian National Health and Medical Research Council Code of Responsible Conduct of Research and the National Statement on Ethical Conduct in Research involving Humans. All results will be de-identified and presented in an aggregate format.
## Patient and public involvement
The specific intent of this project is to inform service-level initiatives and future ACP programmes and policy. People whose records will be audited will not be involved in the audit, the study design or recruitment. As only de-identified information will be collected, it is not possible to provide results to participants. However, study sites will receive de-identified individualised reports including information about their service and how they compare with similar services. The study results will also be provided and discussed with the 'National ACP Engagement Advisory Group', which includes representation from consumers and consumer organisations including National Seniors, Dementia Australia, Palliative Care Australia and the Victorian Cancer Council.
## Ethics and dissemination
Study results will be provided to the participating sites and the Australian Government. No reports will identify any specific participant or site but jurisdictional comparisons will be possible. However, we will exercise caution in reporting jurisdictional or setting prevalence results when such results are primarily driven by only one or two sites.
The results will be highly relevant to clinical practice and policy nationally and internationally; therefore, the findings of this study will also be disseminated through relevant government departments, as well as through various national and international professional bodies, societies and peer-review networks. Findings will be presented at relevant conferences and published in peer-reviewed journals, on the ACPA website and in lay and social media where appropriate. Investigators will review manuscript drafts, abstracts, press releases and any other publications arising from the study. Authorship will be determined in accordance with the International Committee of Medical Journal Editors guidelines.
# Discussion
Effectiveness of national ACD prevalence research requires standardisation of methodology including the accurate classification of ACP documentation, thereby facilitating meaningful measurement and comparison. The current study has built on the findings and key learnings from the pilot feasibility study, particularly with respect to training of data collectors, standardisation of Open access methodology and facilitation of consistency regarding document classification.
# Strengths
The protocol retains the key study design features of the pilot which were shown to be feasible and successful and improves on the identified limitations. The study aims to recruit a diverse range of services increasing the generalisability of the measured outcomes. Data collectors from participating sites are trained in audit methodology and ACD classification using a standardised approach. This should facilitate meaningful comparison of prevalence rates between services. Furthermore, given that data collectors are staff from participating sites, education provided as part of this study has the potential to increase knowledge and awareness at the sites, and therefore contribute to capacity building within these sites.
The National ACD Prevalence Study will generate one of the largest and most comprehensive data sets on ACD prevalence in Australia and internationally. This data set will provide information on the types and characteristics of ACP documentation being used within and across three health sectors that can be used to better understand current practice and inform future ACP strategies and initiatives. The findings will also contribute to the methodology of undertaking research into the prevalence of ACP documentation.
limitations Recruitment via an expression of interest process means that there is a likely selection bias towards sites with an existing interest in ACP. While the aim of this study is to recruit sites across all three health sectors and all eight Australian jurisdictions, it is likely that not all sectors will be similarly represented nor will states and territories, thus limiting generalisability of findings by sector and jurisdiction. As the study focuses on ACD prevalence at the point of care in health and residential aged care services, findings are not generalisable to the wider Australian community. This study uses an audit methodology, and therefore it is not possible to determine quality and extent of ACP conversations known to be an important determinant of a successful ACP programme. Furthermore, given the audit methodology, it will not be possible to determine whether documentation translates into care that is consistent with the person's preferences, the ultimate goal of ACP.
[table] Table 1: Site-level data variables Demographic information Jurisdiction, location, funding (government, not-for-profit, private), record management system used Size of service Number of beds, number of health staff ACP programme/activity Current ACP programme/activity within the service, when this commenced, ACP training available for staff and how ACP is funded and/or implemented Presence/absence of ACP policy, consumer resources and ACD templates/forms. (Sites will be asked to upload their ACP policy documents and/or blank ACD templates/forms used in their service) ACD, advance care directive; ACP, advance care planning. [/table]
[table] Table 2: National advance care Directive prevalence study: data collection training and manual Review of data items and how to collect data Contains documents to assist with study and data collection. These include: ► A flowchart of relevant documentation to the audit. ► A decision tree to help with classification of ACP documentation. ► A list of statutory ACD used in Australia. ► Checklists for Study Leads to prepare for study day. ► A list of contact numbers. ACD, advance care directives; ACP, advance care planning. [/table]
[table] Table 3: Participant data variables [/table]
[table] Table 4: Reliability analysis (n=40) [/table]
|
Idiopathic small bowel diaphragm disease: a case report
Small bowel diaphragm disease is a rare condition usually associated with the prolonged use of non-steroidal antiinf lammatory drugs (NSAID) and that can be mistaken and treated as other pathologies. We describe a case of a 64-year-old man with a prolonged course of pain and subacute bowel obstructions, without any history of NSAID usage, found to have a multiple diaphragmatic stricture in the small bowel.
# Introduction
Small bowel diaphragm (SBD) disease, first described by Lang et al.in the late 1980s, is a rare condition usually associated to the prolonged use of non-steroidal anti-inflammatory drugs (NSAID) and characterized by the presence of multiple diaphragm-like septa that cause the narrowing of the small bowel lumen. Most cases are asymptomatic until they present with acute intestinal obstruction. It is a condition that is difficult to diagnose both preoperatively and intraoperatively. Herein, we describe the case of a 64-year-old man with a prolonged course of pain and subacute bowel obstructions, without any history of NSAID usage, found to have a multiple diaphragmatic stricture in the small bowel.
## Case presentation
On the initial admission, the 64-year-old man had a 1-year history of chronic intermittent abdominal pain and episodes of subacute bowel obstruction. The patient had no history of chronic medication usage.
Initial investigations included magnetic resonance enterography that was normal and a capsule endoscopy that showed rare erosions, aphthoid ulcers and two ulcerated stenosis.
One month later, he developed an acute small bowel obstruction, and a laparotomy was therefore undertaken.
The laparotomy revealed the presence of two transition zones between dilated and non-dilated small bowel. These zones were located within 40 cm proximal to the ileocecal valve, and they corresponded to a few millimeters thin lesions, with no evidence of a thickening in the bowel or mesenteric abnormalities (classic stigmata of stenotic Crohn disease). An enterotomy for one of the two lesions was undertaken, which showed no macroscopic abnormality, and then a stricturoplasty was made.
A diagnosis of ileal Crohn disease was suspected leading to the initiation of a treatment with biotherapy (adalimumab).
A second capsule endoscopy showed the same lesions.
One year later, he was readmitted with small bowel obstruction investigated with an abdominal computed tomography (CT) scan that showed no radiological finding of stenotic Crohn disease or other obstructive etiology.
The small bowel occlusion was resolved with non-operative management. After a multidisciplinary team discussion, an exploratory laparoscopy was decided in order to explore the extraluminal aspects at the obstruction area and execute an enteroscopy in order to ensure the existence of intraluminal unique or multiple lesions.
At laparoscopy, the patient was found to have stigmata from the previous operation without any other extraluminal anomalies on the small bowel or the mesenteric fat. A conversion to a mini-laparotomy was undertaken, and using manual palpation, 2 mm thin strictures were suspected to be the possible zones of the small bowel obstacle.
No other particular anomalies were perceived and a small enterotomy was performed 35 cm above the Bauhin valve near the previous stricturoplasty. The peroperative enteroscopy reached proximally the duodeno-jejunal angle and distally the ileocaecal valve and the caecum allowing the exploration of the entire small bowel. It showed multiple ileal diaphragm including three diaphragms with small ulcers (video). A 60 cm segmental resection was performed removing all the six lesions and the histological features found in the lesion were consistent with the diaphragm disease.
# Discussion
SBD disease, first described by Lang et al.in the late 1980s, is a rare condition. While the long-term use of NSAIDs is considered the single most important causative factor of SBD, many other medications, diseases and processes may cause injuries to the small bowel and result in strictures of the small bowel lumen, such as potassium chloride tablets, surgical anastomoses, radiation, ischemia, Crohn disease, tuberculosis, eosinophilic enteritis, lymphoma, etc.
The disease process of SBD does not affect the full thickness of the bowel and is predominantly found in the ileum. Although rarer, it has also been reported to affect the colon.
Most cases are asymptomatic until they present with acute intestinal obstruction. Other presenting features are irondeficient anemia, acute hemorrhage and an acute abdomen secondary to perforationCT scan and endoscopic techniques are commonly used in preoperative diagnosis; however, CT scanning tends to lack the resolution required to highlight the SBD and generally do not show any abnormal results when used to explore chronic abdominal pain, and it could advocate for small bowel acute strangulation on adhesions when used to explore acute small bowel occlusion. In regards to capsule enteroscopy, it should only be undertaken on patients with iron-deficiency anemia with no obstructive symptoms, to prevent the risk of capsule retention.
On the other hand, surgical intervention generally does not show any extraluminal abnormalities on the small bowel or the mesenteric fat, which leads to an enterotomy. This enterotomy frequently shows no abnormal mucosa, and generally leads to a stricturoplasty for its closure without a good knowledge of the diagnosis.
In contrast, our observation advocates for the role of peroperative enteroscopy in detecting the SBD that could be unique or multiple, which can reduce the small bowel lumen and be responsible for chronic abdominal pain and recurrent small bowel obstructions.
Histological features of the lesions are thickening and chaotic arrangement of muscular bundles in the muscularis mucosae, fibrosis of the lamina propriae, invariable mucosal ulceration at the apex of the diaphragm, numerous vessels, ganglion cells, nerve fibers and eosinophils in the mucosa.
Surgical intervention remains the treatment of choice and will usually involve resection of the affected segment. In cases of extensive or recurrent disease stricturoplasty may be preferred. In some cases the disease can affect the entire length of the small bowel, and it may be that some residual disease has to be left after the initial resection of the lesions. |
Stress, physical activity, and screen‐related sedentary behaviour within the first month of the COVID‐19 pandemic
This study investigated how stress, physical activity and sedentary behaviours, of a small sample of Canadians, changed within the first month (i.e. March/April) of the COVID-19 pandemic and the reasons/barriers associated with such changes. Individuals who regularly wear activity trackers were recruited via social media. Participants (N = 121) completed fillable calendars (March/April 2020) with their step counts and answered an online survey. Separate paired-sample t-tests, one-way ANOVAs and bivariate chi-squares were conducted, in addition to qualitative analysis. Daily (p <.001) and work (p =.003) stress increased, physical activity (measured by step count) decreased (p =.0014), and screen-related sedentary behaviour increased (p <.001) as a result of COVID-19. A decrease in physical activity, as a result of the pandemic, was also associated with a larger increase in work stress, compared with those who self-reported their physical activity to have been maintained or increased (p =.005). The most common reasons/barriers to changes in physical activity behaviours were access/equipment, time and motivation. Findings provide initial evidence of the impact of the COVID-19 pandemic on the health of some Canadians and highlight the need for continued monitoring of the health of Canadians throughout the pandemic.| 455 bs_bs_banner PHYSICAL ACTIVITY AND COVID-19
# Introduction
On 11 March 2020, the World Health Organization (WHO) declared the novel coronavirus (SARS-CoV-2), which causes COVID-19, to be a global pandemic . By late-March, all provinces and territories in Canada had declared a state of emergency or a public health emergency. Subsequently, various federal and provincial public safety measures were enacted in order to limit the spread of the COVID-19. Canadians were asked to practise physical distancing (i.e. remain at least 2 m (6 feet) away from others), to self-isolate and only leave their homes for essential errands and to quarantine if they had been exposed to the virus or returned from international travel. Other measures included the closure of non-essential businesses (e.g. gyms/recreation centres and non-grocery stores), as well as the closure of public spaces (e.g. parks and hiking trails; Ontario . With over 9,000 confirmed COVID-19 cases in Canada by the end of March 2020 and over 90,000 cases by the start of June 2020, the COVID-19 pandemic has resulted in substantial lifestyle changes for many Canadians, especially in regard to stress and mental health , physical activity (Ontario and screen-related sedentary behaviour .
Specifically, more than half of Canadians who filled out Statistics Canada's (2020a) survey 'Impacts of COVID-19 on Canadians: Your mental health' between 24 April and 11 May 2020 reported poorer mental health since the onset of physical distancing/self-isolating associated with the pandemic. Moreover, nearly all participants (88%) reported at least one symptom of anxiety (e.g. depression, grief, panic) in the two weeks prior to completing the survey (Stats Canada, 2020a). One means of combatting poor mental health, especially during the COVID-19 pandemic, could be through physical activity . Yet, lack of access to public spaces such as parks and hiking trails, as well as temporary closures of gyms/recreation centres, has had inevitable implications for physical activity. Despite the known benefits of physical activity for both physical and mental health, <1 in 5 Canadian adults met the current guidelines of 150 min per week of moderate-to vigorous-intensity aerobic exercise (Canadian Society for Exercise Physiology, 2020; Statistics Canada, 2019) prior to COVID-19 pandemic. Thus, the potential for even lower physical activity adherence rates (and its associated mental and physical health implications) as a result of a significant and substantial increase in barriers to physical activity during the pandemic is a probable reality. For example, a recent study byreported a significant decrease in self-reported perceived moderate-to-vigorous physical activity as a result of pandemic-related restrictions among a sample of just over one thousand Canadian adults. Moreover, using objectively measured physical activity data collected through a national physical activity promotion application, moderate-to-vigorous physical activity dropped significantly as a result of the onset of the pandemic, yet seemed to rebound 6 weeks later. Lastly, the Canadian Perspectives Survey Series March/April (Statistics Canada, 2020b), which collected data from 29 March to 3 April 2020, reported that 63% and 68% of Canadians were spending more time on the Internet and watching TV, respectively, compared with before the pandemic.
As a result, a call to action byurged researchers to investigate the associations between physical activity and COVID-19 parameters, including the need to develop methods for measuring physical activity during the pandemic and to identify factors that influence physical K E Y W O R D S COVID-19, pandemic, physical activity, sedentary behaviour, stress, wearable activity trackers activity during such unprecedented times. Therefore, the purpose of this study was to investigate how stress, physical activity and screen-related sedentary behaviour changed within the first month of the COVID-19 pandemic (March/April 2020). Two research questions (RQs) were specifically developed: RQ #1: How did stress, physical activity (objectively and subjectively measured) and screen-related sedentary behaviour levels change during the initial month of the COVID-19 pandemic (March/April 2020)?
RQ #2: What were the main reasons/barriers to the change in physical activity behaviour?
# Materials and methods
## Procedure and participants
The study was listed in the International Network of COVID-19 and Physical Activity Research (INCPAR) Repository (as suggested byyet not registered. Upon receiving University of Windsor Research Ethics Board Clearance (REB# 20-069), participants (18 years of age and older) were recruited during the month of April (2020) via a social media advertisement posted on Facebook, Instagram and Twitter, asking those who regularly wear activity trackers to contact the research team via email. The social media advertisement was originally shared by the researchers on their personal pages, and they encouraged friends/followers, community organisations and the university to repost and share the advertisement as well. Upon contact, an information letter, a survey link and fillable calendar of the month of March were emailed to prospective participants. If after two days the survey had not been completed and/or the calendar not returned, a reminder email went out to the participants. In total, 167 potential participants emailed the research team. Upon completion of the survey and return of the calendar, participants had a chance to win a $25 grocery/gas gift card (1 in 50 chance).
## Measures
Step count Objective physical activity was measured using daily step count (i.e. steps/day), as recorded from their wearable activity tracker (such as those made by Apple, Fitbit, Samsung and Garmin) that was collected via a fillable calendar. Mobile phones with step counts were not accepted as activity trackers.
Participants were asked to go to the application or online portal associated with their activity tracker to retroactively record their step counts for each day of March in the fillable calendar. Although the WHO (2020) declared COVID-19 a pandemic on 11 March, 'pre' and 'post' pandemic average step counts were determined for each participant based on the date each individual began physical distancing/self-isolating (or in some cases, quarantining) as a result of the pandemic (M = 16 March 2020, SD = 4.7 days, range = 13-31 days).
Wearable activity trackers have become popular because of their functionality and potential to provide individualized feedback when paired with a smartphone . Numerous devices and models exist (e.g. Apple Watch, Fitbit, Samsung and Garmin), all with varying capabilities and functionalities, and although differences in validity exist between devices for measuring step count under different/various conditions, the purpose of comparison was not between devices, but rather within people (over time). An assessment of 10 different consumer activity trackers suggested high reliability for test-retest reliability, and given the difficulties/challenges of assessing physical activity in free-living individuals, the use of activity trackers was deemed appropriate for use in the current study.
## Survey
Participants were also asked to complete a survey administered via Qualtrics XM (Provo, UT), which took approximately 20-25 min. All participants provided informed consent upon beginning the survey. Most questions were generally taken and/or adapted from validated and/or national surveys. The survey asked about physical activity behaviours, barriers and changes that occurred as a result of the COVID-19 pandemic, in addition to questions specific to COVID-19, stress/coping, lifestyle behaviours and demographics (e.g. gender, age, employment status and country). As previously mentioned, participants identified the day they began physical distancing/self-isolating as a result of the pandemic (including an option that they had not begun doing so) and this day (specific to each participant) was used to determine pre-and post-COVID-19 comparisons across all variables. Participants also confirmed they usually wear their activity tracking device at least 10 hr/day.
Daily stress was asked using two separate questions (i.e. pre-and post-COVID-19) 'Thinking about the amount of stress in your life, would you say that most days were/are?' with responses graded from 1 (not at all stressful) to 5 (extremely stressful; Statistics Canada, 2020d). Work stress was asked using two separate questions (i.e. pre-and post-COVID-19) 'Would you say that most days at work were/ are' with response options ranging from 1 (not at all stressful) to 5 (extremely stressful) (Statistics Canada, 2020d).
Subjective physical activity time (minutes/week) was measured using the question 'Prior to the COVID-19 pandemic [or "In the past 7 days"], how many times did you participate in moderate or vigorous physical activity? Please note, moderate or vigorous physical activity causes and increase in breathing and heart rate' with response options of 1-7 days/week . This question was followed up with 'On days that you were active, what was the average amount of time (in minutes) per day that you participated in moderate or vigorous physical activity?' . Responses were multiplied to calculate a weekly physical activity time (minutes/week).
Physical activity behaviour change as a result of the COVID-19 pandemic was asked with a newly developed question: 'Since the COVID-19 pandemic, do you think that your physical activity has' with response options of 'decreased a lot', 'decreased a little', 'stayed the same', 'increased a little' and 'increased a lot'. Responses were collapsed into 'decreased', 'stayed the same' and 'increased'. In order to assess reasons/barriers to the change in physical activity, participants were qualitatively asked: 'If there has been a change in your physical activity behaviour, why?' to which they responded open-endedly.
Pre-and post-COVID-19 screen-related sedentary behaviour was measured with two questions (pertaining to weekdays and weekends): 'On a school or workday [or "On a day that was not school or a workday"], how much of your free time do you spend watching TV or a screen on any electronic device while sitting or lying down' (Statistics Canada, 2020d). Response options included '2 hr or less per day', 'more than 2 hr but less than 4 hr', '4 hr to less than 6 hr, 6 hr to less than 8 hr', '8 hr or more per day' and 'was not at school or work'. Data were transformed into a continuous variable:
'60 min (2 hr or less per day)', '180 min (more than 2 hr but less than 4 hr)', '300 min hours (4 hr to less than 6 hr)', '420 min (6 hr to less than 8 hr)' and '540 min (8 hr or more per day)' and then prorated (by 5 or 2 days) and divided by 7 to capture pre-and post-COVID-19 average daily sedentary time (minutes/day).
# Data analysis
Minimal step count data were missing (5%; 197 data points out of a possible 3,751); as such, the average daily step counts were calculated using the step count divided by the total number of days with data. Separate paired-sample t-tests were conducted to determine whether there were statistically significant differences between pre-and post-COVID-19 stress (daily and work), objective step count (steps/day), subjective physical activity (minutes/week) and screen-related sedentary time (minutes/ day) changes. Effect sizes were measured using Cohen's d. Four separate one-way ANOVAs, with the Tukey post hoc testing, were conducted to determine whether there was a statistically significant difference between those that decreased, maintained or increased their physical activity and stress (daily and work), objective step count (steps/day) and sedentary time (minutes/day). Effect sizes were measured using omega-squared. Bivariate chi-square analyses were used to test the association between reasons/barriers to physical activity changes themes and physical activity changes categories (decreased, maintained and increased). Effect sizes were measured using Cramer's V. All statistical procedures were completed using the Minitab 17 computer software (Minitab, State College, PA, USA), with a level of significance set at p <.05. Moreover, given the current study's focus on healthrelated outcomes, no corrections for multiple comparisons were completed, as suggested byand.
Qualitative data were analysed using a thematic analysis. After a familiarization period, two of the authors (SJW and PC) discussed possible themes within the data set and a codebook was mutually agreed upon. The coding of themes was then completed independently with each comment potentially being coded into different themes (thus potentially overlapping). Once all data were coded, the coders met to discuss any differences and agree upon final themes (similar to.
# Results
Among the 167 participants who responded to the advertisement, 142 completed the survey and 132 provided valid pedometer data, resulting in a final data set of 132 participants (107 women, 23 men and 1 cis) from Canada (n = 121), the United Kingdom (n = 8), the United States (n = 2) and New Zealand (n = 1). Due to the small number of international participants (n = 11), they were removed from the current analysis to focus only on Canadian participants (N = 121). describes the participant demographics, in which the majority of the sample was Caucasian (n = 115; 88%), female (n = 96; 80%), with a mean age of 36.2 years (±13.12; ranged from 18 to 77 years). Most participants lived in Ontario (n = 115; 95%) and had a university certificate, diploma or degree at the bachelor's level or above (n = 86; 72%). Prior to COVID-19, the majority of participants worked outside the home (i.e. 68% versus 26% other/student/retired/unemployed and 6% working inside the home), whereas post-COVID-19, the majority were other/student/retired/unemployed (45%) compared with working from inside the home (40%) or working outside the home (15%).
## | 459
bs_bs_banner PHYSICAL ACTIVITY AND COVID-19 RQ#1: Changes in stress, physical activity (objectively and subjectively measured) and screen-related sedentary behaviour during the first month (i.e.
## March/april 2020) of the covid-19 pandemic
Pre-and post-COVID-19 differences between daily stress, work stress, daily step count, self-reported physical activity time (minutes/week) and sedentary time (minutes/week) are presented in . Most notably, step count significantly decreased over the month (p =.004), yet self-reported physical activity time was maintained (p =.450). Furthermore, an equal number of participants reported an increase (n = 55; 45%) or decrease (n = 55; 45%) in physical activity, with the remaining participants reporting no change (n = 11; 10%).
As such, breaks down stress (daily and work), step count and screen-related sedentary behaviour by those reporting that their physical activity increased, maintained or decreased. Daily step
## Rq #2: main reasons/barriers to changes in physical activity behaviour
Qualitatively, eight themes emerged as to why participants had a change in their physical activity behaviour including the following: access/equipment, incidental movement, motivation, time, coping, change in physical activity type, being sick/quarantined and to compensate for increased food consumption. Descriptively, includes the frequency of the themes, with access/equipment (n = 57), time (n = 47) and motivation (n = 33) cited most frequently. Access/equipment, the most cited reason for a change in physical activity (n = 57 times), was more frequently mentioned for those decreasing their physical activity (72%, n = 41) than those who increased (25%, n = 14), p <.001. Participants mentioned 'I do not have access to my typical gyms and do not own weights to simulate previous exercise habits' (male, 29 years), 'Gym closing, inadequate exercise equipment' (female, 27 years) and 'I don't work out nearly as long or hard because I'm not at the gym. The weather has not been the best so it's hard to go for a run outside or even a walk' (female, 23 years). Opposingly, for those who increased their physical activity, access/equipment was noted as 'I bought a bike, so I bike a couple hours when the weather is good' (female, 30 years) and 'My CrossFit membership allowed only 13 classes per month, now I have access 6x [times] a week at home via online classes so I can participate more' (male, 35 years).
PHYSICAL ACTIVITY AND COVID-19
The second most commonly cited barrier/reason to a change in physical activity was time (n = 47) and was cited more frequently among those increasing their physical activity (79%, n = 37) as opposed to decreasing physical activity (21%, n = 10); for example 'I have so much more time throughout the day to be active in ways I enjoy without feeling rushed or having other commitments to fill my time' (female, 25 years), 'I have more time to exercise now. My classes at the fitness centre have discontinued so now I am exercising at home more. I am trying to get my husband out walking more. I have less things to do in my day, so I have more time' (female, 61 years) and 'I have time since I am working from home, I don't have to worry about getting sweaty, I ride the stationary bike while doing zoom calls when no video is required I feel like I need it to create some routine and control in my day' (female, 49 years). However, time was also noted by those who decreased their physical activity. For example, participants stated 'Not enough time, gym is closed, not a lot of outdoor space, and no .005 0.079
Step count childcare' (female, 38 years), 'Change to home workouts, less time management, more time spent taking care of family' (male, 23 years) and 'Not as much free time' (female, 40 years). The third most commonly cited barrier/reason to a change in physical activity was motivation (n = 33) yet was not significantly different those who decreased, maintained or increased their physical activity. Participants mentioned 'Inability to go to the gym, lack of desired equipment, slightly less motivation' (female, 27 years), 'Gym has been closed and I find it hard to work out alone' (female, 31 years) and 'No gym classes so must work out at home to videos. Not as motivating as being in a group class' (female, 52 years). Opposingly, other participants mentioned an increase in motivation such that the 'Gym is no longer open so forced to do things at home -might as well exercise with my family to promote good health (physical and mental). Knowing I can help them keeps me motivated, in addition to having more time to work out as I am no longer working due to COVID-19 (just online school rather than both)' (female, 25 years) and 'I have more time to exercise as I'm not at work. My partner is also home, so I have someone to exercise with' (male, 47 years).
# Discussion
The purpose of this study was to examine how stress, physical activity and screen-related sedentary behaviour of Canadians changed within the first month (March/April 2020) of the COVID-19 pandemic. Overall, the current findings provide initial support for the importance of promoting and facilitating physical activity and limiting screen-related sedentary behaviour during the COVID-19 pandemic. This study also follows a call to action byto explore the effects of the COVID-19 pandemic on physical activity, with the aim to help inform public policy surrounding the current, and future, pandemic(s).
According to Statistics Canada (2020a, 2020c), the COVID-19 pandemic is associated with increases in stress and anxiety, as well as worsening mental health for Canadians, especially for women. Also, as many Canadians are working the frontlines, adapting to new ways of working, and/or having other significant changes in their day to day life, the COVID-19 pandemic has inherently introduced new stressors to the lives of Canadians. Overall, the findings of this study support those of Statistics Canada (2020a, 2020c), as participants reported significantly higher daily and work stress as a result of COVID-19. In addition, when grouping participants based on their self-reported ratings of physical activity (i.e. decrease, maintain and increase), there were no differences between groups in regard to the change in daily stress as a result of the COVID-19 pandemic. However, compared with participants who self-reported their physical activity to have maintained or increased within the first month of the pandemic, those who self-reported a decrease in their physical activity had a significantly greater change (increase) in their work stress. Despite significant increases in both daily and work stress across participants, it appears as though greater work-related increases in stress (e.g. working the frontline, change of work environment, working remotely, financial worries) had more significant impacts on participants' abilities to maintain their pre-COVID-19 levels of physical activity as a result of the pandemic. These findings align with research by Stults-Kolehmainen and Sinha (2014) andwho suggested that higher stress (in this case work stress) is associated with decreases in physical activity. However, this study provides additional insight that it is particularly work stress at the onset of COVID-19 pandemic that interfered with participants' abilities to maintain their physical activity.
Within the first month of the COVID-19 pandemic, significant decreases in participants' step count were observed. Specifically, prior to the COVID-19 pandemic, participants in the current study, most of whom are women, were accumulating more steps per day (i.e. just over 9,500 steps/day) compared with Canadian women (i.e. 8,400 steps/day) innationally representative study. However, as a result of the pandemic, average step count in the current study dropped to just under 8,500 steps/day, much closer to the national average previously determined by. In addition, the current findings support those ofwho examined physical activity during the COVID-19 pandemic across multiple countries (including Canada) and reported a decrease in step count relative to the severity of isolation and quarantine orders throughout the COVID-19 pandemic. However, when participants in the current study were asked to subjectively report the number of physical activity minutes they accrued pre and post the COVID-19 pandemic, no statistical differences were reported. Although such contradictory results could be partially related to recall bias and participants inaccurately reporting their physical activity behaviour from the previous month, it appears as though the discrepancy may also be reflective of participants' reduction in incidental physical activity and non-purposeful activity (e.g. chores, basic hygiene, walking to walk;caused by the public safety measures and restrictions put in place to limit the spread of COVID-19 among Canadians. For example, although there was no evidence of statistical differences in the number of self-reported physical activity minutes pre-and post-COVID-19 pandemic, many participants did note that working from home, leaving the home only for essential errands and/or having no access to recreation centres/public places decreased the amount of low-intensity or incidental physical activity they typically accrued throughout the day prior to the COVID-19 pandemic. As incidental physical activity contributes more to total energy expenditure in some people than dedicated periods of physical activity, the results of the current study highlight the potential damaging and adverse effects that the COVID-19 pandemic is having on the physical health of Canadians, even if they have maintained pre-COVID-19 pandemic intentional physical activity levels.
Results of the current study also demonstrated that screen-related sedentary behaviour levels increased significantly for all participants, with those who reported decreasing their levels of physical activity observing greater increases in sedentary behaviour. These findings are consistent with results from the Canadian Perspectives Survey Series (March/April) released by Statistics Canada (2020b), which evaluated levels of screen-related sedentary behaviour during the COVID-19 pandemic. As screen-related sedentary behaviour is associated with adverse health outcomes, public awareness surrounding the risks associated with sedentary behaviour, especially during the current pandemic, is warranted. The current findings stress the importance and necessity of public health interventions to address screen-related sedentary behaviour during a global pandemic. Public health interventions are especially important when, as in the current pandemic, it is uncertain how long stay at home orders and closures will last and, as a consequence, how long the increase in screen-related sedentary behaviour will persist.
Access to equipment, time and motivation were the three most common reasons/barriers participants used to explain changes in their physical activity behaviour. For participants who reported a decrease in physical activity, access to equipment was the most cited reason/barrier. This finding emphasizes the important role of gyms and fitness trainers continuing to offer their services during the pandemic and identifies a key opportunity for those individuals working in the health and fitness industry. Moreover, some participants who identified access/equipment as a reason/barrier to the increase in their levels of physical activity noted that they were able to purchase additional equipment to allow them to continue to exercise from home or were able to continue paying for a gym membership that switched to providing online workouts that could be performed with limited equipment. It should be noted that most participants in the current study were most likely of moderate-to higher socioeconomic status (i.e. 72% of participants had a university certificate, diploma or degree at the bachelor level or above, and all had already purchased wearable activity trackers), which may have provided them means to purchase new | 465 bs_bs_banner PHYSICAL ACTIVITY AND COVID-19 gym equipment or access paid-for online fitness classes, affording them the opportunity to continue their fitness regimen at home. The ability to continue a gym membership virtually or to purchase new gym equipment is not available for many Canadians and highlights the potential financial implications of the COVID-19 pandemic even further. Time was the second most cited reason/barrier to changes in physical activity and mainly acted as a facilitator for participants who reported increasing their physical activity behaviour. For example, many participants reported that working from home allowed them more flexibility and opportunity to exercise. This finding is unsurprising as lack of time is often noted as a barrier to physical activity.
# Limitations
Despite the need for research examining associations between the COVID-19 pandemic and numerous health outcomes (e.g. stress, physical activity and sedentary behaviour) and this study's contribution to thecall to action, it is not without its limitations. First, the study had a small sample size (N = 121) with the majority of participants being educated, younger adult women living in Ontario, which could be attributed to the nature of data collection of using wearable activity trackers. As such, the results are not generalizable to the Canadian population, but could provide some evidence for how the pandemic is affecting educated, younger adult women living in Ontario.
Users of wearable activity trackers are most often women in their 30s who have pursued higher education. Along those lines, not all participants wore the same brand/type of activity tracker, so it is possible that comparisons between people may not be entirely accurate. Yet, there is evidence supporting validity and inter-device reliability of step counts between different types of wearable activity tracker devices. Third, research bysuggested that individuals wearing activity trackers report greater levels of physical activity compared with those no longer wearing an activity tracker. Therefore, it is likely that our sample population engaged in more physical activity prior to the COVID-19 pandemic than the general population and our findings may not translate to those who do not regularly wear an activity tracker.
Additionally, correcting (or not) for multiple comparisons is a subject of great debate among researchers, especially for those conducting health-related research. The current study did not correct for multiple comparisons, but acknowledge that not all researchers hold the same views regarding this subject.
Also, although participants confirmed typically wearing their activity tracker 10 + hours/day, it would have been beneficial to include the device wear time as a covariate in the quantitative analyses. Finally, participants responded subjectively about their levels of stress, physical activity screen-related sedentary behaviour alongside the objective wearable activity tracker data, with self-report data being subject to recall and social desirability bias. It should also be noted that the transformation of sedentary behaviour from a categorical to continuous variable may have influenced the current results. However, data collection occurred in a timely manner and used reliable and valid survey questions to mitigate such biases as much as possible.
# Conclusion
This study responded to the call to action byby examining how stress, physical activity and screen-related sedentary behaviour changed with the first month of the COVID-19 pandemic for Canadians. Current findings provide initial evidence that daily stress and work stress have increased, that physical activity levels (as measured by step count) have significantly declined as a result of the pandemic, and that screen-related sedentary behaviour has also increased. The study also suggested a decrease in physical activity, as a result of the pandemic, is associated with a larger increase in work stress, compared with those who self-reported their physical activity to have been maintained or increased. Lastly, these findings were contextualized with reasons/barriers to changes in physical activity. Overall, such findings emphasized the importance of promoting and facilitating physical activity and limiting screen-related sedentary behaviour during a pandemic. However, more research is needed to examine the ongoing impact of the COVID-19 pandemic on stress, physical activity (especially using objective measures) and sedentary behaviours, especially as restrictions are lifted and/or Canadians adapt to the pandemic in the long term. |
Flow Cytometric Diagnosis of Paroxysmal Nocturnal Hemoglobinuria: Pearls and Pitfalls – A Critical Review Article
A R T I C L E I N F O A B S T R A C TParoxysmal Nocturnal Hemoglobinuria (PNH) is a rare blood disorder characterized by chronic intravascular hemolysis, thromboses in unusual sites and cytopenias related to bone marrow failure. The diagnosis is based on the Flow Cytometric (FCM) detection of peripheral blood cell clones lacking the surface molecules linked to the GPI anchor, which is altered by mutations.Consensus studies have developed standardized and robust multicolor FCM assays to disclose PNH clones among red cells, neutrophils and monocytes at a high level of sensitivity and accuracy. High-resolution procedures have been also established to detect small PNH clones at a sensitivity level of around 0.01% in red cells and neutrophils. Cell clone type and size have been put into correlation with the clinical presentations of the associated diseases, and recommendations for the clinical follow-up have been established. The recent advent of the therapeutic monoclonal
## A r t i c l e i n f o a b s t r a c t
Paroxysmal Nocturnal Hemoglobinuria (PNH) is a rare blood disorder characterized by chronic intravascular hemolysis, thromboses in unusual sites and cytopenias related to bone marrow failure. The diagnosis is based on the Flow Cytometric (FCM) detection of peripheral blood cell clones lacking the surface molecules linked to the GPI anchor, which is altered by mutations.
Consensus studies have developed standardized and robust multicolor FCM assays to disclose PNH clones among red cells, neutrophils and monocytes at a high level of sensitivity and accuracy. High-resolution procedures have been also established to detect small PNH clones at a sensitivity level of around 0.01% in red cells and neutrophils. Cell clone type and size have been put into correlation with the clinical presentations of the associated diseases, and recommendations for the clinical follow-up have been established.
The recent advent of the therapeutic monoclonal Bruno Brando, Arianna Gatti, Frank Preijers Flow cytometric diagnosis of paroxysmal nocturnal hemoglobinuria antibody Eculizumab has dramatically improved both the quality of life and the life expectancy of the affected patients, further increasing the importance of an accurate FCM detection and monitoring of the clones. The technical features of the FCM diagnostic workup and the many critical aspects of the analytical process are discussed here.
# Introduction
Paroxysmal Nocturnal Hemoglobinuria (PNH) is a rare, acquired clonal blood disorder of the hematopoietic stem cell characterized by a triad of signs and symptoms: chronic intravascular hemolysis with recurrent crises, thromboses in unusual sites and cytopenias related to bone marrow failure (BMF).
Its name was coined early in the 20th century, but after the extensive research made during the last few decades, nowadays the three key terms characterizing the disease may sound somewhat misleading.
While some 90 percent of patients experience hemoglobinuria at some point of their clinical course (1), hemoglobinuria is seen in about one quarter of cases only at diagnosis. Chronic hemolysis with hemoglobinuria occurs not only during nighttime, but 24 hours per day, it is most often asymptomatic and may present as 'paroxysmal' only occasionally.
PNH is also known as 'the most vicious acquired thrombophilic state', probably through the free hemoglobin itselfand the reduced intravascular nitric oxide, although its multifactorial biological basis still remains partly unknown.
The strong association between PNH and BMF syndromes, including myelodysplasia (MDS) and aplastic anemia (AA), is well known and recommendations have been published to include PNH screening in the clinical workup of the BMF syndromes, MDS and AA.
A number of other apparently incoherent clinical signs and symptoms are often experienced by the affected patients, namely fatigue, dyspnea, abdominal pain, pulmonary hypertension, renal insufficiency, dysphagia, erectile dysfunction, thus making the clinical assessment long and difficult, with frequent misdiagnosis and mismanagement.
Except for Direct Antiglobulin Test (DAT)-negative hemolysis tests and cytopenias, which may focus medical attention, ordinary laboratory assessments usually do not provide specific clues for the diagnosis of PNH.
The uncontrolled complement activation, leading to life-threatening hemolysis and thromboses forms the basis of symptomatic PNH. The pathogenic mechanism of PNH resides in the faulty synthesis of the cell membrane protein anchor Glycosyl-Phosphatidyl Inositol (GPI) due to mutations occurring in the PIG-A gene located on chromosome X. GPI is the anchor for many functional cell surface glycoproteins (13): the reduced or absent expression of the GPI-linked complement-regulator molecules on red blood cells (RBC), the Decay Accelerating Factor (DAF, also known as CD55) and the Membrane Inhibitor of Reactive Lysis (MIRL, also known as CD59) are the cause of hemolysis in PNH. In PNH all blood cell lines deriving from a mutated stem cell carry the same genetic and phenotypic defect, and can coexist as 'PNH clones' along with normal, unmutated counterparts. The flow cytometric (FCM) demonstration of the absence of GPIlinked molecules in a sizable fraction of peripheral blood red cells, neutrophils and monocytes is the cornerstone of the diagnostic process of PNH. Three types of clinical presentation of PNH are schematically considered: the classical form, associated to overt hemolysis Flow cytometric diagnosis of paroxysmal nocturnal hemoglobinuria and large PNH cell clones; the PNH associated to BMF/AA with hemolysis and usually small clones, and the subclinical PNH, with no or minimal hemolysis and variable signs of BMF. A more extensive discussion on the clinical spectrum of PNH presentations and the associated laboratory features can be found in the recent literature.
Blood transfusions, anticoagulants and steroids have been the only therapeutic options for PNH for decades. Patients with PNH occurring in the context of BMF take advantage from the specific treatment of the underlying bone marrow disorder. In recent years, the advent of the complement inhibitory monoclonal antibody Eculizumab has dramatically improved the prognosis and the quality of life of patients with symptomatic PNH, although the only cure may reside in allogeneic stem cell transplantation in selected cases.
## Flow cytometric diagnosis of pnh
## The reasons for testing
The classical list of signs and symptoms which may prompt a testing order for PNH includes haemoglobinuria or haemosiderinuria, unexplained DAT-negative haemolysis, aplastic or hypoplastic anaemia, MDS, thrombosis in unusual sites, dystonic symptoms (dysphagia, abdominal pain, erectile dysfunction). However, not all such signs and symptoms have the same diagnostic efficiency in helping the definition of a clinical PNH, especially when occurring separately. In general, when the above listed signs and symptoms combine, the clinical yield increases according to . For instance, testing for an underlying PNH in case of thrombosis alone may account for a clinical yield as low as 0.4 to 2.7%,, whereas if thrombosis is associated with anemia and/or other cytopenias, the diagnostic efficiency may raise up to about 14%.
## Preanalytical issues
The available guidelines on FCM analysis of PNH have not addressed the specific preanalytical requirements of PNH analysis (1) or have provided only very concise indications. In the year 2000 guidelines, the maximum blood sample age for analysis was reported as 7 days for RBC and 48 hours for white blood cells (WBC), with the recommendation to refrigerate samples at 4°C if storage time exceeds 24 hours. The more recent guidelines (17) just recommend less than 48 hour storage time for RBC with no further details.
It is important to consider carefully the peculiar analytical context of FCM analysis of PNH, where the absence rather than the presence of cell surface markers has to be demonstrated. The cell viability in anticoagulated blood samples progressively decreases with storage time, and may introduce unpredictable changes in the cell staining features, both increasing or decreasing artifactually the binding properties of conjugated antibodies or of other reagents.
Pre-analytical procedures, especially ammonium chloride lysis, may kill cells which are already damaged in aged samples. Dead and apoptotic cells can strongly influence PNH clone detection, while also causing non-specific antibody binding, increased autofluorescence and cell aggregation.
To avoid under-or overestimation of clone size a viability dye, i.e. DRAQ7, can be used to exclude dead cells. DRAQ7 is a fluorescent impermeant dye (excited at 630 nm, emitting at >670 nm) that does not enter intact cells. When the cell membrane integrity is compromised DRAQ7 enters the cells and readily binds to nuclei to report cell death, thus improving the assay specificity and reducing artifacts. The usage of erythrocyte bulk lysis procedures by ammonium chloride buffer before staining has been found useful in reducing the background signal and to increase the resolution between positive and negative cell populations.
## The flaer reagent
The Aeromonas hydrophila toxin Aerolysin was demonstrated to be a highly specific ligand for the GPI molecule on white blood cells, thus making it a suitable indicator of GPI-deficient leucocytes in PNH.
An inactivated non-toxic, fluorochrome-conjugated molecular variant was then developed under the acronym FLAER (FLuorescent AERolysin) and extensively applied as a highly effective non-antibody reagent to be used in FCM studies of PNH leucocytes, along with other conjugated monoclonal antibodies (MoAb).
The binding properties of FLAER on blood cells have been accurately studied, and the higher signal resolution obtained by the pre-staining bulk lysis of the sample has been demonstrated.
## Cell staining protocols
The 2018 ICCS/ESCCA guidelineshave provided a fully comprehensive series of technical recommendations on sample preparation, reagent titration, MoAb clone and fluorochrome selection.
Bulk erythrocyte lysis of 1-2 ml of anticoagulated blood with ammonium chloride-containing buffer is useful to enrich cytopenic samples and to concentrate the cell pellet for high-resolution analyses requiring the collection of large numbers of events.
For every cell type to be analyzed, gating antibodies are first used to capture the target populations, then GPI-deficient cells are defined in each subset using the appropriate MoAb combination and FLAER, where applicable. Several possible staining alternatives have been studied and optimized.
It is important to restrict the analysis to mature WBC populations using MoAb reactive with GPI molecules fully expressed by peripheral blood mature cells and not by their progenitors, since these cells usually express GPI-linked markers at a lower density. To determine GPI-deficient RBC clones, of course, a lysis procedure should not be used. RBC are first diluted 1:100 in phosphate buffer (PBS), stained with anti-glycophorin-A (CD235a) for gating and CD59 for GPI-deficiency, then two washes using PBS supplemented with bovine serum albumin (BSA) are mandatory before reading. The FSC-W vs FSC-A pulse analysis of RBC physical parameters can be also of help in defining the RBC gate.
[formula] Becton Dickinson FACSCanto (2-Laser) FLAER- Alexa 488 CD24-PE (Clones SN3, ML5) CD15-PerCP- Cy5.5 (Clone MEM-158) CD64-PE- Cy7 (Clones 10.1, 22) CD14-APC (Clone MφP9) CD45-APCH7 (Clone 2D1) Beckman Coulter Navios (3-Laser) FLAER- Alexa 488 CD24-PE (Clones SN3, ALB9) CD15-PC5 (Clone 80H5) CD64-PC7 (Clones 10.1, 22) CD14-APC700 (Clone RM052) CD45-KO (Clone J33) [/formula]
Thiazole orange pre-staining of reticulocytes within the RBC cluster can be also used to measure the An example of PNH analysis using the recommended guidelines in a highly symptomatic patient, off therapy proportion between normal and GPI-deficient erythropoiesis, which is very useful in evaluating the red cell compartment response to therapy.
[formula] E) l G) [/formula]
The state-of-the-art FCM technique to study PNH clones among WBC is a five-to six-color combination that is adjusted slightly according to instrument brand, optical filter sets and laser types.
For WBC analysis whole blood is preferably bulklysed and washed twice with PBS+BSA before staining. CD45 and side scatter are used along with CD15 and CD64 to gate neutrophils (with exclusion of eosinophils that express CD15 in lower amount) and monocytes, respectively. FLAER and CD24 is the optimal reagent combination to identify GPI-deficient neutrophils, whereas FLAER and CD14 are used to study PNH monocyte clones.
Lymphocytes are not useful in PNH diagnosis, due to their very long lifespan and to an excess variability in the expression of surface GPI-linked molecules. The role of lymphocytes in PNH analysis is currently limited to internal staining controls.
Platelets in PNH have a normal lifespan despite expressing the same GPI-molecule deficiency of RBC, and are not included in routine diagnostic procedures (6).
The reported limited availability or the high cost of FLAER in certain countrieshave prompted the development of non-FLAER based staining protocols, introducing CD157 as a replacement for CD24 and CD14 to determine PNH clones in neutrophils and monocytes, respectively. However, the reduced CD157 expression by eosinophil granulocytes may cause artifacts in case of imprecise gating of CD15+ neutrophils, and genetic or ethnic variants of CD157 have been described which may account for the detection of false PNH clones in some individuals.
Recently, additional multicolor FCM protocols have been described, coupling FLAER to CD157 (7-colors)or adding CD5 and CD19 as a dump channel to reduce background fluorescence (8-colors). The advantages provided by this increase in fluorochrome number and complexity are still to be fully evaluated.
## Instrument setup, optimization and gating syntax
Modern digital flow cytometers are equipped with very efficient setup and calibration bead suspensions that should be used to optimize automatically the photomultiplier voltages, the thresholds, the intra-and inter-laser fluorescence spillover compensation, ensuring instrumental consistency with time and comparable results among laboratories using the same type of equipment. Therefore, if manufacturer's directions are strictly applied, the FCM setup is nowadays easy to standardize, thus making the instrument-related variables almost negligible.
The creation of an appropriate gating syntax, however, is still a manual process that requires skilled operators and sample-to-sample adjustments in order to correct for the unavoidable inter-patient variability in cell cluster positioning.
## The limits of detection -coping with background events
As mentioned, in PNH analysis the lack rather than the presence of cell surface molecules is the indicator of the disease-related abnormalities. The presence of some GPI-negative elements in the various blood cell subsets of healthy subjects is a well-known phenomenon and represents the assay background.
Furthermore, the presence of blasts or other immature cells, as it occurs in MDS, may strongly influence the amount of background events.
## Bruno brando, arianna gatti, frank preijers flow cytometric diagnosis of paroxysmal nocturnal hemoglobinuria
This requires a strict check of the analytical process using a cohort of healthy donors to control and minimize the background event frequency, which in turn vary in the different cell populations and in the different clinical settings. Moreover, it is mandatory to pay attention to the numerous biological and staining artifacts that may generate negatively stained cells and false 'PNH events' in the background, such as the presence of autofluorescent events that come as diagonally shaped or 'angle of dangle' parasite events in fluorescence plots.
Before the advent of FLAER the high background level in healthy controls and other technical limitations made it difficult to enumerate accurately PNH events below 1%. Following the more recent FLAER-based guidelines on high-resolution PNH analysis, a clearer picture of the lower limits of detection (LLOD) and of quantitation (LLOQ) of PNH events can be obtained. These issues have been thoroughly discussed and validated in the recent study by Payne.
It is important to mention here how the cytometry researchers have tried over the years to translate the concepts of lower limit of blank (LLOB), LLOD and LLOQ, typical of chromatographic or spectrophotometric procedures (55, 56), into a technology based on event accumulation. This transition was made -rather simplistically -by analogy with clinical chemistry, but surely without applying the same methodological care, and it is still matter of debate.
With the use of a negative or normal sample, the events falling into the final acquisition window may be taken as the LLOB, and ratioed to their cellular counterpart (the 'clean cell denominator') to obtain the 'blank' level to be subtracted from the analysis of the relevant cell population.
The Poisson's statistics governing rare cellular event analysis mandates the acquisition of at least 100 relevant events to generate measures with the acceptable coefficient of variation of 10%. In the delicate field of high-resolution FCM, however, the collection of 100 events can be sometimes out of reach, so a compromise has been developed, proposing a lower limit of 20 (31), 25or 30 (58) relevant events for the calculation of LLOD in high-sensitivity PNH and leukemia minimal residual disease analyses, keeping a universally accepted limit of 50 events for the LLOQ. Little attention has been paid however to the inherent variance of such a new approach, which dramatically increases with the lowering of the relevant event number. Despite increasing the acquisition to millions of clean cells, the lower the established minimum level of the detectable relevant events, the wider the confidence interval and the variability of the cytometric measure, according to Poisson's statistics.
In the daily practice, assuming 20 events (31) or 30 events (58) as the minimum size of a relevant cell event cluster for detection, and 50 events as the minimum for quantitation, the calculation of LLOD% is: 2,000 (or 3,000) / total number of clean events; and the calculation of LLOQ% is: 5,000 / total number of clean events.
When the same criteria for minimal residual disease studies with high resolution FCM analysis of hematological malignancies are applied (i.e. from 500,000 to >1,000,000 clean cell events as the denominator (58), the median background 'PNH' Type III RBC events in normals are from 3 to 5 per million, with a range from 0 to 17 per million or 0.002-0.007% and a LLOQ around 0.005%. Type II RBC events are often difficult to dissect precisely from Type I and Type III populations, moreover they can be generated by artifacts. As a consequence, Type II detection levels are usually not addressed in high-resolution PNH studies. For CD15+ neutrophils the average background in normals ranges from 2 to 10 per million or 0 to 0.001%
## Bruno brando, arianna gatti, frank preijers
Flow cytometric diagnosis of paroxysmal nocturnal hemoglobinuria. Similar figures in normal subjects have been consistently reported by other groups. For CD64+ monocytes, the low number of events that can be acquired in clinical samples greatly reduces the achievable sensitivity level, with background levels ranging around 0.1%.
## Pnh clone patterns -diagnostic criteria
The total absence of GPI-linked molecules defines a Type III PNH clone, whereas a partial deficiency defines a Type II PNH clone. It is customary to quantify and report the presence of Type II clones in the RBC population only, while WBC Type II (if detectable) and Type III clones are usually summed up in the reporting, with no further details.
RBC PNH clone size may vary remarkably according to the occurrence of hemolytic crises, to the administration of blood transfusions or Eculizumab, so RBC PNH clones may not reflect accurately the disease extent, are often smaller than the respective WBC clones and are not an ideal target for PNH quantitation at diagnosis. The established guidelines state that the absence of two GPI-linked molecules must be demonstrated in WBC lineages to corroborate the diagnosis. Neutrophil and monocyte PNH clones usually tend to be of similar magnitude. In a small percentage of patients, however, marked discrepancies may occur, more often with PNH monocyte clones outnumbering neutrophil clones. PNH clone size and type can also change with time, and can both increase or decrease in an unpredictable fashion, thus indicating the need for a follow up with time. The correlations between PNH cell clones and the associated clinical pictures -also in pediatric patients -have been extensively reviewed, as well as the recommended criteria for the longitudinal follow-up of small clones.
Patients under Eculizumab treatment usually display expanded RBC clones, due to the increased lifespan of GPI-deficient erythrocytes. In such patients the abrupt discontinuation of Eculizumab may increase the intensity and the severity of hemolytic crises.
## The high-resolution pnh analysis and assay precision
Until a few years ago a technical distinction between 'routine diagnostic' and 'high-resolution' PNH analysis was commonplace in the literature and considered in external quality assessment exercises. With the advent of the highly standardized recent protocols such a difference has lost its significance, since using the same sensitive laboratory technique both routine diagnostic and high-resolution analysis of PNH clones can be accomplished accurately. It is well known that PNH+ patients can have a neutrophil PNH clone of <1% in some 42% of cases (31), thus stressing the need for the systematic usage of a very sensitive FCM technique. Small PNH clones are more common in patients with BMF syndromes, may change their size with time and seem positively correlated to the response to immunosuppressive treatments. The term 'High resolution FCM' refers mainly to the collection of an adequate number of cell events (i.e. from 500,000 to 1-2 million), which enables analyses at a sensitivity levels around 10 -4 with LLOQ equal or less than 0.005%. This feature can be easily applied to RBC and neutrophil analyses, but such a level of resolution is almost impossible to achieve in the routine with monocytes, due to their low level in peripheral blood.
As expected in any FCM rare event analyses, the lower the PNH clone percentages the worse the assay precision, ranging from 14 to 25% at the lowest limit of sensitivity. Moreover, the precision in detecting a very small but well resolved Type III clone is usually better that the Flow cytometric diagnosis of paroxysmal nocturnal hemoglobinuria one achieved with larger but less clearly separated Type II clones. The acceptable limit of precision for clinical purposes is not yet well established in the literature. Very small PNH clones, however, do not make a diagnosis of clinical PNH and usually do not indicate the need of therapeutic measures, although a strict monitoring also of small clones over the time is mandatory in any instance.
## Assay results -reporting pnh analysis
In the reporting of PNH analyses it is important to use a standardized terminology, in order to avoid a common shortcoming, namely the overdiagnosis of clinical PNH when only small GPIdeficient clones are detected. The recommended reporting terminology to be used is: 'Presence of a PNH Clone [specify the cell type(s)] ' in case of cell clones >1%; 'Presence of minor [specify the cell type] PNH population(s)' from 0.1% to 1%; 'Presence of rare GPI-deficient cells [specify the cell type]' when <0.1%. The achieved sensitivity level should be specified in the reporting for every cell subset, since it may vary according to the patient's status and the overall cell collection. As mentioned, the presence of Type II clones should be reported for RBC only.
## Screening tests versus diagnostic confirmation
PNH is a very rare disease, with a diagnostic yield ranging from 6% to 14% under the best reason for testing conditions. This means that testing for PNH will give negative results in the vast majority of the checked cases. In the literature, a primary screening test before a diagnostic confirmation assay in selected patients has been repeatedly proposed to rule out the more numerous negative cases. The current 6-or 7-color technical guidelines can be costly, complex and demanding for small peripheral laboratories, thus preventing a diffuse first-level disease surveillance, which should be practiced in every blood transfusion center and hematology laboratory. Due to their favorable features, neutrophils and monocytes may serve better than RBC as the preliminary index population for the screening of new PNH casesin patients with the appropriate reasons for testing. In some countries samples with an initially positive first-level screening test for PNH are then referred to an accredited central facility for diagnostic confirmation, using a state-of-theart multicolor technique. A simplified and inexpensive PNH screening test, which can be performed upon the initiative of the laboratory itself and affordably repeatable over the time, is a cost-effective option especially in resourcerestricted countries. This approach proved valuable in disclosing new PNH cases in local studies, and can be a practical solution to overcome the disappointing level of disease surveillance reported in some settings.
## External quality assessment for pnh
The blood stabilization technology implemented by United Kingdom National External Quality Assurance Schemes (UK NEQAS) has been applied in developing stable PNH samples to be used as process control and in external quality assessment studies. This advancement has greatly facilitated the diffusion of the FCM analysis of PNH and stimulated the participating centers in selecting the optimal procedures, antibody clones and fluorochromes. The periodic surveys made by UK NEQAS, however, still highlight a certain degree of variability in the interpretation and usage of the available guidelines by some participants, which may account for an excess of non-specific assay results in some instances (i.e. PNH clone erroneously detected in healthy donor samples). Two separate PNH schemes -ordinary and high-sensitivity -have been developed initially, and were subsequently merged for the reasons discussed above.
## Pifalls and artifacts
Provided the reasons for testing are clinically appropriate and the recommended techniques are applied, the FCM operators involved in PNH analysis may come through a series of technical artifacts and pitfalls with real-life samples, that require an in-depth expertise for the full control of the technology.
Despite CD55 is traditionally included in the list of GPI-linked RBC molecules, Anti-CD55 MoAbs have been progressively dismissed from FCM panels due to the weak antigen expression. When RBC are stained with anti-CD59, at least two washing steps are then required to eliminate excess fluorochrome and non-specific MoAb binding, which greatly interfere with positive/ negative discrimination.
A typical problem with WBC analysis is represented by the appearance of putative small Type II clones, especially in the monocyte gating, often due to errors in mixing the sample with the reagents, to an imprecise gating or to cell adhesion phenomena. Basophils and NK cells that can be erroneously captured in the neutrophil or in the monocyte gate are characteristically stained dimly by FLAER, and can simulate Type II PNH clones. Myeloid blasts that often circulate in patients with BMF syndromes behave the same way, and can be captured in the monocyte gate, especially if they express CD64. Large platelet aggregates can also be dragged into the monocyte gate, and may account for FLAER-negative events. This artifact can be effectively prevented by bulk lysis and extensive washings.
Another common problem is represented by the acquisition of an insufficient number of cell Flow cytometric diagnosis of paroxysmal nocturnal hemoglobinuria events and by the calculation of improper PNH clone percentages, in the erroneous belief to have achieved a sufficient level of assay sensitivity. The latter problem is particularly evident with monocytes, that in some instances cannot be collected in sufficient number for a true highsensitivity analysis.
## Concluding remarks
Flow Cytometry is today the laboratory technique of choice to provide an accurate, sensitive, standardized and widely applicable diagnosis of PNH. The diagnostic efficiency largely resides in the appropriateness of the reasons for testing, which may be a major conditioning factor in the clinical workup of this rare and elusive disease.
The state-of-the art multicolor FCM technology is mature, robust, covered by international ISO17043-certified external quality assessment schemes, and enables the detection also of very small PNH clones with good accuracy and precision. The implementation of a diffuse screening network in partnership with central, secondlevel reference facilities seems today a rational and affordable model to favor the disclosure of new PNH cases. |
Overweight and Obesity Among Chinese College Students: An Exploration of Gender as Related to External Environmental Influences
[bib_ref] Physical activity, food choice, and weight management goals and practices among U.S...., Lowry [/bib_ref] [bib_ref] College weight gain and behavior transitions: Male and female differences, Cluskey [/bib_ref] [bib_ref] Beijing: China Statistics Press. Department of Disease Control and Prevention, National Ministry..., Debate [/bib_ref] [bib_ref] Physical activity, food choice, and weight management goals and practices among U.S...., Lowry [/bib_ref] [bib_ref] Obesity and socioeconomic status in developing countries: A systematic review, Dinsa [/bib_ref] [bib_ref] Impact of socioeconomic status on prevalence of overweight and obesity among children..., Viswanathan [/bib_ref] [bib_ref] Overweight, obesity and influence of stress on body weight among undergraduate medical..., Soma [/bib_ref] [bib_ref] Prevalence and correlates of being overweight or obese in college, Odlaug [/bib_ref]
# Methods
## Study area and participants
This study reports individual data from the Global Health Professions Student Survey (GHPSS) on Tobacco Control in China GHPSS. While this project primarily focused on smoking and tobacco control, it also was concerned with other aspects of health, including, behavioral issues, mental stress, obesity, and others. Students were recruited at 50 of the 60 universities participating in the program to complete the survey and these 50 universities residing in 42 cities across China and differentiated by regional location. Twenty-two of them were medical universities offering only medical programs and 28 were comprehensive universities offering medical and nonmedical programs. A detailed description of the selection process can be found in .
## Data collection
Individual variables were measured via a structured self-administered questionnaire. The questionnaire was administered during regular class meetings and took approximately 30 min to complete. All responses were anonymous. A common research protocol was utilized across all 50 universities to assure homogeneity of questionnaire administration and data collection techniques. A detailed description of the research protocol can be found in . Regional variables were retrieved from the National Bureau of Statistics database (Department of Comprehensive Statistics of The National Bureau of Statistics, 2014; Department of Urban Social Economic survey of the National Bureau of Statistics, 2014).
## Measures
Dependent variable. Height and weight were measured through self-report. Body mass index (BMI) was calculated by dividing body weight (kg) by squared height (m 2 ). Individuals with BMI scores of 24.0-27.9 kg/m 2 were categorized as overweight, and those with scores of ≥28.0 kg/m 2 were categorized as obese. These categories are consistent with the national standard for overweight and obesity utilized in China (Department of Disease . Self-report measures of overweight and obesity have been shown to be reliable in this population as there is a high correlation with objective measures for these variables (97.1% for the combined prevalence of overweight and obesity, and 98.8% for obesity [bib_ref] Individual and regional association between socioeconomic status and uncertainty stress, and life..., Yang [/bib_ref].
Individual-level independent variables. Sociodemographic questions were included on the questionnaire to determine age, gender ethnicity, father's and mother's occupations, and others.
Life stress was operationally defined as something causing college students to worry on a daily basis. It was measured through a standardized questionnaire designed by [bib_ref] Individual and regional association between socioeconomic status and uncertainty stress, and life..., Yang [/bib_ref]. The instrument had acceptable validity and reliability, and has been used extensively in Chinese research. The questionnaire consisted of eight items covering such issues as financial situation, health status, interpersonal relationships, study stress, and other problems. All items were rated on a 5-point scale where 0 is feeling no stress, 1 is little stress, 2 is some stress, 3 is more stress, and 4 is much more stress. The total score was obtained by summing up the scores for each individual item. Consistent with prior practice, a score of 24 or more signified severe stress [bib_ref] Work stress, life stress, and smoking among rural-urban migrant workers in China, Cui [/bib_ref] [bib_ref] Individual and regional association between socioeconomic status and uncertainty stress, and life..., Yang [/bib_ref].
Regional contextual independent variable. Several independent variables reflected potential regional variation. The first regional variable included in this study was level of economic development, as measured by per capita Gross Domestic Product (GDP) in Yuans. Both the GDP of the original home province (home region GDP) which the students came from and of the city (university city GDP) where their university was located were measured. The above data were obtained from the National Bureau of Statistics (Department of Comprehensive Statistics of National Bureau of Statistics, 2014; Department of Urban Social Economic Survey of the National Bureau of Statistics, 2014). The second regional variable was unemployment rate. Unemployment is defined as when people are without work and actively seeking work. Unemployment is associated with problems in economic growth and development that produce stress for residents .
Unemployment rates in cities where universities are located may also create stressors for college students because of decreased work opportunities and pay levels, as well as feelings of frustration and hopelessness related to future employment, etc. In this study, we included urban unemployment rates as an indicator of regional stress on college students. University city unemployment rate was operationalized as the number of unemployed people (in millions) in the city where the university was located and was based on data obtained from the National Bureau of Statistics (Department of Urban Social Economic Survey of the National Bureau of Statistics, 2014).
# Data analysis
All data were entered into a database using Microsoft Excel. The dataset was then imported into SAS (9.3 version) for statistical analyses. Descriptive statistics were calculated to determine the prevalence of overweight and obesity. A logistic model was utilized to assess associations between the dependent and independent variables. Both unadjusted and adjusted methods were considered in the data analyses, and implemented to examine these associations in males and females, respectively. The unadjusted method used only the key factors of interest as independent variables in the analyses, while the adjusted method added all of the possible confounders listed in Tables 1 and 2 as covariates in the logistic models. SAS survey logistic procedures were applied in the all analyses, using the university as the clustering unit, in order to account for a within-clustering correlation, attributable to the complex sample for unadjusted analysis. Associations were confirmed through application of a multilevel logistic regression model using the SAS GLIMMIX procedure. Series models were built for each primary predictor, with adjustment for the influence of potentially confounding sociodemographic characteristics. Initial analysis started with the Null Model, a three level (individual, original region, and cities where universities were located) model with random intercepts in building the overweight and obesity multilevel logistic regression model. To this base, sociodemographic variables and one primary predictor as fixed main effects were added to form a multilevel model for evaluating the impact of overweight and obesity. Only variables significant in the univariate analysis for the total sample were included in the final analysis. All individual and contextual variables, with categories, are listed in Tables 1 and 2. The first category for each variable served as the referent in the logistic regression analysis (see.
To assess for the presence of multicollinearity, the dependent variable from the logistic regression analysis was used as a dependent variable in a linear regression. The collinearity diagnostic statistics are based on the independent variables only, so the choice of the dependent variable does not matter. The multicollinearity assumption was evaluated during data analysis, using the variance inflation factor (VIF) to demonstrate the absence of multicollinearity among independent variables. Values of VIF exceeding 10 are often regarded as indicating multicollinearity(see [fig_ref] Table 3: Parameter Estimates From Multiple Level Models [/fig_ref].
All analyses were weighted. Weights included: (a) sampling weights, as the inverse of the probability of selection, calculated at university; and (b) post-stratification weights, calculated in relation to sex, based on estimated distributions of this characteristic from a national survey (National Ministry of Education, 2016). The final overall weights were computed as the product of the above two weights [bib_ref] Weighted estimation in multilevel ordinal and binary models in the presence of..., Grilli [/bib_ref]. A nonresponse weight was not utilized because nonresponse rates were so low in this study.
Unadjusted logistic regression analyses were weighted using the overall participant-level weights, and the multilevel analysis was weighted using sampling weight and subject-level weights with post-stratification weights, respectively [bib_ref] Weighted estimation in multilevel ordinal and binary models in the presence of..., Grilli [/bib_ref].
# Results
A total of 12,211 questionnaires were completed by students from 50 universities. Of these responders, 11,954 were fully completed and available for analysis. BMI was calculated for the 11,673 responders who provided complete height and weight data; the remaining 281 (2.7%) had missing data and were eliminated from the analysis. There were no significant differences in demographic characteristics between responders and nonresponders for BMI. Of sample 4,177 (49.1%) were male, 7,496 (50.9%) were female. There were no significantly differences for most variables of the demographic characteristics between male and female (see [fig_ref] Table 1: Demographic Characteristics of Sample [/fig_ref].
The prevalence of overweight and obesity combined was 9.5% (95% CI [7.7, 11.3]) in the overall study sample, 13.9% (95% CI in males, and 6.1% (95% CI [4.1, 8.1]) in females. Overall obesity prevalence alone was 2.2% (95% CI [1.3%, 3.1]), with a prevalence of 3.0% (95% CI [1.5%, 4.4]) among males and 1.6% (95% CI [0.6, 2.5]) among females (see. Subsequent analyses focused solely on the combined sample of students who were classified as overweight and obesity.
The unadjusted logistic analysis showed: Father's occupation, family income, perceived stress in life, home region GDP, and unemployment were associated with being classified as overweight or obese in males. Enrollment in a medical major, father's and mother's occupation, university city GDP, and unemployment were associated with being classified as overweight or obese in females (see.
The final multiple level logistic models showed that higher family income, perceived life stress, home region GDP and university city unemployment were associated with being classified as overweight or obese in males. However, unlike male students, only the regional unemployment was associated negatively with overweight and obesity among females. In demographic characteristics, males with a father's occupation in staff and administration had a higher likelihood of being overweight and obese; females who did not have a medical major were associated with a higher likelihood of being overweight and obese (see.
There was no interaction between family income and original region GDP (Parameter estimate: 0.1080, SE: 0.1452, p > .05), and life perceived stress and unemployment (Parameter estimate: 0.6662, SE: 0.5060, p > .05) among male students.
Analyses for multicollinearity indicated low multicollinearity. VIF values were 1.07, 1.11, 1.81, 1.69, and 1.19 in family income, perceived life stress, university type, home region GDP, and unemployment, respectively among male students. Those values are lower than 10 and well below a value of 2.5 for even a weak model. It was therefore, determined that adjustments to deal with the presence of multicollinearity were not warranted (see [fig_ref] Table 3: Parameter Estimates From Multiple Level Models [/fig_ref]. In the null model, the random effect of original regions and university city were estimated 9.0553 and 1.3152 among female students, and 9.8249 and 1.3150 among male students respectively, but they all declined in the full model (see [fig_ref] Table 3: Parameter Estimates From Multiple Level Models [/fig_ref]. This finding likely reflects significant overweight and obesity variation between original regions and university city regardless of gender.
# Discussion
This study was the first empirical study in China or elsewhere to use a multilevel framework to examine gender differences in overweight and obesity. Results indicated that the prevalence of overweight and obesity among college students was 9.5%, with a specific prevalence of 13.5% in males and 6.1% in females. The overall prevalence of overweight and obesity was lower in this sample than in many countries [bib_ref] Prevalence of overweight/obesity and its associated factors among university students from 22..., Peltzer [/bib_ref] [bib_ref] College weight gain and behavior transitions: Male and female differences, Cluskey [/bib_ref] , but higher than in some Eastern countries [bib_ref] Prevalence of overweight/obesity and its associated factors among university students from 22..., Peltzer [/bib_ref] [bib_ref] Prevalence of obesity among the medical students in Sylhet, Sultana [/bib_ref] [bib_ref] Obesity among university students, Nojomi [/bib_ref]. In any case, college presents a time of transition for young adults that often involves adaptation to a new environment. Many studies of college students reveal weight gains occurring during this time [bib_ref] College freshman stress and weight change: Differences by gender, Economos [/bib_ref] [bib_ref] Vulnerability to freshmen weight gain as a function of dietary restraint and..., Pliner [/bib_ref]. From this perspective, weight and obesity issue of college students should highly be an attention.
Gender differences related to overweight and obesity in college students have not been sufficiently explored. Results from the current study indicate that male students are more susceptible to overweight and obesity than female students in China. This trend is consistent with many studies among college students [bib_ref] Sex differences in dieting trends, eating habits, and nutrition beliefs of a..., Davy [/bib_ref] [bib_ref] Physical activity, food choice, and weight management goals and practices among U.S...., Lowry [/bib_ref] [bib_ref] College weight gain and behavior transitions: Male and female differences, Cluskey [/bib_ref] , but did not consistent with some studies [bib_ref] Changes in body weight and fat mass of men and women in..., Hoffman [/bib_ref] [bib_ref] College freshman stress and weight change: Differences by gender, Economos [/bib_ref]. From individual perspective the gender difference may be mainly influenced from different behavioral styles between men and women.
This study mainly explored gender differences related to external environmental influences on overweight and obesity. Higher family income and higher home region GDP were associated with a higher likelihood of being classified as overweight and obese among male students, but these associations were not found among female students. This trend may indicate male students are more sensitive to external environment factors than female students who seem better able to avoid excess food consumption even in the context of readily available economic resources. Socially, it may be more acceptable for Chinese men to be overweight or obese than it is for females to be overweight or obese. Many people think that strong is the symbol of men's masculinity in China, but they mistakenly link overweight and obesity with strength. In China size matters. As young adults, giving the appearance of strength through large size contributes to social status. For Chinese female college students being small and petite is more highly valued than strength, thus they may be more motivated to control their diet to maintain the preferred body size.
This study showed that higher perceived life stress and regional unemployment were associated with being classified as overweight and obese in males. However, unlike male students, only regional unemployment was associated with overweight and obesity among females, and in this case the relationship was inverse. This may indicate that male students have less behavior control over food consumption and choice under stressful situations than their female counterparts.
Although it is plausible that young adults' interpretations of and response to regional unemployment rates and, therefore their levels of stress, may reflect important gender differences in future career expectations. Other factors are also likely at play and there is a need for additional research to more clearly identify these relationships. Courtenay argued that the socially constructed gender roles have far-reaching influence on the ascriptive guidelines of what is considered appropriate for each gender, such gendered social roles ascribe masculinity with risk-taking behaviors and define normative feminity with risk-avoidance and generally health-protective behaviors. Men who subscribe to their socially constructed gender roles are particularly at risk of overweight and obesity through a lack of concern for body shape and eating habits . Chinese society is dominated by a patriarchy. Performing and manifesting masculinity is deeply rooted in the cultural mandates of any patriarchy, which expects men to be the leaders of society and the heads of their families [bib_ref] A qualitative study on unassisted smoking cessation among Chinese Canadian immigrants, Mao [/bib_ref]. In a parochial society, men are expected to take primary responsibility in maintaining the economic well-being of their society and family, while women are relegated with subordinate roles reserved in the domestic sphere. Thus, men are also particularly at risk of the negative impact of employment instability on their health outcomes. At a structural level, regional unemployment rates are associated with the pace of economic development, labor absorption capacity, and social stability. It is found among college students that the stressors originate from unemployment worries include lack of demand in the market, low or no pay, feelings of frustration and hopelessness about future employment [bib_ref] Analysis of college students' employment, Lou [/bib_ref]. The structural economic strain placed on employment prospect may have an amplified effect on male students due to the socially prescribed gender roles that elevate the necessity of financially supporting a potential family to the level of the existential meaning associated with being a man. As a consequence, male students may compensate this insecurity by unhealthy behaviors that ultimately manifest in eating problems and substance use. This study found there is no interaction between family income and original region GDP, and life-perceived stress and unemployment among male students. This indicates that individual and regional economic status, and stress status, influencing on overweight and obesity are more independent, there is no obvious phenomenon of mutual promotion.
Although the association between psychological stress and unhealthy weight has not been well understood in China, the findings of this study provide a basis for further exploration. Further, these findings may assist in the development and evaluation of new approaches to prevent and treat overweight and obesity by focusing on psychological perspectives. In addition, universities should focus on creating environments which reduce stressors for students and support their overall mental health and well-being.
Participants in this study were mainly college students pursuing a medical degree. Given this specific characteristic and assuming their health knowledge is superior to the general population, the students might be expected to be less prone to overweight and obesity. But the reason why differences between medical students and other students were only found in females is unclear. Perhaps the social pressure to appear big and strong in men is so prominent that male medical students intentionally try to gain weight despite their medical knowledge. As a result, their weights are similar to other college students. Women on the other hand try to remain small and petite using their medical knowledge to assist in this process.
This study presents a number of limitations. First, the study was based on a cross-sectional design, which precludes causal inference. Longitudinal studies are necessary to further confirm these findings in the future. Second, this work only focused on college students. More research needs to be done on those who are not in college and are still facing significant amounts of individual and regional stress in a country undergoing significant changes. Obesity was measured using height and weight. Some participants may have had more muscle mass and thus were classified as overweight or obese when in fact they were not. To control for this limitation, more sophisticated methods for determining obesity would need to be utilized. Finally, further investigation on the impact of external and environmental influences on gender differences in overweight and obesity is essential, and to fully understand the mechanisms underlying the influence of socioeconomic variables and psychological stress on the prevalence of overweight or obesity.
# Conclusion
Our research indicates males were more susceptible to overweight and obesity, and were more easily influenced by external variants than were females in China. Considering these influences, gender sensitive strategies are needed on campuses to enable adoption of healthy foods and physical activity, and support stress management to prevent overweight and obesity. International efforts are underway to transform campus cultures and environments in ways that support the health and wellbeing of students, staff, and faculty, and may be useful in guiding the development and implementation of systemwide action strategies.
## Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
# Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The authors were partly funded by National Nature Science Foundation of China (71490733, 71473221).
[table] Table 1: Demographic Characteristics of Sample.Table 2. Overweight and Obesity Prevalence and Associated Individual and Regional Influences.Note. GDP = gross domestic product; CI = confidence interval, OR = odds ratio; RMB = Ren Min Bi (Chinese Currency). '-' only signifies the age group sectionalization. [/table]
[table] Table 3: Parameter Estimates From Multiple Level Models. [/table]
|
A method for independent component graph analysis of resting‐state fMRI
A method for independent component graph analysis of resting-statefMRI DemetriusRibeirodePaula 1 | ErikZiegler 2 | PubudithaM.Abeyasinghe 1 | TusharK.Das 1 | Carlo Cavaliere 2,3 | MarcoAiello 3 | LizetteHeine 2 | CaroldiPerri 2 | AthenaDemertzi 2,4 | QuentinNoirhomme 2 | VanessaCharland-Verville 2 | AudreyVanhaudenhuyse 5 | JohanStender 6 | FranciscoGomez 7 | Jean-FloryL.Tshibanda 8 | StevenLaureys 2,8 | AdrianM.Owen 9 | AndreaSoddu 1This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.AbstractIntroduction:Independent component analysis (ICA) has been extensively used for reducing task-free BOLD fMRI recordings into spatial maps and their associated timecourses. The spatially identified independent components can be considered as intrinsic connectivity networks (ICNs) of non-contiguous regions. To date, the spatial patterns of the networks have been analyzed with techniques developed for volumetric data.Objective: Here, we detail a graph building technique that allows these ICNs to be analyzed with graph theory.Methods:First, ICA was performed at the single-subject level in 15 healthy volunteers using a 3T MRI scanner. The identification of nine networks was performed by a multiple-template matching procedure and a subsequent component classification based on the network "neuronal" properties. Second, for each of the identified networks, the nodes were defined as 1,015 anatomically parcellated regions. Third, between-node functional connectivity was established by building edge weights for each networks. Group-level graph analysis was finally performed for each network and compared to the classical network.Results:Network graph comparison between the classically constructed network and the nine networks showed significant differences in the auditory and visual medial Correspondence
# | introduction
The evaluation of functional connectivity from resting-state fMRI data is broadly based on two families of analytical methods. Seed-based correlation analysis estimates the relationship between a predefined region (the "seed") and all other voxels around the brain [bib_ref] Functional connectivity in the motor cortex of resting human brain using echoplanar..., Biswal [/bib_ref] [bib_ref] Spontaneous fluctuations in brain activity observed with functional magnetic resonance imaging, Fox [/bib_ref]. A commonly employed alternative is independent component analysis [bib_ref] Consistent resting-state networks across healthy subjects, Damoiseaux [/bib_ref] [bib_ref] High-Dimensional ICA Analysis Detects Within-Network Functional Connectivity Damage of Default-Mode and Sensory-Motor..., Dipasquale [/bib_ref] , a data-driven approach to decompose whole-brain BOLD signal into a number of contributing volumetric spatial maps and their associated time-courses, such that the spatial independence of the components is maximized. ICA has proven to be an effective and robust tool for the isolation of low-frequency resting-state patterns from data acquired at various spatial and temporal resolutions [bib_ref] fMRI resting state networks define distinct modes of long-distance interactions in the..., De Luca [/bib_ref]. The ICA-generated volumetric maps are generally reported as z-scores in order to show the contribution of each component's time-course to the BOLD signal in each voxel. These maps are then commonly analyzed in a voxel-wise manner in order to show between or within-group connectivity differences [bib_ref] Defaultmode network activity distinguishes Alzheimer's disease from healthy aging: Evidence from functional..., Greicius [/bib_ref]. The number of components is generally user-defined and many of the resulting components may be due to non-neuronal activity (such as cardiovascular signal, eye movements, muscle activity in the vicinity of the head and head movement). Classically 10 functional networks can be reliably identified from ICA [bib_ref] Consistent resting-state networks across healthy subjects, Damoiseaux [/bib_ref] [bib_ref] fMRI resting state networks define distinct modes of long-distance interactions in the..., De Luca [/bib_ref] [bib_ref] Spontaneous fluctuations in brain activity observed with functional magnetic resonance imaging, Fox [/bib_ref] by decomposing the signal and to separate neuronal from non-neuronal components. Apart from voxel-wise statistics, graph theory could be applied to the functional networks which provides tools for analysis of network topography and connectivity [bib_ref] Complex network measures of brain connectivity: Uses and interpretations, Rubinov [/bib_ref]. Functional and structural connectome analysis has become increasingly common in neuroscience.
In these methods, the brain is defined abstractly as a network of nodes with edges connecting them. Nodes are often anatomically defined regions, whereas edges typically carry weights describing the correlation, similarity, or degree of connectivity between nodes. Adjacency graphs, in which the edge's existence is binary and carry no representative weight, are also common, as data-driven approaches for brain parcellation [bib_ref] Brain graphs: Graphical models of the human brain connectome, Bullmore [/bib_ref]. Usually, a weighted graph W is created by calculating a connectivity measure between every pair of regions; subsequently, the meaningful properties of the graph are carried out by network analysis methods, such as summary statistics (e.g., degree, small-worldness) and permutation testing . To our knowledge, no graph theory approach has been developed to evaluate the organization properties within brain networks derived from ICA. A graphical method permits to extract properties of a region of interest in the context of the full network. A comparison of graphical properties goes far beyond the comparison of the z values, as commonly done with the comparison of IC spatial maps voxel by voxel, where each voxel is treated as a separate entity and not considered in the context of the full brain spatial map.
The method we present here incorporates pieces of both families of methods and it is a generalization to the entire brain of a previously introduced approach restricted to the default mode network to study functional connectivity changes in patients with disorders of consciousness [bib_ref] Identifying the default-mode component in spatial IC analyses of patients with disorders..., Soddu [/bib_ref]. Following, this approach has been also applied in patients suffering from tinnitus focusing on the auditory network [bib_ref] Auditory resting-state network connectivity in tinnitus: A functional MRI study, Maudoux [/bib_ref] [bib_ref] Connectivity graph analysis of the auditory resting state network in tinnitus, Maudoux [/bib_ref]. We do believe that our method, for the very first time, permits a comparison of IC spatial maps using the advantage of graph theory.
# | methods
In short, we used ICA and machine learning classification to isolate a set of neuronal components [bib_ref] Multiple fMRI system-level baseline connectivity is disrupted in patients with consciousness alterations, Demertzi [/bib_ref] , and then construct weighted graphs for each of these components. As in other functional and structural connectivity mapping methods, we defined our regions based on structural parcellation [bib_ref] The connectome viewer toolkit: an open source framework to manage, analyze, and..., Gerhard [/bib_ref]. As the parcellation concerned cortical areas only, the cerebellum was not here considered as a network of interest, networks with regard to the average degree and the number of edges, while the visual lateral network showed a significant difference in the small-worldness.
Conclusions:This novel approach permits us to take advantage of the well-recognized power of ICA in BOLD signal decomposition and, at the same time, to make use of well-established graph measures to evaluate connectivity differences. Moreover, by providing a graph for each separate network, it can offer the possibility to extract graph measures in a specific way for each network. This increased specificity could be relevant for studying pathological brain activity or altered states of consciousness as induced by anesthesia or sleep, where specific networks are known to be altered in different strength.
## K e y w o r d s
BOLD fMRI, graph theory, independent component analysis, resting state reducing the number of ICA maps analyzed to nine. The edge weighting in the graphs was fairly unique to this method, and depended on a node-based calculation of the contribution of the specific component to the average BOLD signal in that region. We then calculated the weight (dependent on the magnitude and type of contribution) between each pair of nodes that reflects the similarity between them through a common similarity with the time course of the component of interest. This approach provided weighted within-component graphs that are related to distinct and known cognitive functions and could be quantified by using graph theory. In principle, this technique can be applied to any independent component map.
## | subjects
Fifteen (mean age = 43 ± 15, 7 women) healthy subjects, free of psychiatric or neurological history, were studied. The Ethics Committee of the Medical School at the University of Liège approved the study.
Informed consent to participate in the study was obtained from every subject.
## | dataacquisitionandpreprocessing
Functional MRI time series were acquired on a 3T head-only scanner (Siemens Trio, Siemens Medical Solutions, Erlangen, Germany) operated with a standard transmit-receive quadrate head coil. Three hundred multislice T2*-weighted functional images were acquired with a gradient-echo echo-planar imaging sequence using axial slice orientation and covering the whole brain (32 slices; voxel size: The three initial volumes were discarded to avoid T1 saturation effects. The subjects were instructed to close their eyes, relax without falling asleep and refrain from any structured thinking such as counting, singing etc. A high-resolution T1-weighted image was also acquired for each subject (120 slices, repetition time = 2,300 ms, echo time = 2.47 ms, voxel size = 1 × 1 × 1.2 mm 3 , flip angle = 9°, field of view = 256 × 256 mm 2 ). Data preprocessing was performed using Statistical Parametric Mapping 8 (RRID:nif-0000-00343; www.fil.ion. ucl.ac.uk/spm). Preprocessing steps included realignment and adjustment for movement-related effects, coregistration of functional with structural images, segmentation of structural data, spatial and functional normalization into standard stereotactic Montreal Neurological Institute space, and spatial smoothing of the fMRI data with a Gaussian kernel of 8 mm full-width at half-maximum. Further correction for small, large, and rapid motions, noise spikes, and spontaneous deep breaths was applied using ArtRepair (RRID:SCR-005990; cibsr.
stanford.edu/tools/human-brain-project/artrepair-software.html).
# | independentcomponentanalysisand componentclassification
Single-subject independent component analysis was performed using the Infomax algorithm within the Group ICA of fMRI Toolbox (RRID: SCR-001953; http://mialab.mrn.org/software/gift/) with a predefined number of components equal to thirty. The component spatial images were calibrated to the raw data so that the intensity values were in units of percent signal change from the mean. Components were subsequently assigned to the putative intrinsic connectivity networks using a multiple-template matching procedure. This method extends the single-template goodness-of-fit approach by assigning the independent component that maximizes the goodness-of-fit to a binary predefined template while considering all of the RSNs simultaneously. This procedure has been described previously [bib_ref] Multiple fMRI system-level baseline connectivity is disrupted in patients with consciousness alterations, Demertzi [/bib_ref].
For the discrimination between "neuronal" and "non-neuronal" independent components, a binary classification approach using support vector machine (SVM) was performed. The training of the SVM classifier was performed on the fingerprints of the independent components obtained from ICA decomposition with 30 components in 19 independently studied healthy subjects [bib_ref] Multiple fMRI system-level baseline connectivity is disrupted in patients with consciousness alterations, Demertzi [/bib_ref].
The classifier with highest overall classification rate was selected and subsequently used to label neuronal independent components. Independent components that failed to pass the "neuronality" test were excluded. See for a pictorial description of the abovedescribed procedure.
## | regionalparcellation
Segmentation of each subject's T1-weighted image was performed with Freesurfer's automatic segmentation pipeline and the Desikan Killiany atlas [bib_ref] fMRI resting state networks define distinct modes of long-distance interactions in the..., De Luca [/bib_ref]. Further parcellation, using the Lausanne 2008 atlas and its 1,015 individually labeled regions, was done with functions from the Connectome Mapping Toolkit, in order to separate different ICA-derived networks into non-overlapping spatial components [bib_ref] The connectome mapper: An open-source processing pipeline to map connectomes with MRI, Daducci [/bib_ref] [bib_ref] The connectome viewer toolkit: an open source framework to manage, analyze, and..., Gerhard [/bib_ref] [bib_ref] Nipype: a flexible, lightweight and extensible neuroimaging data processing framework in Python, Gorgolewski [/bib_ref]. To facilitate the creation of the functional networks, this type of parcellation was performed for both the original and spatially normalized T1 images for each subject. See
## | functionalnetworkconstruction
In the following, the letters N, P and R will refer to the number of ICA neuronal components (N), the number of BOLD signal time points (P) and the number of parcellated regions (R), respectively.
Let X a N × P matrix storing at the ith row the timecourse of the ith independent component and Y a R × P matrix containing at the ith row the timecourse of the BOLD signal averaged over the voxel belonging to the ith region. The process of identifying regional time-courses across a set of anatomically defined voxels is shown in . We solve the simple equation:
for the N × R matrix β, using the least squares solution [bib_ref] Analysis of fMRI time-series revisited-Again, Worsley [/bib_ref]. This means minimizing the value of:
[formula] (1) Y = Xβ (2) ‖Y − Xβ‖ 2 [/formula]
with an error term ε, for every region and each component:
The regression value β ij describes how the time course of the jth region can be explained by the ith independent component time course plus an error term. The regression values are, however, values of arbitrary size and variance and cannot be interpreted directly. In order to deal with this issue, we chose to transform these values into t-values [bib_ref] Analysis of fMRI time-series revisited-Again, Worsley [/bib_ref] , using the following equation:
where c is a vector indexing each component, and T is the matrix of t-values by region and IC. At this point, we have regional t-values for each component, which can be used to build weighted graphs, . The goal is to choose an appropriate weighting scheme such that edges are strong between regions which both contribute largely, and with similar strengths, to the regional fMRI time course. For this reason, we have chosen a straightforward weighting equation for the degree of connection between two nodes:
where W a,b represents the edge weight between nodes a and b, and t a , t b are the t-values for node a and b, respectively.
Equation 5 produces a zero weight for t a and t b with opposite signs. Here, we restrain the analysis to positive correlation networks only, for which a connection is considered only between nodes with t-values of the same sign. It is also possible to construct the complementary anti-correlation networks, by drawing edges between regions
[formula] (3) ε = Y − Xβ (4) T = c T β √ Var(ε)c T (X T X) −1 c (5) W a,b = |t a | + |t b | − |t a − t b | F I G U R E 1 [/formula]
Processing workflow for functional networks: (a) 1, preprocessing: registration to T1 (SPM8); 2, independent component analysis (GIFT) and 3, neuronality check and template matching. (b) 1, Spatial normalization of T1 (SPM8), 2, Atlas-based segmentation (FreeSurfer) and 3, parcellation (Nipype/CMTK, Lausanne 1,015 atlas). (c) 1, Extract regional fMRI timecourses (Y), 2, least squares solution to estimate β values using as predictors the timecourse from ICA as in (a) and 3, transform β values to t values. (d) 1, Calculate edge weights and draw networks for each independent component and 2, threshold correlation and anti-correlation networks and analyze separately with opposing signs. For creating weighted edges between nodes of opposite sign, the following equation can be used:
## | thresholdingnetworkedges
We estimated the degrees-of-freedom by considering the number of functional image volumes and the number of independent components in our network construction procedure. This led to the equation:
Given the degrees of freedom and the desired significance level (p < .001), we calculated the t-value with which to threshold our graphs. We chose to threshold our networks to remove all t-values that were not in the 99th percentile. This led to the calculation: Following the weighting scheme described in Equation 5, the nodes can have edge weights ranging from zero to the sum of their t-values. For this reason, we chose to discard edges above 2 × t threshold with 2 × t threshold being the maximum value for W a,b given the assumed t threshold . Group level networks, as presented in were created by thresholding each subject weighted matrix W a,b and by subsequently keeping the edges which were present in at least 25% of the subjects.
## | classicalnetworkconnectivitymatrix
With the aim of the creation of a whole-brain connectivity matrix, we reconstructed the functional MRI signal at each voxel by a linear combination of the components deemed "neuronal" by the classifier with weights coming from the corresponding scalar maps of each component. The resulting signal, cleaned off "non-neuronal" noise, was then averaged over all the voxels belonging to the same parcellated region to obtain time courses for each of the 1,015 regions. The W matrix, representing the classical network, was then calculated.
# | networkanalysisandstatistics
Nodes for each network of interest were previously extracted as the non-isolated nodes of the corresponding group level networks previously introduced and presented in . We have considered several network summary statistics for the graphs. We computed the number of edges (E), average degree (k), average number of triangles and small-worldness [bib_ref] Complex network measures of brain connectivity: Uses and interpretations, Rubinov [/bib_ref] as shown in [fig_ref] Table 1: The Sensorimotor was [/fig_ref].
A graph is a formal mathematical representation of a network and each object in a graph is called a node. The number of edges represents the number of connections between each pair of connected nodes. The average degree represents the average number of connections (edges) per node. Isolated nodes (nodes without edges) were discarded and all the properties were calculated for the constellation of connected nodes. The number of triangles represents triplets of nodes in which each node is connected to the two others. The smallworldness was calculated by:
where C and C rand are the clustering coefficients and L and L rand are the characteristic path lengths of the network of interest and the corresponding random network [bib_ref] Network 'small-world-ness': A quantitative method for determining canonical network equivalence, Humphries [/bib_ref]. Smallworldness is an extremely important property of networks. These networks are "highly clustered, like regular lattices, yet have small characteristic path lengths, like random graphs" making them very efficient in information transfer [bib_ref] Collective dynamics of 'small-world' networks, Watts [/bib_ref].
[formula] (6) W AC a,b = |t a | + |t b | − |t a + t b | (7) DOF = N volumes − N components − 1 (8) t threshold = t INV (Confidence, DOF) (9) σ = C∕C rand [/formula]
L∕L rand F I G U R E 2 Sagittal and axial representation of the network and cortical extent of the relative IC for the DMN. The degree of each one of the 1,015 regions is represented by the node's size and orange to yellow gradient. On the right, cortical parcellation for DMN extracted by group-ICA superimposed to a structural T1 image
The properties were calculated at subject-level producing a distribution of graphical values for the network of interest. This method was used to compare the graph properties [fig_ref] Table 1: The Sensorimotor was [/fig_ref] [bib_ref] Whole-brain anatomical networks: Does the choice of nodes matter?, Zalesky [/bib_ref]. Column 3 in [fig_ref] Table 1: The Sensorimotor was [/fig_ref] shows the number of non-isolated nodes for each network. For the number of triangles, the comparison was performed by firstly normalizing the number of triangles by a random network containing the same number of nodes.
To determine which networks are different from the others, we conducted a post hoc paired comparison (Tukey post hoc test as implemented in R), which is designed to evaluate the difference between each pair of networks, .
# | results
## | componentclassification
The template-matching and fingerprint selection criteria identified the predefined neuronal components within the group of healthy controls. The percentage of subjects in which they were found is presented in the second column in F I G U R E 3 Graphical representation of the network and cortical extent of the relative ICs for the left and right executive control networks. The degree of each one of the 1,015 regions is represented by the node's size and color gradient. On the right column, cortical location for each network was extracted by group-ICA superimposed to a structural T1 image
## | networkproperties
The network properties were calculated for the nine independent component networks of interest for all the 15 subjects. The results are shown in [fig_ref] Table 1: The Sensorimotor was [/fig_ref]. [fig_ref] Table 1: The Sensorimotor was [/fig_ref] for graph properties and for a similar ANOVA analysis performed this time over all the 1,015 nodes).
The sagittal and axial graphical representations of the networks, as created from the respective group level matrix W, are shown in They were also presented separately, DMN in , executive control left and executive control right in , auditory and salience in , sensorimotor and visual lateral in and visual medial and visual occipital in . As shown in [fig_ref] Figure 8: all nine overlapping networks and the classical network [/fig_ref] , the nodes with the highest degree of the classical network are also the nodes with the highest average degree shared by the nine different networks. it is also possible to recognize that the clusters of nodes with higher degree tend to be spatially distributed as the high z-value in the corresponding independent component z-map for the spatial patterns of the nine independent components [bib_ref] Consistent resting-state networks across healthy subjects, Damoiseaux [/bib_ref].
# | discussion
We presented a novel method for applying graph theory metrics to resting state fMRI brain networks derived from data-driven ICA. With F I G U R E 4 Graphical representation of the network and cortical extent of the relative ICs for auditory and salience networks. The degree of each one of the 1,015 regions is represented by the node's size and color gradient. On the right column, cortical location for each network was extracted by group-ICA superimposed to a structural T1 image this within-component approach, the between-group topological differences in connective structure of functional networks can be studied with well-established network measures employed in graph theory.
Classical approaches are based on the recognition of RSNs from different ICA processing results, such as multiple template-matching [bib_ref] Multiple fMRI system-level baseline connectivity is disrupted in patients with consciousness alterations, Demertzi [/bib_ref] , high dimensional ICA [bib_ref] High-Dimensional ICA Analysis Detects Within-Network Functional Connectivity Damage of Default-Mode and Sensory-Motor..., Dipasquale [/bib_ref] and fully exploratory network ICA [bib_ref] Fully exploratory network ICA (FENICA) on resting-state fMRI data, Schöpf [/bib_ref] ; once each RSN is estimated, their spatial distributions are usually compared voxel-wise between subjects or groups for the assessment of withinnetwork differences. On the other hand, at the best of our knowledge, there is not yet an effective procedure for the analysis and comparison of graph properties between network components derived from spatial ICA procedures.
A similar approach has been proposed for the analyses of withinnetwork organization. In [bib_ref] Graph Independent Component Analysis Reveals Repertoires of Intrinsic Network Components in the..., Park [/bib_ref] , an alternative method for the estimation of each RSN is employed. As our method, it provides information on the connectivity among each component, but each component is not derived with a conventional spatial-ICA approach.
Therefore, our method offers more flexibility, since it deals with commonly derived ICA components, exploiting the well-established advantage of spatial ICA for the rejection of artifactual components.
Comparison between the nine networks and the classical network did not show, for most of the comparisons, significant differences in the studied graph properties, which indicates similarity in their graph structure. The absence of main differences is also indicating the fact that when restricting to network's regions, connectivity due to the time course behavior representative of the IC of interest is capturing almost entirely the full neuronal behavior. in supplementary material shows clearly that when extending the analysis to the full brain, i.e., in the selected constellation of nodes. Therefore, this approach could be quite appealing to patients suffering severe brain injury, considering that the networks might be highly affected. For these patients, structural information could be used to indicate the partial selection of nodes which we would like to focus on, and simply select the component based on the highest average number of edges for the newly selected constellation of nodes.
# | conclusions
We presented an approach for the analysis of resting state networks carried out by dissecting the connectivity patterns of task-free fMRI data by using ICA. Our results suggest that, by evaluating independent component networks using graph theory instead of using volumetric data, one could take full advantage of graphical methods, which permit the comparison of local properties in the context of the full brain.
Moreover, our approach is the basis for further refinements such as a component clustering based on network topology. Similarly, the initial investigation into the anti-correlation networks for the 15 subjects studied here showed wide variability in structural topology, and it needs to be further investigated. In future, one could selectively connect edges based on the sign of two nodes t-values and explore the differences between graphs created with differing parity. F I G U R E 8 Sagittal and axial representation of all nine networks (default mode network, executive control left, executive control right, visual lateral, visual medial, visual occipital, auditory, sensorimotor and salience) overlapped and classical network, with threshold 0.45. The average degree of each node is represented by the node's size and color gradient
# Acknowledgments
[fig] 3: ; matrix size 64 × 64 × 32; repetition time = 2,000 ms; echo time = 30 ms; flip angle = 78°; field of view = 192 × 192 mm 2 ). [/fig]
[fig] Figure 1b: for an illustrative description of the final parcellation. [/fig]
[fig] Figure 8: all nine overlapping networks and the classical network). [/fig]
[table] Table 1: The Sensorimotor was [/table]
|
Tracheal tube misplacement in the thoracic cavity: A case report
BACKGROUNDPenetrating neck injuries require prompt recognition, diagnosis and management of critical airways. This case demonstrates an emergent situation that a "medical negligence" was avoided with the aid of end-tidal carbon dioxide (ETCO 2 ) waveform.CASE SUMMARYWe report a case of malposition of the endotracheal tube into the right hemithoracic cavity for cervical knife trauma, resulting in pneumothorax. Tube placement was not confirmed during emergency airway management, and the patient was directly transferred to the emergency operation room. Assisted by ETCO 2 and imaging examinations, the anesthetist timely noticed the absence of ETCO 2 waveform and resolved this urgent situation before anesthesia induction.CONCLUSIONThis case emphasizes the necessity of ETCO 2 waveform and/or X-ray confirmation of endotracheal intubation even in emergent situations.
## 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: htt ps://creativecommons.org/Licens es/by-nc/4.0/
# Introduction
Penetrating neck injuries (PNI) are neck injuries penetrating the platysma and represent 5% to 10% of all traumatic events with a high mortality rate due to unprotected airway and proximity to vital vascular structures and trachea . PNI requires anesthesiologists and surgeons to maintain vigilance on timely recognition, diagnosis, and management. Challenging airway management is inevitable. End-tidal carbon dioxide (ETCO 2 ) monitoring non-invasively detects the concentration of CO 2 at the end of expiration. And the appearance of ETCO 2 waveform is regarded as the golden standard for proper intubations. However, intubation without confirmation tends to happen, especially in urgent situations, such as PNI. Here we report a rare case of misplaced endotracheal tube into the thoracic cavity.
## Case presentation
## Chief complaints
A 28-year-old female patient was admitted to the emergency department with knife injury to the neck for 12 h.
## History of present illness
The patient was found to have tracheal injury 1cm below the thyroid cartilage with severe pain, active bleeding (the specific amount of blood loss was unknown) dyspnea, chest distress, shortness of breath and dysphonia. She was managed with compression packing at a local hospital and was transferred to our hospital for further management.
## History of past illness
The patient had a disease-free personal and family history.
## Personal and family history
The patient had not the special personal or family history.
## Physical examination
On admission, the patient was awake with stable vital signs: temperature 36.6 ℃, pulse rate 98 bpm, respiratory rate 22/min and blood pressure 118/70 mmHg. Tissue deficits were identified on the left sternocleidomastoid muscles. Breath sounds were slightly diminished, and dry and moist rales were noticed on both upper lobes.
## Laboratory examinations
Leucocyte count was 12.79 × 10 9 /L, where neutrophils accounted for 86.8%. And other examinations were all normal, such as hematocrit and hemoglobin count. Urine analysis was also normal. Prothrombin, partial thromboplastin times, and d-dimers were within normal ranges. Electrocardiogram showed a sinus rhythm. Volume 9 Issue 34
## Imaging examinations
After the endotracheal tube was inserted, the patient underwent urgent imaging examinations for operation preparation. Computed tomographic angiography (CTA) showed that there was no leakage, occlusion or expansion of cervical blood vessels and branches. Emergent cervical computed tomography (CT) scan revealed extensive gas accumulation in the mediastinum and underneath the cervical tissue, and continuous interruption in the anterior part of trachea. Chest CT scan demonstrated that the right hemithoracic cavity with limited pleural effusion was collapsed by 70% approximately. The patchy lesions and shadows suggested slight infection in the right hemithorax.
## Final diagnosis
The patient was diagnosed with cervical knife trauma and tracheal injuries, which should be managed by emergency operation.
## Treatment
The emergency medicine physician inserted a 6.5# endotracheal tube into the wound, and inflated the cuff to prevent bleeding from the lumen into the ruptured trachea. Right chest tube was placed for preventing suspected pneumothorax. The patient was immediately transferred to the operating room for exploration after CT scan was performed, but without final reading and confirmation.
## Outcome and follow-up
The anesthetist was informed that the airway was secured without aspiration risks. However, the patient was agitated and in respiratory distress with 85% pulse oxygenation on room air. After the tracheal tube was connected with anesthesia circuit, the breathing bag of the anesthesia machine was expanding and shrinking during patient's spontaneous breathing (25/min). However, ETCO 2 waveform was absent. CT scan was reviewed immediately and revealed that the endotracheal tube entered into the right hemithoracic cavity and the right lung was collapsed by 70% approximately due to extensive pneumothorax . The anesthesiologist and otolaryngologist immediately reinserted the tracheal tube, connecting it to the ETCO 2 monitor. The tube placement was confirmed carefully before final fixation. Next, the patient underwent open neck exploration, tracheal endto-end anastomosis, recurrent laryngeal nerve reconstruction and tracheotomy thereafter. The subsequent clinical course was uneventful, the patient was transferred into intensive care unit and discharged after two-week hospitalization.
# Discussion
We presented a case of tracheal tube misplacement into the thoracic cavity for neck injury. Although the emergency medicine physicians promptly evaluated and attempted to manage the airway with an awake endotracheal intubation, the tube was inserted into the thoracic cavity and produced pneumothorax. Because the patient was on spontaneous breathing and a chest tube was placed on the same side, breath sounds were heard bilaterally. The patient was sent immediately to the operating room assuming successful airway establishment. Successful airway management should have been confirmed with clinical evaluation, chest radiography, and ETCO 2 detection.
Iatrogenic tracheobronchial injuries by intubation have been reported, including tracheal laceration[9,10] and subcutaneous emphysema. Incorrect tube sizes and reintubation may contribute to iatrogenic injuries with direct laryngoscopy after endotracheal intubation . Therefore, some studies recommended awake intubation, flexible fiberoptic bronchoscopy, or direct ultrasound visualization to avoid false passage and tracheal injury . When dealing with tracheal trauma from PNI, confirmation of the endotracheal tube placement by ETCO 2 waveform December 6, 2021
Volume 9 Issue 34
Figure 1 X-ray and computed tomography findings after insertion of endotracheal tube and right chest tube. A and B: X ray (A) and computed tomography (CT, B) scan images revealed that there was a tracheal tube inserted into the right hemithorax (orange arrow); C and D: CT scan images showed that right pulmonary markings were absent, and right lung was condensed. The right lung was condensed by 70% and the mediastinum shifted to the right side.
and/or X-ray/CT scan is mandatory.
# Conclusion
Airway establishment is the priority option for tracheal injuries, which was an extremely urgent situation for PNI. Emergent evaluation and treatment are challenging. Negligence is inevitable, especially in emergency situations. This case highlights the role of ETCO 2 waveform and/or chest radiography in confirmation of emergent endotracheal intubation after emergent intubation, especially for junior doctors and emergent physicians.
https://www.wjgnet.com
© 2021 Baishideng Publishing Group Inc. All rights reserved. |
Sexual and Reproductive Health Problems of Female University Students in Iran: A Qualitative Study
Objective: Youth is defined as the time of transition into adulthood and an important period in a person's life. During this period new behavior is learned easier than adulthood. Therefore, special attention has to be necessarily paid to this period in order to promote the health. Addressing adolescent reproductive health issues is also a critical factor Methods: This research was a qualitative study conducted from January 2014 to July 2014. Data from focus group discussions and semi-structured interviews with 25 female students and 10 key members of the university (including university authorities, consultants, reproductive health professionals and university officials) was collected and all interviews were recorded, formulated and classified.Results:The mean age of participants was 22.4 3 years. A total of 8 students majored in geology, 5 majored in chemistry, 3 in statistics, 3 in mathematics, and 6 in biology. 17 had a bachelor's degree, 3 master's degree and 5 doctorate degree. Majority of students (82.4%) were never married and 23 of them lived in dormitories. The following three main themes were extracted from the interviews: Reproduction thought as pregnancy; the taboo of sex; and inappropriate relation between parents and children.Conclusion:Most participants stressed the need to provide reproductive health services for young girls.One of the most important experiences reported by students in this study was the modesty shown about the sexual reproductive health issues. The youth are ashamed to ask for information from adults who reluctant to discuss these issues. Hence, taboos, beliefs and traditions may prevent the youth from accessing the necessary information. In many cultures, parents do not talk about sex with their children and therefore they are useful sources of information(UNFPA, 2008). Research revealed that with an increase in the parent-adolescent communication and conversation the level of sexual risk decreases.The present study showed that the relationship between parents and children is very poor.Since the Iranian society is transforming from a traditional society to a modern one, many families have lost their traditional attitudes and thus children are given more independence. Research revealed that with an increase in the parent-adolescent communication and conversation the level of sexual risk decreases(Diorio et al., 1999;Kotchick et al., 1999). Other studies showed that parental control is also another factor in predicting risky sexual www.ccsenet.org/gjhs
# Introduction
Youth is defined as the time of transition into adulthood and an important period in a person's life. During this period new behavior is learned easier than in adulthood . This fact was stressed in the International Conference on . In this regard, the World Health Organization (WHO) estimates that more than one billion people in the world are between 15-24 years old . Moreover, more than 85% of them live in developing countries. Investment in the health of this age group has played a major role in the development of human communities due to the dual role of women in community health and well-being of future generations as one of the main paths to the achievement of Millennium Development Goals (MDGs) and youth goals [bib_ref] The social context of adolescent health: a Quality study. School of Nursing..., Parvizi [/bib_ref].
In connection with the discussion of health, sexual and reproductive health is an important part of world health [bib_ref] An exploration of knowledge, attitudes and behaviors' of young multiethnic Muslim-majority society..., Ping [/bib_ref] and as a part of human rights it has been approved for public [bib_ref] Knowledge, attitudes and behavior regarding reproductive health of women in seven cities..., Mazloomy Mahmoud Abbad [/bib_ref]. In this regard, many young women have very little reproductive health information [bib_ref] An exploration of knowledge, attitudes and behaviors' of young multiethnic Muslim-majority society..., Ping [/bib_ref]. In addition, it has been reported that the young are facing different sexual and reproductive health problems like unwanted pregnancy, unsafe abortion, and STI including HIV . It is estimated that many women in the world die due to complications during pregnancy and abortion every year. On the other hand, the mortality rate caused by unwanted pregnancy and sexually transmitted diseases (STD) is increasing around the globe .
Despite the concerns announced by the United Nations, 180 member countries, international organizations, and individual adolescents everywhere, the reproductive health concerns of young people are too often neglected [bib_ref] In this generation: sexual and reproductive health policies for a youthful world, Greene [/bib_ref]. Most of the population in our country (i.e. more than half of the population) is under 25years of age. According to the Statistical Center of Iran, a total of 60% of the population is under 25 years and over 50% are under 20 years (http://www.space.ir/barnameh/Baranmeh%20gozashe, 2012). Therefore, neglecting this population and their future reproductive health will bring about irreparable consequences for the adults [bib_ref] Towards adulthood: exploring the sexual and reproductive health of adolescents in South..., Bott [/bib_ref]. Hence, it is necessary to maintain the functionality of the present programs and assess emerging needs and current trends associated with this issue [bib_ref] Reproductive health knowledge, attitudes and practices of Iranian and Afghan men in..., Sadeghipour Roudsari [/bib_ref].
Based on results of the Iranian national Census, of the almost 70-millionpopulation of Iran, about 11.5 million are girls aging between 10-24 years [bib_ref] The experience of puberty in Iranian adolescent girls: a qualitative content analysis, Hagikhani Golchin [/bib_ref].As a result of concerns for this important issue, the United Nations has changed the theme of 2013 World Population Day was teenage pregnancy. This highlighted the significant of the role played by teen girls in positively influencing future generations and stressed the importance of providing them with adequate health care and educational support .
In Iran, girls' health status is often considered to be more important than boys. There are several cultural and social reasons for this point of view. First, a girl's lifespan includes several stages and milestones including infancy, childhood, puberty, marriage, pregnancy, childbearing and menopause [bib_ref] The experience of puberty in Iranian adolescent girls: a qualitative content analysis, Hagikhani Golchin [/bib_ref] [bib_ref] Educational needs of adolescent girls for reproductive health in teachers views, Mirzaei [/bib_ref]. Despite this trend, little is known about the reproductive health needs of young people in Iran. In addition, the few studies conducted on the knowledge, attitude, beliefs and behavior associated with sexual reproductive health of Iranian youth have indicated that the level of knowledge of reproductive health is low [bib_ref] Reproductive Knowledge, Attitudes and Behavior among Adolescent Males in Tehran, Mohammadi [/bib_ref]. It is the responsibility of health researchers to identify the needs for reproductive health promotion and to plan and implement the necessary educational programs that might include prevention of STIs/HIV/AIDS as well as unwanted pregnancies.
In Islamic Republic of Iran, students forma large part of the population were selected to study the reproductive health status of Iranians and students following separation from parents, which exposes them to sexual reproductive health problems like the following: love relationship, physical health problems, alcohol use, weight change, eating problems, AIDS, drug abuse, time management, fear of examinations and serious psychological concerns. Some of these concerns include depression, sadness, mood changes, anxieties and phobias [bib_ref] Reproductive Knowledge, Attitudes and Behavior among Adolescent Males in Tehran, Mohammadi [/bib_ref]. University students in general are the hopes of every country and, unfortunately, there are no comprehensive studies in this field. The existing studies and articles are also limited to quantitative studies of youth. Most experts believe that the range of quantitative methods for the assessment of real needs sweeping capabilities is very limited and qualitative research methods are required to make it good enough. Therefore, research on youth reproductive health status and requirements of Reproductive Health in Iran is important for various reasons. Since college students and alumni form important communities, these communities were selected to explain the reason. The reproductive health status of female students of Lorestan University residing in Khoramabad city (in the southwest of IRAN) was also examined. Khoramabad is the capital of Lorestan province and 22 nd most populated city of Iran. This city is located in the valleys of the Zagros Mountain Range at an altitude of 1147.8 meters. It accommodates a population of about 687167 people with 33% of the population being the youth . Therefore; this study was aimed to systematically explore the sexual and reproductive health problems among female students attending the University of Lorestan.
# Methods
Lorestan University is the oldest higher education institute in Khoramabad. It was established in 1977as the Education Center of Lorestan. Currently, this university has 7 faculties and serves 8239 students. The present study was conducted in the faculty of sciences serving1785 students studying for bachelor's, master's and doctorate degrees (Lu.ac.ir, 2012). A qualitative design based on the content analysis approach was used to conduct this study aim from January 2014 to July 2014. Content analysis is a systematic method of classification and coding using which information can be extracted from text. It is used for revealing a pattern of words, repetition, and structure. It is used to analyze relationship and interactions between words, repetition and structure [bib_ref] Experiences of infertility through the lens of Iranian infertile women: A qualitative..., Behboodi Moghadam [/bib_ref]. Therefore, in this study implicit and explicit concepts were identified based on the explanations of the participants and principles of coding, summarization, categories, and themes were identified. Codes were based on meaning extracted from the participants' descriptions. The meanings were later classified based on the discovered differences or similarities [bib_ref] Experiences of infertility through the lens of Iranian infertile women: A qualitative..., Behboodi Moghadam [/bib_ref] [bib_ref] Qualitative content analysis in nursing research: concepts, procedures and measures to achieve..., Graneheim [/bib_ref].
The study population for this study included female students of Lorestan University. Data were collected through semi-structured interviews. In-depth interviews were conducted by a trained interviewer. Each interview was carried out with 25 female students who were in Khoramabad, Iran. All female students were interviewed in a private room, and each interview lasted 45-90 minutes in average. Two focus group discussions involving eight participants were conducted in two sessions. The sessions were attended by students, Reproductive Health Professionals, consultants and university authorities. The duration of focus group discussion was between 1 and 3 hours. The contents were discussed; the materials were discussed before the arrival of the participants.
The interviews questions were focused on the following themes:
- Explanation of reproductive health
- Explanation of reproductive Health issues - Elaboration on reproductive health problems Each interview was recorded and transcribed verbatim and then the resulting material was analyzed. Point of idea saturation was considered to end in-depth interviews and focus discussions. Concepts were merged based on their themes and a manual thematic framework analysis was also used. The results were summarized and presented in narrative forms.
## Rigor
In this study, various aspects of trustworthiness have been observed. Credibility was established through member checking, peer checking, and prolonged engagement. Member checking was done by asking the respondents to approve the transcripts and emerging codes from the interviews. Research teams consulted with each other to deal with any ambiguities in the coding process, categories and themes researchers independently analyzed the data by identifying and categorizing codes for the subjects' responses to each question, and then the two authors' codes and their latest analysis development as themes were compared. In areas where the two did not agree, definitions were clarified and discussion continued until consensus was reached .For addressing transferability, the complete set of data analysis documents are on file and available upon request.
This study is a part of the first author's doctoral dissertation. The Ethics Committee of Tehran University of Medical Sciences approved the study proposal and corroborated its ethical considerations. All participants were informed about the purposes and the methods employed in this study. The permission for typing and recording the interviews was obtained from the participants. They were informed that participation in the study was voluntary and they could refuse to participate at any time without being deprived of the services delivered to them.
# Results
The mean age of participants was 22.4 3 years. A total of 8 students majored in geology, 5 majored in chemistry, 3 in statistics, 3 in mathematics, and 6 in biology.17 had a bachelor's degree, 3 master's degree and 5 doctorate degree. Majority of students (82.4%) were never married and 23 of them lived in dormitories.
The following three main themes were extracted from the interviews: Reproduction thought as pregnancy; the taboo of sex, inappropriate relation between parents and children.
## Reproduction thought as pregnancy
The majority of participants mentioned reproduction thought as pregnancy. This theme had three sub themes including the following: Lack of awareness and accurate information on reproductive health; lack of coverage of sexual and reproductive health problems by the media; lack of non-medical reproductive health-related books.
## Lack of awareness and accurate information on reproductive health
One of the things that most participants mentioned was the lack of knowledge and information about reproductive health. Almost all students defined reproductive health in association with pregnancy. One of the participants described her experience as follows:
## "i think that reproductive health means pregnancy and the term reproductive health refers to health. in general, my knowledge of this area is very little". (22-year-old girl)
The primary reason for such inadequate understanding of sexual and reproductive health could be the lack of awareness increasing courses in universities.
During the focus group discussions few of the participants stated that they did not clearly know what sexual and reproductive health was about.
Also some of the participants referred to a lack of knowledge on sexual and reproductive health issues especially among students coming from rural parts of the country.
Most students referred to the lack of centers for sexual and reproductive health care in universities.
## Lack of coverage of sexual and reproductive health issues by the media
Many of the students were complaining about the lack of coverage of sexual and reproductive health issues by the media.
Most participants stated that the role of the media in the field of sexual and reproductive health was very insignificant.
One of the participants said:
"The media has a significant impaction the education of youth on sexual and reproductive health issues. Media should provide accurate information on sexual and reproductive issues to the youth". Most students participated in this study believed that reproductive health information obtained through friends is not reliable and cannot be considered are liable source. Therefore, the media can eliminate these sources of non-academic information through proper training.
One of the participants said:
"In our current community there are not many sources of information for young people. Therefore, there are few advertisers and advertisement takes place through the mass media in this regard". (23-year-old girl)
## Lack of non-medical reproductive health textbooks
Nearly all of the participants wanted sexual health education to be offered formally in their curriculum as a formal course similar to other courses. They were interested in taking a formal course on reproductive health in the university.
One of the students said:
"We do not have non-medical books on reproductive health care…I wish we had a specific book for such things". (a 19-year-old girl)
## The taboo of sex
This theme included the following one sub themes: cultural and social environment.
## Cultural and social environment
Most participants in this study suggested that cultural barriers are a crucial factor contributing to the concealment of sexual and reproductive health problems by students. Lack of positive interaction between mothers and daughters for various reasons such as embarrassment, negative attitudes, and fear of shame, and lack of adequate information on these issues regard were among other barriers.
One of the participants described her experience as follows:
## "i'd like to get more information on reproduction, sexual transmitted diseases and risky sexual behavior, etc. we are not comfortable asking about these things at home. so i usually get my information from my friends, the internet, magazines or books". (a 22-year-old girl)
Another participant stated:
"We are not comfortable in the society and if we have a question about sex, we do not know where to go. So our knowledge of sex is poor". Due to cultural and social taboos set by the society, many signs of puberty in girls and their sexual development are not embraced.
One of the students stated:
"I'm really asking questions about sex from my mother. But I am not simply comfortable asking such questions." "My mother believed these things should not be taught to girls until marriage because girls may become rude".
## Inappropriate relation between parent and children
The majority of students believed that family is a very important factor in the development of sexual behavior. This theme included the following three sub themes: ensured independence of children, lack of child control, and too much reliance on children.
One participant stated:
"In some cultures, parents are the influential sources of knowledge, beliefs, attitudes, and values for children and the youth. If young people do not get information from their families, they will seek answers elsewhere and try to find answers through peers, the media or their observations of adult actions."
One of the reasons stated by the majority of students was as follows:
"The freedom given by some families transforming from traditional to modern families to the youth is increasing and is diminishing the role of parents."
# Discussion
Research results indicated that students of the Lorestan University have large sexual and reproductive health needs. According to the results, students' awareness of reproductive health issues such as sexually transmitted diseases (e.g. HIV) and contraceptive and sexual health is inadequate. The lack of knowledge about sex and reproduction among youth is also a factor reported by studies carried out in many developing countries [bib_ref] Barriers to sexual health services for young people in Nepal, Regmi [/bib_ref] [bib_ref] Exploring the socioeconomic dimension of adolescent reproductive health: a multicounty analysis, Rani [/bib_ref].
The results of the present study regarding the most important problem with sexual and reproductive health of the youth were consistent with the results of other studies in other developing countries [bib_ref] Knowledge, attitudes and sources of information regarding HIV/AIDS in Iranian adolescents, Yazdi [/bib_ref] [bib_ref] Knowledge and attitude towards HIV/AIDS among Iranian students, Tavoosi [/bib_ref].In this regard, it is essential to enhance the knowledge of the youth and increase the participation of experts and experienced teachers in developing educational materials that comply with the cultural and religious valuesof society .Accordingly, young people's hall benefit from sexual and reproductive health education and should not be deprived of this right.
The most important finding of this study was the significance of the lack of coverage of sexual and reproductive health issues by the media, as the media plays an important role in raising awareness about sexual and reproductive health . The study by revealed that the media plays an important role in enhancing their reproductive health of adolescents in sub-Saharan Africa. Lack of non-medical reproductive health-related books was another problem referred to by the students. Based on the young student' opinions, not only the existing textbooks fail to provide comprehensive sexual and reproductive health information, but also there is no access to any appropriate book teaching sexual and reproductive health materials. By the Statement issued by the International Conference on Population and Developmentin 1994, it is necessary to provide access to reproductive health information and government services to the youth. Sexual issues and lack of proper information in this regard were among the most important concerns reported by most students. The majority of the student wanted to know more about topics such as STDs and high risk sexual behaviors, which are considered to be taboos in Iran due to the cultural and social environment governing the country. A qualitative study in Iran concluded that considering the Iranian cultural and religious background, families and religious beliefs play an important role in reducing high-risk sexual behaviors among Iranian adolescents [bib_ref] Adolescents' perspectives on addiction: qualitative study, Parvizi [/bib_ref]. Most students in this study reported that community does not influence sexual issues. behavior in adolescents [bib_ref] Impact of perceived parental monitoring on adolescent risk behavior over 4 years, Li [/bib_ref]. Another study reported that a warm relationship between parents and children as well as parent control can considerable contribute to the prevention of behavior risk-taking in female adolescent .
# Conclusion
This study showed that the knowledge of students about sexual and reproductive health is very low and that the society and family do not have a significant role in addressing the reproductive health problems of the youth.
The following recommendations are thus provided based on the research findings:
- University authorities should provide opportunities for students to address many reproductive health issues and problems.
- Universities should increase its number of counselors to solve many of the sexual and reproductive problems faced by students.
- University authorities are strongly recommended to consider provision of friendly health services.
- There should also be a charter for sexual and reproductive health rights of female students in universities.
AcknowledgementsThis study was financially supported by the Tehran University of Medical Sciences. The authors wish to express their sincere gratitude to the study participants without whom this study could not have been conducted.Conflict of InterestsThere are no conflicts. |
Copper-based fluorinated reagents for the synthesis of CF2R-containing molecules (R ≠ F)
Over the years, the development of new methodologies for the introduction of various fluorinated motifs has gained a significant interest due to the importance of fluorine-containing molecules in the pharmaceutical and agrochemical industries. In a world eager to eco-friendlier tools, the need for innovative methods has been growing. To address these two challenges, copper-based reagents were developed to introduce CF 2 H, CF 2 R F , CF 2 CH 3 , CF 2 PO(OEt) 2 and CF 2 SO 2 Ph motifs on a broad range of substrates. Copperbased fluorinated reagents have the advantage of being inexpensive and generally in situ generated or prepared in a few steps, which make them convenient to use. In this review, an overview of the recent advances made for the synthesis of fluorinated molecules using copper-based fluorinated reagents will be given.1051
# Introduction
In a society in which fluorinated molecules are playing a pivotal role in pharmaceutical and agrochemical industries as well as in materials science, the quest for innovation in the organofluorine chemistry field is of high importance. In that context, the development of new strategies is an important driving force , offering efficient and original tools to introduce a fluorine atom or a fluorinated moiety of unique properties. Despite the tremendous advances made in that field, key synthetic challenges remain to synthesize fluorinated scaffolds. Among the different developed strategies to ravel synthetic issues, the use of inexpensive and readily available copper-based fluorinated reagents appeared over the years as a powerful tool in various transformations for the introduction of fluorinated moieties. Such strategy has already demonstrated a significant synthetic value for the trifluoromethylation of various compounds [bib_ref] Emerging Areas in Copper-Mediated Trifluoromethylations: Catalytic and Oxidative Processes, Jouvin [/bib_ref]. In contrast, available reagents for the incorporation of a CF 2 R (R = H, alkyl, R F , FG; FG = functional group) moiety remain restricted, despite the potential of these functionalized fluorinated moieties. In this review, the main contributions in the field of copper-based reagents for the introduction of CF 2 H, CF 2 FG, CF 2 Me and CF 2 R F moieties over the last 5 years (period of 2014-2019) will be summarized. The design and the elaboration of either pre-formed or in situ-generated copper-based reagents was an efficient tool in several reactions. Note that only transformations involving the use of such copper-based reagents will be depicted and copper-catalyzed reactions are therefore beyond the scope of this review.
## Review copper-based difluoromethylating reagents
In this section the key advances made to access copper-based difluoromethylating reagents are summarized. The CF 2 H moiety, a well recognized alcohol and thiol bioisoster, is particularly attractive due to its unique features. Besides, this residue is present in several bioactive compounds such as Deracoxib and Thiazopyr. In comparison with trifluoromethylcopper complexes, the difluoromethylcopper ones are less stable as demonstrated by the work of Burton in 2007. Investigations on the in situ synthesis of difluoromethylcopper from a difluoromethylcadmium source at low temperature and the study of its reactivity with various classes of compounds such as allylic halides, propargylic halides and tosylates, iodoalkynes and reactive alkyl halides were realized. It was established that CuCF 2 H readily decompose into 1,1,2,2-tetrafluoroethane and cis-difluoroethylene. From this pioneer work, attention was paid either to the design of new synthetic pathways for the synthesis of a well-defined copper-based reagent or to new tools for the in situ generation of an active CuCF 2 H species and its application in several transformations.
## Pre-defined difluoromethylating reagents
In the quest for well-defined and isolable MCF 2 H species, Sanford depicted for the first time in 2017 the synthesis and characterization of isolable difluoromethylcopper(I) complexes. The latter were prepared in a two-step sequence starting from the corresponding (NHC)CuCl as precursors in the presence of NaOt-Bu followed by the addition of TMSCF 2 H (Scheme 1). The latter was prepared in a one step synthesis after reduction of the Ruppert-Prakash reagent with sodium borohydride. The key of success was the use of bulky IPr and SIPr ligands to stabilize the organometallic species. Indeed, in the case of IPr as a ligand, the complex was stable in solution at room temperature for at least 24 hours. The reactivity of the complex was then studied in stoichiometric reactions with aryl iodides and iodonium salts. The difluoromethylation reaction was smoothly carried out at 90 °C with electron-rich and electron-poor aryl iodides. However, the reaction was more efficient with electron-poor aryl iodides (Scheme 1). It is important to highlight that, in the course of their study for the synthe-sis of a stable and isolable (NHC)CuCF 2 H complex and the study of its reactivity, Sanford and co-workers demonstrated the possibility to develop a catalytic version of the reaction through the in situ generation of the active (IPr)CuCF 2 H, starting from (IPr)CuCl. Scheme 1: Synthesis of the first isolable (NHC)CuCF 2 H complexes from TMSCF 2 H and their application for the synthesis of difluoromethylated arenes from aryl iodides. a Yields were determined byF NMR with fluorobenzene as the internal standard. b Reaction carried out at 120 °C.
In situ-generated copper-based difluoromethylating reagents Although the review focused on the 2014-2019 period, a brief overview of seminal major advances should be given. In 2012, Hartwig and co-worker studied the difluoromethylation reaction of aryl and vinyl iodides by a copper-mediated transformation using TMSCF 2 H as the fluorinated source. In this work, CuCF 2 H was suggested as the active species to promote the expected transformation. They highlighted that the formation of a cuprate species: Cu(CF 2 H) 2 − , favoured by the presence of an excess of TMSCF 2 H, might act as a reservoir of the unstable and reactive CuCF 2 H species. Xu and Qing reported a similar strategy for the difluoromethylation of electron-poor (hetero)aryl iodides at room temperature, using only 2.4 equivalents of TMSCF 2 H. Note that the use of a strong base (t-BuOK) and 1,10-phenanthroline as a ligand was crucial in their system. In 2012, Prakash also studied the in situ generation of CuCF 2 H from n-Bu 3 SnCF 2 H, the presence of DMF being the key to stabilize the CuCF 2 H intermediate(Scheme 2). From these seminal works, a handful of reports was then published by different research groups. In 2014, the group of Goossen astutely reported the in situ generation of the CuCF 2 H complex starting from TMSCF 2 H, CuSCN and CsF as an activator in DMF. This approach was successfully applied in a Sandmeyer-type difluoromethylation reaction (Scheme 3). Starting from (hetero)aryldiazonium salts, a panel of difluoromethylated arenes and heteroarenes was obtained (26 examples, up to 84% yield). Note that the transformation was also carried out starting from 4-methoxyaniline followed by the in situ formation of the corresponding diazonium salt.
In the same vein, the authors used this in situ generation of a CuCF 2 H species to access high value-added difluoromethylthiolated molecules starting from organothiocyanates. With this approach, they then developed a one pot, two-step sequence (generation of the organothiocyanates followed by the difluoromethylation step) for the functionalization of alkyl bromides, alkyl mesylates, aryldiazonium saltsas well as electron-rich arenes(Scheme 4).
In 2015, the group of Qing investigated the oxidative difluoromethylation reaction of terminal alkynes with TMSCF 2 H via a copper-mediated reaction. Using a stoichiometric amount of CuI, in the presence of t-BuOK and 9,10-phenanthraquinone, the functionalization of a panel of (hetero)aromatic and aliphatic terminal alkynes was achieved (Scheme 5). A good functional group tolerance was observed as alkynes bearing a cyano, ester, bromo or amino group among others were suitable substrates. Based onF NMR studies, the authors suggested the following mechanism: first the in situ generation of a CuCF 2 H introduce them onto molecules was studied over the last years and the major advances will be summarized in this section.
An in situ-generated copper-based CF 2 PO(OEt) 2 reagent
As a bioisostere of the phosphonate group, a lot of attention was paid to the difluoromethylphosphonate residue as well as the development of efficient methodologies to introduce it onto molecules. In that context, main contributions were made by the groups of Poisson and Goossen. In the course of their study regarding the synthesis of difluoromethylphosphonate-containing molecules, Poisson and co-workers investigated the in situ generation of a CuCF 2 PO(OEt) 2 species and its application to functionalize various classes of compounds. The active species was prepared from TMSCF 2 PO(OEt) 2 , a copper salt and an activator. Note that the TMSCF 2 PO(OEt) 2 was easily prepared from the commercially available BrCF 2 PO(OEt)and TMSCl under basic conditions. The access to CF 2 PO(OEt) 2 -containing arenes was obtained after a Sandmeyer-type reaction (Scheme 7, reaction a). The reaction was efficient, al-Scheme 7: Synthesis of difluoromethylphosphonate-containing molecules using the in situ-generated CuCF 2 PO(OEt) 2 species.
though heteroaryl diazonium salts were reluctant in this reaction. To overcome these limitations, hypervalent iodinated species were used as substrates. The copper-mediated reaction with λ 3 -iodanes demonstrated a large functional group tolerance and was efficiently applied to the synthesis of CF 2 PO(OEt) 2 -containing (hetero)arenes, alkenes and alkynes (Scheme 7, reactions b-d). Later on, the same group depicted the Pd-catalyzed introduction of the CF 2 PO(OEt) 2 residue on (hetero)aryl iodidesby using an in situ-generated copper-based reagent Finally, the Poisson's group developed a methodology for the Ullman cross-coupling reaction between the in situ-generated CuCF 2 PO(OEt) 2 and aryl iodides containing a coordinating group (e.g., CO 2 CH 3 , COCH 3 , NO 2 ), at the ortho-position of the halide. This reaction broadened the portfolio of CF 2 PO(OEt) 2 -containing molecules leading to the corresponding compounds in good to excellent yields (Scheme 9). Note that the versatility of this methodology was further proved through its application to disulfideswith moderate to good yields.
Scheme 9: Synthesis of difluoromethylphosphonate-containing molecules using in situ-generated CuCF 2 PO(OEt) 2 species by an Ullman cross-coupling. CG = coordinating group.
Poisson and co-workers also reported the reaction of the CuCF 2 PO(OEt) 2 reagent with α-diazocarbonyl derivatives. Depending on the copper salt used for the generation of the copper reagent, the reaction with α-diazocarbonyl derivatives provided either the α-fluorovinylphosphonate, in a stereoselective fashion, or the SCF 2 PO(OEt) 2 derivatives. In the same vein, the reaction of the CuCF 2 PO(OEt) 2 species, generated from CuSCN, with α-bromoketones provided the α-SCF 2 PO(OEt) 2 -containing ketones(Scheme 10).
In 2019, the group of Goossen developed an approach to access SCF 2 PO(OEt) 2 -containing arenes based on a Sandmeyer thiocyanation reaction followed by a Langlois-type nucleophilic substitution of the cyano group by the CF 2 PO(OEt) 2 residue. Several (diethylphosphono)difluoromethylthiolated products were obtained and this report further showcased the potential of using a copper-based reagent for the introduction of fluorinated moieties as this reaction involved the in situ generation of a suitable CuCF 2 PO(OEt) 2 species (Scheme 11).
Scheme 10: Synthesis of (diethylphosphono)difluoromethylthiolated molecules using in situ-generated CuCF 2 PO(OEt) 2 species. Phen: 1,10-phenanthroline.
## An in situ-generated copper-based cf 2 ch 3 reagent
A strong interest was dedicated to the CF 2 CH 3 residue, an important moiety in medicinal chemistry. Among the different approaches developed to synthesize CF 2 CH 3 -containing molecules, Wang, Hu and co-workers demonstrated the possibility to use 1,1-difluoroethylsilane (TMSCF 2 CH 3 ) as a precursor for the in situ generation of the corresponding CuCF 2 CH 3 species. The synthetic utility of this copper-based reagent was illustrated through the 1,1-difluoroethylation of diaryliodonium salts, leading to the corresponding (1,1-difluoroethyl)arenes in moderate to high yields (Scheme 13). The transformation turned out to be functional group tolerant and even heteroaromatic compounds were functionalized.
## Copper-based cf 2 r f reagents
Due to the importance of perfluorinated moietiesand since their synthesis could not be achieved from the fluorination of the corresponding alkyl chains like in case of perfluoroalkyl arenes, several research groups investigated the synthesis of CF 2 R F -containing molecules via the use of perfluoroalkyl copper species. Before 2014, key contributions were made by the groups of Kremlev, Tyrra, Hartwig [66,67] and Grushinas briefly summarized below. These major advances paved Scheme 14: Pioneer works for the pentafluoroethylation and heptafluoropropylation using a copper-based reagent. Phen = 1,10-phenanthroline. a 19 F NMR yields determined using 1,3-bis(trifluoromethyl)benzene as the internal standard.
Scheme 13: Copper-mediated 1,1-difluoroethylation of diaryliodonium salts by using the in situ-generated CuCF 2 CH 3 species.
the way towards the synthesis of important pentafluoroethylated and more generally perfluoroalkylated molecules. Kremlev, Tyrra and co-workers depicted the in situ generation of a CuCF 2 CF 3 species by mixing Zn(CF 3 )Br·2DMF and CuBr, and its application for the functionalization of (hetero)aryl halides (Scheme 14).
In the course of their studies to develop stable and well-defined copper reagents for perfluoroalkylation reactions [66], Hartwig developed in 2011 the (Phen)CuCF 3 and (Phen)CuCF 2 CF 2 CF 3 complexes from inexpensive reagents. Indeed, when mixing (CuOt-Bu) 4 , 1,10-phenanthroline and the corresponding TMSR F , the perfluoroalkyl copper complexes were isolated for the first time . One year later, they demonstrated that these copper-based reagents ((Phen)CuCF 2 R F , R F = F, CF 3 and CF 2 CF 3 ) were efficient in a two-step sequence reaction (borylation/perfluoroalkylation) allowing the functionalization of either sterically hindered arenes or aryl bromides with the CF 2 CF 3 and CF 2 CF 2 CF 3 moieties (Scheme 14, b). In 2013, the group of Grushin reported the synthesis, characterization and application of a copper-based pentafluoroethylating reagent (Scheme 14). Using the cost-efficient pentafluoroethane as a precursor, the (K(DMF) 2 )(t-BuO)Cu(CF 2 CF 3 ) complex was prepared either from the pre-isolated (K(DMF))Cu(Ot-Bu) 2 or in situ from CuCl, t-BuOK in DMF in a nearly quantitative yield. The copper reagent was used for the pentafluoroethylation of a panel of (hetero)aryl iodides and bromides (up to 99% 19 F NMR yield) and its synthetic utility was further demonstrated with the functionalization of different classes of compounds (benzyl and vinyl bromides, 4-biphenylboronic acid, phenylacetylene for instance).
From these pioneering reports of perfluoroalkylation (trifluoromethylation, pentafluoroethylation and heptafluoropropylation), several groups studied the synthesis and/or the application of copper-based reagents in various transformations as depicted in this section. This latter will be organized into two sub-sections depending if the CuR F -reagent was well-defined or in situ generated.
## Well-defined pentafluoroethylating reagents
In 2014, a report from Hartwig dealt with the copper-mediated perfluororalkyaltion of (hetero)aryl bromides using the previously developed PhenCuR F. Although the trifluoromethylation reaction was mainly studied, the methodology was efficiently extended to the pentafluoroethylation of various heteroarenes such as pyridine, pyrimidine and quinolone derivatives, for instance, when the PhenCuCF 2 CF 3 complex was used as the pentafluoroethyl source (24 examples, up to 99% 19 F NMR yield and up to 93% isolated yield, Scheme 15). Note that a complete mechanistic study was recently reported to explain the reactivity of this well-designed complex.
In 2015, Grushin reported the generation of four well-defined CuC 2 F 5 complexes, namely (Ph 3 P) 2 CuCF 2 CF 3 , (bpy)Cu-CF 2 CF 3 , (IPr*)CuCF 2 CF 3 and (Ph 3 P)Cu(Phen)CF 2 CF 3 . The reactivity of the latter was studied for the synthesis of pentafluoroethyl ketones from acyl chlorides. Indeed, the pentafluoroethylation of a large panel of acyl chlorides was achieved illustrating the synthetic utility and the efficiency of the newly designed (Ph 3 P)Cu(phen)CF 2 CF 3 reagent (Scheme 16).
Huang and Weng and co-workers reported the synthesis of airstable perfluorocarboxylatocopper(I) complexes and their use in the perfluoroalkylation of (hetero)aryl halides. By mixing t-BuOCu, in situ generated from CuCl and t-BuONa, with 1,10phenanthroline, followed by a reaction with perfluorocarboxylic acids, four (Phen) 2 Cu(O 2 CCF 2 R F ) complexes were synthesized (R F = CF 3 , CF 2 CF 3 , CF 2 CF 2 CF 3 and CF 2 CF 2 CF 2 CF 3 ). The reaction was efficient to 97% yield), showed a good functional group tolerance (i.e., cyano, ester, ketone) and even heteroarenes such as pyridine, Scheme 15: Pentafluoroethylation of (hetero)aryl bromides using the (Phen)CuCF 2 CF 3 complex.F NMR yields were determined using 4-trifluoromethoxyanisole as the internal standard. a Isolated yields.
Scheme 16: Synthesis of pentafluoroethyl ketones using the (Ph 3 P)Cu(phen)CF 2 CF 3 reagent.F NMR yields were given using 1,3-bis(trifluoromethyl)benzene as the internal standard.
quinoline and quinoxaline were functionalized with the four fluorinated moieties (Scheme 17).
Scheme 17: Synthesis of (Phen) 2 Cu(O 2 CCF 2 RF) and functionalization of (hetero)aryl iodides.
In situ-generated copper-based pentafluoroethylating reagents. Note that the methodology was also applied to the functionalization of a vinylboronic acid and a vinyl bromide (Scheme 18). In 2015, Grushin and co-workers further investigated the functionalization of vinyl halides with CuR F reagents generated from inexpensive fluoroform (R F = CF 3 ) and pentafluoroethane (CF 3 CF 2 H). Both trifluoromethylation and pentafluoethylation of vinyl bromides and iodides were efficiently achieved in high yields under mild reaction conditions. Noteworthy, the transformation turned out to be functional group tolerant and highly chemo-and stereroselective (Scheme 20).
The group of Hu studied the fluoroalkylation of aryl halides. Indeed, a copper(0)-mediated reductive cross-coupling reaction between the iodobenzene and various 2-bromo-1,1,2,2-tetrafluoroethyl derivatives (RCF 2 CF 2 Br) was developed presumably involving a RCF 2 CF 2 Cu species (Scheme 21).
In 2015, Yagupolskii and co-workers investigated the synthesis of perfluoroalkylcopper reagents. Depending on the reaction conditions they were able to access to perfluoroorganolithium copper species or perfluroalkylcopper derivatives from iodoperfluoroalkanes in reaction with either n-BuLi or copper powder, respectively (Scheme 22).
In 2017, the group of Hu offered an original synthetic route to the generation of the PhenCuCF 2 CF 3 reagent. Indeed, they demonstrated that the Ruppert-Prakash reagent was a suit- able source for the generation of tetrafluoroethylene in the presence of a catalytic amount of NaI. Then, the cupration of the tetrafluoroethylene led to the formation of the expected Phen-CuCF 2 CF 3 reagent (Scheme 23). This constituted a complementary approach to the existing ones for its synthesis, as it avoided the use of TMSCF 2 CF 3 or CF 3 CF 2 H. This copperbased reagent was then used for the pentafluoroethylation of iodoarenes. The transformation was efficient and turned out to be functional group tolerant. The same group extended their protocol to the functionalization of aryldiazonium salts. Very recently, a similar protocol was applied to the pentafluoroethylation of (hetero)aryl halides as well as alkenyl iodides derived from natural compounds (e.g., glycals, nucleosides and nucleobases).
In 2018, Hu and co-workers reported a complementary approach for the pentafluoroethylation of aryl iodides using TMSCF 3 for the formation of CuCF 2 CF 3. They suggested that in the presence of CuCl, KF and TMSCF 3 , the corresponding CuCF 3 species will be formed and a subsequent homologation step involving a putative copper difluorocarbene will allow the formation of the CuCF 2 CF 3 species. With this tool in hand, a panel of aryl iodides was functionalized (Scheme 24).
## Scheme 24:
Generation of a CuCF 2 CF 3 reagent from TMSCF 3 and applications.
# Conclusion
This review aims at providing an overview of the recent advances made since 2014 for the construction of CF 2 R-containing molecules (R ≠ F) using versatile and efficient copperbased reagents. Groundbreaking advances were made in the synthesis of well-defined copper-based reagents and innovative strategies were developed to generate in situ CuR f complexes from various precursors. Unprecedented transformations were successfully achieved using these copper-based reagents and these efficient synthetic tools opened new perspectives in the very active research field of organofluorine chemistry. Nevertheless, this field is still in its infancy and milestones towards copper-based difluoromethylating reagents are expected in the upcoming years. |
Pharmacokinetic interactions and tolerability of rosuvastatin and ezetimibe: an open-label, randomized, multiple-dose, crossover study in healthy male volunteers
# Introduction
Globally, cardiovascular disease (CAD) is a major cause of morbidity and death. Dyslipidemia, hypertension, diabetes mellitus (DM), and obesity are important risk factors for CAD. 1,2 Among them, dyslipidemia due to elevated total and low-density lipoprotein cholesterol (LDL-C) and reduced high-density lipoprotein cholesterol is closely associated with coronary artery disease, atherosclerosis, and cerebrovascular disease.
Cholesterol-lowering drugs, such as 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase inhibitors (statins) reduce the risk of CAD. Rosuvastatin is a fully synthetic HMG-CoA reductase inhibitor that is more effective at reducing LDL-C levels than other statins when used at a low dose; thus, it is the most potent drug for reducing cardiovascular risk. [bib_ref] Rosuvastatin: beyond the cholesterol-lowering effect, Cortese [/bib_ref] However, statin monotherapy does not always reduce LDL-C levels to acceptable levels in patients with severe hypercholesterolemia, particularly those at high cardiovascular risk. [bib_ref] Statin: within-group comparisons, statin escape and combination therapy, Tikkanen [/bib_ref] Therefore, to meet the target LDL-C levels, it may be necessary to combine stains with an additional drug that acts via a different mechanism.
Ezetimibe is a selective cholesterol and phytosterol absorption inhibitor that targets gastrointestinal cholesterol in the small intestine; this drug reduces plasma cholesterol in humans by 15%-20%. [bib_ref] Ezetimibe: an update on its clinical usefulness in specific patient groups, Hammersley [/bib_ref] Ezetimibe administration is effective for both mono-and combination therapy with HMG-CoA reductase inhibitors in patients with hypercholesterolemia. [bib_ref] Ezetimibe in hypercholesterolemia, Farnier [/bib_ref] Ezetimibe monotherapy is the recommended treatment for primary hypercholesterolemia in patients in whom statin therapy is contraindicated or in those that cannot tolerate statins. A previous study shows that a fixed-dose combination of ezetimibe and rosuvastatin is significantly better than rosuvastatin alone in terms of reducing LDL-C and triglyceride levels, and that the rate of reduction is even greater in patients with DM or metabolic syndrome. [bib_ref] Effect of fixed-dose combinations of ezetimibe plus rosuvastatin in patients with primary..., Kim [/bib_ref] An early trial demonstrated additional reductions in LDL-C levels (by 12%-19%) when ezetimibe was administered with atorvastatin. [bib_ref] Effect of ezetimibe coadministered with atorvastatin in 628 patients with primary hypercholesterolemia, Ballantyne [/bib_ref] Daily combined administration of ezetimibe plus rosuvastatin led to a significant fall in LDL-C when compared with rosuvastatin alone. [bib_ref] Pharmacodynamic interaction between ezetimibe and rosuvastatin, Kosoglou [/bib_ref] The potential therapeutic interaction between ezetimibe and HMG-CoA reductase inhibitors offers significant clinical benefits provided that this drug combination is both safe and well tolerated. [bib_ref] The effect of fluvastatin on the pharmacokinetics and pharmacodynamics of ezetimibe, Reyderman [/bib_ref] [bib_ref] Ezetimibe: an update on the mechanism of action, pharmacokinetics and recent clinical..., Sweeney [/bib_ref] The primary purpose of this study was to examine the pharmacokinetic (PK) interactions and tolerability of rosuvastatin and ezetimibe when taken orally over 10 consecutive days by healthy Korean volunteers.
## Subjects and methods subjects
Healthy male volunteers aged 19-45 years with a body mass index of 19-28 kg/m 2 were eligible for the study. All subjects were considered to be in good health based on medical history, physical examinations, vital signs (blood pressure, heart rate, and body temperature), 12-lead ECG, clinical laboratory tests (hematology, blood chemistry, and urinalysis), serology (hepatitis B surface antigen, hepatitis C virus antibodies, and HIV antigen/antibodies), and urine drug screening (amphetamines, methamphetamines, barbiturates, cocaine, opiates, benzodiazepine, cannabinoids, and methadone) within 4 weeks before the first administration of the study drug. Subjects with a known allergy or hypersensitivity to rosuvastatin or ezetimibe, or with a history of drug abuse, were excluded.
## Study design
The study was designed as an open-label, randomized, multiple-dose, 3-treatment, 3-period, 6-sequence, and crossover clinical trial [fig_ref] Figure 1: study design of clinical trial [/fig_ref]. All subjects were randomly assigned to 1 of 6 sequences and received 1 of 3 different treatments every 24 hours over 10 days: rosuvastatin 20 mg (Treatment A), ezetimibe 10 mg (Treatment B), or rosuvastatin 20 mg plus ezetimibe 10 mg (Treatment C). All treatments were given under fasted conditions along with 240 mL of water. After administration of the study drug(s), subjects were required to fast for 4 hours. Following a 2-week washout interval, subjects received one of the other treatment regimen. This was repeated until they had received all 3. Subjects were admitted in the Clinical Trial Center (CTC) at Asan Medical Center (AMC) from Days 9 to 11 (24 hours after the last dose) during each treatment period. On Days 12 and 13, all subjects visited the CTC
## 817
DDi of rosuvastatin and ezetimibe to assess the tolerability and PK of rosuvastatin or ezetimibe. The schedule for the second and third treatment periods was the same as that for the first period. Follow-up visits were performed within 11 days after the last treatment.
For PK analysis, sequential blood samples were collected prior to (Days 1 and 9) and 0, 0.25, 0.5, 1, 1.5, 2, 3, 4, 6, 8, 12, 24, 48, and 72 hours after the Day-10 dose. All blood samples used to determine the concentration of rosuvastatin, total ezetimibe, and free ezetimibe were drawn into EDTA tubes and separated by centrifugation at 1,800 g for 8 minutes at 4°C. Samples were then stored at -70°C until analysis.
Tolerability was assessed throughout the study by: 1) measuring vital signs, 2) 12-lead ECG, 3) clinical laboratory tests (hematology, blood chemistry, and urinalysis), 4) physical examinations, and 5) monitoring of adverse events (AEs). AEs were recorded in terms of symptoms and signs, duration, intensity, relationship to the study drug, action taken, outcome, and seriousness.
The study protocol was approved by the Korean Ministry of Food and Drug Safety and the Institutional Review Board of the AMC, Seoul, Republic of Korea. The study was conducted at the CTC of the AMC from June to September 2014. All subjects provided written informed consent before undergoing screening tests. The trial was registered at ClinicalTrials.gov (identifier number NCT02127320).
# Analytical methods
## Measurement of rosuvastatin concentrations
The plasma concentration of rosuvastatin was measured using a validated liquid chromatography method with tandem mass spectrometric detection (LC-MS/MS) (Thermo Fisher Scientific, Waltham, MA, USA). The analytical column was Nanospace SI-2 C 18 column (75×2.1 mm × 3.0 μm, Shiseido, Tokyo, Japan), and the mobile phase comprised acetonitrile (A), deionized water (B), and formic acid (C) (A:B:C=45:55:0.1, v/v/v). The flow rate was 0.2 mL/minute. For rosuvastatin and the internal standard (rosuvastatin-d 6 ), the precursor-to-production reactions monitored were m/z 482.18 → 258.16 and 488.20 → 264.20, respectively. This assay had a lower limit of quantitation (LLOQ) of 0.5 ng/mL (signal to noise ratio .5), and calibration curves covered the concentration range of 0.5-300 ng/mL (R 2 .0.995).
## Measurement of ezetimibe and total ezetimibe concentrations
The plasma concentration of free and total ezetimibe (free ezetimibe plus ezetimibe glucuronide) was measured using a validated LC-MS/MS (Sciex API 4000 LC-MS/MS system, Framingham, MA, USA). The analytical column was ACQUITY UPLC C 18 column (75×2.0 mm × 3.0 μm, Milford, MA, USA), the mobile phase comprised acetonitrile (A) and 5 mM ammonium acetate (B) [A:B=65:35, v/v]. The flow rate was 0.2 mL/minute. For ezetimibe and the internal standard (ezetimibe-d 4 ), the precursor-to-production reactions monitored were m/z 408.3 → 271.1 and 412.3 → 275.1, respectively. The LLOQ was 0.2 and 0.5 ng/mL for free and total ezetimibe (signal to noise ratio .5), respectively. The calibration curves of free and total ezetimibe covered the concentration range 0.2-200 ng/mL and 0.5-500 ng/mL, respectively (R 2 .0.995). All plasma analyses were performed at BioCore Co., Seoul, Republic of Korea.
## Pk assessment and statistical analysis
The PK of rosuvastatin, free and total ezetimibe in each subject was analyzed using a non-compartmental method with WinNonlin ® software 6.3 (Pharsight Co., Princeton, NJ, USA). All analyses were based on actual sampling times. The peak plasma concentration at steady-state (C max,ss ) and the time taken to reach C max,ss (T max,ss ) were determined from observed values. The terminal elimination rate constant (λ z ) was estimated by linear regression of the terminal log-linear portion of the plasma concentration-time curves. The t 1/2β for each participant was calculated as ln(2)/λ z .
All statistical analyses were performed using SAS ® software (v 9.3; SAS Institute Inc., Cary, NC, USA) and Phoenix ® WinNonlin 6.3. Demographic data and PK parameters were summarized using descriptive statistics. For the comparison of PK characteristics between rosuvastatin or ezetimibe alone with combined administration of rosuvastatin and ezetimibe, C max,ss and (AUC τ,ss ) were log-transformed and tested by analysis of variance. The mean differences and 90% CIs were back-transformed to obtain geometric mean ratios and CIs for those ratios. Fisher's exact test was used to compare the frequency of AEs between 2 drugs. A P-value ,0.05 was deemed significant. 12
# Result study participants
Of the 24 healthy Korean male volunteers enrolled, 21 completed the study. Three subjects withdrew informed consent for personal reasons. One was before first hospitalization and 2 after second hospitalization. All subjects were included in the tolerability assessment, whereas only the subjects who completed blood sampling as scheduled were included in PK analysis. Participant demographics, including age, height, weight, and body mass index are summarized in [fig_ref] Table 1: Demographic characteristics of the study participants [/fig_ref]
# Pk analysis
To evaluate the PK drug-drug interactions between rosuvastatin and ezetimibe, the PK profiles of rosuvastatin, total and free ezetimibe were separately assessed [fig_ref] Figure 2: Mean [/fig_ref].
Of the 24 subjects who were administered the study drugs, 1 was excluded from all PK analyses due to delay in administering ezetimibe. Finally, 23 subjects were included in the PK analysis of rosuvastatin, and 20 were included in the PK analysis of ezetimibe.
The PK parameters of rosuvastatin in the absence and presence of ezetimibe are shown in [fig_ref] Table 2: PK comparisons of rosuvastatin, total and free ezetimibe [/fig_ref]. For rosuvastatin, the steady-state mean ratios of single versus combination doses (90% CI) for AUC τ,ss and C max,ss were 1.076 (1.019-1.136) and 1.099 (1.003-1.204), respectively.
For total ezetimibe, the steady-state mean ratios (90% CI) of AUC τ,ss and C max,ss were 1.055 (0.969-1.148) and 0.996 (0.873-1.135), respectively, while those for free ezetimibe were 1. [fig_ref] Figure 1: study design of clinical trial [/fig_ref] , respectively.
## Tolerability
Ten subjects experienced a total of 28 AEs; of these, 23 were considered to be "possibly related" and 2 "probably related" to the study drug. Among the 25 events considered to be related to the study drug, the most common in the ezetimibe alone and ezetimibe plus rosuvastatin group were abdominal pain (4 subjects, 4 events) and headache (3 subjects, 4 events; [fig_ref] Table 3: summary of adverse events after administration of rosuvastatin and ezetimibe [/fig_ref]. No significant changes in hepatic and renal function test, such as ALT, AST, creatine kinase, serum creatinine, and BUN were observed (data not shown). Laboratory tests revealed microscopic hematuria (occult blood) in 2 subjects:
## 819
DDi of rosuvastatin and ezetimibe in 1 during the second period after rosuvastatin-only treatment and in the other one at the follow-up visit. No clinically significant abnormalities were found with respect to vital signs, ECGs, and physical examinations. No volunteers dropped out due to AEs and no serious AE occurred during the entire course of the study. All AEs were mild in severity, and resolved without any sequelae.
# Discussion
The present study examined the potential PK interaction between rosuvastatin and ezetimibe, and their tolerability by healthy male subjects taking multiple oral doses over a 10-day period. Once absorbed, ezetimibe is extensively metabolized in the intestine to yield its glucuronide conjugate; therefore, it exists mainly as ezetimibe glucuronide in the body. Cholesterol absorption inhibiting activity assays in a rat model demonstrate that both ezetimibe and ezetimibe glucuronide are pharmacologically active and that ezetimibe glucuronide is more potent than ezetimibe of the 2. [bib_ref] Comparison of the activity and disposition of the novel cholesterol absorption inhibitor,..., Van Heek [/bib_ref] Here, we show that the 90% CIs for the geometric mean ratios of single versus combination treatment for C max,ss and AUC max,ss were 0.80-1.25; the exception was C max,ss of free ezetimibe (90% CI, 1.038-1.346). Recently, a PK study in healthy subjects using rosuvastatin (study part A) and ezetimibe (study part B) was reported; however, the study design was 7 daydosing, 2-treatment, 2-period, 2-sequence crossover study with 2 treatment parts. 14 Although the study design was different from the present study, the coadministration increased both C max,ss and AUC max,ss of free ezetimibe (90% CI of C max,ss [fig_ref] Figure 1: study design of clinical trial [/fig_ref] gMr of C max,ss (90% ci) [fig_ref] Figure 1: study design of clinical trial [/fig_ref] Notes: Values are presented as arithmetic mean ± sD (cV, %). a T max at day 10. b n=20. Two subjects were excluded because R 2 ,0.90, and 1 subject was excluded because of %aUc extrapolation .20%. c n=19. One subject was excluded because of R 2 ,0.90 and/or %aUc extrapolation .20%. d n=12. eight subjects were excluded because of R 2 ,0.90 and/or %aUc extrapolation .20%. e n=18. Two subjects were excluded because of R 2 ,0.90. Abbreviations: C max,ss , concentration at steady-state; CV, coefficient of variation; GMR, geometric least square mean ratio; T max,ss , time taken to reach c max,ss . and AUC max,ss , 0.994-1.285 and 1.094-1.341, respectively). These results suggest that we might need larger sample size when comparing a fixed-dose combination product with separate tablets.
[formula] Visual impairment 0 (0) 1 (1) 0 (0) 1 (1) abdominal discomfort 0 (0) 3 (3) 1 (1) 4 (4) Diarrhea 0 (0) 1 (1) 0 (0) 1 (1) Dyspepsia 0 (0) 0 (0) 1 (1) 1 (1) nausea 0 (0) 1 (1) 0 (0) 1 (1) Vomiting 0 (0) 1 (1) 0 (0) 1 (1) chest pain 1 (2) 0 (0) 0 (0) 1 (2) Feeling hot 0 (0) 1 (1) 0 (0) 1 (1) sensation of foreign body 0 (0) 1 (1) 0 (0) 1 (1) Upper respiratory tract infection 1 (1) 0 (0) 0 (0) 1 (1) Blood uric acid increased 0 (0) 0 (0) 1 (1) 1 (1) Myalgia 0 (0) 1 (1) 0 (0) 1 (1) Pain in extremity 0 (0) 1 (1) 0 (0) 1 (1) Dizziness 0 (0) 1 (1) 0 (0) 1 (1) headache 0 (0) 1 (2) 2 (2) 3 (4) hematuria 2 (2) 0 (0) 1 (1) 2 (3) Urticaria 0 (0) 0 (0) 1( [/formula]
A previous single-dose study of 20 mg ezetimibe reported that the C max,ss of free ezetimibe was higher (5.2 ng/mL) than that reported herein (5.0 ng/mL; combination treatment, [fig_ref] Table 2: PK comparisons of rosuvastatin, total and free ezetimibe [/fig_ref] ; however, no AEs or clinically significant changes were noted on physical examinations. [bib_ref] Disposition of the selective cholesterol absorption inhibitor ezetimibe in healthy male subjects, Patrick [/bib_ref] As the pharmacological effect of the ezetimibe conjugate is more potent than that of the parent drug, and there is a small amount of free ezetimibe (,8%) in the body, the increase in the C max,ss value of free ezetimibe may have no impact on treatment effects or on toxicity.
Kosoglou et al reported that the T max of free ezetimibe was 6.0 hours in individuals administered 10 mg rosuvastatin plus 10 mg ezetimibe, and 10 mg of ezetimibe alone for 14 consecutive days; the subjects were mainly Caucasian (87%) with LDL-C levels $130 mg/dL. [bib_ref] Pharmacodynamic interaction between ezetimibe and rosuvastatin, Kosoglou [/bib_ref] These data are consistent with the results presented herein (5.4 and 7.3 hours, respectively). When we compared the T max,ss value of free ezetimibe with the time of AE onset, we found no clear association between them, suggesting that AEs were not related to the C max,ss . Taken together, rosuvastatin and ezetimibe were well tolerated when administered either alone or concomitantly. We found that the LDL-C level significantly decreased in rosuvastatin-only and combination (rosuvastatin plus ezetimibe) groups (by 43.7% and 56.9%, respectively), but not significantly (9.7%) in the group taking ezetimibe alone, after the 10-day multiple dosing regimen. In all treatment groups, the LDL-C levels returned to baseline after a 2-week washout. Although administration of 10 mg of ezetimibe reduced LDL-C levels significantly in volunteers (n=8) with hypercholesterolemia, 1 of 8 subjects in the ezetimibe-only group achieved a 35%-50% reduction compared with half (4/8) of those in the combination group (ezetimibe and fluvastatin). [bib_ref] The effect of fluvastatin on the pharmacokinetics and pharmacodynamics of ezetimibe, Reyderman [/bib_ref] This suggests that ezetimibe alone has limited efficacy to control hypercholesterolemia. A recent population pharmacodynamic model of Japanese patients supported the beneficial effects of ezetimibe when combined with rosuvastatin. [bib_ref] Population pharmacodynamic analysis of LDL-cholesterol lowering effects by statins and comedications based..., Kakara [/bib_ref] Therefore, ezetimibe is expected to have synergistic effects in patients with hypercholesterolemia.
Previous reports demonstrated multiple peaks in the plasma concentration-time profiles of total and free ezetimibe; this was due to enterohepatic recycling after an oral dose of ezetimibe. [bib_ref] Pharmacokinetics and safety of ezetimibe/simvastatin combination tablet, Chu [/bib_ref] [bib_ref] A population pharmacokinetic model that describes multiple peaks due to enterohepatic recirculation..., Ezzet [/bib_ref] Here, we also observed multiple peaks in PK plots for total and free ezetimibe; these peaks coincided with mealtimes and are thought to be due to enterohepatic recycling [fig_ref] Figure 2: Mean [/fig_ref]. This suggests that food intake triggers the enterohepatic recycling of total ezetimibe by emptying the gallbladder. Further studies based on population PK modeling would help to elucidate the impact of enterohepatic circulation on PK and its inter-individual variation quantitatively. Finally, this study would be important evidence to develop fixed-dose combination tablet to increase patient compliance. [bib_ref] Comparison of pharmacokinetics and safety of a fixed-dose combination of rosuvastatin and..., Min [/bib_ref] [bib_ref] Pharmacokimetics and bioequivalence of a rosuvastatin/ezetimibe fixed-dose combination tablet versus single agents..., Kang [/bib_ref] Conclusion Rosuvastatin (20 mg) and ezetimibe (10 mg) showed no clinically significant PK interactions. Concomitant administration of these 2 drugs was well tolerated by healthy male subjects taking multiple doses over 10 days.
[fig] Figure 1: study design of clinical trial (open-label, multiple-dose, three-treatment, three-period, six-sequence crossover study). Notes: (A) rosuvastatin 20 mg. (B) ezetimibe 10 mg. (C) rosuvastatin 20 mg plus ezetimibe 10 mg. Abbreviation: seQ, sequence. Drug Design, Development and Therapy 2018:12 submit your manuscript | www.dovepress.com Dovepress Dovepress [/fig]
[fig] Figure 2: Mean (sD) plasma concentration-time curves for rosuvastatin (A), total ezetimibe (B), and free ezetimibe (C). [/fig]
[table] .: Drug Design, Development and Therapy 2018:12 submit your manuscript | wwwdovepresscom [/table]
[table] Table 1: Demographic characteristics of the study participants (n=24) [/table]
[table] Table 2: PK comparisons of rosuvastatin, total and free ezetimibe [/table]
[table] Table 3: summary of adverse events after administration of rosuvastatin and ezetimibe [/table]
|
Dynamics of the excised base release in thymine DNA glycosylase during DNA repair process
Thymine DNA glycosylase (TDG) initiates base excision repair by cleaving the N-glycosidic bond between the sugar and target base. After catalysis, the release of excised base is a requisite step to terminate the catalytic cycle and liberate the TDG for the following enzymatic reactions. However, an atomistic-level understanding of the dynamics of the product release process in TDG remains unknown. Here, by employing molecular dynamics simulations combined with the Markov State Model, we reveal the dynamics of the thymine release after the excision at microseconds timescale and all-atom resolution. We identify several key metastable states of the thymine and its dominant releasing pathway. Notably, after replacing the TDG residue Gly142 with tyrosine, the thymine release is delayed compared to the wild-type (wt) TDG, as supported by our potential of mean force (PMF) calculations. These findings warrant further experimental tests to potentially trap the excised base in the active site of TDG after the catalysis, which had been unsuccessful by previous attempts. Finally, we extended our studies to other TDG products, including the uracil, 5hmU, 5fC and 5caC bases in order to compare the product release for different targeting bases in the TDG-DNA complex.
# Introduction
Thymine DNA glycosylase (TDG) plays an essential role in correcting mismatched/damaged DNA base pair (bp) by cleaving the N-glycosidic bond between the sugar and target base through the base excision repair pathways [bib_ref] The enigmatic thymine DNA glycosylase, Cortázar [/bib_ref] [bib_ref] Initiation of base excision repair: glycosylase mechanism and structure, Mccullough [/bib_ref]. TDG specifically recognizes the double-stranded DNA containing U·G or T·G mismatches caused by the deamination of the cytosine or 5-methylcytosine (5mC), thus preventing the C→T mutation. More recently, TDG has also been found to be involved in an active DNA demethylation process by removing the oxidized products of the 5mC and 5-hydroxymethylcytosine (5hmC), namely the 5formylcytosine (5fC) and 5-carboxycytosine (5caC), respectively, which are generated by ten-eleven-translocation (Tet) proteins [bib_ref] Tet proteins can convert 5-Methylcytosine to 5-Formylcytosine and 5-Carboxylcytosine, Ito [/bib_ref] [bib_ref] Genome-wide analysis reveals TETand TDG-dependent 5-methylcytosine oxidation dynamics, Shen [/bib_ref] [bib_ref] TET-mediated active DNA demethylation: mechanism, function and beyond, Wu [/bib_ref] [bib_ref] Thymine DNA glycosylase can rapidly excise 5-formylcytosine and 5-carboxylcytosine: potential implications for..., Maiti [/bib_ref]. This epigenetic modification conducted by TDG has profound influences on maintaining the genome integrity and correctly programming the embryogenic development [bib_ref] Role of base excision "Repair" enzymes in erasing epigenetic marks from DNA, Drohat [/bib_ref] [bib_ref] Thymine DNA glycosylase is essential for active DNA demethylation by linked deamination-base..., Cortellino [/bib_ref] [bib_ref] Embryonic lethal phenotype reveals a function of TDG in maintaining epigenetic stability, Cortázar [/bib_ref] [bib_ref] Genome-wide profiling of 5-formylcytosine reveals its roles in epigenetic priming, Song [/bib_ref].
Despite of recognizing a broad range of substrates, TDG adopts a universal base-flipping mechanism to carry out the catalytic reaction, thereby the target bases can be flipped out of the DNA duplex to reach into the active site of TDG, as observed in several TDG-DNA complexes [bib_ref] Structural basis for excision of 5-formylcytosine by thymine DNA glycosylase, Pidugu [/bib_ref] [bib_ref] Thymine DNA glycosylase specifically recognizes 5-carboxylcytosine-modified DNA, Zhang [/bib_ref] [bib_ref] Structural basis of damage recognition by thymine DNA glycosylase: key roles for..., Coey [/bib_ref] [bib_ref] Lesion processing by a repair enzyme is severely curtailed by residues needed..., Maiti [/bib_ref]. For example, in a recent determined TDG-DNA complex (PDB ID: 5hf7) (16), a flipped dU analog is inserted into a binding pocket formed by several TDG residues, including Tyr152, Ala145, Pro153 and Asn191 that can form direct contacts with the uracil, and His151 and Asn140 that indirectly interact with the uracil through water molecules. Notably, the void room left by the flipped base is occupied by a key TDG residue Arg275 from an intercalation loop region (residues 270-281), which prevents the return of the flipped base back to the DNA duplex (see. Finally, after catalysis, the cleaved base is released from the active site, creating an abasic (AP) site within the DNA. Therefore, the product release is a requisite step to terminate the catalytic cycle and liberate the enzyme for subsequent reactions. Recently, several studies have been devoted to capture the ternary product complex for the TDG using crystallographic methods. have recently reported the crystal structures of the 5-hydroxymethyluracil (5hmU) and 5caC bound TDG-DNA complexes [bib_ref] Excision of 5-hydroxymethyluracil and 5-carboxylcytosine by the thymine DNA glycosylase domain: its..., Hashimoto [/bib_ref] [bib_ref] Activity and crystal structure of human thymine DNA glycosylase mutant N140A with..., Hashimoto [/bib_ref]. Nonetheless, Malik et al., using an improved crystalliza- The modeled ternary product complex. TDG is shown in gray and DNA chain is shown in orange cartoon. A zoomed-in view of the active site is provided on the right panel, with the thymine represented in green sticks, the key residues that form direct/indirect contacts with the thymine shown in gray sticks, and the AP site in orange sticks. The Pro-rich loop (residues Pro153 to Asn157) and the gating helix (residues Leu143 to Lys148) regions are shown in wheat and cyan, respectively. The intercalated residue Arg275 is highlighted in blue sticks. The references to color are only for the online version. tion condition, failed to trap any type of cleaved bases in TDG, suggesting that the product is prone to release from the active site right after the reaction and exhibits low binding affinity with the TDG-DNA complex [bib_ref] Thymine DNA glycosylase exhibits negligible affinity for nucleobases that it removes from..., Malik [/bib_ref]. Based on the high-resolution structures, Malik et al. also proposed a possible product releasing pathway along a solvent-filled channel [bib_ref] Thymine DNA glycosylase exhibits negligible affinity for nucleobases that it removes from..., Malik [/bib_ref].
Although the crystallographic studies have provided atomistic-level understandings of the structural features of the TDG-DNA complexes [bib_ref] Crystal structure of human thymine DNA glycosylase bound to DNA elucidates sequence-specific..., Maiti [/bib_ref] [bib_ref] Crystal structure of a G:T/U mismatch-specific DNA glycosylase: mismatch recognition by complementary-strand..., Barrett [/bib_ref] [bib_ref] Crystal structure of thymine DNA glycosylase conjugated to SUMO-1, Baba [/bib_ref] , the detailed dynamics of the product release process remains unknown, e.g. is it possible for the product to release through the proposed solvent-filled channel? How does the release process actually happen? In addition, since the ternary product complexes for the TDG system have not been determined yet, we wonder whether a TDG variant exists that can potentially trap the cleaved base in the active site after the catalysis. However, due to the transient dynamics, it is still challenging for the current experimental techniques to directly capture the product release process in TDG. Molecular dynamics (MD) simulation has become a powerful theoretical and computational tool to reveal the dynamic properties of the bio-molecules at an atomistic resolution [bib_ref] Molecular dynamics simulations of biomolecules, Karplus [/bib_ref]. Moreover, a Markov state model (MSM) can be constructed based on a large number of MD simulations to extract both thermodynamic and kinetic properties for certain conformational changes taken place at a relatively long timescale (i.e. microsecond or longer) that is inaccessible by performing a long unbiased MD simulation [bib_ref] MSMBuilder: statistical models for biomolecular dynamics, Harrigan [/bib_ref] [bib_ref] Markov state models of biomolecular conformational dynamics, Chodera [/bib_ref] [bib_ref] Markov models of molecular kinetics: generation and validation, Prinz [/bib_ref] [bib_ref] Markov state models provide insights into dynamic modulation of protein function, Shukla [/bib_ref] [bib_ref] Transition networks for modeling the kinetics of conformational change in macromolecules, Noé [/bib_ref] [bib_ref] Describing protein folding kinetics by molecular dynamics simulations. 1. Theory †, And [/bib_ref]. By discretizing the conformational phase space into many sub-states based on certain metrics, i.e. the root mean square deviation (RMSD), one can construct an MSM using a coarsegrained lag time Δt and then build a transition probability matrix T where each entry T ij represents the transition probability from state i to state j after a certain lag time Δt; the transition only depends on the current state i but not the preceding states (i.e. the markovian process). In this manner, one can investigate the comparatively long timescale dynamics by propagating the transition probability matrix to a given time of interest according to the following equation:
[formula] P (n t) = [T ( t)] n P (0) [/formula]
By diagonalizing the transition probability matrix T, the stationary distribution can be found as the right eigenvector, corresponding to the eigenvalue of 1. The entries of the vector correspond to the probabilities of each state in the stationary distribution. Other eigenvalues <1 can be used to derive the kinetic properties by calculating the implied timescales (see 'Materials and Methods' section 6 for more details). In recent years, the MSM method has been used to study many critical bio-molecular events, including protein folding, protein-ligand recognition and protein conformational dynamics in general [bib_ref] Molecular dynamics simulations reveal ligand-controlled positioning of a peripheral protein complex in..., Ryckbosch [/bib_ref] [bib_ref] Elucidating conformational dynamics of multi-body systems by constructing Markov State Models, Zhu [/bib_ref] [bib_ref] Multiensemble Markov models of molecular thermodynamics and kinetics, Wu [/bib_ref] [bib_ref] Bridge helix bending promotes RNA polymerase II backtracking through a critical and..., Da [/bib_ref] [bib_ref] Protein conformational plasticity and complex ligand-binding kinetics explored by atomistic simulations and..., Plattner [/bib_ref] [bib_ref] Allostery through the computational microscope: cAMP activation of a canonical signalling domain, Malmstrom [/bib_ref] [bib_ref] A jump-from-cavity pyrophosphate ion release assisted by a key lysine residue in..., Da [/bib_ref] [bib_ref] Millisecond dynamics of RNA polymerase II translocation at atomic resolution, Silva [/bib_ref] [bib_ref] Structure-guided simulations illuminate the mechanism of ATP transport through VDAC1, Choudhary [/bib_ref] [bib_ref] Atomistic folding simulations of the five-helix bundle protein (6−85), Bowman [/bib_ref] [bib_ref] Dynamics of an intrinsically disordered protein reveal metastable conformations that potentially seed..., Qiao [/bib_ref] [bib_ref] Cloud-based simulations on Google Exacycle reveal ligand-modulation of GPCR activation pathways, Kai [/bib_ref] [bib_ref] Dynamics of pyrophosphate ion release and its coupled trigger loop motion from..., Da [/bib_ref] [bib_ref] Rapid equilibrium sampling initiated from nonequilibrium data, Huang [/bib_ref] [bib_ref] Discovery of a regioselectivity switch in nitrating P450s guided by molecular dynamics..., Dodani [/bib_ref] [bib_ref] T7 RNA polymerase translocation is facilitated by a helix opening on the..., Da [/bib_ref].
In this work, we investigated the dynamics of the product release process in TDG at an atomistic resolution by constructing an MSM based on hundreds of MD simulations with an aggregated simulation time of ∼13 s. We explored the thymine release through a solvent-filled channel, which is flanked by a Pro-rich loop and a gating helix domain (residues Gly138 to His158, see. Our MSM reveals four metastable states (S1-S4) during the thymine release and predicts that the overall product release process takes place at a timescale of ∼10 s. Notably, our MD simulations captured a key intermediate state, which can be stabilized by a pre-existing TDG conformation, indicating a conformational selection mode of the binding between TDG and the product. Moreover, for comparative studies, we performed potential of mean force (PMF) calculations to reveal the relative releasing kinetic rates for several alternative bases (e.g. uracil, 5hmU, 5fC and 5caC) by employing the thymine as a reference system. Finally, we explored potential strategies to delay the product release from the active site of TDG, by replacing a glycine located in the releasing channel with a bulkier residue. Our work thus aims at revealing the structural basis of the product release in TDG and providing rational experimental design strategies.
# Materials and methods
## Modeling of the tdg-dna complexes bearing one g·t mismatched bp before and after the catalysis
We first built the TDG-DNA complex containing one G·T mispair before the cleavage based on one recently resolved crystal structure (PDB ID: 5hf7) [bib_ref] Structural basis of damage recognition by thymine DNA glycosylase: key roles for..., Coey [/bib_ref]. We replaced the Fluorine atom of the flipped 2'-fluoroarabino analogue of dU with H atom, and kept all the DNA chains containing 24 bps. Then, the dU nucleotide was modified to dT by adding one methyl-group on the 5-site of the pyrimidine ring. The missing side chains of several TDG residues were added. This final model of the TDG-DNA complex was subject to energy minimization using Gromacs-4.6 package to relieve the steric clashes (see Supplementary. We then used this minimized TDG-DNA structure to construct the ternary product complex by directly cleaving the N-glycosidic bond of the flipped dT nucleotide, followed by energy minimization (see Supplementary. The sugar conformation of the AP site after the nucleophilic attack by one water molecule was modeled as an ␣-form rather than a -form as suggested by previous studies [bib_ref] A mechanistic perspective on the chemistry of DNA repair glycosylases, Stivers [/bib_ref]. Although the -form can be an alternative conformation after the product release [bib_ref] Structural basis of damage recognition by thymine DNA glycosylase: key roles for..., Coey [/bib_ref] , an ␣-form is used here since we are focused on the overall product release process starting right after the catalytic reaction. The comparison of these above two minimized TDG-DNA complexes, before and after the catalysis, indicates subtle differences in the active site regions (see Supplementary. Finally, the minimized structure of the ternary product complex was used for the subsequent MD simulations.
## Setup of the md simulations and equilibration of the ternary product complex
All the MD simulations were performed using GROMACS 4.6 package [bib_ref] GROMACS 4: algorithms for highly efficient, load-balanced, and scalable molecular simulation, Hess [/bib_ref] [bib_ref] GROMACS: fast, flexible, and free, Van Der Spoel [/bib_ref] [bib_ref] GROMACS: a message-passing parallel molecular dynamics implementation, Berendsen [/bib_ref]. The ternary product complex was centered in a triclinic box with a minimal system and box wall distance set to 7.0Å. The whole system was solvated in a simple point-charge (SPC) water box, and in order to neutralize the system and ensure an ionic concentration of 0.15M, we added 93 Na + and 44 Cl − ions. The final system contains 51 935 atoms. We employed the AMBER99sb force field with PARMBSC0 corrections for the nucleic acid to describe the whole system [bib_ref] Single-stranded DNA within nanopores: conformational dynamics and implications for sequencing; a molecular..., Guy [/bib_ref] [bib_ref] Comparison of multiple Amber force fields and development of improved protein backbone..., Hornak [/bib_ref] [bib_ref] Determination of alkali and halide monovalent ion parameters for use in explicitly..., Joung [/bib_ref] [bib_ref] Molecular dynamics simulations of the dynamic and energetic properties of alkali and..., Joung [/bib_ref]. The amber force fields of the thymine and other TDG products, including the uracil, 5fC, 5hmU and 5caC, were generated using the Antechamber module implemented in the AmberTools package [bib_ref] Development and testing of a general amber force field, Wang [/bib_ref] , and the RESP charges were calculated after a full optimization of each base using the Hartree-Fork methods under the basis set of 6-31G* (see [fig_ref] Figure 2: The MSM reveals four metastable states for the thymine release in TDG [/fig_ref] for the complete charge parameters of all the above bases). A switched Van der Waals type was used with a cutoff of 12Å. We set a cutoff for the short-range electrostatic interactions to 12Å and treated the long-range electrostatic interactions using the Particle-Mesh Ewald sum-mation method [bib_ref] A smooth particle mesh Ewald method, Essmann [/bib_ref]. LINCS algorithm was used to constrain all the chemical bonds [bib_ref] LINCS: a linear constraint solver for molecular simulations, Hess [/bib_ref].
Starting from the above modeled ternary product complex, we first performed energy minimization using the steepest decent method, followed by a 500-ps restrained MD simulation by constraining all the heavy atoms of the system (protein, nucleic acid and the thymine). We then performed temperature annealing by increasing the temperature from 50 K to 310 K within 200 ps. Finally, we equilibrated the whole system by performing one unbiased 100-ns NVT MD simulation with a time step of 2 fs at 310 K using the velocity rescaling thermostat [bib_ref] Canonical sampling through velocity rescaling, Bussi [/bib_ref]. The final snapshot from the above MD trajectory was used for the subsequent pulling simulations.
## Performing steered md (smd) simulations to obtain the initial product release pathways
Next, in order to obtain an initial product releasing pathway, we employed Steered MD (SMD) to pull the thymine out of the TDG active site using the last snapshot of the above 100-ns MD trajectory. A close examination on the TDG structure reveals no alternative channel or pocket from the TDG surface for the product to release except for the solvent-filled channel as identified by former studies [bib_ref] Thymine DNA glycosylase exhibits negligible affinity for nucleobases that it removes from..., Malik [/bib_ref] (see Supplementary. To determine the dominant product releasing pathway, we conducted two sets of SMD simulations: with and without pre-defined pulling directions. For the first setup, we pulled the center of mass (COM) of the thymine along the abovementioned solventfilled channel toward two opposite directions using two DNA P atoms as the reference points, respectively, (denoted as. In contrast, the product can be readily released from the solvent-filled channel in all three parallel SMD simulations toward the ref a (see Supplementary. In addition, we designed a second SMD setup by only increasing the distance between the COM of the thymine and the C ␣ atoms of two TDG residues Asn191 and Ile134, respectively, without defining any pulling direction in order to automatically determine which releasing path the product would use (denoted as ref c in Supplementary. Consistent with the results from the first SMD setup, the product is unambiguously released through the solvent-filled channel in all above SMD simulations (see Supplementaryand F). Taken together, our results confirm that the TDG product is determined to release along the solvent-filled channel, as also suggested by former experimental studies [bib_ref] Thymine DNA glycosylase exhibits negligible affinity for nucleobases that it removes from..., Malik [/bib_ref].
Moreover, former quantum-mechanical/molecularmechanical studies have indicated that the protonated form of the His151 is critical for the catalytic reaction in TDG [bib_ref] Mechanism of the glycosidic bond cleavage of mismatched thymine in human thymine..., Kanaan [/bib_ref]. To evaluate the role of the His151 in product release process, we designed three different TDG systems with varied His151 forms, namely the ⑀-histidine form (HIE), the protonated form (HIP) and the ␦-histidine form (HID) (see Supplementary Figures S1, S4A and B). For each system, using the same pulling parameters, we then performed one set of SMD simulations to exam how different His151 forms might affect the product release process. As shown in Supplementary [fig_ref] Figure 4: Direct observations of the thymine release from the S3 state to the... [/fig_ref] , the results show that for all three systems, the product can be readily released from the active site within 20 ns, suggesting that different His151 forms have no substantial influences on the product release process. We thus used the HIE form of the His151 throughout this work. Next, we performed two additional SMD simulations along the ref a using two different pulling force constants, 0.1 and 0.2 kJ/mol/Å 2 . We finally performed geometric clustering for all the conformations derived from the SMD simulations along the solvent-filled channel and chose representative conformations as the starting structures to shoot more unbiased MD simulations.
## Geometric clustering of the smd conformations
Based on the initial thymine release pathways, we then performed geometric clustering by at first projecting all the high-dimensional SMD conformations onto a lowdimensional space using time-structure independent component analysis (tICA) [bib_ref] Identification of slow molecular order parameters for Markov model construction, Pérezhernández [/bib_ref] [bib_ref] Improvements in Markov State Model construction reveal many non-native interactions in the..., Pande [/bib_ref] [bib_ref] Slow dynamics of a protein backbone in molecular dynamics simulation revealed by..., Naritomi [/bib_ref] [bib_ref] Slow dynamics in protein fluctuations revealed by time-structure based independent component analysis, Naritomi [/bib_ref]. tICA can efficiently capture the slowest dynamics for a certain conformational change in a system by constructing a time-lagged correlation matrix. This information, however, is hard to be extracted using other related dimension-reduction techniques (e.g. the principle component analysis, or PCA). In recent years, tICA has been successively applied to study many dynamic processes when constructing the MSM for bio-molecules [bib_ref] MSMBuilder: statistical models for biomolecular dynamics, Harrigan [/bib_ref] [bib_ref] Markov state models provide insights into dynamic modulation of protein function, Shukla [/bib_ref] [bib_ref] Bridge helix bending promotes RNA polymerase II backtracking through a critical and..., Da [/bib_ref] [bib_ref] T7 RNA polymerase translocation is facilitated by a helix opening on the..., Da [/bib_ref]. Different metrics, i.e. distance, contact number etc, can be used as the order parameters for the tICA. In current work, we selected a total of 513 distance pairs between the heavy atoms of the thymine and several other TDG/DNA motifs (see Supplementary , specified as following: heavy atoms of thymine -C ␣ atoms of TDG residues K107 to L124. heavy atoms of thymine -C ␣ atoms of TDG residues G138 to H158. heavy atoms of thymine -3 sidechain heavy atoms of TDG residue N191. heavy atoms of thymine -7 sidechain heavy atoms of TDG residue Y152. heavy atoms of thymine -8 P atoms of DNA nucleotides out of the active site.
We then projected each SMD conformation onto the top four slowest tICs, followed by clustering the projected conformations into 100 classes using the K-centers algorithm implemented in the MSMbuilder package [bib_ref] MSMBuilder: statistical models for biomolecular dynamics, Harrigan [/bib_ref] [bib_ref] Using generalized ensemble simulations and Markov state models to identify conformational states, Bowman [/bib_ref]. Finally, from each cluster, the center conformation was selected as the representative structure for the subsequent unbiased MD simulations.
## Seeding unbiased md simulations
In addition to the above 100 representative conformations, we selected 25 additional conformations in which the thymine located near to the exit of the release channel in order to ensure enough transitions could be observed. Then, the above chosen 125 conformations were directly subject to the unbiased 100-ns NVT MD simulation using the same MD setup as used to equilibrate the initial ternary product complex. Finally, we collected a total of 125 MD trajectories (100 ns each) with an aggregated simulation time of 12.5 s and used this final simulation dataset to construct the MSM.
## Construction and validation of the msm
We employed a splitting and lumping procedure to construct the MSM. We first did geometric clustering for all the conformations from the above 125 100-ns MD simulations (in total of 1 250 000 MD snapshots) into different number of microstates by performing tICA projection followed by K-centers clustering using the same distance metric as described before (see Supplementary . Then, to dissect the detailed thymine release mechanism in TDG, we further lumped a 500-state MSM into four macrostates using the PCCA+ method (67), detailed as below:
Clustering MD conformations into different number of microstates. To evaluate the effects of microstate number and the correlation lag-time on the stability of the MSM, we performed tICA projections by constructing the timelagged correlation matrix using three different correlation lag-times: 20, 30 and 40 ns, and for each case, we clustered the MD conformations into 500, 600, 700 and 800 states, respectively. Then, for each clustering setup (in total of 3 × 4 different clustering), we constructed the transition count matrix C in which each entry C ij was obtained by counting the transition numbers for each microstate pair (i and j) from the unbiased MD simulations. This raw count matrix was then symmetrized and converted to a transition probability matrix T by normalizing each entry by the sum of the row.
Next, we calculated the eigen-function of the TPM from which the kinetic properties were obtained by calculating the implied timescales according to the following equation:
[formula] τ k = − τ ln μ k (τ ) [/formula]
where, μ k is the k-th eigenvalue of the transition probability matrix T with lag time , each implied timescale curve indicates the average transition time between two sets of conformations. Therefore, we can readily estimate at what timescale the slowest transition, in most cases also the principal conformational change of interest, might take place. For each clustering, the slowest implied timescale curves are all converged at about tens of microseconds (see Supplementary [fig_ref] Figure 6: The thymine release in the G142Y mutant TDG is delayed in comparison... [/fig_ref] , suggesting our MSM is robust to the different choices of either the microstate number or correlation lag-times.
To reveal the metastable states involved in the thymine release process, we chose the 500-state MSM at the lag-time of 20 ns as a representative model and further lumped the 500 microstates into four macrostates using the Robust Perron Cluster Cluster Analysis (PCCA+) method implemented in MSMbuilder 2.7 package [bib_ref] Robust Perron cluster analysis in conformation dynamics, Deuflhard [/bib_ref].
Validation of the MSM. To exam if our MD simulations are sufficient to build a reliable MSM, we sub-sampled the original MD simulations by choosing subsets of the conformations from each MD trajectory, with each subset data containing an aggregated simulations time of (125 × 50 ns), (125 × 60 ns), (125 × 70 ns), (125 × 80 ns), (125 × 90 ns) and (125 × 100 ns), respectively. Then for each subset data, we used a correlation lag-time of 20 ns for the tICA projection and clustered the corresponding MD conformations into 500 states. We then projected each set of conformations onto the same top two tIC vectors as that for the complete dataset (see Supplementary . The results show a clear convergence of the MD samplings along the two slowest tICs, suggesting that our MD simulations are fairly sufficient to build a reliable MSM. Next, to exam if the kinetics (i.e. the timescale at which the slowest event takes place) would be changed or not, we constructed the MSM for each subset data and calculated the implied timescales with the errors estimated by bootstrapping the trajectory-list 100 times. The results clearly show that, for each subset data, the slowest transition times are consistently converged to about tens of microseconds (see Supplementary , indicating that the MSM is reliable to predict reasonable kinetic properties. To further validate our MSM, we performed the Chapman-Kolmogorov test for several microstates that are the most populated from our MD simulations. The results show that the evolution of the residence probability of the microstates observed from the raw MD simulations data is in good agreement with the values predicted by the MSM (see Supplementary .
## Calculations of the mean first massage time (mfpt) and stationary distributions.
We finally calculated the mean first massage time (MFPT) for each transition and the equilibrium population for each state. To estimate the corresponding errors, we generated 100 trajectory lists with each containing 125 randomly chosen trajectories from the original MD simulations. Then for each trajectory list, we constructed the corresponding TPM from which a 10-ms long Monte Carlo (MC) simulation was then generated. This MC trajectory is so long that the stationary distribution for each state and the MFPT value for each transition can be readily obtained. Finally, the mean and standard deviation were calculated by averaging over the results from the above 100 bootstrapping.
## Setup of the pulling simulations for the pmf calculations
To evaluate the relative kinetic rates of the product release between the varied systems (including the uracil, 5fC, 5caC, 5hmU in wild-type (wt) TDG, and the thymine in G142Y mutant TDG) and the thymine release in wt TDG that is served as a reference system, we carried out the PMF calculations for each of the product release processes along the dominant releasing path. The PMF shows the free energy changes along a certain reaction coordinate (RC). Comparisons of the free energy landscapes between different systems can thus provide the relative kinetic rates for certain transitions. Here, we chose the distance between the COM of each product and one backbone P atom of the DNA chain as the RC for the PMF calculations (defined as d rc ), which is the same reference point we used for the initial SMD simulations (see Supplementary. We then performed a number of constant velocity SMD (cv-SMD) simulations for each of the above six systems to calculate the PMF for the product release along the d rc using the Jarzynski's equality that evaluates the free energy difference between two states from the work done through many realizations of the process connecting the two states (68):
[formula] e −β[G(x t )−G(x 0 )] = e −βW 0 [/formula]
where,  = 1/k B T (k B is the Boltzmann constant, and T is the temperature in kelvin), x denotes a time-dependent RC evolving from an initial value of x 0 at time t = 0 to x t = x(t) at a certain time t, and the angular bracket averages over an ensemble of trajectories starting from an equilibrated distribution of the initial state and arriving to the final state corresponding to x t . Employing the same protocol as described by previous simulation studies using the Jarzynski method (69,70), we divided the product releasing channel into two sections, with the d rc value, serving as the RC x here, decreased from ∼20Å to 15Å (section I) and then from 15 A to 9Å (section II). Notably, these two sections encompass the S1→S2 and S2→S3 transitions for the thymine release in the wt TDG, respectively, as seen from the calculated d rc value based on our MSM (the mean d rc value is ∼16 A for S2 and ∼9Å for S3, see Supplementaryand the main text for the results of the MSM).
We, at first constructed the ternary product complexes for each product based on the thymine system. For each case, we then performed one 100-ns MD simulations to fully relax the structure and the final MD snapshots were used as the input structures for the cv-SMD simulations. Then, for each of the six system, namely the thymine (in both wt and G142Y mutant TDG), uracil, 5fC, 5caC and 5hmU systems, we performed eight 5-ns cv-SMD simulations for each section (8 × 2 × 6 = 96 simulations in total, which amounts to ∼0.5 s simulation time in aggregation for the PMF calculations), with a pulling rate v = ∼1Å/ns and a force constant k = 30 kcal/mol/Å. We applied the force on the COM of the product and calculated the external work using the following equation:
[formula] w(t) = t 0 v f (τ )dτ [/formula]
where, the f(τ ) is the applied external force at time τ . We finally resorted to the second order cumulant expansion of the external work to calculate the free energy changes [bib_ref] Energetics of glycerol conduction through aquaglyceroporin GlpF, Jensen [/bib_ref] [bib_ref] Mechanism of gating and ion conductivity of a possible tetrameric pore in..., Yu [/bib_ref] :
[formula] G(x t ) = W(t) 0 − β 2 [ W(t) 2 0 − W(t) 0 2 ] where, we define W(t)=w(t) − k(x t − x 0 − vt) 2 /2. [/formula]
## Generating a base-flipping pathway for the g142y mutant tdg-dna complex using targeted md (tmd) simulations
To exam whether the G142Y substitution might potentially affect the base-flipping process, we performed the TMD simulation to obtain a base-flipping pathway for the dT nucleotide. To achieve this, we at first built one interrogating complex and one recognition complex for the TDG-DNA system in which the mispaired dT was in an intrahe-lical and an extrahelical form, respectively. The above modeled TDG-DNA complex prior to the catalysis serves as the model of recognition complex (see Supplementary. To model the interrogating complex, we employed the double-stranded DNA chains from the above recognition complex and constructed an intrahelical T·G-mispair containing DNA duplex (see the Supplementaryfor the energy minimized DNA structure). We then relaxed the DNA backbones by performed one 7 ns MD simulations and the last snapshot was used for modeling the final TDG-DNA binding complex (see the Supplementary. Notably, we adopted the TDG conformation that forms non-specific complex with the DNA duplex where the intercalated residue Arg275 is lying along the minor groove (PDB ID: 2rba) [bib_ref] Crystal structure of human thymine DNA glycosylase bound to DNA elucidates sequence-specific..., Maiti [/bib_ref]. This model was subjected to the energy minimization and served as the final model of interrogating complex. Finally, we introduced the G142Y mutation to both the interrogating complex and recognition complex (see the Supplementaryand B), and performed TMD simulations to obtain the base-flipping pathway by targeting the interrogating complex to recognition complex and constraining the backbone P atoms of two DNA ends using a force constant of 500 kcal/mol/Å 2 (see Supplementary. Several discontinuous regions were selected as the targeting regions using a pulling force constant of 5 kcal/mol/Å 2 , including 16 P atoms of several DNA nucleotides locating at the middle part of the DNA chain; the heavy atoms of the mispaired dT nt and one of its adjacent nucleotide; the heavy atoms of the intercalated residue Arg275; the C ␣ atoms of the TDG residue from Cys276 to Glu303 (see Supplementary. The amber force field 99SB with the parambsr0 correction was used to describe the TDG-DNA system.
# Results
## Structural features of the ternary product complex
We built the ternary product complex based on the crystal structure of the TDG-DNA complex prior to the catalysis (PDB ID: 5hf7) [bib_ref] Structural basis of damage recognition by thymine DNA glycosylase: key roles for..., Coey [/bib_ref] , which was at first subject to an energy minimization (see Supplementary. Then, by directly cleaving the N-glycosidic bond between the sugar and the base, we obtained a locally stabilized ternary product complex (see, Supplementaryand 'Materials and Methods' section 1 for more details of the model construction). In this structure, the thymine can form direct packing interactions with the residue Tyr152, while the same interaction can also be found before the catalysis (see Supplementary. The 5-methyl group of the thymine locates in a hydrophobic pocket formed by residues Pro153 and Ala145 (see. In addition, the thymine O2 atom forms a hydrogen bond (HB) with the Ile139 backbone N-H; the N3 and O4 atoms can form two HBs with the side chain of residue Asn191 by expelling one crystal water molecule as observed in the crystal structure (PDB ID: 5hf7); the O4 atom can also indirectly interact with the His151 through one water molecule. Moreover, we treated the sugar group at the AP site, after a nucleophilic attack by a water molecule, as an ␣-form rather than a from, as implicated by previous studies. Taken together, in general the thymine can retain all the major interactions with TDG before and after the catalysis (see Supplementary. This modeled ternary product complex was finally used for investigating the product release process.
## Msm reveals four metastable states during the thymine release in tdg
To determine the dominant product releasing pathway, we have performed two sets of SMD simulations starting from one well-equilibrated ternary product complex. We are able to confirm that the solvent-filled channel flanked by the Prorich loop (residues Pro153 to Asn157) and the gating helix (residues Leu143 to Lys148) region is the major product release pathway (see 'Materials and Methods' section 3 for more details). Four metastable states have been clearly identified from the MSM during the thymine release in TDG, namely S1-S4 states (see [fig_ref] Figure 2: The MSM reveals four metastable states for the thymine release in TDG [/fig_ref]. At a glimpse, the product passes through the releasing channel by at first visiting the S1 and S2 states (see [fig_ref] Figure 2: The MSM reveals four metastable states for the thymine release in TDG [/fig_ref]. Then the release pathway is promptly broadened up so that the thymine can be directly released to the bulk water regions. Accordingly, the S3 and S4 states can be viewed as the released states. Specifically, in the S1 state, the most populated state among the four, the thymine remains stable in the active site and forms a similar interacting network with the neighboring residues or water molecules as observed in the initial ternary product complex, except that two additional waters enter into the active site and form two extra HBs with the N1 and O2 atoms of the thymine (compare. Then, the S1 state can transit to S2, the least stable state. In S2, the thymine loses its interactions with the active site residues Tyr152, Ile139 and Asn191, while it is stuck within a region surrounded by the Pro-rich loop, the gating helix, and the DNA backbone. The thymine can actually form two HBs with the Met144 and Ala145 backbone N-H, one HB with the P-O − group at the AP site, and can also maintain hydrophobic contacts with the side chains of residues Pro153, Phe111 and Met144 (see [fig_ref] Figure 2: The MSM reveals four metastable states for the thymine release in TDG [/fig_ref]. Notably, several water molecules can be found in the active site, occupying the void region left by the released thymine and forming water chains connecting the thymine and several active site residues, e.g. His151 and Asn191 (see [fig_ref] Figure 2: The MSM reveals four metastable states for the thymine release in TDG [/fig_ref]. This water-mediated interacting network greatly compensates the energy cost after the thymine leaves the active site. From S2, the thymine can directly transit to either the S3 or the S4 state, while in the S3 state, the thymine has already reached at the interface between TDG and the bulk region, though in some of the configurations, the thymine can still form direct contacts with either the DNA chains or the Nterminal region of TDG (see [fig_ref] Figure 2: The MSM reveals four metastable states for the thymine release in TDG [/fig_ref]. In one representative structure shown in [fig_ref] Figure 2: The MSM reveals four metastable states for the thymine release in TDG [/fig_ref] where the thymine locates at the exit of the release channel, the N4 atom of the thymine can form one HB with the backbone oxygen of Pro153, and the other polar atoms are fully occupied by forming HBs with the solvent waters. In S4, however, the thymine is fully exposed to the solvents (see Supplementary.
Our MSM predicts that the overall product release process in TDG occurs at a timescale of ∼10 s, with the S1→S2 and S2→S3/S4 transitions taken place at compa- . The four-state kinetic network derived from the MSM. The size of each circle is roughly proportional to the corresponding equilibrium population: 42.2 ± 5.0% (S1); 0.8 ± 0.1% (S2); 28.4 ± 3.7% (S3); 28.6 ± 3.8% (S4). The MFPT for the forward transition is also provided above each arrow. (C-E). Representative conformations for the S1 (C), S2 (D) and S3 (E) states. The structure was randomly selected from the most populated microstate for each macrostate. Key residues and water molecules that interact with the thymine are shown in sticks, and the HBs between the thymine and neighboring molecules are highlighted with dashed lines. Refer tofor other representations. rable kinetic rates (see [fig_ref] Figure 2: The MSM reveals four metastable states for the thymine release in TDG [/fig_ref]. As noted before, the S1→S2 transition results in impairing several HBs between the thymine and active site residues, i.e. Asn191, Ile139 and Tyr152. On the other hand, the contribution of the solvent waters to the stability of the thymine in the S2 state remains constant because the number of water molecules surrounding the thymine are comparable for the S1 and S2 states (see, which suggests that the thymine in S2 tends to return back to the active site due to its unstable features. Notably, the S2→S3 transition greatly increases the solvent accessible areas of the thymine owning to the sudden widening of the exit channel (see [fig_ref] Figure 2: The MSM reveals four metastable states for the thymine release in TDG [/fig_ref] , reflected from a significant increase of the surrounding water molecules in the S3 and S4 states (see. This solvation process significantly compensates for the energy loss between the thymine and TDG residues, i.e. Pro153, Phe111 and Met144. Finally, we found that as long as the thymine reaches to the S3 state, the subsequent release process becomes relatively fast, taking place at about hundreds of nanoseconds. Interestingly, from some of the unbiased MD simulations, we indeed observed direct transitions of the thymine from S3 to the bulk region (see next section).
Consistently, the free energy landscape of the MD conformations along the two slowest tICs clearly shows that the first tIC corresponds to the S1→S2→S3 or S1→S2→S4 transitions that take place at a timescale of ∼10 s, according to the MFPT calculations (see [fig_ref] Figure 2: The MSM reveals four metastable states for the thymine release in TDG [/fig_ref]. We then calculated the correlation coefficients between the first tIC and each of the order parameters used for constructing the MSM, consisting of 513 distance pairs. We pinpoint the exact distance pair that has the largest anti-correlation coefficients with the first tIC, i.e. the distance between the C ␣ atom of the TDG residue Gly138 and the O2 atom of the thymine (see Supplementary , which reflects how much the product has been released away from the active site. On the other hand, the second tIC is mainly attributed to the S3→S4 transition, which occurs at about hundreds of nanosecond (see.
It is also worth to note that although the S3 and S4 states can be further lumped into a single state to form a threestate MSM due to their comparatively fast transitions (compare Supplementaryand B), they are biologically distinguishable. First, the S3 state is much closer to the interface between the narrow channel and the solvents compared to the S4 state (see [fig_ref] Figure 2: The MSM reveals four metastable states for the thymine release in TDG [/fig_ref] , hence, there is a higher probability of the product to form direct contacts with TDG in S3. This idea is supported by the observation that the average distance between the thymine and the active site residue Asn191 is apparently smaller in S3 than that in S4 (see Supplementary. Moreover, the probability to fully solvate the thymine in S3, i.e. with 5, 6 or 7 surrounding waters, is noticeably lower than that in S4 (see, further suggesting that the product experiences different biological environments between S3 and S4. Accordingly, to illustrate a complete process of the product release in TDG, we adopted the four-state MSM.
## Direct observation of the thymine release from the s3 to solvents within 100-ns md simulations
After arriving at the S3 state (see [fig_ref] Figure 2: The MSM reveals four metastable states for the thymine release in TDG [/fig_ref] , the thymine can readily release to the solvents, which has been directly observed from our 41 independent 100-ns MD simulations (see [fig_ref] Figure 4: Direct observations of the thymine release from the S3 state to the... [/fig_ref]. According to the direction of the thymine release, we classify these 41 MD trajectories into two groups, with each representing one thymine release pathway (path . Particularly, in path A, captured in 28 out of 41 MD trajectories, the thymine releases 'upward' in the viewpoint represented in [fig_ref] Figure 4: Direct observations of the thymine release from the S3 state to the... [/fig_ref] , either above (red and blue paths in [fig_ref] Figure 4: Direct observations of the thymine release from the S3 state to the... [/fig_ref] or even in the perpendicular direction (green, cyan and purple paths in [fig_ref] Figure 4: Direct observations of the thymine release from the S3 state to the... [/fig_ref] to the extending direction of the DNA chain. On the other hand, in path B, observed in 13 MD trajectories, the thymine releases 'downward', either along the minor groove direction of the DNA chain (green and purple paths in [fig_ref] Figure 4: Direct observations of the thymine release from the S3 state to the... [/fig_ref] or by interacting with the N-terminus domain of TDG (cyan, red and blue paths in [fig_ref] Figure 4: Direct observations of the thymine release from the S3 state to the... [/fig_ref]. Notably, in both pathways, prior to releasing to the solvents, the thymine can visit a region surrounded by several TDG residues, i.e. Arg110, Phe111, Met144 and Pro155, and the minor groove of the DNA chain (see [fig_ref] Figure 2: The MSM reveals four metastable states for the thymine release in TDG [/fig_ref] and black dashed circles in [fig_ref] Figure 4: Direct observations of the thymine release from the S3 state to the... [/fig_ref]. After diffusing around the abovementioned region at a nanosecond timescale, the thymine can finally release to the solvents through one of the above pathways. In summary, as long as the product arrives at the exit of the releasing channel, the subsequent releasing process takes place almost immediately.
## The relative movements between tdg and dna chains facilitate the product release in a conformational selection mode
Our MSM indicates that the S2 state is a key intermediate state before the thymine releases to the solvents where several residues, i.e. Pro153 from the Pro-rich loop, along with Met144 and Ala145 from the gating helix region, can form direct contacts with the thymine (see [fig_ref] Figure 2: The MSM reveals four metastable states for the thymine release in TDG [/fig_ref]. To see whether TDG or DNA underwent any structural changes to stabilize the thymine in the S2 state, we measured the relative distance between the Pro-rich loop and the gating helix domain for each macrostate. Notably, in S2, the distance between the COM of the gating helix and that of the Pro-rich loop region (denoted as d1 in lies predominately around the high-value range (∼12Å) compared to other three states, in which the d1 largely remains at a low value (∼9Å) (see , C, D and E). The significant variation of d1 suggests that the above two TDG domains undergo an opening motion when the thymine reaches to the S2 state. Moreover, to examine the relative movements between TDG and the DNA chains, we calculated the distance between the gating helix domain and the P atom at the AP site on the DNA (denoted as d2 in . Again, the gating helix region shows notable movements relative to the DNA chain in the S2 state (d2 ∼ 12Å), higher than that in the other states (d2 ∼10Å; see , C, D and F).
The above results indicate that the opening motion of the gating helix and the Pro-rich loop in TDG, coupled with the relative movements between the gating helix and the DNA chain, can potentially enlarge the releasing channel, which in turn facilitates the thymine transiting to the S2 state after it leaves the active site. It is also worth to note that in the S1 state, even though the gating helix region is predom- . The dynamics of the Pro-rich loop and the gating helix regions play key roles in stabilizing the thymine in the S2 state via a conformational selection mechanism. (A-D) Structure highlights on the Pro-rich loop and the gating helix regions for each metastable state from the S1 to S4 states, respectively. The same representative structures from [fig_ref] Figure 2: The MSM reveals four metastable states for the thymine release in TDG [/fig_ref] were used for the analyses. For each state, the starting minimized TDG-DNA complex (in transparent) was superimposed as a reference. (E and F) Projection of the MD conformations onto two RCs: one is the first tIC, and the other one is the distance d1 (denoted in A) between the COMs of the gating helix and the Pro-rich loop regions (E), or the distance d2 (also denoted in A) between the COM of the Pro-rich loop and the P atom at the AP site (F), with the location of each metastable state labeled on the plot. inately in the closed state, it is also able to explore the openstate configurations, i.e. with both d1 and d2 values reaching to ∼12Å, comparable to those in the S2 state (see and F). These findings suggest that the above opening motion is an intrinsic property of TDG, thereby transition of the thymine from S1 to S2 can selectively stabilize the transiently accessed open-state of the gating helix, implying a conformational-selection recognition mode. Since the S2 state is the requisite state for the thymine to finally release to the solvents, the above structural re-arrangement of the TDG domains and the DNA chain becomes a key regulator for the product release.
## The product release in the g142y mutant tdg is delayed compared to the wt
Our model suggests that the thymine can release through a narrow, solvent-filled channel (see [fig_ref] Figure 2: The MSM reveals four metastable states for the thymine release in TDG [/fig_ref]. It is expected, therefore, that if the release channel is blocked by certain residues with bulky side chains, the product release process would be inhibited, so that one can potentially trap the product in the active site. Following this idea, we attempted to design a mutant TDG with bulkier residues lying within the product release channel. Notably, in the S2 state, Pro153 and Gly142 are relatively close to the thymine compared to other TDG residues (see [fig_ref] Figure 6: The thymine release in the G142Y mutant TDG is delayed in comparison... [/fig_ref]. Owing to a smaller residue size of glycine, we thus pinpoint Gly142 as a potential candidate for further mutagenesis studies. In addition, in Uracil-DNA glycosylase (UDG), which is structurally related to TDG, the counterpart residue of Gly142 is a tyrosine (1), and people have successively captured one excised uracil in the active site of UDG in the crystal form [bib_ref] Uracil-DNA glycosylase-DNA substrate and product structures: conformational strain promotes catalytic efficiency by..., Parikh [/bib_ref]. We therefore constructed a G142Y mutant TDG system for further studies (see [fig_ref] Figure 6: The thymine release in the G142Y mutant TDG is delayed in comparison... [/fig_ref].
To evaluate the relative kinetic rates of the thymine release between the wt and the G142Y systems, we carried out the PMF calculations for the thymine release along the solvent-filled channel in both systems (refer to the 'Materials and Methods' section 7 for the details of the PMF setup). We chose the distance between the COM of the thymine and one backbone P atom of the DNA chain as the RC for the PMF calculations (defined as d rc ). As shown in [fig_ref] Figure 6: The thymine release in the G142Y mutant TDG is delayed in comparison... [/fig_ref] , for the wt system, we observe two transition events, with the d rc value decreased from ∼20Å to ∼15Å and from ∼15Å to ∼10Å (see [fig_ref] Figure 6: The thymine release in the G142Y mutant TDG is delayed in comparison... [/fig_ref] , which correspond to the S1→S2 and S2→S3 transitions, respectively, as observed from our MSM (compare [fig_ref] Figure 6: The thymine release in the G142Y mutant TDG is delayed in comparison... [/fig_ref] and Supplementary. Moreover, the intermediate state identified from the PMF landscape demonstrates very similar structural features to the S2 state (see [fig_ref] Figure 6: The thymine release in the G142Y mutant TDG is delayed in comparison... [/fig_ref]. That is, the thymine is trapped between the Pro-rich region and the gating helix region by forming two HBs with the backbone N-H of Met144 and Ala145, one HB with the P-O − group at the AP site, and forming nonpolar contacts with Pro153. The consistencies between the MSM and the PMF results suggest that our PMF calculations are capable of providing a reasonable free energy profile for the product release along the solvent-filled channel.
Compared to the wt system, the free energy barrier for the S1→S2 transition in the G142Y system is increased by ∼7.3 k B T, which corresponds to ∼1000-fold ( G # ≈ e −7.3 ) decrease in the transition rate (see [fig_ref] Figure 6: The thymine release in the G142Y mutant TDG is delayed in comparison... [/fig_ref] , suggesting that the G142Y substitution has profound influences on the product release process. The captured intermediate state shows that the Tyr142 has to rotate away from the releasing channel in order to allow the product release, and formsstacking with the thymine (see [fig_ref] Figure 6: The thymine release in the G142Y mutant TDG is delayed in comparison... [/fig_ref]. In addition, for the S2→S3 transition, the barrier heights are similar to each other for these two systems, suggesting that after arriving at the S2 state, the subsequent kinetic rates would be similar for both systems. Taken together, the PMF calculations suggest that the G142Y mutation in TDG can delay the product release in comparison with the wt. shown to clamp the thymine in the S2 state. For the S1 and S2 states, the average distance between the COM of the thymine and the C ␣ atom of each residue from Gly138 to His158 was calculated. The lighter blue color indicates a shorter distance. (B) The PMF curves along the d rc for the wt (in blue) and G142Y mutant (in orange) systems. The barrier height changes of the mutant system in reference to the wt TDG system for the S1→S2 and S2→S3 transitions (denoted as Δh [bib_ref] Embryonic lethal phenotype reveals a function of TDG in maintaining epigenetic stability, Cortázar [/bib_ref] and Δh 23 , respectively) are ∼7 k B T and 1 k B T, respectively. In the background, the mean and fluctuation of the d rc values calculated based on the MSM are represented by colored boxes for S1 in green, S2 in magenta and S3 in blue (see the right panel in Supplementary
## Comparisons of product release for different bases in tdg
In addition to the G·T mismatch, TDG also targets on other types of damaged DNA nucleotides, such as G·U, G·5hmU, G·5fC and G·5caC base pairs [bib_ref] Divergent mechanisms for enzymatic excision of 5-formylcytosine and 5-carboxylcytosine from DNA, Maiti [/bib_ref] [bib_ref] Excision of 5-halogenated uracils by human thymine DNA glycosylase. Robust activity for..., Morgan [/bib_ref]. To estimate the releasing kinetics for the above four nucleotides, we performed the PMF calculations for each base along the same releasing path (d rc ). We, at first constructed the ternary product complex for each base system based on the original thymine system. For each complex, we then performed one 100-ns MD simulations to fully relax the structure and the final snapshots were used as the input structures for the pulling simulations. As shown in , for both uracil and 5hmU systems, the product forms very similar interaction networks with the TDG residues, as observed for the thymine, by forming HBs with Asn191 and Ile139. On the other hand, the formyl oxygen of 5fC and carboxyl-group of 5caC can form an additional HB with the Tyr152 backbone N-H, and form nonpolar contacts with the methyl group of Ala145. Moreover, owing to the cytosine ring, several key interactions present for the thymine, uracil, and 5hmU systems, are partially (i.e. Asn191) or completely destroyed (i.e. His151) in the 5fC and 5caC systems (see . Finally, for each of the four ternary product complexes, we employed the same pulling protocol and parameters as used for the thymine system to calculate the PMF.
The PMF landscapes reveal two major transition events for the uracil, 5fC, and 5caC as observed for the thymine, however, only one dominant transition is observed for the 5hmU (see . In specific, during the S1→S2 tran-sition, both the uracil and 5caC demonstrate lower transition barriers than that for the thymine (Δh 12 < 0), suggesting faster releasing rates for these two products (see . The 5fC, however, exhibits a similar barrier height to the thymine (see . On the other hand, for the S2→S3 transition, the 5fC and 5caC show comparable barrier heights with the thymine, but the uracil demonstrate an even lower transition barrier. Therefore, in summary, comparing to the thymine, both the uracil and 5caC have faster releasing rates while the 5fC exhibits a similar releasing rate. Furthermore, from the structural perspectives, the observed intermediate state that is equivalent to the S2 state of the thymine shows that the uracil can form two HBs with the G142 backbone N-H and the P-O − at the AP site, demonstrating distinct structural features from the S2 state of the thymine (see . In contrast, for the 5fC and 5caC, the intermediate state can form one HB with the Ala145 backbone N-H via the formyl group in 5fC and the carboxyl group in 5caC. In addition, owning to the negative charge of the 5caC, two water molecules can form HBs with the carboxyl group of the base and the 5fC can form one HB with the P-O − at the AP site (see .
Interestingly, the 5hmU demonstrates quite different dynamic properties compared to all other bases. Remarkably, we identified an alternative binding site of the 5hmU in TDG at d rc = ∼18Å comparing to its initial location at d rc = ∼20Å (marked as blue star in . In this structure, the HBs between Asn191 and 5hmU that are present in the initial ternary model are switched to the backbone atoms of the residues Asn140 and Gly142 (compare . (A) Initial structures of four ternary product complexes for the uracil, 5hmU, 5fC and 5caC. These conformations are the input structures for the PMF calculations. The excised base is shown in green stick; key HBs are highlighted with dashed lines. (B) Comparisons of the PMF landscapes between the thymine (in blue) and each of the four altered base systems (in orange) along d rc . The relative barrier heights for the S1→S2 transition in the uracil, 5hmU, 5fC and 5caC systems, comparing to the thymine system (Δh 12 ), are ∼−5 k B T, −2 k B T, −1 k B T and −3 k B T, respectively. The relative barrier heights for the S2→S3 transition (Δh 23 ) in the uracil, 5fC and 5caC systems are around ∼-5 k B T, −1 k B T and −1 k B T, respectively. The mean and fluctuation of the d rc values calculated based on the MSM are represented with colored boxes for S1 in green, S2 in magenta and S3 in blue. The location of the key intermediate state is marked by colored stars. (C) The key intermediate states for each system (marked as colored stars in B). The products are shown in purple sticks for uracil, 5fC and 5caC but green sticks for 5hmU. The key HBs are indicated with dashed lines. 7A and C), and the 5-hydroxylmethly group forms one HB with the P-O − at the AP site. Furthermore, different from other four bases, only one transition event was captured for the 5hmU as d rc decreased from ∼18Å to ∼10Å, with a transition barrier of similar height to that of the S1→S2 transition for the thymine.
In conclusion, the four different bases exhibit either a faster (e.g. uracil, 5caC and 5hmU) or a similar (e.g. 5fC) product releasing rate comparing to the thymine. Moreover, different bases share a common releasing mechanism by which the base can interact with both the Pro-rich loop and the gating helix regions. However, owing to their unique chemical structures, the specific contacting points vary for different bases.
# Discussion
Close examination of the TDG-DNA complex reveals only one most likely product release pathway filled with solvent waters as observed in the recent crystal structure [bib_ref] Structural basis of damage recognition by thymine DNA glycosylase: key roles for..., Coey [/bib_ref]. Previous experimental studies have suggested that the cleaved bases exhibit very weak binding affinities with TDG, and the following releasing process takes place very fast. Here, we employed extensive MD simulations (with an aggregated simulation time of ∼13 s) combined with the MSM construction to reveal the dynamics of the product release in TDG at an atomistic resolution. We identify four metastable states during the thymine release in TDG (S1-S4 states), with S1 termed as the bound state, S3 and S4 as the released states. Essentially, the S2 state is a key intermediate state that can be transiently stabilized by the Pro-rich loop and the gating helix regions of TDG and several water molecules. As long as the thymine reaches to the S3 state, the subsequent release process takes place very fast and has been directly captured from our unbiased 100-ns MD simulations. Our MSM suggests that the overall release process of the thymine in TDG occurs at ∼10 s. Given that TDG belongs to the family II of the UDG superfamily, similar product release pathways can be readily deduced for other UNG members that are structurally related to TDG.
## Energetics and role of solvent waters during the thymine release
As shown in [fig_ref] Figure 2: The MSM reveals four metastable states for the thymine release in TDG [/fig_ref] , the releasing pathway of the thymine is a very narrow, solvent-filled channel. The transition of the thymine from the S1 to S2 state is an enthalpy unfavorable process because it needs to break several interactions with the TDG residues in the active site, i.e. Asn191, Ile139, His151 and Tyr152. On the other hand, the number of the water molecules surrounding the thymine is comparable in the S1 and S2 states (see. Therefore, the S2 state is energetically less stable than the S1 state. Further release of the thymine from S2 to S3 greatly increases the solvent accessible area of the thymine and breaks most of its interactions with the TDG residues. Notably, the prompt increase of the entropy plays an important role in promoting the product release from the active site to the solvents, as reflected from the second tIC shown in. After arriving at the S3 state, the solvation become the main driving force to control the dynamics of the thymine by forming up to six HBs with the polar groups of the thymine (see. Accordingly, the solvents play an important role in facilitating the product release in TDG during the whole release process: first by mediating the interactions between the thymine and the active site residues during the S1→S2 transition; then by solvating the thymine once it leaves the narrow channel.
## Intrinsic motions of tdg promotes the product release
Our findings suggest that the intrinsic dynamics of TDG and its coupled motions with the DNA chains are important for the product release. The Pro-rich loop and the gating helix regions of TDG maintain an intrinsic opening motion even when the thymine is still trapped in the S1 state (see . This opening motion can significantly enlarge the release channel, therefore, to accommodate the dissociating thymine. Interestingly, the transition of the thymine from S1 to S2 can take advantage of the above open-state TDG by interacting with both the gating helix and the AP site of the DNA duplex, suggesting a conformational selection mechanism of recognition. Although the S2 state is less stable comparing to other metastable states, it is the key intermediate state connecting the bound state and the released states (see [fig_ref] Figure 2: The MSM reveals four metastable states for the thymine release in TDG [/fig_ref]. In this regard, the conformational selection from S1 to S2 for the TDG and thymine interaction is an efficient mechanism to promote the product release from the active site.
## The g142y mutation slows down the product release in tdg
Former experimental efforts have been devoted to trapping the product in the active site of the TDG-DNA complex [bib_ref] Excision of 5-hydroxymethyluracil and 5-carboxylcytosine by the thymine DNA glycosylase domain: its..., Hashimoto [/bib_ref] [bib_ref] Activity and crystal structure of human thymine DNA glycosylase mutant N140A with..., Hashimoto [/bib_ref] [bib_ref] Thymine DNA glycosylase exhibits negligible affinity for nucleobases that it removes from..., Malik [/bib_ref]. However, due to the transient product releasing event, the structure basis for the product-bound TDG-DNA complex remains unclear. This work designed a mutant G142Y TDG in which, comparing to the wt TDG, the thymine release process is delayed by the side chain of the replaced Tyr142 that locates within the release channel (see [fig_ref] Figure 6: The thymine release in the G142Y mutant TDG is delayed in comparison... [/fig_ref]. The PMF calculations show that the thymine release in the G142Y mutant TDG exhibits ∼1000-fold reduction in the releasing rate comparing to the wt system, reaching a releasing timescale of ∼10 ms. The findings shed light on further experimental trials along this route to potentially capture the excised base in the active site of the TDG-DNA complex.
Notably, our control simulations suggest that the G142Y substitution has no substantial effects on either the TDG structure or the dynamics of the base-flipping process. We compared the 100-ns MD simulations for the wt and G142Y TDG-DNA complexes. For both systems, the TDG structure reaches to a well-equilibrated conformation, as reflected from the converged RMSD values for the TDG backbones (see Supplementary. In comparison, the G142Y system is shown to exhibit ∼1Å higher RMSD fluctuations compared to the wt system. Further comparisons of the MD snapshots at 100 ns for these two systems demonstrate that the above structural difference mainly lies in the gating helix region where the G142Y substitution is introduced (see Supplementary. Notably, the above minor structural perturbation has no effects on either the secondary structures of the gating helix domain or the location of the residue Tyr142 within the product release channel. Moreover, we performed the TMD simulation to exam whether the G142Y substitution might potentially affect the base-flipping process (see Supplementaryand B and the 'Materials and Methods' section for more details of the model construction and the TMD setup). By targeting the interrogating complex to the recognition complex, it is shown that the residue Tyr142 is not lying on the base-flipping pathway; it is thus unlikely to make impacts on the dynamics of the base-flipping process (see Supplementaryand D and Supplementary Movie for the complete base-flipping process).
## Relative product releasing rates for different bases in tdg-dna complex
By constructing the MSM, we obtained the product releasing timescales for the thymine in the TDG-DNA complex.
To estimate the kinetic rates for other bases, we employed the thymine as a reference and performed PMF calculations to evaluate the relative product releasing rates for different bases along the same releasing channel, from which we get to know the releasing timescales for other products. The PMF landscape for the thymine successively captures the main transition events (e.g. S1→S2 and S2→S3) and key intermediate state that have been revealed by the MSM. The consistencies suggest that the PMF setup is appropriate to derive the free energy landscapes for different bases. Our results show that due to the lack of one methyl group, the uracil has a faster releasing rate than the thymine. Similarly, the 5caC also releases faster than the thymine, which is likely caused by the stabilizing effects from the water molecules owning to the carboxyl group of the 5caC. In comparison, the 5fC exhibits a similar releasing rate to the thymine, and the corresponding intermediate state also demonstrates a similar interaction network with the TDG-DNA complex to that for the thymine system, i.e. by forming two HBs with the Ala145 N-H and AP P-O − groups (see . In contrast, the 5hmU can form one HB with the AP P-O − group due to the presence of the 5-hydroxymethyl group right after the catalysis, which is not observed in other base species. This specific interaction leads to a rather different binding site of the 5hmU in TDG, which is captured by our pulling simulations.
[fig] Figure 1: (A) Schematic illustration of the DNA repair process in TDG. (B) [/fig]
[fig] Figure 2: The MSM reveals four metastable states for the thymine release in TDG. (A) For each of the S1 (green), S2 (purple), S3 (blue) and S4 (gray) states, random conformations of the thymine were chosen according to the stationary distributions. For each thymine conformation, only C2 atom is shown in sphere. The active site residue Tyr152 is also shown as a reference point. (B) [/fig]
[fig] Figure 3: (A) The free energy profile of the MD conformations projected onto the two slowest tICs, with each state from S1 to S4 labeled on the plot. (B) Solvents actively facilitate the thymine release. Four polar groups from the thymine, either serving as a hydrogen donor (HD) or acceptor (HA), were chosen for the analysis. The number of water molecules were then calculated using a distance cutoff of 3Å between two heavy atoms. Finally, for each state, the probability of different number of water molecules surrounding the thymine is plotted. [/fig]
[fig] Figure 4: Direct observations of the thymine release from the S3 state to the solvents from multiple 100-ns MD simulations. According to the releasing direction, two groups of release pathways are shown (A and B). For each group, five representative MD trajectories are demonstrated using different colors (blue, red, green, black and purple), in which the starting conformations of the MD trajectories are highlighted within dashed black circles, and the last MD snapshots (released states) are represented with spheres. [/fig]
[fig] Figure 6: The thymine release in the G142Y mutant TDG is delayed in comparison with the wt TDG. (A) Two TDG residues, Gly142 and Pro153, are [/fig]
|
Effect of Metoprolol Succinate in Patients with Stable Angina and Elevated Heart Rate Receiving Low-Dose β-Blocker Therapy
Aims: β-blockers are underused in Chinese patients with coronary heart disease. The prescribed dose is often low. The aim of this study was to investigate the effect of metoprolol succinate doses of 95 mg and 190 mg on heart rate (HR) control, as well as drug tolerance, in Chinese patients with stable angina, low-dose β-blocker use and unsatisfactory HR control. Methods: This was a multicenter, randomized, open-label, parallel-group trial in 15 clinical sites. Patients with stable angina, taking low-dose β-blockers (equivalent to metoprolol succinate 23.75-47.5 mg/day), and having a resting HR of ≥ 65 bpm were enrolled and randomized to either the metoprolol 95-mg group or the 190-mg group. The change in 24-h average HR from baseline recorded by Holter monitoring and the percentages of patients with resting HR controlled to ≤ 60 bpm were compared between the two groups. Results: Two hundred thirty-one patients entered the intent-to-treat population for the main analysis. The change in 24-h average HR from baseline was -0.62 ± 0.66 bpm in the 95 mg group and -2.99 ± 0.62 bpm in the 190 mg group (p = 0.0077) after 8 weeks of treatment. The percentages of patients with resting HR controlled to ≤ 60 bpm were 24.1% (95% CI: 16.35%, 31.93%) and 40.0% (95% CI: 31.05%, 48.95%), respectively (p = 0.0019). Only 4 and 2 of the patients, respectively, discontinued the study drugs because of hypotension or bradycardia.Conclusions:The metoprolol succinate dose of 190 mg is superior to the 95 mg dose in terms of HR control, in Chinese patients with stable angina, low-dose β-blocker use and unsatisfactory HR control. Both doses were well tolerated.
# Introduction
Elevated heart rate (HR) is an independent risk factor for all-cause death and cardiovascular events in the general population [bib_ref] Heart rate and cardiovascular mortality: the Framingham Study, Kannel [/bib_ref] and in patients with various cardiovascular diseases, including hypertension and coronary artery disease [bib_ref] Heart Rate Working Group. Resting heart rate in cardiovascular disease, Fox [/bib_ref]. Current guidelines in China and abroad recommend that the objective of stable angina treatment is to reach a target resting HR of 55-60 bpm, if tolerated, or 50 bpm for patients who have severe angina without symptoms of bradycardia.
However, HR is not optimally controlled in patients with coronary heart disease (CHD) in clinical practice. According to the Prospective Observational Longitudinal Registry of Patients with Stable Coronary Artery Disease, 44.0% of the total 33,438 CHD patients and 41.1% of the 24,910 patients receiving various doses of β-blockers had a resting HR of > 70 bpm . In China, the management of HR among CHD patients is also barely satisfactory; one national survey found that the average resting HR of patients with chronic angina pectoris who were on and off β-blockers was 75 bpm and 73 bpm, respectively .
β-blockers are among the most effective therapies for CHD. The clinical benefit and improvement in prognosis following β-blocker treatment in CHD patients have been associated with a reduction in resting HR . In patients who have stable angina without contraindications, β-blocker therapy is recommended as a strategy to improve disease outcome and symptoms. Although β-blockers have been widely available in China for many years for the treatment of cardiovascular diseases, they are underused among CHD patients in common clinical practice. According to the Prospective Urban Rural Epidemiological study, only 6.8% of CHD patients in China had been prescribed β-blockers, which was far lower than the rate in the United States and in European countries (45.5%) . On the other hand, a national survey conducted mostly in tertiary hospitals revealed a more favorable finding, wherein 61.1% of all patients were on β-blockers .
Physicians' knowledge and experience in using β-blockers can have some impact on the underutilization of these drugs and, consequently, on the incidence of elevated HR. A possible reason for their lack of prescription is the concern that Chinese patients may not have the same tolerance of the higher dose of β-blockers recommended in Western countries. However, these concerns are mostly based on personal experience rather than verified evidence. Thus, a study to address the issue of β-blocker underuse and poor HR control in Chinese CHD patients is warranted.
We report herein the findings of our study involving Chinese patients with stable angina and elevated HR who were taking a low dose of β-blockers. The patients were administered different doses of metoprolol succinate (95 mg vs. 190 mg), and their HR as well as other safety parameters were assessed.
# Methods
## Study design
This multicenter, randomized, open label, parallel group study was reviewed and approved by the Ethics Committee of Peking University First Hospital. The study was performed in accordance with the principles of the Declaration of Helsinki. Patients were enrolled into the study from October 2010 to September 2011.
The study included a one-week run-in period and a randomized eight-week treatment period. During the run-in period, all subjects were treated with 47.5 mg metoprolol succinate. The subjects were then randomized into two groups: 1) a 95-mg group receiving 47.5 mg/day metoprolol succinate for two weeks before dose titration to 95 mg/day; and 2) a 190-mg group receiving 95 mg/day metoprolol succinate for two weeks before dose titration to 190 mg/day. Dose titrations were performed only in patients who could tolerate the first dosage without exhibiting symptoms of bradycardia-systolic blood pressure (SBP) < 100 mmHg or HR < 45 bpmaccording to a 12-lead electrocardiogram recorded after the first two weeks of randomized treatment. If a patient could not tolerate a given dose level, as confirmed by the investigator at any visit, the dose was reduced to a lower level and the patient continued to receive that dose until the end of the study, provided that the final dose was > 47.5 mg. Any patient who could not tolerate a dose of 47.5 mg was excluded from the study After the treatment period, the drug was administered for another two weeks to ensure that the subjects were well advised concerning the type and dose of β-blocker they would have to take after the study.
## Patients
The inclusion criteria of study were: 1) provision of informed consent prior to participation in the study; 2) Chinese patients aged 18-75 years; 3) resting HR ≥ 65 bpm; 4) diagnosed with stable angina for at least 1 month and with stable angina pectoris symptoms within 2 weeks prior to enrollment; 5) left ventricular ejection fraction (LVEF) ≥ 50% according to echocardiography; 6) on β-blockers for at least 4 weeks at a dosage equivalent to metoprolol succinate 23.75-47.5 mg/day. This population was intended to simulate the patients we commonly encounter in clinical practice who have been prescribed a low dose of β-blockers and yet exhibit elevated HR.
The main exclusion criteria included: 1) significant clinical, laboratory, or electrocardiographic abnormalities that would place the subject at undue risk (in the Investigator's opinion), including renal impairment (serum creatinine > 2.0 mg/dL), serum ALT or AST > 3 × upper limit of reference range, serum potassium < 3.0 mEq/L, serum sodium ≤ 130 mEq/L, acute or chronic hepatitis or cirrhosis (clinical diagnosis), uncontrolled hyperthyroidism (clinical diagnosis); 2) SBP ≥ 180 mmHg, or < 100 mmHg at enrolment; 3) PR interval > 0.24 s at enrollment; 4) second-or higher-degree atrioventricular block; 5) bundle branch block (complete left bundle branch block and bifascicular block or trifascicular block); 6) symptomatic bradycardia or sick sinus syndrome; 7) poorly controlled diabetes mellitus (fasting plasma glucose > 10 mmol/L or 180 mg/dL or symptomatic hypoglycemia within the past three months); 8) unstable angina or Prinzmetal's angina; 9) patients with unstable, non-compensated heart failure (pulmonary edema, hypoperfusion, or hypotension); 10) systolic heart failure with reduced LVEF; 11) severe stable symptomatic heart failure (New York Heart Association class IV); 12) cardiogenic shock; 13) hemodynamically significant disorder of cardiac valves; 14) atrial fibrillation; 15) under continuous or intermittent inotropic therapy acting through β-receptor agonism; 16) serious peripheral vascular disease with risk of gangrene; 17) asthma or chronic obstructive pulmonary disease; 18) pheochromocytoma; 19) history of allergy or intolerance to any of the components of the study drugs; 20) malignancy or life expectancy less than 6 months because of malignancy; 21) history of alcohol or drug abuse; 22) use of a medication contraindicated for use with a β-blocker agent, such as barbiturate, propafenone, or verapamil; 23) inability to be off antiarrhythmic agents (class I, III, and IV, except for dihydropyridine calcium channel blockers); 24) pregnancy or nursing an infant; 25) female of childbearing potential, unless using the following forms of birth control during this study: intra-uterine device, implantable progesterone device, progesterone intramuscular injection, oral contraceptive, or barrier method plus spermicide; 26) athlete.
## Outcome assessment
The primary parameter of treatment efficacy was the 24-h average HR after eight weeks of treatment. Other parameters included: 1) the 24-h average HR after two weeks of treatment; 2) the change in 24-h average HR within groups after two and eight weeks of treatment; 3) the percentage of patients whose resting HR was lower than 60 bpm after two and eight weeks of treatment; 4) the number of patients who discontinued the treatment owing to an adverse event (AE); and 5) the changes in fasting plasma glucose, total cholesterol, and triglyceride levels. Holter monitors were used to record the 24-h average HR.
# Statistical analysis
To determine the difference in 24-h average HR between the two treatment groups after two and eight weeks' treatment, a mixed model of repeated measures analysis was performed on the change in 24-h average HR. This approach assumed that missing observations were missing at random, and utilized all the available data for analysis. The model included treatment, study center, baseline 24-h average HR, visit, and treatment-by-visit interaction as explanatory variables. The study center was regarded as a random effect, while all other variables were treated as fixed effects. To model the covariance structure, an unstructured covariance matrix was used. If the algorithm did not converge for the data set, a first-order autoregressive covariance structure was employed. The primary contrasts of interest were the treatment differences between the 95-mg and 190-mg groups during the eight weeks of treatment.
For comparisons of continuous variables within groups, the paired t-test was used. Other continuous safety parameters were analyzed using the same method as used for analyzing the change in the 24-h average HR.
The percentage of patients whose resting HR was controlled to ≤ 60 bpm was analyzed in both intent-to-treat (ITT) and per-protocol populations for the two treatment groups after two and eight weeks of treatment. The comparisons were performed using a logistic regression model with factors fitted for treatment, treatment center, and baseline HR, and were presented as odds ratios (OR) with 95% confidence intervals (CIs). For comparisons within groups, McNemar's test was used.
Categorical variables were summarized by frequency and percentage of population. The incidence and severity of AEs were analyzed for each body system and preferred term for both treatment groups. All other safety data, including physical examination, vital signs, and laboratory data, were analyzed using descriptive statistics and frequency distributions.
All statistical analyses were performed using SAS software (version 9.1.3; SAS Institute, Cary, NC, USA).
# Results
## Baseline characteristics
As shown in [fig_ref] Figure 1: Flow chart illustrating the selection process [/fig_ref] , 317 patients were screened and 251 patients were randomized into one of the two treatment groups. For the 66 patients who were not enrolled into the study, the most common reasons for screening failure were voluntary discontinuation (n = 34) and incorrect enrollment (n = 27). Of the 251 enrolled patients, 223 patients (88.8%) completed the treatment course while 28 patients (11.2%) discontinued the treatment, including 12 patients (9.6%) in the 95-mg group and 16 patients (12.7%) in the 190-mg group. Twenty enrolled patients had missing HR data and were thus excluded from the analysis, leaving 231 patients included in the ITT analysis (95-mg group, n = 116; 190-mg group, n = 115).
There were no significant differences in the baseline data between the two treatment groups [fig_ref] Table 1: Characteristics of ITT patients in the two treatment groups [/fig_ref].
## Heart rate control
Based on the ITT analyses, the baseline 24-h average HR was 70.9 ± 8.19 bpm and 71.4 ± 8.02 bpm for the 95-mg and 190-mg groups, respectively; after eight weeks of treatment, the 24-h average HR was 70.9 ± 8.71 bpm and 68.6 ± 8.40 bpm for the 95-mg and 190-mg groups, respectively. The least square mean of the difference between the two groups was 2.36 bpm (95% CI: 0.63, 4.09) after eight weeks of treatment (p = 0.0077; . The difference between the baseline HR and the HR after eight weeks was -0.62 bpm (95% CI: -1.94, 0.69) and -2.99 bpm (95% CI: -4.23, -1.75) in the 95-mg and 190-mg groups, respectively. There was a statistically significant change in 24-h average HR from before to after the eight-week treatment in the 190-mg group, but not in the 95-mg group.
The percentages of patients (ITT) with resting HR controlled to ≤ 60 bpm after two weeks of treatment were 21.6% (95% CI: 14.07%, 29.03%) and 24.3% (95% CI: 16.50%, 32.19%) in the 95-mg and 190-mg groups, respectively, with the difference between the two groups being 2.8% (95% CI: −8.04%, 13.64%; p > 0.05). Values for the same parameter after eight weeks were 24.1% (95% CI: 16.35%, 31.93%) and 40.0% (95% CI: 31.05%, 48.95%) in the 95-mg and 190-mg groups, respectively, with the difference being 15.9% (95% CI: 4.00%, 27.73%; p = 0.0019). A significant increase in the percentage of patients with resting HR controlled to ≤ 60 bpm was observed only in the 190-mg group [fig_ref] Table 3: Proportion of patients with resting HR ≤ 60 bpm during randomized treatment... [/fig_ref].
## Safety parameters
During the treatment period, the percentage of patients who reported AEs in the 190-mg group (16.1%) was slightly higher than that in the 95-mg group (12.2%). Similarly, drug-related AEs were also marginally more common in the 190-mg group (9.8%) than in the 95-mg group (7.9%). The percentages of patients with serious AEs were 0.7% and 2.7% in the 95-mg and 190-mg groups, respectively. No death occurred in the study.
The percentages of patients who discontinued the drug therapy because of an AE were similar in both groups (95-mg group, 4.3%; 190-mg group, 4.5%). Only six patients (2.2%) discontinued the treatment because of hypotension (SBP < 100 mmHg); four of them were from the 95-mg group while two were from the 190-mg group. Two patients (both from the 95-mg group) withdrew from the study because of bradycardia; one developed sinus bradycardia and one had an HR of < 45 bpm.
There was no significant difference between the two groups in the changes in fasting plasma glucose, total cholesterol, and triglyceride levels from baseline [fig_ref] Table 4: Changes in blood chemistry from baseline after 8 weeks' treatment [/fig_ref].
# Discussion
The association between increased HR and overall cardiac mortality or other adverse outcomes in CHD patients has been well established by previous studies [bib_ref] Heart Rate Working Group. Resting heart rate in cardiovascular disease, Fox [/bib_ref]. A sub-analysis of the BEAUTIFUL study [bib_ref] Relationship between ivabradine treatment and cardiovascular outcomes in patients with stable coronary..., Fox [/bib_ref] showed that HR reduction by medication is related to better clinical outcomes in selected patients with CHD. Furthermore, a meta-regression analysis demonstrated a statistically significant relationship between resting HR reduction by β-blockers and clinical benefits, including reductions in cardiac death (p < 0.001), all-cause death (p = 0.008), sudden death (p = 0.015), and recurrent non-fatal myocardial infarction (p = 0.024). Each 10 bpm reduction in HR is estimated to reduce the relative risk of cardiac death by 30%.
Researchers have long discussed the ideal range for target HR. The Framingham Heart Study showed that CHD mortality rates rose considerably with increasing HR, with the lowest mortality rate observed in patients with an HR < 65 bpm [bib_ref] Influence of heart rate on mortality among persons with hypertension: the Framingham..., Gillman [/bib_ref]. Moreover, according to an analysis of the Treating to New Targets (TNT) study based on a Cox proportional hazard model, a nadir HR of 52.4 bpm was associated with the lowest rate of AEs [bib_ref] HR in patients with coronary artery disease -the lower the better? An..., Bangalore [/bib_ref]. Based on established clinical evidence, the guidelines for the treatment of stable angina recommend a target resting HR of 55-60 bpm [bib_ref] Heart Rate Working Group. Resting heart rate in cardiovascular disease, Fox [/bib_ref]. Hence, we set the target HR at 60 bpm in the present study.
The effectiveness of β-blockers in controlling HR is well established. According to one report, atenolol, bisoprolol, and metoprolol were more effective than other β-blockers in managing resting HR in CHD patients . Previous studies of multiple patient groups have demonstrated that changes in both 24-h average HR and resting HR were related to the dosage of β-blocker administered [bib_ref] Heart rate distribution and predictors of resting heart rate after initiation of..., Zhao [/bib_ref] [bib_ref] Pharmacokinetic and pharmacodynamic comparison of metoprolol CR/ZOK once daily with conventional tablets..., Lücker [/bib_ref]. In a study involving Chinese CHD patients who were β-blocker-naive, 42.5% and 79.4% of the patients achieved a HR of less than 70 bpm after one and two months of metoprolol succinate treatment, respectively [bib_ref] Heart rate distribution and predictors of resting heart rate after initiation of..., Zhao [/bib_ref]. Baseline HR and metoprolol succinate dosage were independent predictors of the resting HR two months after β-blocker therapy. The present study revealed that 190 mg metoprolol succinate has a better effect in reducing HR than a 95 mg dose, as seen by both the 24-h average HR and the percentage of patients with resting HR < 60 bpm after eight weeks of treatment. In general, the higher dose of the drug was well tolerated by patients. Although the data exhibited the same trend seen in earlier studies, the decrease in HR induced by the drug was less than previously reported values.
The change in HR following metoprolol succinate treatment observed in this study was relatively small, especially in the 95-mg group. There may be several reasons why the drug was less effective in these patients. Firstly, to simulate the common clinical population of patients who are on low-dose β-blocker medication and have an elevated HR, only patients who were on β-blocker therapy at a dose equivalent to 23.75-47.5 mg metoprolol succinate with HR ≥ 65 bpm were included in the present study. For these patients, who can tolerate a low dose of metoprolol succinate well, a higher dose of the drug may be needed to better block the β-receptors. It has been shown that the cardiac autonomic nervous system is impaired in CHD patients who have increased sympathetic nerve excitability [bib_ref] Circadian rhythms of frequency domain measures of heart rate variability in healthy..., Huikuri [/bib_ref].
Secondly, a higher dose of metoprolol succinate may result in a higher threshold for angina, allowing for increased physical activity and thus a higher daytime HR, which will increase the 24-h average HR. Moreover, as suggested by the MERIT-HF study, the dose-response relationship of metoprolol succinate may exhibit an exponential correlation, which partly explains the lack of efficacy of the 95 mg dose.
The possible difference in β-blocker tolerability between patients from China and Western countries is one of the reasons why physicians are reluctant to prescribe a high dose of these drugs in China. In the present study, metoprolol succinate was well tolerated by the majority of patients, without extreme HR reduction. Furthermore, patients treated with the higher dose of metoprolol succinate did not report additional AEs. The incidence of AEs during the study was comparable to that at 12 weeks of treatment with 5-7.5 mg ivabradine b.i.d or 12.5-25 mg atenolol b.i.d in Chinese patients with stable angina, where the reported incidence of all AEs was 27.4% [bib_ref] The efficacy and safety of ivabradine hydrochloride versus atenolol in Chinese patients..., Li [/bib_ref] , further supporting the proposition that 190 mg metoprolol succinate can be safely prescribed for Chinese patients.
There are several limitations to our investigation. Although this study was innovative in using the 24-h average HR as the primary outcome parameter, unlike previous studies it was not designed to obtain detailed HR information during the day. Additionally, β-blockers have different effects on resting and exercise HR; thus, the efficacy and tolerability of different drug doses would be better measured with additional parameters, such as the lowest and highest HR during the 24-h period. Furthermore, a group of patients who can tolerate only a low dose of β-blockers was not included in this trial. There may be differences in the effect of β-blockers on patient outcomes between those who are sensitive to these drugs and those who are not, independently of HR control. Hence, further studies are needed to address these questions.
# Conclusion
A metoprolol succinate dose of 190 mg was superior to a 95 mg dose in controlling the HR of stable angina patients with elevated HR who were on a low-dose β-blocker therapy; the treatment was well tolerated by these patients. Only the administration of 190 mg metoprolol succinate achieved the treatment goal of controlling HR to ≤ 60 bpm in this patient population. Close follow-up and careful dose titration will be needed to achieve the goal for HR control in these patients.
## Abbreviations
HR: heart rate; bpm: beats per minute; CHD: coronary heart disease; SBP: systolic blood pressure; LVEF: left ventricular ejection fraction; AE: adverse event; ITT: intent-to-treat; OR: odds ratios; CI: confidence interval Acknowledgement This study was sponsored by AstraZeneca (Study code: D4022L00008; ClinicalTrials.gov identifier: NCT01213173).
[fig] Figure 1: Flow chart illustrating the selection process [/fig]
[table] Table 1: Characteristics of ITT patients in the two treatment groups [/table]
[table] Table 3: Proportion of patients with resting HR ≤ 60 bpm during randomized treatment compared with baseline (ITT set) [/table]
[table] Table 4: Changes in blood chemistry from baseline after 8 weeks' treatment [/table]
|
Role of Polar Phonons in the Photo Excited State of Metal Halide Perovskites
The exciton binding energies presented in the paper have been calculated using 'semi-cubic' unit cells. These structures have been obtained by allowing all atoms to relax (keeping the experimental lattice shape and volume fixed) using damped relaxation algorithms with a 4 × 4 × 4 k-point grid. The bromine and chlorine structures were constructed from the resulting high-symmetry iodine unit cells.For these structures, in the first step all degrees of freedom were relaxed (including cell shape and volume) at a high energy cut off of 600 eV. Hereafter, a simulated annealing was performed, linearly cooling the structure from 800K to 500K in 50000 steps. Approximately every ∼1000 steps a snapshot was taken from the trajectory. For these structures the cell shape as well as the internal coordinates were relaxed and the lowest global energy structure was determined.To confirm that the lowest energy structures for the primitive unit cells were correctly determined, a second molecular dynamics run was started from the first lowest energy structure. The same annealing procedure was used with the volume and cell shapes fixed to those of the yet best structures. More than 15 structures were picked from the MD and fully relaxed (including volume and cell shape). In only two cases, an even lower energy structure than in the first cycle was found, however, these two structures were only 15 meV lower in energy than in the first cycle. We are therefore confident that we have determined the lowest energy structures for the primitive unit cells reliably. Furthermore, all structures were carefully checked for instabilities in the vibrational frequencies and no instabilities were found, whereas virtually all initial structures constructed from experimental data alone exhibited such instabilities, even after careful relaxation. The resulting structures have been attached to this document.Model dielectric functionA local model dielectric function (ε m ) [1] has been used in the BSE calculations to converge the exciton binding energies on dense k-point grids.A local function makes the screened Coulomb kernel diagonal (G = G ) in the screened Coulomb potential and allows us to calculate the high resolution plots ofFig. 1 (right)in the paper. The parameter ε m comes from DFPT calculations on a shifted 8×8×8 k-point grid and the screening length parameter (λ) was fitted to match the diagonal (G = G ) part of dielectric function from the GW 0 calculations on the shifted 6 × 6 × 6 k-point grid, seeTable 1. This approximation is also used in Ref.[2], however in that work a different functional was used. This approximation works very well, in particularly in the low energy part as is illustrated inFig. 1. The imaginary part of the dielectric function calculated with the 'normal' GW-BSE matches with the one calculated with the model BSE (mBSE) method. Both calculations were performed using a 4 × 4 × 4 k-point grid with the same KS orbital basis and GW 0 quasiparticle energies. 0 1 2 3 4 5 ω (eV) Im ε GW-BSE GW-mBSE Figure 1: Imaginary part of ε(ω) calculated with the GW-BSE and GW-(model)BSE method. (MAPbI 3 ) 1
The exciton binding energies presented in the paper have been calculated using 'semi-cubic' unit cells. These structures have been obtained by allowing all atoms to relax (keeping the experimental lattice shape and volume fixed) using damped relaxation algorithms with a 4 × 4 × 4 k-point grid. The bromine and chlorine structures were constructed from the resulting high-symmetry iodine unit cells.
For these structures, in the first step all degrees of freedom were relaxed (including cell shape and volume) at a high energy cut off of 600 eV. Hereafter, a simulated annealing was performed, linearly cooling the structure from 800K to 500K in 50000 steps. Approximately every ∼1000 steps a snapshot was taken from the trajectory. For these structures the cell shape as well as the internal coordinates were relaxed and the lowest global energy structure was determined.
To confirm that the lowest energy structures for the primitive unit cells were correctly determined, a second molecular dynamics run was started from the first lowest energy structure. The same annealing procedure was used with the volume and cell shapes fixed to those of the yet best structures. More than 15 structures were picked from the MD and fully relaxed (including volume and cell shape). In only two cases, an even lower energy structure than in the first cycle was found, however, these two structures were only 15 meV lower in energy than in the first cycle. We are therefore confident that we have determined the lowest energy structures for the primitive unit cells reliably. Furthermore, all structures were carefully checked for instabilities in the vibrational frequencies and no instabilities were found, whereas virtually all initial structures constructed from experimental data alone exhibited such instabilities, even after careful relaxation. The resulting structures have been attached to this document.
## Model dielectric function
A local model dielectric function (ε m ) [bib_ref] An efficient method for calculating quasiparticle energies in semiconductors, Bechstedt [/bib_ref] has been used in the BSE calculations to converge the exciton binding energies on dense k-point grids.
[formula] ε −1 (k + G) = 1 − (1 − ε −1 m )e −(2π|k+G|) 2 4λ 2 [/formula]
A local function makes the screened Coulomb kernel diagonal (G = G ) in the screened Coulomb potential and allows us to calculate the high resolution plots of [fig_ref] Figure 1: Imaginary part of ε [/fig_ref] (right) in the paper. The parameter ε m comes from DFPT calculations on a shifted 8×8×8 k-point grid and the screening length parameter (λ) was fitted to match the diagonal (G = G ) part of dielectric function from the GW 0 calculations on the shifted 6 × 6 × 6 k-point grid, see [fig_ref] Table 1: Parameters of the model dielctric function [/fig_ref]. This approximation is also used in Ref. [bib_ref] Efficient O(N 2 ) approach to solve the bethe-salpeter equation for excitonic..., Fuchs [/bib_ref] , however in that work a different functional was used. This approximation works very well, in particularly in the low energy part as is illustrated in [fig_ref] Figure 1: Imaginary part of ε [/fig_ref]. The imaginary part of the dielectric function calculated with the 'normal' GW-BSE matches with the one calculated with the model BSE (mBSE) method. Both calculations were performed using a 4 × 4 × 4 k-point grid with the same KS orbital basis and GW 0 quasiparticle energies. : Eigenvalue of the first exciton with e-h interactions (circles) and in the independent particle picture (squares). The exciton binding energy is the difference between these two levels. The inset shows a zoom-in for the levels calculated at dense k-meshes. The green and blue lines are linear fits to the last three data points. (Example: FASnI 3 )
## K-point convergence
The calculated exciton binding energies have been carefully converged with respect to the k-point grid density. Coarse meshes result in exciton binding energies up to an order too large. The exciton binding energies presented in the paper are the result of a linear extrapolation to an infinite k-point grid [bib_ref] Efficient O(N 2 ) approach to solve the bethe-salpeter equation for excitonic..., Fuchs [/bib_ref] as is illustrated in .
All unit cell structures have been added and formatted as Crystallographic Information File (CIF) files. The CIF formatted files can, for instance, be viewed with the freely available atomic structure visualization package VESTA (http://jp-minerals.org/vesta/).
## %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%-------iodines-------%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
[fig] Figure 1: Imaginary part of ε(ω) calculated with the GW-BSE and GW-(model)BSE method. (MAPbI 3 ) [/fig]
[table] Table 1: Parameters of the model dielctric function. Fitted screening length parameter (λ) and the "ion clamped" static dielectric constant (ε ∞ ) calculated using DFPT on a shifted 8 × 8 × 8 k-point grid. [/table]
|
Correlation between initial tumour volume and treatment duration on Dabrafenib: observation study of subjects with BRAF mutant melanoma on the BRF112680 trial
Background: Planar-based measurements of lesions in metastatic melanoma have limitations in estimating tumor burden of a patient and in predicting response to treatment. Volumetric imaging might add predictive value to Response criteria in Solid Tumor (RECIST)-measurement. Based on clinical observations, we explored the association between baseline tumor volume (TV) and duration of treatment with dabrafenib in patients with metastatic melanoma. We have also explored the prognostic value of TV for overall survival (OS) and progression free survival (PFS).Methods: This is a retrospective, chart-review of primary source documents and medical imaging of a cohort of patients participating in the BRF112680 phase 1 clinical trial at the Prince of Wales Hospital. TV was quantified by contouring all the measurable baseline target lesions in the standard manner for radiation planning using Voxxar™ software. We used Cox regression models to analyse associations between TV and duration of treatment with dabrafenib and between TV, PFS and OS.Results: Among 13 patients of BRAF 112680 trial, 10 were included in the retrospective analysis. Target lesion sum volume ranged from 0.3 to 1065.5 cm 3 (cc), with a median of 27.5 cc. The median PFS and OS were 420 days (range 109-1765) and 1680 days (range 390-2940), respectively. The initial TV was inversely correlated with duration of treatment with dabrafenib (rho − 0.6; P 0.03). In multivariate analysis, TV was a predictor for OS ). Patients with tumour volume above the median had significantly lower OS of 6months compared to 56-months survival for patients with smaller volumes; P = 0.019.Conclusions: TV is a predictor for treatment duration and is prognostic of OS and PFS in patients with metastatic melanoma. These findings need to be validated prospectively in clinical trials.
# Background
The BRAF inhibitors have significantly changed the outcomes for patients with BRAF mutant metastatic melanoma by improving survival and reducing symptoms by inducing tumour response, when compared with chemotherapy.
In early and intermediate stage melanoma, it has been shown that tumour volume (TV) estimated by Cavalieri's method, is a practical and reproducible variable, which quantifies tumour burden in accordance with tumour biology and has demonstrated predictive value in multivariate analyses. However, in metastatic melanoma (MM), the unidimensional response evaluation criteria in solid tumors (RECIST) and bi-dimensional variables or WHO measurements (which are based on clinical examination) are the most commonly used surrogates of TV for assessing response. Recently, studies have shown that CT tumor volume measurements using segmentation tools have been consistently more reproducible than diameter measurements used in RECIST, and reflect the entire TV rather than planar measurements on one axial plane image. Yet, using TV parameter is still evolving and need to be prospectively validated in the clinical trials.
In 2009, the BRF112680 phase 1 clinical trial -first in human-investigating the safety and tolerability of dabrafenib was opened at our institution (Prince Of Wales Hospital, NSW). A total of 13 patients with metastatic cancers, mainly melanoma, were recruited. Three patients who had MM remained on dabrafenib monotherapy for more than 4 years, compared with the median PFS of 5-6 months in reported studies. An observation was made that these long-term survivors had qualitatively lower volumes of disease at presentation than other patients. This study aimed to explore if this clinical observation does indeed represent a signal of initial tumour volume being a predictive factor for response to treatment. Our primary objective was to study the association between measured total initial TV and the duration of treatment on DF. The secondary objectives were its association with overall survival (OS), progression free survival (PFS).
# Methods
## Study design
This is a retrospective, chart review of primary source documents and medical imaging of a cohort of patients participated in the BRF112680 phase-1 clinical trial at the Glaxo Smith Kline (GSK) Medicines Research Unit at Prince of Wales Hospital. The patients with BRAF-V600E mutation positive, unresectable or metastatic melanoma were selected. Total initial TV was defined as the sum of measured volumes of all metastatic lesions at all metastatic sites seen on the baseline pre-enrolment CT scan of that individual. Duration of treatment was defined as the interval between commencement and permanent cessation of dabrafenib on trial. PFS was the interval between first dose of dabrafenib and date of disease progression or death due to any cause. Progression was defined as ≥20% increase in the smallest sum of study or new lesions of new lesions as per RECIST 1.1 criteria. Overall survival was defined as the interval between the first dose of dabrafenib and death due to any cause, end of study period or lost to follow up. The phase 1 study, BRF112680-A Phase I, Open-Label, Multiple-Dose, Dose-Escalation Study to Investigate the Safety, Pharmacokinetics, and Pharmacodynamics of the BRAF Inhibitor GSK2118436 in Subjects with Solid Tumors was approved by the Bellberry Human Research Ethics Committee (B65/09) and signed informed consent were obtained from all patients that included the publication of the study results in medical journals. The protocol of this further analysis of the role tumour volume and treatment duration was separately assessed and approved by the South Eastern Sydney Local Health District Human Research Ethics Committee (LNR/14/ POWH/539) which found that this study involved no further disclosure than which was expected by the patients when consenting to the original study.
## Study population
This retrospective study was performed on 13 patients with BRAF-mutant metastatic cancer enrolled in one centre of the BRF112680 trial. The phase I trial commencing in 2009 was conducted with an accelerated dose titration design. This centre enrolled the first few patients in this first-in-man trial of dabrafenib. Three patients had colorectal cancer and had been excluded from the study. The inclusion criteria (as per the BRF 112680 study) were; histologically confirmed solid tumour for which no curative treatment was available, 18 years or older, Eastern Cooperative Oncology Group (ECOG) performance status of 1 or less (patients with an ECOG status of 2 could be enrolled with the approval of the study's medical monitor), life expectancy of 3 months or longer, absence of known progressing or unstable brain metastases, adequate hematologic, hepatic, and renal function. The presence of a BRAFV600E mutation was mandatory. The exclusion criteria were; age < 18 years, patients with BRAF wild-type melanoma or with unknown BRAF mutation status, patients previously treated with BRAF inhibitors, patients with moderate or severe hepatic impairment. Upon successful escalation of dose cohort, patients on lower dose levels were given the increased dose, which for the patients in this trial, was subsequently 100 mg TID or 150 mg BD.
## Management of data
Patients enrolled in BRF112680 trial were assigned a study number. However, as all patients had baseline CT scans in the hospital, they were also given a hospital medical record number through which archived medical images were stored. Tumour volume measurements were done on patient's identified scans but were calculated by the investigator who was not involved in the BRF 112680 phase 1 trial, to reduce observer bias, and entered into the TV worksheet. Data extracted from the source documents was collected by investigators involved in the phase I trial using the BRF112680 study number. Once primary data collection was completed and verified, the work sheets containing the patient's identifiers had the primary identifiers erased, leaving only the study number. Any further analysis was done on the de-identified but re-identifiable data. The total initial tumour volume was measured using Focal 4D™ software. The baseline staging CT for each patient was loaded from the hospital PACS™ to Voxxar™ software and imported into a workstation containing the Focal 4D™ software. By using the formal radiological and PET reports, we identified the organs involved with metastatic disease for each patient. The sites and the number of organs involved in each patient (eg. liver, lung, lymph nodes, soft tissues, bones, brain) were also recorded.
Tumour volume was calculated by contouring the lesions in the standard manner as used by Radiation Oncologists to outline the gross tumour volume for radiation planning:
Drawing a contour for each lesion, using the same colour label for any lesions seen in the same organ. After contouring all lesions in an organ, the software will automatically calculate a tumour volume of that organ. The number of metastases in each organ will also be recorded automatically. The sum of the tumour volumes in all organs identified to have metastatic disease provided the overall initial tumour volume for the patient.
# Statistical methods
Data were verified, coded by the researcher and analysed using IBM-SPSS™ 21.0 (IBM-SPSS Inc., Chicago, IL, USA). Descriptive statistics: means, standard deviations, medians, IQR and percentages were calculated.
Predictors of the TV were tested using multivariable linear regression analysis (likelihood Ratio Test). The prognostic effect of the various parameters on clinical outcome was tested using the Kaplan-Meier method with the log-rank test was applied to compare survival curves. Multivariate analysis was carried using the Cox regression model for OS and PFS. A significant p value was considered when it is equal or less than 0.05.
# Results
## Patients' characteristics
From the total of 13 patients enrolled in the BRF 112680 phase I trial at our site, three patients were excluded from this analysis as they had non-melanoma diagnoses. The final series comprised 10 patients, eight men and two women (median age 61.5 years, range 28-81 years). The main clinical, biological, and radiological characteristics of patients are shown in Tables 1 and 2. Collection of serum lactate dehydrogenase (LDH) was not mandated in the clinical trial. Results were available for eight patients, six (75%) of them has level above the normal upper limit. Comorbidities were documented in six patients. None of the patients had brain metastases on trial commencement.
Target lesion sum volume ranged from 0.3 to 1065.5 cm 3 (cc), with a median of 27.5 cc. Two patients had only one organ involved. Organs of involvement are depicted in.
There was an inverse correlation between the initial TV and the duration of dabrafenib treatment; rho was − 0.612 p = 0.03. Likewise, there was negative correlation between the initial TV and the overall survival time; rho was − 0.636 p = 0.024.
The median PFS and OS were 420 days (range 109-1765) and 1680 days (range 390-2940), respectively. In the PFS multivariate analysis, after adjusting for age and sex, the TV showed a signal as a potential predictor of PFS; HR was 8.76 (CI 1.05-43.58) p = 0.048. Also, the serum LDH level presented a signal as a potential predictor of PFS; HR was 1.05 (CI 1.004-4.27) p = 0.047. When dichotomized by the TV median, patients with volume above the median had significantly lower OS of 6-months compared to 56-months survival for patients with smaller volumes; P = 0.019. Similarly, patients with volume above the median had significantly lower PFS of 4-months compared to patients with smaller volumes who lived longer than the median PFS; P = 0.013.
# Discussion
The existing prognostic variables for advanced melanoma, such as the number and sites of metastatic diseaseare semi-quantitative. Quantitative imaging, such as volumetric CT, may represent a useful in-vivo biomarker, allowing adaptive trial designs and facilitating management decisions in clinical settings.
Our study has confirmed TV as an independent predictive factor for duration of dabrafenib treatment. To our knowledge, there are no previous studies which have rigorously examined the correlation between TV and duration of anti-BRAF treatment as a surrogate for clinical benefit. In addition, this case series provides a signal for TV as having prognostic value for PFS and OS, with both P values of 0.04. Some recent studies have reported the impact of baseline tumor burden as assessed by CT scans on survival. For example, in the KEYNOTE-1 study, the sum of the maximum diameters of target lesions on the baseline scan (baseline tumor size) was used as an index of tumor burden. A higher burden was associated with significantly worse OS. In another study, that semi-automated analysis of MRI-tumor volume (initial and final change) was a better predictor of recurrence free survival than the longest dimension measured manually by RECIST or WHO guidelines in breast cancer patients treated with neoadjuvant chemotherapy. There are limitations to using tumor volumes and more broadly all anthropomorphic measurements as predictive biomarkers. For example, it was noted that diameter-based RECIST measurements may not be reflective of the actual changes in the tumor mass or being reflective of the appearance of new lesion within complex masses. Hence, static tumor volume is not an adequately useful biomarker. Hitherto, the dynamics of TV growth rate rather than a static volume measurement has been proposed as an emerging tool to predict disease response in many solid tumours including lung cancer[22], head and neck tumor, and adult high grade gliomas. For example, in a cohort of EGFR-mutant advanced non-small-cell lung cancer (NSCLC) patients treated with erlotinib or gefitinib, the 8-week tumor growth rate has been shown to correlate with survival. A greater tumor volume decrease at 8 weeks was associated with longer OS (p = 0.01). Similarly, in another series of 42 ALK-mutant lung cancer treated with crizotinib, the higher 8-weeks volume reduction was significantly associated with longer survival. Taken together, there is a mounting evidence suggesting that volumetric image analysis adds value to clinical trial science in terms of sensitivities and precision particularly in geometric challenges such as spiculated masses, infiltrative tumors that lack clear margins or tumours of small size.
# Conclusions
Although our study is limited by the small number of patients analysed and the retrospective design, but it has confirmed the observation of a significant inverse relationship between tumor volume and clinical benefit with treatment on dabrafenib. If prospectively validated, initial tumour volume may represent a predictive factor to consider when evaluating results of oral kinase inhibitors and a prognostic factor for clinicians regarding duration of clinical benefit. |
A Randomized Double-Blind, Placebo-Controlled Study on Anti-Stress Effects of Nelumbinis Semen
Citation: Ye, M.; Bae, H.; Park, S.; Lew, J.; Kim, K.S.; Shim, I. A Randomized Double-Blind, Placebo-Controlled Study on Anti-Stress Effects of Nelumbinis Semen. Int.
# Introduction
Depression is a common mood disease characterized by a low mood with loss of interest or pleasure in usual activities [bib_ref] The antidepressant effect of Gastrodia elata B1. on the forced-swimming test in..., Chen [/bib_ref]. Depression contributes to global disease burden. The development of antidepressant medications for treating depression has shifted toward serotonin as a key target [bib_ref] Antidepressants Targeting the Serotonin Reuptake Transporter Act via a Competitive Mechanism, Apparsundaram [/bib_ref]. Nevertheless, serotonin is not the only factor that plays a role in depression. Based on new research, depression can occur due to various factors such as increased stress hormones, abnormal activities in certain parts of the brain or the immune system, and inflammation. [bib_ref] Research and treatment approaches to depression, Wong [/bib_ref]. Accordingly, identifying effective and safe treatments with clear pharmacological mechanisms is urgently needed. Many antidepressant medications are tolerated but side effects suggest that antidepressants are not completely effective [bib_ref] Tolerability issues during long-term treatment with antidepressants, Cassano [/bib_ref]. Thus, better antidepressants with greater efficacy and fewer side effects are needed. Although the etiology of depression is unknown, numerous systematic reviews and research papers have been published about the effects of herbal medicine on depression [bib_ref] Control of stress-induced depressive disorders by So-ochim-tang-gamibang, a Korean herbal medicine, Choi [/bib_ref].
The Beck depression inventory (BDI), created by Aaron T. Beck in 1998, has become one of the most widely used psychometric methods for detecting depression in normal populations and in different psychiatric patient cohorts which confirm the validity of the Beck depression inventory. We used the Beck depression inventory (BDI), a 21-question multiple-choice self-report inventory, as one of the most widely used psychometric tests for measuring the severity of depression. Its development marked a shift among mental health professionals who had viewed depression from a psychodynamic perspective without considering a patient's own thoughts.
We used the stress response inventory (SRI), commonly used in clinical practice and psychiatric research [bib_ref] Development of the Stress Response Inventory and its application in clinical practice, Koh [/bib_ref]. The SRI was devised to score mental and physical symptoms that occurred during the preceding two weeks which might have influenced the status of mental stress level. The SRI includes 39 items focusing on emotional, somatic, cognitive, and behavioral stress responses. SRI scores can be categorized into seven stress factors: tension, aggression, somatization, anger, depression, fatigue, and frustration [bib_ref] Development of the Stress Response Inventory and its application in clinical practice, Koh [/bib_ref].
Studies of the antidepressant effects of physical exercise based on neurophysiological responses to such exercise activity are scarce. Electroencephalography (EEG) is a reliable tool reflecting upper cognitive functions and mental or psychological states. It is generally considered that higher spectral activity is correlated with arousal, cognitive processing, or emotional activity [bib_ref] Detection of stress/anxiety state from EEG features during video watching, Giannakakis [/bib_ref]. Electrical brain activity measured with EEG shows a desynchronized pattern during stress, with strong excited emotions and beta frequencies being dominant. In a relaxed, non-stressful state, EEG contains an alpha activity [bib_ref] Psychological phenomena and the electroencephalogram, Lindsley [/bib_ref]. Although mechanisms underlying EEG generation are not fully understood, frontal and prefrontal cortices are assumed to play a key role in various rhythmical EEG activities [bib_ref] Timofeev, I. Cellular basis of transition between sleep and electrographic seizures, Steriade [/bib_ref]. EEG brainwaves are divided into four bands: delta, theta, alpha and beta. The power spectrum of the beta waves over the alpha waves has a correlation with the level of stress, and the numerical equation of the ratio is given to indicate the state of the subject directly [bib_ref] Assessment of Surgeon's Stress Level and Alertness Using EEG during Laparoscopic Simple..., Duru [/bib_ref].
Nelumbinis semen (NS) has been widely used in Korean traditional medicine to treat insomnia, anxiety, and postmenopausal depression. The use of NS for depressive mood has been reported previously. A previous study has found that NS can exert an antidepressantlike effect in rats, based on a forced swim test. The preclinical study suggested that NS could advance local dopaminergic and cholinergic or norephinergic neurotransmission via activation of cAMP formation in the cortex and hippocampus [bib_ref] The anti-depressant effect of Nelumbinis semen on rats under chronic mild stress..., Kang [/bib_ref]. Recent studies have found that NS has an antidepressant effect on forced swim-induced depression-like symptoms [bib_ref] Mental health service use among South Africans for mood, anxiety and substance..., Seedat [/bib_ref] and chronic mild stress (CMS)-induced depression-like symptoms [bib_ref] The anti-depressant effect of Nelumbinis semen on rats under chronic mild stress..., Kang [/bib_ref] in rats. These results indicate that NS could serve as an alternative medication to treat persons suffering from depression. Therefore, this study will extend the appraising process of WNS's efficacy and safety in improving depression, anxiety, and EEG asymmetry in adults diagnosed with depression through a randomized, double-blind, placebo-controlled clinical trial. Furthermore, we examined the ratio of α/β activity using EEG to provide an assessment of stress.
# Methods
## Study design and participants
In this study, 46 participants were randomly recruited. Permission to perform the study was granted by the Ethical Review Board of Human Research of the College of Medicine at Catholic University (Seoul, Korea) after ethical approval by its ethical committee, with code number KCMC07MS274. Each participant submitted prior written consent. This study was designed as a double-blind, placebo-controlled, randomized clinical trial conducted in South Korea. All participants were Koreans. Participants were aged 18-65 years. All participants were screened for BDI. They were divided into three groups (a placebo-treated group, a 2.4 g WNS-treated group, and a 4.8 g WNS-treated group) based on computergenerated random selections.
Criteria for patient inclusion in this study were: those who were ready and capable of conforming with the study protocol, those who were capable of understanding and selecting Korean and English alphabets, those who were capable of communicating with clinical trial performers fully, those who were capable of complying with all tests and medical examinations demanded by the study protocol, those who were capable of sufficiently understanding the study purpose, and those who were willing to indicate consent by adding their names or seal. Exclusion criteria were: those who had a fasting glucose level of over 126 mg/dl, systolic blood pressure of over 140 mm Hg or diastolic blood pressure of 90 mm Hg, blood thyroid-stimulating hormone below 0.3 or over 4.0, hemoglobin below 13.0 g/dl for men or 12.0 g/dl for women, abnormal liver function exceeding 1.5 times normal levels of ALT or AST, infections in upper airways or chronic diseases, as well as mentally ill patients.
## Study medications
We used hard gelatin capsules containing 400 mg of WNS. A dry extract of the herbal medicine was manufactured by Sun Ten pharmaceutical company in Taiwan. Participants received treatment (12 capsules per day taken twice daily) for 2.4 g of WNS or 4.8 g of WNS, or a placebo, for two weeks. Subjects in the placebo group received capsules containing an equivalent amount of starch and lactose.
Results from participants experiencing 11 effective days (80% of 14 days) of treatment were used for the final analysis, i.e., 11 days after taking the capsules. Patients with protocol violations were excluded from the final analysis. Study drugs were labeled sequentially by the manufacturing department of the sponsor based on a randomization list. They were provided to investigators in blocks of three. All investigators and personnel who were actively involved in this trial were blinded to group assignment until the database was closed. After completion of this trial, drug supplies that were not used were returned to the monitor.
## Measures
## Beck depression inventory (bdi)
BDI is a 21-item self-rated inventory. Each item was rated with a set of four possible answer choices of increasing intensity. When the test was scored, a value of 0 to 3 was assigned for each answer and then the total score was compared to a key to determine the depression's severity. BDI had three subscales: (1) negative attitudes towards self (A), (2) performance impairment (B), and (3) somatic disturbances (C). The highest score on each of these 21 questions was 3 and the highest possible total score for the whole test was 63. The lowest possible score for the whole test was zero. Only one score per question (the highest rated if more than one was circled) was added. An index score ranging from 5 to 9 was considered to be within the normal range. A score of 10 to 15 indicated mild depression. A score of 16-23 meant moderate depression and a score of 24-63 indicated severe depression. A score below 4 was recognized as possible denial of depression or faking good health. Thus, participants who received scores below 4 were excluded.
## Stress response inventory
The SRI scale consisted of a total of 39 response items under seven subscales (tension, aggression, somatization, anger, depression, fatigue, and frustration). When the test was scored, a value of 0 to 4 was assigned for each answer. The total score was then compared to a key to determine severity of the stress.
## Electroencephalogram recording and analyses
Brain wave activity was assessed by a trained professional researcher using EEG spectra with a PolyG-1 (LAXTHA, Daejeon, Korea). Electrodes were attached to cortical sites on the participants' prefrontal lobes (Fp1, Fp2) according to the international 10/20 system. Among several types of EEG waveforms, including α and β waves, only the α and β waves were measured since they were key indicators of brain arousal and relaxation. EEG measurement time varied depending on the purpose of the study and characteristics of participants. EEG data were quantitatively analyzed using Telescan 2.98 (Laxtha Inc., Daejeon, Korea). Sixty seconds of raw EEG data were analyzed for each measurement (after excluding the first and last 10 s of recording). In addition, only EEG waves between 5 and 50 Hz were analyzed. To correct EEG differences between subjects due to variations in scalp thickness and tension levels, a relative band power was calculated from the absolute band power using the ratio of the absolute band power at a certain frequency. Relative alpha power (10-13 Hz) and relative beta power (20-30 Hz) were then analyzed.
# Statistical analysis
Data were statistically analyzed using one-way analysis of variance (ANOVA) followed by an LSD post-hoc test. Statistical significance in the same subject group was determined using a paired t-test, and the gender comparison was determined using a chi-square test.
All of the results are presented as mean ± standard error of mean (SEM). IBM SPSS Statistics 23.0 for Windows was used for analysis of the statistics. Statistical significance was set at p < 0.05.
# Results
## Participants' characteristics at entry
A total of 70 people participated in this trial with their permission at the start. After applying inclusion and exclusion criteria, 58 volunteers were found to be eligible for the study. These 58 volunteers underwent a BDI test. Of them, 46 had BDI score above 5. Finally, these 46 volunteers were randomly assigned to treatment groups of 2.4 g WNS (14 cases), 4.8 g WNS (17 cases), and placebo (15 cases). The study flowchart is shown in [fig_ref] Figure 1: The study flowchart. [/fig_ref]. These three groups did not differ in age or sex distribution after randomization. All 46 volunteers initially randomized completed this study. However, for the final analysis, only results from participants experiencing 11 effective days (80% of 14 days) were used. Based on this criterion, five volunteers in the placebo group, three volunteers in the 2.4 g WNS group, and six volunteers in the 4.8 g WNS-treated group were excluded. Thus, results of the remaining 32 volunteers were finally analyzed. [fig_ref] Table 1: Characteristics of randomized subjects [/fig_ref] shows characteristics of these subjects. Values are presented as mean ± SEM; p-values were calculated using χ 2 test unless otherwise specified.
## Wns's effect on bdi and sri scores
When the test was scored, participants who received scores below 4 were excluded since these scores are below usual scores for normal, recognized as possible denial of depression or faking good health. After two weeks of intervention, the BDI score was significantly reduced in the 2.4 g and 4.8 g WNS-treated groups as compared with that in the placebo group (p < 0.05) (Table 2 and [fig_ref] Figure 2: Differences in BDI mean scores between WNS and placebo groups [/fig_ref]. BDI subscale A (negative attitudes towards self) was substantially shorter in the 2.4 g and 4.8 g WNS-treated group (p < 0.05) than that in the placebo group [fig_ref] Figure 3: Differences in BDI subscale A mean scores of WNS and placebo groups [/fig_ref]. BDI subscale C showed a trend toward a shorter decrease in the WNS 2.4 g group than that in the placebo group (p < 0.05) [fig_ref] Figure 4: Differences in BDI subscale C mean scores between WNS and placebo groups [/fig_ref]. SRI mean scores were significantly reduced in the WNS 2.4 g treatment group (p < 0.05) [fig_ref] Table 3: Changes in total SRI scores and SRI subscale scores in patients after... [/fig_ref] and [fig_ref] Figure 5: Differences in SRI mean scores between WNS and placebo groups [/fig_ref]. Paired t-tests revealed that SRI E was substantially shorter in the 2.4 g WNS-treated group (p < 0.05). These results suggest that WNS 2.4 g treatment can reduce depression. Thus, WNS treatment may help depressed patients recover from depression in clinics.
## Wns's effect on bdi and sri scores
When the test was scored, participants who received scores below 4 were excluded since these scores are below usual scores for normal, recognized as possible denial of depression or faking good health. After two weeks of intervention, the BDI score was significantly reduced in the 2.4 g and 4.8 g WNS-treated groups as compared with that in the placebo group (p < 0.05) (Table 2 and [fig_ref] Figure 2: Differences in BDI mean scores between WNS and placebo groups [/fig_ref]. BDI subscale A (negative attitudes towards self) was substantially shorter in the 2.4 g and 4.8 g WNS-treated group (p < 0.05) than that in the placebo group [fig_ref] Figure 3: Differences in BDI subscale A mean scores of WNS and placebo groups [/fig_ref]. BDI subscale C showed a trend toward a shorter decrease in the WNS 2.4 g group than that in the placebo group (p < 0.05) [fig_ref] Figure 4: Differences in BDI subscale C mean scores between WNS and placebo groups [/fig_ref]. SRI mean scores were significantly reduced in the WNS 2.4 g treatment group (p < 0.05) [fig_ref] Table 3: Changes in total SRI scores and SRI subscale scores in patients after... [/fig_ref] and [fig_ref] Figure 5: Differences in SRI mean scores between WNS and placebo groups [/fig_ref]. Paired t-tests revealed that SRI E was substantially shorter in the 2.4 g WNS-treated group (p < 0.05). These results suggest that WNS 2.4 g treatment can reduce depression. Thus, WNS treatment may help depressed patients recover from depression in clinics.
## Resting eeg asymmetry
The analysis carried out on EEG power spectral content showed differential effects for each EEG band with marked topographical differences. As depicted in [fig_ref] Figure 7: Asymmetry values with depressive disorder after WNS and placebo treatment [/fig_ref] , the alpha-beta ratio was significantly increased in both WNS 2.4 g and 4.8 g treatment groups.
## Resting eeg asymmetry
The analysis carried out on EEG power spectral content showed differential effects for each EEG band with marked topographical differences. As depicted in [fig_ref] Figure 7: Asymmetry values with depressive disorder after WNS and placebo treatment [/fig_ref] , the alpha-beta ratio was significantly increased in both WNS 2.4 g and 4.8 g treatment groups.
## Resting eeg asymmetry
The analysis carried out on EEG power spectral content showed differential effects for each EEG band with marked topographical differences. As depicted in [fig_ref] Figure 7: Asymmetry values with depressive disorder after WNS and placebo treatment [/fig_ref] , the alpha-beta ratio was significantly increased in both WNS 2.4 g and 4.8 g treatment groups.
# Discussion
# Discussion
Depression is a common and serious mood disorder that is a worldwide problem for humans because of its significant association with disorders such as sleep disturbances, low self-esteem, guilty feelings, and suicidal tendencies [bib_ref] Mental health service use among South Africans for mood, anxiety and substance..., Seedat [/bib_ref]. However, depression is considered a complex disorder. Mechanisms underlying its pathogenesis remain unclear. Existing treatments for depression are unsatisfactory. Searching for an effective treatment for patients suffering from depression is a major challenge. The aim of this study was to verify antidepressant effects of NS on symptoms of depression and changes caused on EEG frontal asymmetry.
Nelumbinis semen has been recognized as a traditional natural medicine as a remedy. Although NS effects and mechanisms are still largely unknown, NS has been used to cure diverse mental diseases in traditional medicine. Several studies have reported the pharmacological efficacy of NS in various disease models. A recent research study has suggested that NS possesses pharmacological effects and various activities, such as hypoglycemic, antidiarrheal, antimicrobial, diuretic, antipyretic, anti-inflammatory, hepatoprotective, anti-proliferative, and antioxidant activities. Several studies have also reported anti-depression effects of NS in various depressant animal models. In previous studies, we have found that NS has a clear antidepressant effect in that it can reduce the immobility time of rats in forced swim tests [bib_ref] The anti-depressant effect of Nelumbinis semen on rats under chronic mild stress..., Kang [/bib_ref] and reverse decreases of sucrose intake and serotonin (5-HT)1A receptor binding in the hippocampus (5-HT1A hetero-receptors) induced by CMS. NS contains diverse kinds of alkaloids [bib_ref] Alkaloids of plumula Nelumbinis, Wang [/bib_ref]. Among these alkaloids, neferine is known for its anti-arrhythmic actions [bib_ref] Effects of neferine on heart electromechanical activity in anaesthetized cats, Li [/bib_ref] and tranquilization action [bib_ref] Specific binding of kainic acid to purified subcellular fractions from rat brain, Nieto-Sampedro [/bib_ref] , while isoquercetin is known to possess spasmolytic action [bib_ref] Quercetin glycosides in Psidium guajava L. leaves and determination of a spasmolytic..., Lozoya [/bib_ref]. In particular, anonaine, asimilobine, isoquercetin, hyperoside, lirinidine, and nornuciferine elements of NS are known to have anti-depression effects as they can increase 5-HT concentration by neurotransmitter reuptake inhibition and increase 5-HT receptor-binding potential [bib_ref] Flavonoids from Hypericum perforatum show antidepressant activity in the forced swimming test, Butterweck [/bib_ref] [bib_ref] Nelumbinis Semen reverses a decrease in hippocampal 5-HT release induced by chronic..., Kang [/bib_ref] [bib_ref] Isoquinoline derivatives isolated from the fruit of Annona muricata as 5-HTergic 5-HT1A..., Hasrat [/bib_ref] [bib_ref] Asimilobine and lirinidine, serotonergic receptor antagonists, from Nelumbo nucifera, Shoji [/bib_ref]. These effects might be mediated by components in NS such as serotonin receptor agonists anonaine and asimilobine [bib_ref] Alkaloids of Nelumbo lutea (Wild.) pers. (Nymphaeaceae), Zelenski [/bib_ref].
Our previous studies have reported that NS might possess an antidepressant effect by enhancing serotonin level.
Previous work has also suggested that quantitative sleep EEG can predict clinical responses to antidepressants [bib_ref] Comment: What's Basic About the Brain Mechanisms of Emotion?, Ledoux [/bib_ref]. However, whether underlying mechanisms of these states could be applied to EEG signals remains unclear. Some researchers have suggested a need to pay more attention to the prefrontal cortex [bib_ref] Covariations of EEG asymmetries and emotional states indicate that activity at frontopolar..., Papousek [/bib_ref]. A previous study has stated that the EEG of a person under stress displays a decrease in alpha [bib_ref] The anti-depressant effect of Nelumbinis semen on rats under chronic mild stress..., Kang [/bib_ref] [bib_ref] Mental health service use among South Africans for mood, anxiety and substance..., Seedat [/bib_ref] activity and an increase in EEG amplitude in [bib_ref] Isoquinoline derivatives isolated from the fruit of Annona muricata as 5-HTergic 5-HT1A..., Hasrat [/bib_ref] [bib_ref] Asimilobine and lirinidine, serotonergic receptor antagonists, from Nelumbo nucifera, Shoji [/bib_ref] [bib_ref] Alkaloids of Nelumbo lutea (Wild.) pers. (Nymphaeaceae), Zelenski [/bib_ref] [bib_ref] Comment: What's Basic About the Brain Mechanisms of Emotion?, Ledoux [/bib_ref] Hz and high beta (23-35 Hz) ranges [bib_ref] Rumination in posttraumatic stress disorder, Michael [/bib_ref]. The present study showed that after intervention with NS, the ratio of α/β activity was increased by 51% in the 4.8 g treatment group and 26% in the 2.4 g treatment group compared to that in the placebo group. Our findings provide evidence for the assumed antidepressant effect of NS by changing the ratio of α/β activity in EEGs of the frontal cortex.
BDI and SRI are appropriate and convincing evaluations used to detect many symptoms of depression and anxiety, respectively [bib_ref] Psychometric properties of the Beck Depression Inventory-II in low-income, African American suicide..., Joe [/bib_ref]. Both study groups showed enhanced scores from baseline using both assessment tools. Taking WNS treatment reduced BDI and SRI mean scores in these patients. To the best of our knowledge, this is the first study to report the effect of WNS treatment on clinical status of patients with depression. The 2.4 g WNS group showed greater improvements based on SRI. Our results showed that WNS significantly reduced BDI and SRI mean scores in depressed patients, suggesting that decreases in BDI and SRI mean scores might be related to the therapeutic effect of WNS. It should be noted that treatment with WNS 2.4 g resulted in pronounced improvement of BDI and SRI scores and tended to be slightly larger than that produced by WNS 4.8 g. The exact reasons for these effects were unknown. However, efficacy outcomes from this BDI or SRI suggest that there may be a dose ceiling effect at 2.4 g and this dose may produce a saturating level for depressive symptoms while good drug effects are observed at 2.4 g. There is a need for further research on the use of lower dosages such as 0.6 or 1.2 g/kg in order to find out the most effective and appropriate dosage in the future.
In previous studies, we found that the most effective dose of NS was 400 mg/kg among 100, 200, 400, or 1000 mg/kg of NS in animal studies [bib_ref] Nelumbinis Semen reverses a decrease in hippocampal 5-HT release induced by chronic..., Kang [/bib_ref] [bib_ref] Assessment of anti-depressant effect of nelumbinis semen on rats under chronic mild..., Chung [/bib_ref]. Based on these animal studies, 2.4 g and 4.8 g/kg were selected in the present study.
Although our study has some limitations, this is the first study that evaluated the association between NS and depression. In addition, assessment of the ratio of α/β activity in the EEG could be used for the estimation of depressive symptoms. In addition, the molecular mechanism underlying the antidepressant effect of WNS requires further study before firm conclusions can be drawn. In summary, treatment with WNS can attenuate depression in depressive patients.
# Conclusions
In conclusion, the anti-stress effects of WNS are associated with reduced mean scores of BDI and SRI and increased ratios of α/β activity in the EEG of patients with depression. These results support the development of a novel antidepressant established on the pharmacological action of nelumbinis semen. Based on the results of the present study, we conclude that WNS has potential for treating depression.
[fig] of 11, Figure 1: Int. J. Environ. Res. Public Health 2022, 19, x FOR PEER REVIEW 5 The study flowchart. [/fig]
[fig] Figure 1: The study flowchart. [/fig]
[fig] Figure 2: Differences in BDI mean scores between WNS and placebo groups. (A) Placebo group, (B) 2.4 g of WNS group, (C) 4.8 g of WNS group. [/fig]
[fig] Figure 3: Differences in BDI subscale A mean scores of WNS and placebo groups. (A) Placebo group, (B) 2.4 g of WNS group, (C) 4.8 g of WNS group. [/fig]
[fig] Figure 4: Differences in BDI subscale C mean scores between WNS and placebo groups. (A) Placebo group, (B) 2.4 g of WNS group, (C) 4.8 g of WNS. [/fig]
[fig] Figure 5: Differences in SRI mean scores between WNS and placebo groups. (A) Placebo group, (B) 2.4 g of WNS group, (C) 4.8 g of WNS group. [/fig]
[fig] Figure 6: Differences in SRI E mean scores between WNS and placebo groups. (A) Placebo group, (B) 2.4 g of WNS group, (C) 4.8 g of WNS group. [/fig]
[fig] Figure 7: Asymmetry values with depressive disorder after WNS and placebo treatment. (A) 2.4 g of WNS group, (B) placebo group, (C) 4.8 g of WNS. [/fig]
[fig] Author: Contributions: Conceptualization, M.Y., H.B. and I.S.; Data curation, M.Y. and S.P.; Formal analysis, M.Y.; Investigation, H.B., K.S.K. and I.S.; Methodology, M.Y., K.S.K. and I.S.; Project administration, H.B., K.S.K. and I.S.; Software, M.Y.; Validation, M.Y. and J.L.; Writing-original draft, M.Y.; Writing-review & editing, M.Y., K.S.K. and I.S. All authors have read and agreed to the published version of the manuscript. Funding: This research was supported by a grant from Kyung Hee University in 2021 (KHU-20210141), the National Research Foundation of Korea (NRF-2021R1A2C1093825) and VTBIO Co., LTD. Institutional Review Board Statement: The permission to perform this study was granted by the Ethical Review Board of Human Research of the College of Medicine at Catholic University (Seoul, Korea) after ethical approval by ethical committee with code No. KCMC07MS274. [/fig]
[table] Table 1: Characteristics of randomized subjects. [/table]
[table] Table 2: Changes in total BDI scores and BDI subscale scores in patients after WNS treatment for two weeks. [/table]
[table] Table 3: Changes in total SRI scores and SRI subscale scores in patients after WNS treatment for two weeks. [/table]
|
The breast cancer somatic 'muta-ome': tackling the complexity
ER = estrogen receptor; GI = genomic instability; HD = homozygous deletion; SNP = single-nucleotide polymorphism.
Acquired somatic mutations are responsible for approximately 90% of breast tumours. However, only one somatic aberration, amplification of the HER2 locus, is currently used to define a clinical subtype, one that accounts for approximately 10% to 15% of breast tumours. In recent years, a number of mutational profiling studies have attempted to further identify clinically relevant mutations. While these studies have confirmed the oncogenic or tumour suppressor role of many known suspects, they have exposed complexity as a main feature of the breast cancer mutational landscape (the 'muta-ome'). The two defining features of this complexity are (a) a surprising richness of low-frequency mutants contrasting with the relative rarity of high-frequency events and (b) the relatively large number of somatic genomic aberrations (approximately 20 to 50) driving an average tumour. Structural features of this complex landscape have begun to emerge from follow-up studies that have tackled the complexity by integrating the spectrum of genomic mutations with a variety of complementary biological knowledge databases. Among these structural features are the growing links between somatic gene disruptions and those conferring breast cancer risk, mutually exclusive coexistence and synergistic mutational patterns, and a clearly nonrandom distribution of mutations implicating specific molecular pathways in breast tumour initiation and progression. Recognising that a shift from a gene-centric to a pathway-centric approach is necessary, we envisage that further progress in identifying clinically relevant genomic aberration patterns and associated breast cancer subtypes will require not only multi-dimensional integrative analyses that combine mutational and functional profiles, but also larger profiling studies that use second-and third-generation sequencing technologies in order to fill out the important gaps in the current mutational landscape.
## The copy-number 'muta-ome'
The most prominent feature in the breast cancer copynumber muta-ome is amplification of the HER2 locus, present in about 10% to 15% of all breast tumours. It is remarkable that since the discovery of this amplification no further ERBB2-like oncogene has been conclusively identified.
Although two recent large-scale (145 and 171 tumours) genome-wide profiling studies combining high-resolution copy-number and matched gene expression data have confirmed candidate oncogenes in well-known regions of recurrent amplification (notably, 8p12, 8q24, 11q13-14, 17q21-24, and 20q13), none of these appears to be as frequently amplified as ERBB2 and they rarely exhibit amplification profiles that clearly point at a specific genomic location or target [bib_ref] Genomic and transcriptional aberrations linked to breast cancer pathophysiologies, Chin [/bib_ref] [bib_ref] High-resolution acgh and expression profiling identifies a novel genomic subtype of er..., Chin [/bib_ref]. Instead, the amplification profiles are complex and multi-modal, suggesting that multiple targets may coexist within these regions. This identification problem is compounded by the fact that a relatively high proportion of variation at the gene expression level (approximately 20%) is driven by copy-number changes; thus, focusing on regions of expression bias that are driven by underlying amplifications (so-called 'hotspots') still leaves an unmanageably large number of targets. Nevertheless, by focusing within these hotspots on genes that are also druggable, Chin and colleagues [bib_ref] Genomic and transcriptional aberrations linked to breast cancer pathophysiologies, Chin [/bib_ref] prioritised a smaller set of eight targets, FGFR1, IKBKB, PROSC, ADAM9, FNTA, ACACA, PNMT, and NR1D1, including ERBB2. Confirming the robustness of these findings, all of these were also found to reside in amplification hotspots in an independent breast cancer cohort [bib_ref] High-resolution acgh and expression profiling identifies a novel genomic subtype of er..., Chin [/bib_ref] (A.E. Teschendorff and C. Caldas, unpublished data). In spite of this agreement, the two studies were discordant when hotspots were associated with clinical outcome, mirroring the disagreements of initial gene expression studies. Thus, whereas in [bib_ref] Genomic and transcriptional aberrations linked to breast cancer pathophysiologies, Chin [/bib_ref] associations with survival and recurrence were restricted to the amplicons on 8p11-12 and 17q11-12, in [bib_ref] High-resolution acgh and expression profiling identifies a novel genomic subtype of er..., Chin [/bib_ref] outcome-associated amplicons were found on 8q22.3, 8q24.3, 8q24.11-13, and 11q14. As explained in [bib_ref] High-resolution acgh and expression profiling identifies a novel genomic subtype of er..., Chin [/bib_ref] , this disagreement is most likely due to substantial differences in the clinical characteristics of the two cohorts, yet the lessons learned from mRNA expression microarray studies also suggest that larger studies Viewpoint The breast cancer somatic 'muta-ome': tackling the complexity Andrew E Teschendorff 1 and Carlos Caldas 2,3 may help to resolve such discrepancies. In fact, given the increased complexity of genomic breast cancer profiles relative to mRNA profiles, much larger sample sizes (approximately 500 to 1,000) might be needed before substantial overlap between prognostic copy-number signatures is found. In addition, given that among copynumber aberrations it is high-level amplifications that seem to carry most of the prognostic significance [bib_ref] Genomic and transcriptional aberrations linked to breast cancer pathophysiologies, Chin [/bib_ref] [bib_ref] High-resolution acgh and expression profiling identifies a novel genomic subtype of er..., Chin [/bib_ref] , the development of a future prognostic copy-number signature, similar to the expression signatures currently being tested in clinical trials, would benefit from higher-resolution studies such as those using single-nucleotide polymorphism (SNP) arrays or next-generation sequencing technologies, as these have the capability to detect more focal (that is, less than 20 to 200 Kb) amplifications and homozygous deletions (HDs). Indeed, a recent high-resolution 317,503 to 555,351 SNP array study of 45 breast tumours was able to detect a larger repertoire of copy-number amplifications and HDs, some smaller than 250 Kb [bib_ref] Integrated analysis of homozygous deletions, focal amplifications, and sequence alterations in breast..., Leary [/bib_ref]. Thus, in addition to identifying focal aberrations encompassing known oncogenes (CCND1, CCNE1, and FGFR2) and tumour suppressors (CDKN2A and PTEN), they uncovered a number of other important genes with likely oncogenic or tumour suppressor roles (PCDH8, MRE11A, and HOXA3) [bib_ref] Integrated analysis of homozygous deletions, focal amplifications, and sequence alterations in breast..., Leary [/bib_ref]. The higher-resolution SNP array also allowed Leary and colleagues [bib_ref] Integrated analysis of homozygous deletions, focal amplifications, and sequence alterations in breast..., Leary [/bib_ref] to estimate an average number of 18 copy-number aberrations (7 HDs and 11 high-level amplifications) that were altered significantly above the background mutation rate, implicating about 24 genes that are driving tumour progression.
Larger-scale profiling studies are also needed in order to more fully characterise the genomic landscape of breast cancer. So far, four main genomic subtypes that roughly correlate with the well-known intrinsic subtypes defined from gene expression studies have been identified [bib_ref] Genomic and transcriptional aberrations linked to breast cancer pathophysiologies, Chin [/bib_ref] [bib_ref] High-resolution acgh and expression profiling identifies a novel genomic subtype of er..., Chin [/bib_ref]. Specifically, one subtype (called 'simple') consisted almost entirely of estrogen receptor-positive (ER + )/luminal-A tumours and is characterised mainly by a relatively simple genomic profile defined by a gain of 1q and 16p and a simultaneous loss of 16q. Another genomic subtype ('simple amplifier') was characterised mainly, but not exclusively, by amplifications on 11q13-14 and 17q11-13 and mapped roughly to the ER + /luminal-A and ER -/HER2 + expression subtypes, respectively. Thus, amplifications on 11q13-14 and 17q11-13 occurred in a largely mutually exclusive fashion, suggesting that these tumours are using distinct oncogenic mechanisms. A third genomic subtype ('complex amplifier') showed the highest degree of genomic instability (GI) and complex rearrangements, including frequent amplifications at the 8q24 and 8p12 loci. This genomic subtype is that of worst prognosis, consisting mainly of ER -/basal and ER + /luminal-B tumours. A fourth genomic subtype consisted mainly of ER -/basal tumours but was characterised, surprisingly, by a low GI profile ('flat') [bib_ref] High-resolution acgh and expression profiling identifies a novel genomic subtype of er..., Chin [/bib_ref]. It is doubtful, however, whether such classifications will have much relevance as it is likely that the clinically relevant phenotypes are determined by specific combinations of different types of somatic mutations. Thus, to identify breast cancer genomic subtypes that are more relevant, we envisage that it will be necessary to derive and analyse multi-dimensional mutational profiles.
## The point 'muta-ome'
Two recent landmark sequencing studies [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] have also revealed the remarkable complexity of the point mutational landscape of breast cancer. In addition to the high-frequency somatic point mutations in TP53 (53%), PIK3CA (26%), and CDH1 (21%) (the so-called gene 'mountains'), these studies have confirmed a surprisingly larger number of other genes that also appear to be more frequently mutated than what can be accounted for by chance, albeit at much lower frequencies than TP53 or PIK3CA. Specifically, for the 11 breast tumours considered in the discovery screen of [bib_ref] The genomic landscapes of human breast and colorectal cancers, Wood [/bib_ref] , a total of 1,137 RefSeq genes contained a somatic mutation and 167 of these were also validated in an independent screen of 24 samples. When a number of different estimation procedures for passenger mutation rates were used, it was found that approximately 120 genes were more frequently mutated than the passenger rate, suggesting that these genes are more likely to carry driver mutations (so-called CAN genes). This translates into an average of 101 non-synonymous somatic mutations in RefSeq genes per breast tumour, and approximately 14 of these are thought to be located in CAN genes [bib_ref] The genomic landscapes of human breast and colorectal cancers, Wood [/bib_ref]. It could be argued that this is a gross overestimate due to the highly tumourigenic nature of the breast samples considered (all were ERcell lines mostly representing metastases); however, it is also plausible that it would increase as larger-scale sequencing studies get completed. Combining the published estimate with the average number of 24 driver amplifications/HDs affecting any given breast tumour and assuming mutual exclusivity of mutation type, this would implicate an average of approximately 40 CAN genes in any given tumour.
Given that estimates of passenger mutation rates are only approximate and that a high mutation rate may reflect mechanisms that are only indirectly related to tumour genesis or progression, it could be expected that many of the identified CAN genes are false-positives. Nevertheless, several of the CAN genes identified in [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] (notably, IKBKB, IKBKA, CHD5, STK11, STK6, and BRAF) have been independently implicated in breast or other tumours, whereas other genes (for example, ATM and FGFR2) have been associated with germline mutations and an increased risk of breast cancer. Moreover, in an independent bioinformatics study, several breast cancer CAN genes (GAB1, NLE1, and CNTN6) were rediscovered as members of protein interaction subnetworks with combined expression levels that correlated with clinical outcome [bib_ref] Network-based classification of breast cancer metastasis, Chuang [/bib_ref] , consistent with the idea that mutations in these genes are disrupting important signalling pathways and thus fuelling malignant progression. Taken together, these observations provide strong support for the direct causal role of the identified CAN genes in tumour progression and, for some, probably tumourigenesis also.
## Massively parallel sequencing: a fusion 'muta-ome'?
The emerging pattern of a few highly prevalent CAN genes ('mountains') with a much larger number of low-frequency events ('hills') parallels the aberration landscape pattern observed at the copy-number level. Thus, a question posed by these studies is whether the relative scarcity of cancer gene 'mountains' is a genuine feature of the cancer muta-ome or whether it merely reflects the inability of the technology to detect further high-frequency mutations. While the latter scenario seems to be highly unlikely in the case of point mutations [bib_ref] The genomic landscapes of human breast and colorectal cancers, Wood [/bib_ref] , more focal (=20 to 50 kB) amplifications and deletions will be discovered using higher-resolution SNP arrays and massively parallel sequencing, as already demonstrated with SNP arrays by Leary and colleagues [bib_ref] Integrated analysis of homozygous deletions, focal amplifications, and sequence alterations in breast..., Leary [/bib_ref] , yet it also seems unlikely that these will constitute novel gene 'mountains'. Another possibility is that the missing highfrequency events involve other types of genomic rearrangements such as balanced translocations or inversions, which are not detectable using copy-number analysis but which may be discovered using massively parallel paired-end sequencing or transcriptome sequencing, as illustrated recently in lung cancer [bib_ref] Identification of somatically acquired rearrangements in cancer using genome-wide massively parallel paired-end..., Campbell [/bib_ref] , prostate cancer [bib_ref] Transcriptome sequencing to detect gene fusions in cancer, Maher [/bib_ref] , and breast cancer [bib_ref] Transcriptomeguided characterization of genomic rearrangements in a breast cancer cell line, Zhao [/bib_ref] [bib_ref] A sequence-level map of chromosomal breakpoints in the MCF-7 breast cancer cell..., Hampton [/bib_ref] cell lines. Using a combined strategy of short-and long-read transcriptomic sequencing of a metastatic prostate cancer cell line, Maher and colleagues [bib_ref] Transcriptome sequencing to detect gene fusions in cancer, Maher [/bib_ref] identified a variety of mechanistically different and novel gene fusions, including a novel read-through gene fusion chimaera SLC45A3-ELK4, which interestingly was also found to be aberrantly expressed in 7 out of 20 (35%) metastatic prostate cancer cell lines, 6 of which were negative for ETS gene fusions. Similarly, longread transcriptomic sequencing of a breast cancer cell line (HCC1954) identified not fewer than seven chimaeric transcripts arising mostly from intragenic to intergenic gene fusion events leading to protein truncation. Notably, among the seven rearrangements, one, t(4;11)(q32;q21), causes truncation of the DNA-binding domain of MRE11A, a candidate tumour suppressor involved in DNA damage repair, whereas two other fusions, t(5;8)(q35.3;q24.21) and t(5;8)(p15.33;q24.21), implicate the known cancer genes NSD1 and PVT1 and the 8q24.21 breast cancer susceptibility region. Although these findings are very encouraging, large-scale sequencing studies will be needed to evaluate the prevalence of these genomic fusion aberrations in breast cancer. If these studies were to confirm that specific inter-and intra-chromosomal gene fusions are highly frequent events in breast cancer (as already demonstrated in prostate cancer), it would vindicate the longheld belief that this type of rearrangement constitutes the most prevalent category of somatic aberrations in all tumours.
## Tackling the complexity: structural features of the 'muta-ome'
The broad picture emerging from the large-scale lowresolution array-based comparative genomic hybridisation studies [bib_ref] Genomic and transcriptional aberrations linked to breast cancer pathophysiologies, Chin [/bib_ref] [bib_ref] High-resolution acgh and expression profiling identifies a novel genomic subtype of er..., Chin [/bib_ref] and the smaller-scale higher-resolution sequen-cing studies [bib_ref] The genomic landscapes of human breast and colorectal cancers, Wood [/bib_ref] [bib_ref] Transcriptomeguided characterization of genomic rearrangements in a breast cancer cell line, Zhao [/bib_ref] [bib_ref] A sequence-level map of chromosomal breakpoints in the MCF-7 breast cancer cell..., Hampton [/bib_ref] is that of a highly complex genomic landscape, not only in terms of the sheer numbers of genes that may be directly implicated, but also in terms of the different types and flavours of genomic rearrangements that seem to be causally involved. Clearly, building a vast catalogue of somatic mutations in breast cancer [bib_ref] The catalogue of somatic mutations in cancer (cosmic), Forbes [/bib_ref] constitutes only the first step on the long road to identifying novel drug targets and developing therapies that are more effective. Although the catalogue is still incomplete, it is comprehensive enough that various structural features are starting to emerge.
First, point mutations that are activating or inactivating often occur in genes that are also commonly amplified or deleted [bib_ref] Integrated analysis of homozygous deletions, focal amplifications, and sequence alterations in breast..., Leary [/bib_ref] [bib_ref] Somatic mutations affect key pathways in lung adenocarcinoma, Ding [/bib_ref]. This observation is important for it identifies genes that are disrupted by multiple, and possibly equivalent, mechanisms and that therefore are more likely to drive tumour progression. Thus, it may be possible to categorise all somatic mutations into activating or inactivating aberrations and to carry this simplified picture forward when analysing the systems-level implications of these aberrations. A similar approach was taken by Leary and colleagues [bib_ref] Integrated analysis of homozygous deletions, focal amplifications, and sequence alterations in breast..., Leary [/bib_ref] , who combined the point and copy-number mutomes to identify pathways targeted for disruption.
A second structural feature to emerge has been the mutual exclusivity pattern of mutations [bib_ref] High-throughput oncogene mutation profiling in human cancer, Thomas [/bib_ref] , often involving genes in a common pathway or genes exhibiting high sequence similarity [bib_ref] A multidimensional analysis of genes mutated in breast and colorectal cancers, Lin [/bib_ref] , as well as cooperative or synergistic interactions, often involving physically interacting proteins such as the SMAD2 and SMAD3 receptors in colorectal cancer [bib_ref] Patterns of somatic mutation in human cancer genomes, Greenman [/bib_ref]. In the case of breast cancer, there are some clear mutual exclusivity patterns emerging (for example, those of PIK3CA and TP53 mutations; A.E. Teschendorff and C. Caldas, unpublished data) and independent hints of further patterns such as those of PIK3CA and RAS [bib_ref] High-throughput oncogene mutation profiling in human cancer, Thomas [/bib_ref] [bib_ref] Pi3k pathway alterations in cancer: variations on a theme, Yuan [/bib_ref] as well as hints of cooperation or coexistence such as that involving HER2 amplification and PTEN loss (including loss of PTEN expression) [bib_ref] Pi3k pathway alterations in cancer: variations on a theme, Yuan [/bib_ref]. Although in most of these studies sample sizes are still too small for the results to be conclusive, the strong mutual exclusivity and cooperative patterns observed in, for example, a sequencing study of 188 lung tumours [bib_ref] Somatic mutations affect key pathways in lung adenocarcinoma, Ding [/bib_ref] suggest that similar mutational patterns will be seen in breast cancer.
Finally, a third feature to emerge is the evidently non-random distribution of mutations and affected genes in relation to specific protein domains, gene functions, and molecular pathways. In the case of somatic mutations, a strong association with spectric repeat and fibronectin domains, GTPase activation, extracellular matrix, and cell-cell adhesion as well as calcium ion-binding functions was observed [bib_ref] A multidimensional analysis of genes mutated in breast and colorectal cancers, Lin [/bib_ref] [bib_ref] Functional classification analysis of somatically mutated genes in human breast and colorectal..., Chittenden [/bib_ref]. This reinforces the crucial role that cytoskeletal, including cell-cell adhesion, dysregulation plays in malignant progression of breast cancer, although it remains to be seen how this may depend on the subtype and stage of the specific tumours studied [bib_ref] A multidimensional analysis of genes mutated in breast and colorectal cancers, Lin [/bib_ref] [bib_ref] Functional classification analysis of somatically mutated genes in human breast and colorectal..., Chittenden [/bib_ref]. In the context of pathways, somatic mutations were found to be enriched in many pathways known to be pathogenic in breast cancer, including interferon signalling, cell-cycle checkpoint (G 1 -S), BRCA1/2 (breast cancer 1/2, early onset)-related DNA repair, and phosphatidylinositol 3 kinase (PIK3) and v-akt murine thymoma viral oncogene homolog 1 (AKT) and transforming growth factorbeta (TGF-β) signalling [bib_ref] A multidimensional analysis of genes mutated in breast and colorectal cancers, Lin [/bib_ref] [bib_ref] Functional classification analysis of somatically mutated genes in human breast and colorectal..., Chittenden [/bib_ref]. In an integrative strategy, incorporating both point and copy-number aberrations in the enrichment analysis, Leary and colleagues [bib_ref] Integrated analysis of homozygous deletions, focal amplifications, and sequence alterations in breast..., Leary [/bib_ref] identified further important signalling pathways such as those of Notch, epidermal growth factor receptor (EGFR), fibroblast growth factor (FGF), and v-erb-b2 erythroblastic leukemia viral oncogene homolog 2 (ERBB2). Thus, these results should help to direct increasingly focused efforts that aim to develop therapies tailored to restoring these disrupted pathways. The study of Lin and colleagues [bib_ref] A multidimensional analysis of genes mutated in breast and colorectal cancers, Lin [/bib_ref] also illustrates nicely the added insights gained by adopting a more systems-level approach, mapping mutations onto well-curated protein interaction networks in order to identify subnetworks more prone to disruption. Specifically, they were able to show that genes targeted by point mutations are more prone to occupy hubs in the corresponding protein interaction networks as opposed to genes that are not. Interestingly, of the 83 somatically mutated proteins in breast cancer [bib_ref] Velculescu VE: The consensus coding sequences of human breast and colorectal cancers, Sjöblom [/bib_ref] , over half (59) were shown to be part of a large interaction cluster involving TP53, BRCA1, PIK3R1, and NFKB. Thus, a full interpretation and understanding of how the approximately 30 to 40 CAN genes driving a given tumour cooperate to disrupt important signalling pathways, that themselves exhibit significant cross-talk, will undoubtedly benefit from networkbased approaches that try to give meaning to the aberrations in the context of the whole interaction network. Some further insights in this direction were recently offered by Taylor and colleagues [bib_ref] Dynamic modularity in protein interaction networks predicts breast cancer outcome, Taylor [/bib_ref] , who showed that tumours are preferentially disrupted at 'intermodular' hubs, representing proteins that mediate signals between different cellular functions.
The structural features of the complex muta-ome are only beginning to emerge, and full elucidation of these features will be crucial for translating the vast and growing catalogue of somatic mutations into improvements for therapy and eventual cures. Given the heterogeneity of breast cancer, accomplishing this goal will require upscaling the current Genome Atlas Research Projectto generate multidimensional high-resolution mutational and functional profiles of hundreds if not thousands of tumours.
(page number not for citation purposes)
AcknowledgementsAET is supported by a Heller Research Fellowship.Competing interestsThe authors declare that they have no competing interests. |
Examining the readiness of best evidence in medical education guides for integration into educational practice: A meta-synthesis
Background To support evidence-informed education, health professions education (HPE) stakeholders encourage the creation and use of knowledge syntheses or reviews. However, it is unclear if these knowledge syntheses are ready for translation into educational practice. Without understanding the readiness, defined by three criteria-quality, accessibility and relevance-we risk translating weak evidence into practice and/or providing information that is not useful to educators. Methods A librarian searched Web of Science for knowledge syntheses, specifically Best Evidence in Medical Education (BEME) Guides. This meta-synthesis focuses on BEME Guides because of their explicit goal to inform educational practice and policy. Two authors extracted data from all Guides, guided by the 25-item STructured apprOach to the Reporting In healthcare education of Evidence Synthesis (STORIES). Results Forty-two Guides published in Medical Teacher between 1999 and 2017 were analyzed. No Guide met all STORIES criteria, but all included structured summaries and most described their literature search (n = 39) and study inclusion/exclusion (n = 40) procedures. Eleven Guides reported the presence of theory and/or educational principles, and eight consulted with external subject matter experts. Accessibility to each Guide's full-text and supplemental materials was variable. Discussion For a subset of HPE knowledge syntheses, BEME Guides, this meta-synthesis identifies factors that support readiness and indicates potential areas of improvement, such as consistent access to Guides and inclusion of external subject matter experts on the review team. This analysis is useful for understanding the current readiness of HPE knowledge syntheses and informing future reviews to evolve so they can catalyze translation of evidence into educational practice.What this paper addsBy analyzing BEME Guides, a collection of HPE knowledge syntheses, this meta-synthesis identifies and characterizes factors, related to the quality, accessibility, and relevance of the Guides, that support readiness for translation into educational practice. Our findings highlight positive
# Introduction
In health professions education (HPE), researchers have argued that knowledge syntheses are as important as primary studies [bib_ref] Enhancing the impact of BEME systematic reviews on educational practice, Gordon [/bib_ref]. Thus, it is unsurprising that their numbers have been on the rise, with most HPE journals accepting knowledge syntheses, and foundations and professional groups offering funding for their creation (e.g., the Gold Foundation). Additionally, for almost 20 years, the Best Evidence in Medical Education (BEME) Collaboration has supported HPE researchers in conducting and disseminating knowledge syntheses, which are known as BEME Guides.
Despite support for their creation and swelling numbers [bib_ref] Evidence-based arguments in support of medical education reform, Fernandez [/bib_ref] , knowledge syntheses, especially systematic reviews, have been criticized not only by those who try to use them, but also by those who publish them [bib_ref] Enhancing the impact of BEME systematic reviews on educational practice, Gordon [/bib_ref] [bib_ref] Best Evidence Medical Education (BEME) review process redesigned for methodological rigor, shorter..., Corral [/bib_ref] [bib_ref] Supporting evidence-informed teaching in biomedical and health professions education through knowledge translation:..., Tractenberg [/bib_ref] [bib_ref] The mass production of redundant, misleading, and conflicted systematic reviews and meta-analyses, Ioannidis [/bib_ref]. Norman argued that many reviews end up as exercises in 'bean counting' and often fail to provide conclusive, usable evidence for practitioners. Furthermore, Gordon suggested that systematic reviews often lack relevance and fail to incorporate educational elements critical for educational practice and policy [bib_ref] Enhancing the impact of BEME systematic reviews on educational practice, Gordon [/bib_ref]. Despite these criticisms, our field has yet to evaluate HPE knowledge syntheses to determine whether or not they are ready for translation into educational practice and policy.
HPE is not unique in its lack of understanding about the readiness of knowledge syntheses [bib_ref] Enhancing the usability of systematic reviews by improving the consideration and description..., Hoffmann [/bib_ref]. In clinical medicine, for example, Glasziou reported that systematic reviews of non-pharmacological interventions lacked details to translate findings into practice, such that it was impossible to determine which versions of treatments to use [bib_ref] What is missing from descriptions of treatment in trials and reviews?, Glasziou [/bib_ref]. More recently, a study of over 50 systematic reviews of stroke interventions identified that 80% were missing key details, such as intervention procedures and materials . In education, this might be akin to a knowledge synthesis evaluating the use of simulation, but not clarifying the strategies that may be deployed for optimal teaching and assessment within simulation.
Without understanding the readiness of our knowledge syntheses, defined by three criteria-quality, accessibility and relevance-we risk translating weak evidence into practice and/or providing information to educators that is not usable, both of which have the potential to frustrate teachers and learners and impair teaching and learning practices. Thus, this article is a first attempt to identify and characterize factors that support the readiness of HPE knowledge syntheses for translation into educational practice. In this study we chose to focus on BEME Guides because they are 'designed to assist individual teachers, institutions, and national bodies to make informed decisions about educational practice and policy'. In other words, Guides are explicitly meant to inform educational practice and policy and are described as 'user-friendly' such that practitioners can assess and apply them in a manner appropriate for their own criteria and context.
# Method
We analyzed HPE knowledge syntheses, specifically BEME Guides (hereafter referred to as 'Guides'), to estimate their readiness for integration into practice. While we recognize that the Guides comprise a minority of HPE knowledge syntheses and that not all knowledge syntheses aim to address the knowledge-to-practice gap, we focused on the Guides based on their specific mission to inform practice and policy and to be useful to teachers, researchers and policy makers. Additionally, we focused on Guides because their authors are required to follow structured instructions for the creation of the review, which we believed would allow us to draw comparisons across the Guides.
We convened an author team of experts in HPE, knowledge syntheses, knowledge translation, clinical medicine, information science, scholarly communication, and health policy. Our team also has experience as researchers and as teachers of health professional learners in classrooms and clinics. The research team was convened to maximize expertise in HPE educational practice and research, as well as expertise in knowledge synthesis.
To identify Guides, on 1 May 2017, we searched Web of Science using 'BEME' as a keyword. We downloaded the identified citations into an Excel spreadsheet. We also checked the BEME website (https://www. bemecollaboration.org/) to confirm that we had retrieved all the Guides. Based on titles and abstracts, we excluded non-review citations (e.g., articles about BEME's future, protocols, or articles on how to write a Guide; see . From Medical Teacher, we obtained the full-text of all included reviews in PDF form. When necessary and as possible, we also obtained the text of online supplements from the Medical Teacher and/or the BEME websites.
Data extraction was guided by the STructured apprOach to the Reporting In healthcare education of Evidence Synthesis (STORIES) Statement [bib_ref] STORIES statement: publication standards for healthcare education evidence synthesis, Gordon [/bib_ref]. The STORIES Statement, which includes 25 elements, is a consensus statement on the baseline reporting requirements for knowledge syntheses. In addition to coding the 25 elements, coders kept descriptive notes with an emphasis on identifying exemplars. We selected the STORIES Statement, instead of a more generic methodological tool, such as AMSTAR [bib_ref] Development of AM-STAR: a measurement tool to assess the methodological quality of..., Shea [/bib_ref] , because STORIES is designed specifically for HPE by HPE researchers and includes items specific to education (e.g., identifying the use of pedagogical strategies). Additionally, it supports the varied question types HPE researchers pose, which often go beyond determining an intervention's effectiveness. For example, one Guide characterized why underperforming learners do not fail [bib_ref] The failure to fail underperforming trainees in health professions education: A BEME..., Yepes-Rios [/bib_ref] and another described what physicians need to know about ecosystems [bib_ref] What do tomorrow's doctors need to learn about ecosystems?-A BEME Systematic Review:..., Walpole [/bib_ref]. Lastly, we selected the STORIES Statement since it does not focus on one specific type of knowledge synthesis but instead is flexible and applicable to all knowledge synthesis types.
Two reviewers extracted data from all the Guides. LM coded all the Guides, while CN, HCC, and NT each independently coded 14. Discussions between the coders occurred before, mid-way through, and after coding completion. At each point, the authors compared codes for shared understanding. If two coders disagreed, a third acted as a tiebreaker. Upon overall consensus, LM compiled all data extractions into a master spreadsheet, which was emailed to the entire study team.
# Results
We identified 42 Guides published in Medical Teacher between 1999 and 2017. Thirty-six Guides were described as systematic reviews , three as reviews [bib_ref] A BEME (Best Evidence in Medical Education) systematic review of the use..., Barrett [/bib_ref] [bib_ref] What features of educational interventions lead to competence in aseptic insertion and..., Cherry [/bib_ref] [bib_ref] The effectiveness of selfassessment on the identification of learner needs, learner activity,..., Colthart [/bib_ref] , two as realist reviews [bib_ref] Tools for structured team communication in pre-registration health professions education: a Best..., Buckley [/bib_ref] [bib_ref] A review of longitudinal community and hospital placements in medical education: BEME..., Thistlethwaite [/bib_ref] , and one as a review with quality grading of articles [bib_ref] Teaching and learning communication skills in medicine-a review with quality grading of..., Aspegren [/bib_ref]. Most authors rationalized conducting their Guides due to a lack of knowledge synthesis covering their specific topic. However, there was limited justification as to why a specific synthesis type was selected and whether or not this selection aligned with the Guide's research question (See Tab. 1 for an overview of the review characteristics; See Tab. 2 for methodological details). To structure the reporting of our findings, we use the concept of readiness criteria as defined by three criteria: quality, accessibility and relevance. These criteria are derived from the first BEME Guide published by Harden.
## Quality
We considered quality in relation to how each Guide fulfilled the 25 elements of the STORIES Statement. All reviews fulfilled the following four criteria: contained a structured summary, provided rationale for conducting the review, summarized details of the included studies, and presented findings in light of stated objectives. The least satisfied criteria (n = 8 Guides) pertained to research teams contacting external individuals familiar with the topic. Below we highlight several key criteria. Detailed information for each criteria is available in Supplement 1: https://figshare. com/s/590bcc368c7609280aec.
## Research questions
All Guide authors but one [bib_ref] Teaching and learning communication skills in medicine-a review with quality grading of..., Aspegren [/bib_ref] stated research questions, generally featured in a Guide's introduction section and/or set off with bullet points. Fifteen guides reported the Pop- [bib_ref] Are journal clubs effective in supporting evidence-based decision making? A systematic review...., Harris [/bib_ref]. However, several Guides addressed how to characterize concepts by asking questions, such as 'What factors have been studied that may be associated to medical-career decision making?' [bib_ref] Dynamics of career choice among students in undergraduate medical courses. A BEME..., Querido [/bib_ref].
## Literature searching
Thirty-nine Guide authors provided a rationale for how they conducted their literature searches. To a slightly lesser degree, authors detailed their search processes (n = 36) and provided the full search terms for at least a single database (n = 34). Information professionals were involved in conducting the searches for 16 Guides.
## Inclusion/exclusion
All Guide authors, except one [bib_ref] Teaching and learning communication skills in medicine-a review with quality grading of..., Aspegren [/bib_ref] , described methods for judging the inclusion and exclusion of identified studies and performed data extraction in duplicate. In 36 guides, authors included a flow diagram, such as recommended by the Preferred Reporting Items of Systematic Reviews and Meta-Analyses (PRISMA) Guidelines [bib_ref] Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement, Moher [/bib_ref].
## Qualitative and quantitative analysis
Twenty-three Guides utilized and rationalized their use of qualitative approaches to synthesizing primary evidence, including thematic analysis, qualitative description, and narrative description. Twenty-nine Guides described quantitative methods used to synthesize primary evidence and discussed how they considered the heterogeneity of included studies. Four Guides included meta-analyses [bib_ref] Which professional (non-technical) competencies are most important to the success of graduate..., Cake [/bib_ref] [bib_ref] BEME systematic review: predictive values of measurements obtained in medical schools and..., Hamdy [/bib_ref] [bib_ref] Impact of an intercalated BSc on medical student performance and careers: a..., Jones [/bib_ref] [bib_ref] The effects of audience response systems on learning outcomes in health professions..., Nelson [/bib_ref]. Heterogeneity of the data was the main reason provided for not performing meta-analyses. For example, Reeves pointed out, 'Due to the heterogeneity of interventions (differing curriculum content, duration of courses, participating professional groups) and study designs (quasiexperimental, exploratory, action-oriented) a meta-analysis of studies was not possible [bib_ref] A BEME systematic review of the effects of interprofessional education: BEME Guide..., Reeves [/bib_ref].'
## Accessibility
We considered accessibility in terms of readers' ability to access, either in hard-copy or online, the full-text of a Guide and its supporting materials. Detailed online accessibility information for each article is available in Supplement 1 (Available at: https://figshare.com/ s/590bcc368c7609280aec.) Additionally, we examined the steps authors took (e.g., inclusion of practice points) to make Guides cognitively accessible for a variety of readers, including busy clinician-educators who may not be engaged in HPE research as a primary duty.
The full-texts of all Guides are available by subscription in Medical Teacher. Of the included Guides, 26 were freely accessible from the BEME website. These free versions tended to be the author's final draft, meaning that readers may encounter documents with author 'track changes' and tables and figures missing or located at the end of the manuscript. On the Medical Teacher website, 11 reviews were freely available as of 24 March 2018, and several of them overlapped with those featured on the BEME website. As of 24 March 2018, 11 Guides were not publicly accessible, all of which were published between 2015-2017, thereby comprising the majority of recently published Guides. For most Guides, components of the methods and results, such as search strategies, data abstraction tools, and results tables, were made available on the BEME and/or Medical Teacher websites with neither location containing all the materials.
Guides included or made available supplemental materials to increase the cognitive accessibility for readers at differing levels of engagement with HPE research. For example, all Guides featured publicly accessible abstracts. Currently, Medical Teacher requires structured abstracts under 200 words. Reviews published from 2008-2017 followed this convention. However, reviews published prior to 2008 were significantly longer and included robust method sections and, in many cases, 'headline results' or 'highlight points' set off with bullet points. These brief points connected readers with key findings for practice, making the material more readily accessible. For example, one review included the practice point 'educational feedback is the most important feature of simulation-based medical education [bib_ref] Features and uses of high-fidelity medical simulations that lead to effective learning:..., Issenberg [/bib_ref].' Additionally, within the body of 35 Guides, authors included practice points. Practice points, generally featured in the Guide's introduction section, are 4-6 bulleted points, offset in a box, designed to provide readers a brief snapshot of major findings. For example one review noted, 'It is key that portfolio implementation is welldesigned and sustained, with high-level organizational support, to ensure uptake' [bib_ref] The effectiveness of portfolios for post-graduate assessment and education: BEME Guide No..., Tochel [/bib_ref] and another that 'The Education for Sustainable Healthcare framework can guide curricula and teaching development [bib_ref] What do tomorrow's doctors need to learn about ecosystems?-A BEME Systematic Review:..., Walpole [/bib_ref]. External to the Guides themselves, 34 authors included a 'BEME Spotlight.' A BEME spotlight is a two-page structured summary of the Guide's main conclusions and recommendations. To host these materials, and others, each Guide has a webpage on the BEME website (https://www. bemecollaboration.org/Published+Reviews/). Page contents vary by Guide, but generally included links to the Guide's full-text (if publicly available), the review protocol, an executive summary, PowerPoint slides from conference presentations, conference posters, and the Medical Teacher website. For example, the BEME site for O'Dunn-Orto's Guide on the musculoskeletal exam [bib_ref] Teaching musculoskeletal clinical skills to medical trainees and physicians: a Best Evidence..., O'dunn-Orto [/bib_ref] features links to its abstract, a publicly-accessible MS Word version of the fulltext, a conference poster, a PowerPoint presentation, and the associated Medical Teacher article. While the BEME and Medical Teacher websites indicate via links and/or references that these materials exist, if a practitioner were to access a Guide via PDF, there is no indication in the PDF document that these supplemental materials exist.
## Relevance
Relevance is the 'degree to which something is related or useful to what is happening or being talked about [58]'. Because relevance has a personal dimension, and because a given teaching approach may be relevant to one educator and not to another, we did not judge each Guide's relevance, per se. Instead, we report on the elements provided that might inform readers in making their own relevance judgments based on their individual practice situations. Elements related to relevance included: implications for researchers and educators, inclusion of stakeholders, inclusion of educational theories and/or models, and currency.
## Implications for researchers and/or educators
Forty Guides included sections focused on implications for future research that were generally located before the Guide's general conclusion. Fewer Guides reported implications for education/practice (n = 27). For example, the Guide on role modelling included the implication: 'Role modelling should be explicit in clinical teaching, as it is important for teachers to make an intentional effort to articulate what aspects they are modelling [bib_ref] Doctor role modelling in medical education: BEME Guide No, Passi [/bib_ref].'
## Inclusion of stakeholders
Authors from a variety of backgrounds participated in the Guides, including, but not limited to, clinicians (n = 37 Guides), librarians (n = 16 Guides), trainees (n = 11 Guides) and statisticians (n = 2). Author teams ranged from a single author [bib_ref] Teaching and learning communication skills in medicine-a review with quality grading of..., Aspegren [/bib_ref] to 15 authors [bib_ref] A BEME systematic review of UK undergraduate medical education in the general..., Park [/bib_ref]. The average author team had six members. Nineteen Guides explicitly described how authors' content and/or methodological expertise benefited the Guide. For example, Issenberg described targeted recruitment of team members with content knowledge across a variety of simulation modalities, skills in educational measurement, and expertise in research methods [bib_ref] Features and uses of high-fidelity medical simulations that lead to effective learning:..., Issenberg [/bib_ref]. External to the author team, 25 Guides engaged subject matter and synthesis experts as consultants. For example, Buckley enlisted a professional translator to translate non-English primary studies and a librarian to design and execute the literature searches [bib_ref] Tools for structured team communication in pre-registration health professions education: a Best..., Buckley [/bib_ref].
## Inclusion of theory and educational principles
Eleven Guide authors reported on theories, conceptual frameworks, and/or educational principles that underpinned the individual studies that they had analyzed. For example, Steinert reported how the studies she included adhered to principles of teaching and learning and cited adult learning and experiential learning as organizing concepts [bib_ref] A systematic review of faculty development initiatives designed to improve teaching effectiveness..., Steinert [/bib_ref]. For Guides that did not mention theory or educational principles, it was generally unclear if they were not reported because they were absent from the individual studies or if the review authors chose not to cover these elements. In one Guide, the author notes that theory was absent from the studies analyzed [bib_ref] A BEME (Best Evidence in Medical Education) systematic review of the use..., Barrett [/bib_ref].
## Currency
The oldest Guide, focused on communication skills, was published in 1999 with its most recent study analyzed in 1998 [bib_ref] Teaching and learning communication skills in medicine-a review with quality grading of..., Aspegren [/bib_ref]. In contrast, the most recent Guide, published in 2017, analyzed studies up until 2014 [bib_ref] Effective methods of teaching and learning in anatomy as a basic science:..., Losco [/bib_ref]. Three Guides had been formally updated [bib_ref] How can experience in clinical and community settings contribute to early medical..., Dornan [/bib_ref] [bib_ref] A systematic review of faculty development initiatives designed to improve teaching effectiveness..., Steinert [/bib_ref]. For example, in one instance, 10 years after initial publication, the author revised her earlier Guide on faculty development [bib_ref] A systematic review of faculty development initiatives designed to enhance teaching effectiveness:..., Steinert [/bib_ref] [bib_ref] A systematic review of faculty development initiatives designed to improve teaching effectiveness..., Steinert [/bib_ref]. The two other Guides were updated 4 [bib_ref] What has changed in the evidence for early experience? Update of a..., Yardley [/bib_ref] and 9 years [bib_ref] A BEME systematic review of the effects of interprofessional education: BEME Guide..., Reeves [/bib_ref] after the original Guides were published. In the revised Guides, the authors explicitly referenced and built upon their earlier work. In 2013, Cherry posted to the BEME website a brief, publicly accessible 'follow-up' report of her review on emotional intelligence 1 year after its publication in Medical Teacher [bib_ref] What impact do structured educational sessions to increase emotional intelligence have on..., Cherry [/bib_ref].
# Discussion
We analyzed 42 BEME Guides to identify and characterize factors that support the readiness of knowledge syntheses for use by HPE practitioners. In our findings, we have highlighted positive practices, such as the universal inclusion of structured summaries. However, despite being considered 'the gold standard' of educational reviews [bib_ref] Best evidence medical and health professional education (BEME) collaboration: a moving spotlight, Gordon [/bib_ref] , no Guide met all the STORIES Statement criteria and there was considerable room for improvement in relation to accessibility and relevance. We now consider readiness criteria in light of the existing literature, highlight best practices, and where appropriate provide suggestions for future HPE knowledge syntheses.
## Quality
Practitioners infrequently apply generic evidence to their educational practice without first considering its fit with their local context [bib_ref] Lost in knowledge translation: time for a map?, Graham [/bib_ref]. Clearly stated research questions can help practitioners to quickly determine a review's focus and determine its relevance for their practice. We found that the Guides, with one exception, presented their research questions; however, the minority called out the PICO elements. Common in clinical research and a requirement for reviews following the PRISMA Guidelines [bib_ref] Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement, Moher [/bib_ref] , PICO elements provide a structured representation of the research question, which may facilitate a reader's ability to quickly identify its relevance to their context. While we recognize that the 'kaleidoscopic nature' of HPE research [bib_ref] Are we talking the same paradigm? Considering methodological choices in health education..., Gordon [/bib_ref] may give rise to a variety of types of research questions and the use of diverse methodologies, the majority of Guides were systematic reviews, a synthesis type for which PICO questions are quite appropriate. Recently, the BEME Collaboration has promoted the use of a variety of types of knowledge synthesis and provides resources to help researchers explore and adopt the knowledge synthesis methodology most appropriate for their research question.
Most Guides reported the details of how they executed literature searching and determined the inclusion and exclusion of studies. By reporting these details, authors provided transparency in their methods thus allowing readers to judge the strengths and weaknesses of the review. Including these details also allows future researchers to fully replicate and update the Guide, thereby avoiding inefficiencies and waste [bib_ref] The PRISMA statement for reporting systematic reviews and meta-analyses of studies that..., Liberati [/bib_ref].
## Accessibility
In fields such as public health and clinical medicine, practitioners report that the inability to access full-texts of information in a timely and cost-effective manner is a major barrier to uptake of evidence into practice [bib_ref] Information transfer: what do decision makers want and need from researchers?, Dobbins [/bib_ref]. Fortunately, the full-text of almost 75% of the included Guides was publicly accessible in some form. However, more recent reviews require subscription privileges or payment of $54.00 for 24-hour access. For many educational practitioners, this is a barrier that can potentially impede integrating current evidence into practice. We suggest review authors exercise their right, as described by Medical Teacher's publisher, Taylor & Francis, to deposit a copy of their original manuscript (prior to peer review) to a preprint server. Preprint servers provide practitioners with barrier-free, immediate access to Guide content. Additionally, preprint servers are indexed by search engines, such as Google Scholar, making them findable by a broad audience of educators. Lastly, recent research has reported that articles with an associated preprint actually have higher Altmetric scores and citation rates than those without [bib_ref] Altmetric scores, citations, and publication of studies posted as preprints, Serghiou [/bib_ref].
The creation of targeted derivative products of the Guides, such as Practice Points and BEME Spotlights, aligns well with HPE practitioners' requests for straightforward syntheses and brief presentations of empirical evidence [bib_ref] Affordances of knowledge translation in medical education: a qualitative exploration of empirical..., Onyura [/bib_ref]. In other disciplines, similar approaches have been linked with increased understanding of evidence and improvements in knowledge uptake and application to practice [bib_ref] Tailored and targeted health communication: strategies for enhancing information relevance, Kreuter [/bib_ref] [bib_ref] The effectiveness of evidence summaries on health policymakers and health system managers..., Petkovic [/bib_ref] [bib_ref] Translating evidence into population health improvement: strategies and barriers, Woolf [/bib_ref]. Therefore, these materials should be made maximally visible and readily accessible. Going forward, Medical Teacher editors might consider embedding links to these materials on their website or directly referencing them within the review's full-text. Additionally, the editors might follow Academic Medicine's example and assemble these valuable resources as an e-book.
## Relevance
In HPE, the transferability of educational approaches and innovations from one institution to another and between learners has been questioned [bib_ref] It's NOT rocket science: rethinking our metaphors for research in health professions..., Regehr [/bib_ref]. To address this issue, researchers have suggested that theory-informed approaches enable practitioners to understand why things work and under what conditions, thus making it possible to transform 'off the shelf' approaches into approaches appropriate for their own educational contexts [bib_ref] Theories in medical education: towards creating a union between educational practice and..., Gibbs [/bib_ref] [bib_ref] Knowledge translation and implementation science in health professions education: time for clarity, Thomas [/bib_ref] [bib_ref] Is transferring an educational innovation actually a process of transformation?, Varpio [/bib_ref]. Therefore, it is recommended that descriptions of educational interventions report on the principles, theories, and/or philosophies upon which those interventions are based. This idea is quite similar to the advice of Varpio et al., who noted, 'While an educational innovation's techniques may seem to be surface structures, they are realizations of deeper fundamental principles. The fundamental principles are themselves realizations of the innovation's foundational philosophy. When techniques and/or principles are modified to a context, it is important to analyze if the modifications continue to uphold the innovation's philosophy [bib_ref] Is transferring an educational innovation actually a process of transformation?, Varpio [/bib_ref].' Unfortunately, a minority of Guides reported these elements. We believe this is an area ripe for improvement, which could, in turn, facilitate greater transferability. BEME guidelines state that a Guide should be updated within 3 years of publication. Currently, few Guides meet this criterion suggesting some findings may no longer be relevant. BEME suggests that authors, in addition to publishing a formal update, post a supplement to the original review that highlights important studies published since the completion of the original review. We identified a single supplementary update [bib_ref] What impact do structured educational sessions to increase emotional intelligence have on..., Cherry [/bib_ref]. It is the Guide author's prerogative to update their work (or not); however, BEME reserves the right to permit other researchers to update a Guide. In the future, BEME might consider partnering authors with graduate programs in HPE to enlist students help with increasing the timeliness of Guide updates. Nevertheless, we should also acknowledge that the timing of necessary updates may vary, depending on how much new information is available and whether that new information makes a meaningful difference. Some topics may have no new studies accrue for many years, while others may need very frequent updates [bib_ref] A surveillance system to assess the need for updating systematic reviews, Ahmadzai [/bib_ref] [bib_ref] Updating systematic reviews: an international survey, Garritty [/bib_ref].
## Study strengths and limitations
We focused on a subset of HPE reviews that are highly curated and well-supported. Therefore, these findings may not generalize across HPE reviews. Future research should consider expanding our approach to examine HPE reviews more broadly. We framed our analysis of the Guides using the STORIES Statement. However, future researchers may want to more specifically focus on evaluating the overall quality of the Guides using another targeted tool, such as the AMSTAR Checklist.
Notwithstanding these limitations, we believe this metasynthesis has a number of important strengths. To our knowledge, it is the first analysis of BEME Guides, which are created for use in educational practice. In addition, we undertook a systematic search and independently engaged in extracting data from all Guides. Lastly, we engaged in this research with a diverse team of stakeholders who have varied expertise and backgrounds. We hope this fact increases this meta-synthesis' relevance to a variety of stakeholders. Ultimately, we are optimistic that this analysis will help those who write and publish future Guides, and similar publications, continue to evolve such that knowledge syntheses can more fully catalyze the translation of evidence into educational practice.
Funding Candace Norton was supported in part by her appointment to the NLM Associate Fellowship Program sponsored by the National Library of Medicine and administered by the Oak Ridge Institute for Science and Education.
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http:// creativecommons.org/licenses/by/4.0/), which permits unrestricted 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: Review characteristics [/table]
[table] Table 2: Inclusion of methodological detailsMaterials were considered available in full-text if they were present in the available PDF version from Medical Teacher b Materials were considered available in supplement only if they were not featured in the PDF version of the article, but were available on either the BEME or Medical Teacher website ulation, Intervention, Comparison, Outcome (PICO) elements of their questions. Most Guides focused on intervention effectiveness (e.g., 'Is the journal club effective in supporting evidence-based decision making?' [/table]
|
Xtrapol8 enables automatic elucidation of low-occupancy intermediate-states in crystallographic studies
Unstable states studied in kinetic, time-resolved and ligand-based crystallography are often characterized by a low occupancy, which hinders structure determination by conventional methods. To automatically extract structural information pertaining to these states, we developed Xtrapol8, a program which (i) applies various flavors of Bayesian-statistics weighting to generate the most informative Fourier difference maps; (ii) determines the occupancy of the intermediate states by use of methods hitherto not available; (iii) calculates extrapolated structure factors using the various proposed formalisms while handling the issue of negative structure factor amplitudes, and (iv) refines the corresponding structures in real and reciprocal-space. The use of Xtrapol8 could accelerate data processing in kinetic and time-resolved crystallographic studies, and as well foster the identification of drug-targetable states in ligand-based crystallography.
O nce reserved to a handful of proteins and specialized laboratories, time-resolved and kinetic crystallography (TRX, KX) are on the verge of widespread adoption. This momentum is owed mostly to the advent of room-temperature serial crystallography, pioneered at X-ray free electron lasers (XFEL) [bib_ref] Advances and challenges in time-resolved macromolecular crystallography, Brändén [/bib_ref] yet swiftly implemented at synchrotrons where ease of access and a larger user-base hold promise for groundbreaking studies 2 . A main limitation of TRX and KX remains that full occupancy of the triggered state is hardly ever attained in the crystalline macromolecule, resulting in co-existence with the reference state. Low occupancy may also poison data collected from crystalline ligand-protein complexes, obscuring ligand identification and conformational changes undergone by the protein upon binding.
Based on the assumption that structure factor phases hardly vary upon reaction initiation and progression, and provided that the two datasets are isomorphous, the calculation of Fourier difference electron density maps is a convenient means to highlight the largest structural differences between a reference and a triggered state dataset [bib_ref] Structure of human deoxy cobalt haemoglobin, Fermi [/bib_ref] [bib_ref] Jr Isomorphous difference methods, Rould [/bib_ref] :
[formula] F triggered obs À F reference obs ; φ reference calcð1Þ [/formula]
This assumption holds true for the vast majority of TRX and KX experiments, where only limited conformational changes are observed, and as well for binding of a small molecule to a crystalline protein. It may further remain an acceptable approximation for larger structural changes in which a substantial number of atoms in the unit cell are involved, provided that the isomorphism between the two dataset remains high 5 (e.g., ref. [bib_ref] Dynamics and mechanism of a light-driven chloride pump, Mous [/bib_ref]. The information content of Fourier difference maps can be improved by Bayesian-statistics weighting of structure factor amplitude (SFA) differences [bib_ref] Improved estimation of structure-factor difference amplitudes from poorly accurate data, Ursby [/bib_ref] [bib_ref] A molecular movie at 1.8 A resolution displays the photocycle of photoactive..., Ren [/bib_ref] , yet these maps can be featureless when structural changes or the triggered-state occupancy are small, or if a mixture of triggered states accumulates [bib_ref] Automated electron-density sampling reveals widespread conformational polymorphism in proteins, Lang [/bib_ref] , whose overlaying positive and negative difference peaks cancel each other. Furthermore, Fourier difference maps may allow modeling of the triggered state structure, but not its refinement, limiting both insights and reproducibility. By use of extrapolation methods, whereby SFA differences are inversely scaled to the occupancy of the triggered state and summed with reference SFAs, the hypothetical SFAs for the triggered state present at full occupancy can be estimated, enabling both refinement of the triggered state structure, and to identify potential co-existing intermediate states:
[formula] F extrapolated ¼ w α F triggered obs À F reference obs þ F additionalð2Þ [/formula]
where α is the reciprocal occupancy of the triggered state (α = 1/ occupancy) and w a potential weighting factor. With the advent of serial crystallography, and the consequential blooming of TRX studies, the demand for extrapolation methods has risen, but these remain obscure to the vast majority of crystallographers for three main reasons. Firstly, the chosen occupancy is determined but hardly ever justified in absence of methods to estimate the triggered state occupancy based on crystallographic data only. Secondly, each lab resorting to structure factor extrapolation makes use of its own library of custom-written scripts, generally assembled over years of practice in TRX and KX, making results not easily reproducible. Thirdly, various flavors of extrapolation exist whereby users have the possibility to exploit Bayesianstatistics to downweigh SFA difference outliers, refine phases from the reference state structure against extrapolated SFAs (ESFAs), and recalculate the figure of merit of these phases prior to generation of extrapolated maps. Here, we introduce Xtrapol8, a new software aimed at resolving these issues, and discuss its design and usage. Briefly, provided a triggered state dataset, and a reference dataset and model, Xtrapol8 firstly calculates a weighted Fourier difference electron density map to unbiasedly visualize differences between the reference and triggered state. Secondly, Xtrapol8 computes optionally-weighted ESFAs, and carries out real-space and reciprocal-space refinements in an automated fashion. Finally, Xtrapol8 provides estimates of the triggered state occupancy via two independent methods hitherto not available.
The difference-map method estimates the occupancy based on difference electron density maps, whereas the distance-analysis method relies on the automatically refined (reciprocal-space and realspace) structures. We showcase Xtrapol8's utility by application to recently published KX data collected on the photoconvertible fluorescent protein mEos4b [bib_ref] Fixation-resistant photoactivatable fluorescent proteins for CLEM, Paez-Segala [/bib_ref]. In the Supplementary Information we further evaluate the versatility of Xtrapol8 by revisiting a variety of previously-published challenging TRX or KX studies.
# Results and discussion
Written in python, Xtrapol8 requires only the CCP4 [bib_ref] Overview of the CCP4 suite and current developments, Winn [/bib_ref] and Phenix 12 suites to run. Only standard packages and the cctbx toolbox [bib_ref] Crystallography & NMR system: a new software suite for macromolecular structure determination, Brünger [/bib_ref] are used, which are automatically loaded from the phenix.python environment. This facilitates usage as well as transfer between laboratory computers and data collection centers (XFELs, synchrotrons). [fig_ref] Figure 1: Design of Xtrapol8 [/fig_ref] shows Xtrapol8's overall outline and recommended usage. The minimal required input are two data files in mtz or mmcif format, associated to the reference (ground, native, apoprotein) and the triggered (derivative, ligandbound, intermediate, excited) state, respectively; a model associated to the reference state in pdb or mmcif format; additional files required for proper handling of ligands (such as crystallographic information files, or custom modification files), and a list of possible occupancies for the triggered state [fig_ref] Figure 1: Design of Xtrapol8 [/fig_ref].
Depending on the user's experience, many additional parameters can be adjusted including resolution cutoffs, scaling parameters, weighting schemes, and parameters for difference map exploration or structure refinement. Xtrapol8 takes four main steps when run in the 'calm-andcurious' mode [fig_ref] Figure 1: Design of Xtrapol8 [/fig_ref] : (1) Reading of input files and quality assessment, and preparation of the files needed for the next steps;
(2) Calculation of the optionally-weighted Fourier difference map, integration of peaks and assignment to reference model atoms; (3) Calculation of ESFAs using a variety of formalisms for a user-defined range of possible occupancies, integration of peaks in the extrapolated difference maps and assignment to reference model atoms, and optional automatic structure refinements; (4) Estimation of the triggered state occupancy and assignment of the most likely ESFAs and electron density maps for the triggered state structure. The program can also be run in a 'fast-and-furious' mode, in which the refinements in steps 3 are omitted and only performed in a fifth additional step at the automatically determined occupancy of the triggered state. In the 'Fo-Fo map only' mode, only the first and second step are carried out.
Graphical user interface. Xtrapol8 can be controlled through the command line or via a graphical user interface (XtrapolG8) written in wxPython as a frontend to the command-line version. The 'Configure' tab features three input panels arranged according to the different steps (Input/Output, FoFo/Extrapolation, Refinement) with a selection of options and parameters that are hidden/shown depending on the user-selected expertise level and Xtrapol8 mode (Fo-Fo map only, fast-and-furious, calm-and-curious; [fig_ref] Figure 2: The graphical user interface allows facilitated parameter control [/fig_ref] , Supplementary Figs. 1-4). The 'Results' tab also consists of three panels. The first panel shows the progression of the Xtrapol8 run by tailing the log file. The second panel displays the general output figures, e.g., the plots of R iso and CC iso (see below) as a function of resolution, of the fraction of negative ESFAs as a function of intermediate state occupancy, or the results of the occupancy determinations and refinements . The last panel shows plots specific to a given type of ESFAs and occupancy, accessible through two dropdown menus, e.g., the plots of the number of negative ESFAs or of ESFA signal-to-noise ratio as a function of resolution. Coot can be directly opened from XtrapolG8 and will load the Fourier difference and extrapolated maps, and the refinements' output . For the automatically determined occupancy, an all-atom distance difference matrix is also computed between the refined extrapolated-structure and the reference model, whereby the plotted value for each residue corresponds to the sum of the changes in distance for all constitutive atoms.
Data handling and quality assessment. The input data files should contain merged data in the form of SFAs or intensities (Supplementary Methods). In the case where only the latter are available, they will be converted into SFAs using the French-Wilson algorithm [bib_ref] On the treatment of negative intensity observations, French [/bib_ref] in truncate (CCP4) [bib_ref] How good are my data and what is the resolution?, Evans [/bib_ref]. Afterwards, pointless [bib_ref] How good are my data and what is the resolution?, Evans [/bib_ref] will be run to verify consistent indexing between the triggered and reference datasets, and if needed, will be used to reindex the triggered dataset in respect to the reference dataset. The two datasets are then Steps specific to the 'fast-and-furious' (default options: q-weighting of difference map and ESFAs, rescue of negative ESFAs using the truncate method, occupancy determination based on the difference-map method; additional step 5: refinement in both reciprocalspace and real-space at the automatically determined occupancy) and 'calm-and-curious' modes are boxed in purple and green, respectively. trimmed so as to retain only reflections present in both sets of observed data and lying within the optionally user-defined resolution limits. In the case where no resolution limit is defined, the program will use all data present in the input data files.
Prior to difference map and ESFA calculations, the modelderived calculated SFAs are scaled anisotropically to the reference data using mmtbx.fmodel [bib_ref] Bulk-solvent and overall scaling revisited: faster calculations, improved results, Afonine [/bib_ref] [bib_ref] On the handling of atomic anisotropic displacement parameters, Grosse-Kunstleve [/bib_ref] while scaling of the triggered to the reference data is performed anisotropically or isotropically by scaleit [bib_ref] Overview of the CCP4 suite and current developments, Winn [/bib_ref]. At this step, statistical indicators informing on the isomorphism between the input data sets are calculated, which inform users on the probability that both sensible Fourier difference maps and ESFAs can be calculated. Specifically, the Rfactor (R iso ) between the two datasets is calculated using:
[formula] R iso ¼ ∑ F reference obs À F triggered obs ∑ F reference obs þ F triggered obs =2ð3Þ [/formula]
where F reference obs and F triggered obs are the SFAs of the reference and triggered states, respectively. An overall R iso value lower than 0.10 indicates strong isomorphism between the two data sets, but values up to 0.25 can still lead to useful difference and extrapolated electron density maps [bib_ref] Jr Isomorphous difference methods, Rould [/bib_ref]. The isomorphism-indicating correlation coefficient (cc iso ) is calculated using:
[formula] cc iso ¼ ∑ F reference obs À F reference obs F triggered obs À F triggered obs h i ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffi ∑ F reference obs À F reference obs 2 ∑ F triggered obs À F triggered obs 2 rð4Þ [/formula]
Experimental Fourier difference maps calculation. On the assumption that the structural differences between the reference and triggered state(s) are small-i.e., only a few specific atoms change position, with no rotation or translation of the molecule in the unit cell-the phases of the triggered and the reference states hardly differ, which enables investigators to extract information on the triggered state by calculation of a Fourier difference map (Eq. 1). A simple subtraction between the two sets of SFAs gives a similar weight to all of them, regardless of their measurement precision. Ursby and Bourgeois 7 suggested the use of q-weighting, whereby the magnitude of the difference between SFAs and the relative errors on their measurements are used to downweigh unlikely SFA differences. The use of Bayesian statistics to improve Fourier difference maps was also proposed by Terwilliger and Berendzen 18 . Later, Ren et al. introduced a simpler weighting scheme 8 , to which we refer to as k-weighting in analogy to qweighting. In Xtrapol8, the user has the choice to calculate qweighted, k-weighted or non-weighted Fourier difference maps, but the default is set to q-weighting. In the k-weighting scheme, the extent to which outliers are downweighed can be adjusted by an additional parameter, that has been set to 0.05 19 or 1.0 8,20 depending on reports, with the first one being the default in Xtrapol8.
Difference map analysis. An important feature of Xtrapol8 is the map-explorer module, which will search for peaks in difference maps, integrate and assign them to the closest reference model atom within a given radius, and generate an unbiased list of residues whose positions change relative to those in the reference model. Based on this list, the residues with the highest integrated peak volumes can be used to evaluate the occupancy of the triggered state in a later stage of the program. The stringency in the selection of difference peaks can be adjusted-not only in terms of the maximal and minimal heights for peak selection and integration, respectively, but as well in terms of Z-score filtering and search radius. Default values for first three parameters are ± 4 r.m.s.d., ± 3 r.m.s.d. and 2, while the search radius is defaulted to the highest resolution.
Extrapolated structure factor amplitudes (ESFAs). Extrapolated structure factor amplitudes are estimates of the SFAs that the triggered state would have given rise to if it had been present at full occupancy in the crystal. The general formula for calculating ESFAs is presented in Eq. 2. The weighting factor w and the F additional term have found different interpretations in the literature, explaining that various extrapolation approaches exist. Firstly, F additional was initially derived from the reference model (i.e., F reference model or F reference calc depending on whether or not bulksolvent modeling is included) [bib_ref] Difference refinement: obtaining differences between two related structures, Terwilliger [/bib_ref] , but use of measured reference state SFAs (F reference obs ) was later put forward by Genick and coworkers [bib_ref] Structure-factor extrapolation using the scalar approximation: theory, applications and limitations, Genick [/bib_ref] [bib_ref] Structure of a protein photocycle intermediate by millisecond time-resolved crystallography, Genick [/bib_ref]. Whether the F reference obs should be preferred to F reference model is still under debate [bib_ref] Early-stage dynamics of chloride ion-pumping rhodopsin revealed by a femtosecond X-ray laser, Yun [/bib_ref] , illustrating that it likely depends on the experimental case. Secondly, one can question whether or not the q/k-weighting scheme used in Fourier difference map calculations to reduce the weight of uncertain SFA differences should also be applied in the calculation of ESFAs (weighting factor w) leading to yet another nuance. Finally and perhaps more subtly, initial extrapolated electron density maps are being calculated using the phases of the reference model, but the associated figure of merit (m) can be chosen to represent the phase agreement between the ESFAs and reference state model, or originate from the reference state [bib_ref] Structure-factor extrapolation using the scalar approximation: theory, applications and limitations, Genick [/bib_ref] [bib_ref] Structure of a protein photocycle intermediate by millisecond time-resolved crystallography, Genick [/bib_ref]. As the latter has been proposed by Genick et al. [bib_ref] Structure-factor extrapolation using the scalar approximation: theory, applications and limitations, Genick [/bib_ref] , we refer to these as the (q/k)Fgenick ESFA type. Overall, this leads to nine different types of ESFAs and corresponding extrapolated maps [fig_ref] Table 1: Types of ESFAs calculated by Xtrapol8 and corresponding map coefficients [/fig_ref] , reflecting the variety of approaches that have been followed depending on the experimental case and laboratory since introduction of extrapolation methods.
The parameter α is inversely related to the occupancy of the triggered state, but its correct value remains unknown until occupancy determination has been carried out. Two methods are available in Xtrapol8 to estimate the triggered state occupancy, either based on the raw extrapolated maps or on the refined extrapolated models. Hence, extrapolated SFAs, maps and models are calculated for each of the user-supplied occupancy values, and occupancy determination is performed at the end of the Xtrapol8 run. By default, and in the fast-and-furious mode, only the qF extr type of ESFAs is calculated. Nevertheless, all nine types of ESFAsor a subset thereof-can be calculated in a single run in the calmand-curious mode, leading to a maximum of 9 independent occupancy estimations by each of the two occupancy determination methods-or 18 estimates.
For each occupancy, extrapolated SFAs and map coefficients are used in three subsequent refinement steps. Firstly, a reciprocal-space refinement of the reference state input model is carried out against the ESFAs, allowing to account for small changes in atomic positions and thereby ameliorate the phases and associated figure of merit m, which in turn will result in clearer mF extrapolated − DF calc and 2mF extrapolated − DF calc electron density maps. The latter and the corresponding refined model are then used in automatic real-space refinement, enabling to model conformational changes too large to have been accounted for by reciprocal-space refinement. In addition, real-space refinement of the reference model is also carried out against the initial extrapolated electron density map (calculated before reciprocalspace refinement and update of phases and figure of merits), which offers an unbiased view of the information content in the extrapolated data. The user has the possibility to perform these automatic reciprocal-space and real-space refinement using either phenix.refine [bib_ref] Towards automated crystallographic structure refinement with phenix. refine, Afonine [/bib_ref] and phenix.real_space_refinement 27 or refmac [bib_ref] REFMAC5 for the refinement of macromolecular crystal structures, Murshudov [/bib_ref] and coot 29 , respectively. An important limitation in streamlining the refinement steps is that they are performed starting from the reference model, which may or may not be adequate to accurately refine the triggered state. With the exception of waters, which can be automatically updated during reciprocal-space refinement when using phenix.refine, no atoms can be removed or added during the Xtrapol8 run. For this reason, it is advisable that automatic refinement is not carried out in studies where covalent bonds are being formed or broken, or if the detection and localization of a ligand is of interest. At the command line, the automatic refinement steps can be disabled by setting the refinement.run_refinement parameter to "False". When using the graphical user interface XtrapolG8, users need to un-check the "Perform refinement with" option in the Refinement panel of the Configure tab.
Density modification can aid in meliorating the phase information and cleaning up the electron density maps, offering a better base for real-space refinement. Regardless of the suites of programs used for reciprocal-space and real-space refinement, density modification is performed using the CCP4 program dm [bib_ref] Overview of the CCP4 suite and current developments, Winn [/bib_ref]. We indeed found that this program is generally unsensitive to the often non-ideal distribution of ESFAs and their standard deviations.
Negative ESFAs. An often-overlooked issue in structure factor extrapolation is the presence of negative ESFAs. Their origin lies in large negative values of the weighted difference amplitudes (first term in Eq. 2) that cannot be compensated by the additional amplitude term (second term in Eq. 2). For a small occupancy, and thus a large α-value, the percentage of negative ESFAs can become large, which will result in decreasing the true completeness (i.e., the completeness of the positive reflections only) below a critical value of 90%. Indeed, negative SFAs are not handled by refinement programs so that the refinement is carried out against incomplete data, resulting in weak electron density maps and non-converging refinement. In Xtrapol8, we propose different approaches to correct for this issue. In the first approach, the negative reflections are removed, on the ground that the error in the measurements of the associated reflections is too large to lead to a reliable estimate 5 . However, rejecting these may strongly alter the resulting extrapolated maps while biasing the result toward the positive difference amplitudes. In the second approach, the artificial negative reflections are set to zero, under the assumption that true values of these physically-impossible negative ESFAs are weak, but this treatment implies losing important information on their relative strength [bib_ref] On the treatment of negative intensity observations, French [/bib_ref] and again biasing the result toward the positive amplitudes. In the third approach, the negative ESFAs can be replaced by their corresponding values in either the observed F reference obs À Á or calculated F reference model À Á reference datasets, under the assumption that the strong negative difference amplitude is a consequence of measurement errors in F triggered obs and that the triggered state closely resembles the reference state. Predictably, however, this approach will bias results towards the reference state, which can lead to an unreliable occupancy estimate and alteration of the extrapolated maps. Importantly, usage of this option should mirror the ESFA-calculation strategy [fig_ref] Table 1: Types of ESFAs calculated by Xtrapol8 and corresponding map coefficients [/fig_ref] , i.e., negative ESFAs should be replaced with F reference obs in the case where (q/k)F extr and (q/k)F genick ESFAs are used, but with F reference model when ESFAs are calculated using the (q/k)F extr,calc method. Finally, just as with intensities originating from a diffraction experiment with a fully occupied crystal, also the extrapolated intensities, calculated as the square of the ESFAs multiplied by their initial sign, are supposed to follow the Wilson distributions [bib_ref] The probability distribution of X-ray intensities, Wilson [/bib_ref]. Structure factors can be re-calculated from the extrapolated intensities by the use of an algorithm taking the French-Wilson algorithm 14 into account, e.g., truncate [bib_ref] How good are my data and what is the resolution?, Evans [/bib_ref]. This approach yields the lowest R work and R free values in reciprocalspace refinement and the most defined electron density around atoms in the model-as judged by higher CC mask and CC volume after real-space refinement-hence it is the default. This option has yet the drawback that all structure factors are being re-calculated, not just the negative ones. Hence, in cases where negative ESFAs amount to less than 5 % of all ESFAs, other strategies may become preferable.
Occupancy estimation. Determination of the occupancy of the triggered state is one of the main features of Xtrapol8. Indeed, the ESFAs vary as the occupancy changes (Eq. 2), hence incorrect estimation of the occupancy could lead to under-or overrefinement of structural features in the triggered state. Xtrapol8 features two complementary methods for the occupancy estimation, both fully based on the X-ray data without any external assumption or prior knowledge about the triggered state, and with only minimal expectations on the behavior of the structure factors and map coefficients. Both methods can make use of all atoms in the asymmetric unit; of atoms displaying the strongest peaks in the Fourier difference map (as identified by map-explorer); or of a user-defined selection of residues, ligands and waters. The first method, the difference-map method, relies on the analysis of the initial extrapolated difference map, i.e., that obtained before refinement of the structure (mF extrapolated,occ − DF calc ; last column in [fig_ref] Table 1: Types of ESFAs calculated by Xtrapol8 and corresponding map coefficients [/fig_ref]. When calculated using the correct reciprocalspace occupancy (α) value, this map strongly resembles the Fourier difference map, with positive and negative peaks pointing to appearing and disappearing features with respect to the reference model, respectively. Peaks in the mF extrapolated,occ − DF calc maps grow in height as the reciprocal occupancy α increases towards the correct value, and then decrease upon further augmentation of α due to an increase in the map noise level. The occupancy that yields the extrapolated difference map with the highest signal-tonoise (S/N) ratio is retained as the most probable one. In practice, we first integrate all peaks in the difference map; use the sum of the integrated absolute values of all peaks as a measure of the noise level in the map (N); utilize a Z-scoring approach to single out the most significant peaks and use the sum of their integrated absolute values as a measure of the signal level in the map (S); plot the normalized S/N ratio as a function of α; and pick as the correct occupancy that for which S/N is maximized.
The difference-map method is most closely related to the experimental data. However, it only provides solid results when the Fourier difference map, and thus the extrapolated difference maps, contains features that can be clearly distinguished from noise. In cases where this condition is not fulfilled, the peaks in the Fourier and mF extrapolated,occ − DF calc maps are small, hence the occupancy determination based on the maximization of the S/ N ratio may fail. This might happen when multiple species occupy the same site and reduce each other's signal, or if it was not possible to collect enough data for the reference or triggered state. In such cases, users are advised to check if the default parameters used for the analysis-i.e., maximal and minimal heights for peak selection and integration, and Z-score filteringare not too stringent for their data, and re-run the analysis using slightly modified parameters.
If occupancy determination by the difference-map method nonetheless fails, users can still rely on the distance-analysis method, which estimates the occupancy based on the evolution of differences in interatomic distances between the reference model and the models automatically refined at increasing α values. Indeed, we found empirically that atoms which undergo positional changes in the triggered state move from their initial position in the reference model to their final position in the extrapolated structures with a sigmoidal dependence-i.e., after a slow increment, the distance difference increases steadily with α and then levels at a plateau value indicating that only noise is being added to the maps upon further increasing α. Thus, an estimate of the α value can be extracted by (1) fitting a sigmoid to the distance changes in function of α for each pair of atoms (i, j) (d (i, j) = f(α)); (2) extracting α i,j as the α-value closest to the point where those shifts are 99% of the plateau; and (3) plotting the distribution of α i,j and retaining the average 〈α i,j 〉 and maximum in the distribution (Supplementary Methods). Inasmuch as that the difference signal contributes systematically to the ESFAs (as opposed to noise), the distance-analysis method will provide a reliable estimate of the occupancy-even in cases where the difference signal is too weak to visually standout in the Fourier difference map.
The 'fast-and-furious' and 'calm-and-curious' modes. In the default 'calm-and-curious' mode, Xtrapol8 performs calculations in the most thorough way possible. For rapid data evaluation (e.g., during data collection) or to estimate the parameters for a next run, calculations can be expedited using the 'fast-and-furious' mode. In this mode, default parameters will be used for most arguments, i.e., only the q-weighted Fourier difference map, and ESFAs and map coefficients of the qF extr type will be calculated [fig_ref] Table 1: Types of ESFAs calculated by Xtrapol8 and corresponding map coefficients [/fig_ref] , and refinements will only be performed at the occupancy estimated by the difference-map method [fig_ref] Figure 1: Design of Xtrapol8 [/fig_ref]. As a consequence, occupancy determination using the distance-analysis method is not possible in the 'fast-and-furious' mode. As most of the Xtrapol8 runtime is occupied by the refinement steps, this mode reduces the overall runtime greatly, allowing crystallographers in the midst of an experiment to take decisions swiftly. Xtrapol8 can also be run in a 'Fo-Fo map only' mode in cases where users only need to calculate a Fourier difference map [fig_ref] Figure 1: Design of Xtrapol8 [/fig_ref].
Output. The output of an Xtrapol8 run is stored in a directory whose name can either be defined by the user or automatically called 'Xtrapol8'. This directory will contain the Fourier difference map, a log file, a Pymol session, and various figures meant to help in the evaluation of data quality and interpretation of results . The log file contains the most important information, such as statistics on the input files, the location of output files, the number of negative ESFAs and the results of the occupancy estimation. A subdirectory is created for each tested occupancy value, with the name indicative of whether ESFAs were q-weighted, k-weighted or not weighted. These subdirectories will contain the ESFAs as well as the results from refinements. For each type of ESFAs, a coot session is provided in the subdirectory corresponding to the estimated occupancy.
Good users' practice. The main ambivalence with extrapolation is that in order to calculate the most plausible set of ESFAs, the occupancy must be known. However, occupancy estimations based on X-ray data can only be carried out after the initial extrapolated electron density maps (difference-map method) or refined extrapolated models (distance-analysis method) have been produced. Therefore, the scheme outlined in is advised to obtain optimal results in the shortest time. Running Xtrapol8 in the 'calm-and-curious' mode with improper settings would indeed result in a waste of (computation) time, notably when the program is used to monitor the progress and success of an experiment. Most appropriate is therefore that users run Xtrapol8 in the 'Fo-Fo only' or 'fast-and-furious' mode until good parameters are found. A first good sign is that scaling statistics look reasonable, i.e., R iso (Eq. 3) and CC iso (Eq. 4) display decent values (R iso < 0.20; CC iso > 0.80) even in the highest resolution shell (R iso < 0.40; CC iso > 0.60). In the case they are not, users should check their data processing and merging statistics and the unit cell parameters of the reference and triggered datasets in order to increase their chances of getting useful results.
Ideally, the Fourier difference map should cover specific residues (e.g., active sites residues) and the 2mF extrapolated − DF calc electron density maps should be of acceptable quality (i.e., cover all atoms in the model without breaks in the electron density nor spurious peaks in the solvent channels). It may be, however, that no peaks are visible in the Fourier difference maps, in which case users should proceed with extrapolation so as to verify whether or not information is present in the 2mF extrapolated − DF calc and mF extrapolated − DF calc electron density maps. When this is not the case, the scaling parameters and resolution boundaries may need to be altered.
Finding the correct occupancy of the triggered state may require optimization. If features in the Fourier difference map are clear, the default difference-map method with preset parameters should provide proper results for the estimated occupancy. A wider range of possible occupancies should be tried, and parameters for the map-explorer altered if no clear maximum emerges in the plot of mF extrapolated,occ − DF calc S/N ratios as a function of α. Features in the difference maps may be genuinely weak, in which case the difference-map analysis will not allow to discriminate between tested occupancy values, yet the distanceanalysis method may offer solace. This method however relies on the real-space refined models, and thus requires to run Xtrapol8 in the 'calm-and-curious' mode. Alternatively, users that have already run Xtrapol8 in the 'fast-and-furious' mode can use the refiner.py script (Supplementary Methods) to refine models and apply the distance-analysis methodeither starting from the reference model or a manually-modified model. Additional details concerning the optimization of the results can be found in the Supplementary Methods.
Test case: kinetic crystallography data which allowed identification of a long-lived dark state in mEos4b. The green-to-red photoconversion of mEos4b is the basis of its usage as a marker in photo-activated localization microscopy (PALM). Yet the existence of reversible dark-states, which form upon excitation of the redemitting state, has long limited its application in single-particletracking (spt) PALM. Based on KX experiments, whereby crystalline mEos4b in the green-emitting state was first slowly converted Recommended usage of Xtrapol8. Suggested workflow for efficient usage of Xtrapol8. Users are advised to first run the program in the 'Fo-Fo map only' mode in order to evaluate the height of peaks in the difference Fourier map. Secondly, it is recommended to run Xtrapol8 in 'fast-and-furious' mode (purple boxes) to obtain a crude estimate of the occupancy based on the difference-map method, and a first characterization of the triggered state structure. Finer exploration can then be carried out using the 'calm-and-curious' mode (green boxes), which will refine the occupancy determination based on the difference-map method, produce refined structures for all tested occupancies and ESFA strategies, and enable orthogonal occupancy determination by the distance-analysis method. Evaluation criteria are shown in blue, user actions are depicted in yellow.
to the red-emitting state and then illuminated by a 561-nm laser before freeze-trapping, a long-lived dark state was identified whose characterization enabled the design of a new data collection scheme suited for the recording of long tracks in spt-PALM 31 . In the reference (red-on) and triggered (red-off) datasets, two and three states co-exist, respectively, illustrating how complex investigated structural dynamics may be in TRX and KX studies .
We first ran Xtrapol8 in the 'Fo-Fo map only' mode to ascertain both isomorphism of the data and occurrence of conformational changes in the triggered dataset. The isomorphism between the reference (PDB entry 6GP0) and triggered dataset (PDB entry 6GP1) is high, with an overall R iso of 0.106 (highest resolution shell R iso = 0.261; 2.5% increase in unit cell volume; . The q-weighted Fourier difference map (F redÀoff obs À F redÀon obs ) shows strong features on the chromophore and surrounding residues up to a resolution of 1.5 Å . Running Xtrapol8 in the 'fast-and-furious' mode, we tested eight occupancies in the 0.1-0.7 range. Occupancy estimation using the difference-map method pointed to the triggered state (red-off) displaying an occupancy between 0.3 and 0.4 . Altogether, Fourier difference map calculation, estimation of the occupancy and production of reciprocal and real-space models of the triggered state took 20 min on a mid-range laptop.
Subsequently, Xtrapol8 was run in the 'calm-and-curious' mode. By testing 25 occupancies, the final occupancy was determined to be 0.325 using the difference-map method. The initial extrapolated mF extrapolated -DF calc and 2mF extrapolated -DF calc map confirmed the occurrence of large structural changes in the chromophore and surrounding residues . Some of these could not be modeled automatically during reciprocal and real-space refinements, requiring manual intervention to model chromophore isomerization and accompanying conformational changes. The final triggered state model was characterized by R work /R free and CC mask values of 20.91/23.81% and 91.48%, respectively (compared to values of 15.55/19.00 and 94.67 %, respectively, for mEos4b in the reference red-on state) [bib_ref] Mechanistic investigation of mEos4b reveals a strategy to reduce track interruptions in..., De Zitter [/bib_ref]. Similar results were obtained when other types of ESFAs and weighting schemes were used [fig_ref] Table 1: Types of ESFAs calculated by Xtrapol8 and corresponding map coefficients [/fig_ref] ; [bib_ref] A molecular movie at 1.8 A resolution displays the photocycle of photoactive..., Ren [/bib_ref] [bib_ref] Automated electron-density sampling reveals widespread conformational polymorphism in proteins, Lang [/bib_ref]. Electron density features were only less pronounced in the case of (q/k) Fgenick extrapolated maps 5 or k-weighted extrapolated maps with a high k-scale outlier rejection factor. This observation suggests that recalculating figures of merit for each set of ESFAs benefits extraction of structural features for the triggered states, enabling the observation of lower occupancy structural changes. The use of maximum likelihood weighted maps is also likely beneficial, as it allows one to take into account not only errors on phases (m ref or m) but also those on the measurement and estimation of SFAs (D). Specific to mEos4b, similar results were obtained with all possible treatments of negative ESFAs implemented in Xtrapol8, i.e., when they were rejected, set to 0, replaced by F reference obs or F reference calc , or rescued by use of the truncate method [fig_ref] Figure 1: Design of Xtrapol8 [/fig_ref]. This is likely due to their low amount in ESFAs calculated for an occupancy of 0.325 (ranging from 3.0 to 11.2% depending on the ESFA calculation strategy). Xtrapol8 enables successful extraction of the mEos4b red-off state. The models in cyan and green are mEos4b in the red-on and red-off state, respectively, as downloaded from the PDB (PDB entry 6GP0 and 6GP1, with the only difference that features of green mEos4b were omitted) with the chromophore indicated in yellow (reference state, a, b) and magenta (triggered state, d-e). a traditional 2mF obs − DF calc electron density after rigid body refinement of the red-on state model in the red-off data indicates the absence of signal for the red-off state. b q-weighted Fourier difference map (F redÀoff obs À F redÀon obs ) superposed on the red-on model. c the difference-map method maximizes the information content of extrapolated difference maps and estimates the occupancy to be 0.325 (magenta dashed line). d-e Initial q-weighted 2mF extrapolated − DF calc (d) and mF extrapolated − DF calc (e) electron density maps calculated for an occupancy of 0.325 superposed on the red-off model. f The distance-analysis method, which approximates the correct occupancy by fitting a sigmoid to the evolution of interatomic distances as a function of all tested occupancies and then retrieves the occupancy value for which 99% of the plateau is reached, returns an occupancy estimate of 0.380.
The manually-modified triggered state model was finally subjected to automatic reciprocal and real-space refinements against all sets of ESFAs and maps calculated for different occupancies, respectively, enabling the performance of the distance-analysis method for occupancy estimation to be verified. To this end we used the refiner.py script (Supplementary Methods), which allows the relaunching of all refinements and offers the possibility to run the distance-analysis method based on the refined models. The occupancy was thereby estimated to be 0.38 , offering orthogonal confirmation for the occupancy determined by the difference-map method. The distance-analysis method was hardly sensitive to the number of atoms used for the estimation, yielding similar results when either all protein atoms or exclusively atoms with strong difference map peaks were used. Xtrapol8 enables extraction of low-occupancy states in time-resolved and ligand-based crystallography. In the Supporting results and discussion, we evaluate the versatility and performance of Xtrapol8 by revisiting other time-resolved (TRX), kinetic (KX) and ligand-binding crystallographic studies. For each case the Fourier difference map superposed on the reference state (with carbon atoms of the proteins and ligands colored in cyan and yellow, respectively) and extrapolated electron density map superposed on the triggered state model (with carbon atoms of the proteins and ligands colored in green and magenta, respectively) are shown. a A temperature-dependent KX study was conducted on the covalent complex of acetylcholinesterase with a non-hydrolysable substrate analog, whereby X-rays were used to radiolytically cleave bonds, including disulfide bridges and the bond tethering the substrate analog to the catalytic serine [bib_ref] Shoot-and-trap: Use of specific x-ray damage to study structural protein dynamics by..., Colletier [/bib_ref]. By use of extrapolation, deeper insights could be obtained revealing two binding poses for the radiolytically produced carbocholine product, trapping of CO 2 from radiolysis of buried acidic residues. b A TR-SFX study was conducted on the reversibly fluorescent protein rsEGFP2 with aim to determine the structure of the excited state that preludes isomerization and off-to-on fluorescence switching 24 . Xtrapol8 allowed obtaining similar results as published earlier for the 1 ps time delay dataset. In the Supplementary results and discussion, we show that Xtrapol8 further allows extending the results obtained at the 3 ps time delay. c Comparison of the performance of PanDDA and Xtrapol8 in revealing the electron density of a small compound in a fragment-screening study (BAZ2BA-538) required high-end expertise in crystallography and extensive data processing, yet could be addressed within hours by use of Xtrapol8 [fig_ref] Figure 1: Design of Xtrapol8 [/fig_ref]. We show in at least two cases (see rsEGFP2 and Shoot-and-Trap test cases) that superior results could have been obtained by the use of Xtrapol8, and how beneficial the proper handling of negative ESFAs can be for modeling and refinement. By enabling automatic elucidation of lowoccupancy intermediate states, thereby minimizing the time required to extract meaningful results, Xtrapol8 may accelerate discoveries in TRX, KX and ligand-binding studies. Additionally, it should benefit systematicity and reproducibility of extrapolated map calculation and triggered state structure refinement.
# Conclusion
We here introduced Xtrapol8, a software aimed at standardizing Fourier difference map calculation and application of extrapolation methods. Unique to Xtrapol8 is that it tackles a variety of issues related to Fourier difference maps and extrapolation methods, most notably the need to (i) weight structural observations to maximize the information contents of difference SFAs; (ii) rescue negative ESFAs which can result in sub-optimal refinement and electron density maps; and (iii) determine the triggered state occupancy based solely on X-ray data. As Xtrapol8 offers the possibility to calculate all types of ESFAs at once, and to compare results, it should increase reproducibility while allowing users to make informed decisions as to the method best suited for their project. A level of customization is offered on most important parameters, but defaults are carefully set and Xtrapol8 can be run from the command line or via a GUI, so that adequate results are within reach for experienced and novice users.
# Methods
We applied Xtrapol8 on several sets of published TRX, KX and ligand-binding data. These examples were run using Phenix 12 v.1.19 and the associated cctbx modules 32 , CCP4 11 v.7.1 and Coot 0.8.9 or 0.9.4, unless stated differently. Figures were prepared using Pymol 2.5 33 .
Supplementary methods. Additional methods concerning the refiner.py script and details covering data handling and quality assessment, good user's practice and occupancy determination are provided in the Supplementary Methods section.
Statistics and reproducibility. Bayesian weighting of difference SFAs and ESFAs (q/ k-weighting) is performed as described in ref. [bib_ref] Improved estimation of structure-factor difference amplitudes from poorly accurate data, Ursby [/bib_ref] and ref. [bib_ref] A molecular movie at 1.8 A resolution displays the photocycle of photoactive..., Ren [/bib_ref]. To find the highest peaks in the difference maps (difference-map method), a Z-scoring approach is applied to the selected peaks if they follow a normal distribution (α = 0.05), otherwise all selected peaks are used for occupancy determination. For the distance-analysis occupancy estimation method, only those distances are maintained that are within the 2-6 Å range and to which a logistic function L 1 þ e Àk αÀα 0 ð Þ À1 (with L being the maximum value, k the steepness and α 0 the α -value of the sigmoidal inflection point) can be fitted with an R 2 of at least 0.95 and a χ 2 of 0.5, and with boundaries set to allow for an occupancy of the triggered state between 1 and the maximum sampled occupancy.
[fig] Figure 1: Design of Xtrapol8. Xtrapol8 roadmap. The four main steps followed by Xtrapol8 are depicted in blue. User input are highlighted by yellow boxes, with obligatory input further highlighted in bold. [/fig]
[fig] Figure 2: The graphical user interface allows facilitated parameter control. Panel of the GUI in which the user can alter the options concerning the Fourier difference maps and ESFAs, with zoom-in on some input fields. Details on all input and output tabs and panels is provided inSupplementary Figs. 1-5. [/fig]
[table] Table 1: Types of ESFAs calculated by Xtrapol8 and corresponding map coefficients. [/table]
|
Stuck in the middle: a systematic review of authorship in collaborative health research in Africa, 2014–2016
[bib_ref] The characteristics of highly cited researchers in, Confraria [/bib_ref] [bib_ref] Authorship trends in the Lancet global health, Iyer [/bib_ref]
# Introduction
Despite increases in African research-scientists in Africa more than doubled their publications in science, technology, engineering and math from 2003 to 2012 1 -there is currently a great deficit in resources for African health research. Africa has only 198 researchers per million inhabitants, compared over 4000 in the UK and the US 2 and no country in Africa has achieved the African Union goal of spending 1% of the gross domestic product on research and development. Less than 10% of global investments in health research is spent in developing countries, where more than 90% of preventable mortality occurs. [bib_ref] The mismatch between the health research and development (R&D) that is needed..., Viergever [/bib_ref] In this context, researchers in low-income and middle-income countries (LMICs), and sub-Saharan Africa specifically, collaborate What are the new findings?
► Collaborating with high-income country researchers decreases local representation, particularly in first and last positions, and the lowest representation occurs when collaborating with individuals from a top US university. ► Collaborating with individuals from other African institutions increases local representation in resulting papers.
What do the new findings imply?
► The power dynamics within global health research collaborations must be addressed to increase inclusion and leadership of local authors. ► Investing in South-South partnerships could provide opportunities to increase technical resources for researchers while maintaining the individual's leadership of the resulting work. extensively with high-income country (HIC) researchers on local studies that may be vital to health and policy in their countries. Yet such collaborations can lead to reports of African data with no or very few local authors, meaning that 'the professional recognition and social capital that comes with publication' 7 eludes the African contributors to those projects. These inequities undermine science and the careers of scientists from Africa. [bib_ref] The trouble with inequalities in global health partnerships: an ethical assessment, Fourie [/bib_ref] Several bibliometric analyses over the last decade have used authorship to quantify the equity of partnerships on papers about health in LMICs. While some of these articles report high levels of LMIC representation, most found that LMIC authors are profoundly underrepresented, representing less than 50% of first authors and in some cases, less than 50% of all authors. 11-14 Some of these papers have shown that LMIC author representation decreases in the prime first or last positions compared with overall representation. We revisit this issue with a new lens-if collaborations with external colleagues are key for global health research, how do such collaborations drive overall authorship and publication structure? We focused on five different types of geographic collaborations that we identified a priori as being of interest, assessing authorship dynamics on papers that included collaborators from: (1) top ranking US universities, (2) the USA and Canada, (3) Europe, (4) other African countries and finally (5) a mixture of geographic locales outside of the country of the paper's focus. Here, we present results on how these collaborations affect inclusion of local authors, particularly in prime first or last authorship positions.
## Bmj global health
# Methods
## Identification of relevant articles
We searched PubMed for articles that were published between 2014 and 2016 with the term 'health' in either the title or abstract and 'Africa South of the Sahara' indicated as a Medical Subject Heading (MeSH) term. We selected PubMed and this time period since PubMed began recording author affiliation data in 2014. The PubMed search results were downloaded in XML format. We then used Statistical Analysis System (SAS 9.4, University Edition) software and developed a custom algorithm to scrape and clean the XML file. This algorithm allowed for the extraction of the following data: title, abstract, key words, type of publication, author names, author order, author affiliations and journal names.
To determine whether the article was eligible for analyses, we developed a computer algorithm in Stata V.14 to search and record any country names captured in the title, objectives, methods, results, conclusion, key words and abstract. We also searched for prominent city names or region/province names and linked these to the appropriate country. Extracted articles were removed from the database for the following reasons: no country was named in any of the searched domains or none of the named countries were located in sub-Saharan Africa (considered countries and regional groupings are listed in the online supplementary appendix). In our analysis, we included only articles that focused on a single country and so papers that had multiple countries identified in the searched domains were also removed from the database. For each step of systematic removal of articles, we randomly sampled 10% of the removed articles to confirm ineligibility (reported in [fig_ref] Figure 1: Process of identifying eligible papers and author affiliations for bibliometric analysis of... [/fig_ref]. In the processes of reviewing the subset of articles that required manual collection of author affiliation details or during data cleaning, we identified an additional 326 papers that did not meet eligibility criteria and were therefore removed from our dataset. For quality assessment, we also sampled 200 papers included in the final dataset to confirm eligibility for inclusion and author affiliations for those papers (figure 1).
## Classification of author affiliations and paper collaboration types
For eligible papers, we used a computer algorithm to search author affiliations to determine the country(s) of the author's affiliation. If the affiliation field was missing for any author or if we detected errors in the extracted affiliations-most notably, cases in which the corresponding author's affiliation was assigned to all authors on the paper or cases in which the affiliations of all authors on the paper were assigned to each of the authors-the paper was reviewed and relevant details were manually entered into a Microsoft Access database. This database was merged with the primary database. At this stage, some additional papers that were not accessible online or through requests to the corresponding author were removed (figure 1).
For the final dataset, we classified both authors and papers. We classified each author as being primarily BMJ Global Health affiliated with (1) the same country as the paper's focus, [bib_ref] Investing in health R&D: where we are, what limits us, and how..., Simpkin [/bib_ref] another African country, (3) the USA, Canada or Europe or (4) another country not yet classified (North Africa, Asia, Australasia or South America). If an author had multiple countries associated with their affiliations, we used the following logic: (1) If any affiliation country was the USA, Canada or Europe, then the author was classified as being primarily affiliated with the USA, Canada or Europe. [bib_ref] Investing in health R&D: where we are, what limits us, and how..., Simpkin [/bib_ref] If not, if any affiliation country was the same country as the paper's focus, the author was classified as being primarily affiliated with that country.
(3) If not, if any affiliation was another African country, the author was classified as being primarily affiliated with another African country. (4) Finally, any author remaining unclassified was considered to be primarily affiliated with another country not yet classified. By this logic, affiliations located in the USA, Canada or Europe overrode affiliations located in the same country as the paper's focus. (For example, author BHG often has USA and Rwanda affiliations, and in this classification, she would be classified as being from the USA.) In a sensitivity analysis, we reversed the logic so that affiliations located in the country of the paper's focus and affiliations with other African countries overrode affiliations with the USA, Canada or Europe. (Here, author BHG would be classified as being from Rwanda.)
We classified papers into one of five geographic collaborative types. For the first four types, we looked at:papers that included any authors from a top US university;
(2) paper included any authors from the USA or Canada;
(3) paper included any authors from Europe and (4) paper included any authors from an African institution outside of the country of the paper's focus (figure 2). To disentangle the effects of each collaborative type on each other, we created categories with no overlap between these groups and then added fifth collaborative category, namely papers that included collaborators from multiple geographies outside of the country of the paper's focus. While for author-level analyses, authors with multiple affiliations are classified as having a single locale as described above, in paper collaboration type, we allowed all affiliations for each author to contribute to determination of paper collaborative type. To identify authors from a top US university, we used affiliation with one of the top 20 US graduate schools of public health, as identified by the 2015 US News and World Report rankings, 17 as a proxy. An author with any affiliation with a university on this list was classified as being from a top US university, regardless of whether the author was at the school of public health-for example, since Harvard Chan School was on the 2015 US News and World Report top 20 list, any author with a Harvard-affiliation, such as an affiliation with Harvard Medical School, would be categorised as being from a top US university. The schools included were: Boston University, Columbia University, Emory University, George Washington University, Harvard University, Johns Hopkins University, Ohio State University, Tulane University, University of Alabama at Birmingham, University of California at Berkeley, University of California at Los Angeles, University of Illinois at Chicago, University of Iowa, University of Michigan at Ann Arbour, University of Minnesota at Twin Cities, University of North Carolina at Chapel Hill, University of Pittsburgh, University of South Florida, University of Washington and Yale University.
## Analyses
We describe paper characteristics, including the region of the paper's focus, whether the paper is a single-author or multiauthor paper, the disease area of focus, the journal's coverage (an African national, an African regional journal or an international journal), the journal's impact factor and the paper's collaborative type. We identified the paper's disease area by searching the title, abstract and keywords for disease areas identified in Chapter 4 of the 2014 African Regional Health ReportHIV/AIDS; tuberculosis; malaria; epidemic/pandemic prone diseases (cholera, Ebola, Marburg, influenza, yellow fever, meningococcal meningitis, Lassa fever, polio); neglected tropical diseases (lymphatic filariasis, helminthiasis, trachoma, schistosomiasis, onchocerciasis, dracunculiasis, yaws, leprosy, trypanosomiasis, Buruli ulcer and leishmaniasis); non-communicable diseases (cardiovascular disease, chronic respiratory disease, BMJ Global Health ); cancer; sickle cell disease; mental and neurological disorders (mental illness, mental disorder, depression, psychiatric disorder) and road traffic injuries. It is possible for papers to have more than one disease area categorisation and for some papers, the paper's focus was outside of the disease areas covered by the 2014 African Regional Health Report. We extracted the 2016 impact factor listed on journal websites. If no impact factor was listed on the website or if the impact factor listed was not from 2016, we used the impact factor listed in the Journal Citation Reports (Clarivate Analytics 2017). If the Journal Citation Reports also did not have the 2016 impact factor, the most recent available impact factor was used. If impact factor was not reported on the journal's website or in Journal Citation Reports, it was listed as 'not available'. We assess the relationship between the paper characteristics and collaborative type using χ² tests.
We present the per cent of authors classified as being from the country of the paper's focus, overall and stratified by authorship position (first, last and middle). We base authorship positions on the author order, with first authors being the first name listed and last authors being the last name listed. We did not account for multiple first or last authors as this information is not currently captured in PubMed. For papers with a single author, we counted that author as the first author and did not report a last author for that paper. For all analyses, we present overall paper and author trends stratified by the paper's geographic collaboration type. We also report the per cent of papers that have no authors from the country of the paper's focus, overall and stratified by collaborative type. We compare results across collaborative types using a χ² tests. All analyses were conducted in Stata V.14.1. The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.
## Patient and public involvement
Neither patients nor the public were involved in this research.
# Results
## Description of included papers
Our initial search yielded 9040 articles, and after applying the various exclusion criteria, yielded a final database of 7100 health research papers focused on a single sub-Saharan African country, corresponding to 43 429 authors (figure 1). Of these authors, 3574 had two or more countries of affiliation. For the papers included, 4299 papers had author affiliation details based on data extracted from PubMed and 2801 had author affiliation details manually extracted. Error rates were less than 7% at all validation steps (figure 1). [fig_ref] Figure 3: Geographic distribution of African health research papers and local authors between 2014-2016 [/fig_ref] maps, by country, the number of papers, the per cent of papers with at least one local author and the per cent of papers with a local first and/or last author (numbers provided in the online supplementary appendix). The largest proportion of papers were from East Africa (n=3092, 43.6%), followed by West Africa (n=2045, 28.8%) (table 1). A small fraction of papers (n=281, 4.0%) had only a single author. A fifth of papers (n=1455) were related to HIV/AIDS and very few (n=22, 0.3%) to sickle-cell disease. The majority of papers (n=6331, 89.2%) were published in international journals, most in journals with an impact factor between 1 to 2.5 (n=3104, 43.7%) or 2.5 to 4 (n=2017, 28.4%).
Authorship structure by geographic collaboration type Of the 7100 papers, 4846 (68.3%) had an author from either the USA, Canada, Europe or another African country outside of the country of the paper's focus. Of all papers, 1192 (16.8%) had authors with an affiliation with a top US university, 2617 (36.9%) had an author from the USA/Canada with 1425 (54.5%) of these having no author with an affiliation with a top US university, 2620 (36.9%) had an author from Europe and 879 (12.4%) had an author from another African country outside of the country of the paper's focus. For the non-overlapping collaborative categories (figure 2), 830 (11.7%) of all papers had collaborators from top US universities only, [fig_ref] Table 2: Association of paper characteristics with collaborative groupings on health research from Africa,... [/fig_ref] : Geographic collaboration type was significantly associated with all assessed paper characteristics. Papers from Southern and West Africa and single author papers were significantly more likely to be authored exclusively by local authors (39.1% of Southern Africa papers, 38.4% of West Africa papers, p<0.001; 49.8% of single author papers vs 31.0% of multiple author papers, p<0.001). Papers published in African journals were more likely to only have local coauthors or to have coauthors from another African institution only, whereas the proportion of those papers with authors from a top US university, the USA/Canada only or Europe only were significantly lower as compared with those published in international journals (p<0.001). Over half of the papers (60.6%) published in journals without impact factors were written with local coauthors only, compared with 4.8% of papers published in journals with an impact factor over 10 (p<0.001).
Structure Papers without any authors from the country of focus Overall, 13.5% (n=958 out of 7100) of papers had no authors from the country of focus and every country included in this analysis had some papers with no local representation, ranging from 5.7% of papers from South Africa (n=63 out of 1097) to 68.4% of papers from Somalia (n=13 out of 19) (figure 3B, online supplementary appendix 1). The per cent of papers with no authors BMJ Global Health 44.0 *Note, it is possible for a paper to have more than one area of focus or to cover a topic not captured on this list. †P value calculated using Fisher's exact test. ‡Possible for a single author paper to be considered as having authors from multiple locales if that single author has multiple affiliations from different locales.
## Table 2 continued
# Discussion
During the 3-year period ending in 2016, we found over 50% of all coauthors and of first authors on papers about health in Africa were from the country of the paper's focus. However, of concern, we found that author representation and position, along with the paper's characteristics, is vulnerable to the types of collaboration. If any coauthors were from the USA, Canada or Europe, then overall representation of local authors dropped, BMJ Global Health particularly in the prime first and last author positions. This was most frequent when collaborators were from a top US university or from multiple locales. These results are supported by other analyses that show that researchers from the poorest countries are less likely to be in prominent author positions when engaged in international collaborations 6 that papers with HIC last authors are significantly less likely to have LMIC first authors [bib_ref] Patterns of authorship on community health workers in low-and-middle-income countries: an analysis..., Schneider [/bib_ref] and that US-funded clinical trials are less likely to have LMIC authors in first or last positions. [bib_ref] Equity in global health research in the new millennium: trends in first-authorship..., Kelaher [/bib_ref] We also found that over one in eight publications including coauthors from the USA, Canada or Europe had no coauthors from the country of the paper's focus, showing that publishing 'about Africa without Africa' remains an ongoing issue despite longstanding efforts to call attention to it. [bib_ref] Addressing inequalities in research capacity in Africa, Volmink [/bib_ref] [bib_ref] Partnering with local scientists should be mandatory, Shuchman [/bib_ref] Several factors may explain our findings. HIC investigators may be more likely to do the work of a first or last author, such as writing the first draft of a paper or overseeing the research (often remotely), because they have technical advantages over their LMIC counterparts. Many African countries have nascent research infrastructures and university training programme, limiting the number of qualified collaborators to lead projects. Beyond technical skills, HIC researchers typically have greater economic and academic resources including extensive institutional infrastructure to lead project administration; stronger and deeper research networks, including greater representation on journal editorial boards and more 'cultural capital' in the forms of accepted credentials and styles of speech. Taken together, these advantages give HIC partners a greater chance at prime authorship along with more control over the nature of the collaboration; yet there are several ways to equalise the playing field. HIC researchers and their institutions could create more opportunities for local collaborators to contribute in ways that lead to prime authorship. [bib_ref] Growing research in global surgery with an eye towards equity, Hedt-Gauthier [/bib_ref] Further, there are calls for HIC researchers to invest in the individual and institutional capacity building of their partners, [bib_ref] Resurrecting the triple threat: academic social responsibility in the context of global..., Manabe [/bib_ref] including investments in skills that promote their leadership of research endeavours. [bib_ref] Claim your space: leadership development as a research capacity building goal in..., Airhihenbuwa [/bib_ref] HIC institutions could incentivise African research leadership by changing their promotion policies and funders could look at African leadership and partners' investments in building capacity during the grant selection process. [bib_ref] Genesis of EDCTP2, Zumla [/bib_ref] [bib_ref] A better measure of research from the global South, Lebel [/bib_ref] [bib_ref] With input from the AHI PHIT partnership collaborative. research capacity building integrated..., Hedt-Gauthier [/bib_ref] Further, others have suggested changing the current norms of authorship or how they are applied, [bib_ref] Applying the ICMJE authorship criteria to operational research in low-income countries: the..., Zachariah [/bib_ref] [bib_ref] Authorship ethics in global health research partnerships between researchers from low or..., Smith [/bib_ref] [bib_ref] Best practice to order authors in Multi/ Interdisciplinary health sciences research publications, Smith [/bib_ref] [bib_ref] Swap outdated authorship listings for contributorship credit, Smith [/bib_ref] moves that would recognise the more diverse roles often taken by LMIC partners or remove pressures to exclude collaborators due to the rewards inherent to authorship.
African scientists might also pursue an equity agenda by turning down some collaborations and seeking others. An interesting finding from our analyses is that Africa-based researchers may more readily hold a prime authorship position if they engage in South-South collaborations with researchers from other African countries. Similar to others, [bib_ref] South-South research collaboration of countries in the southern African development community (SADC), Boshoff [/bib_ref] [bib_ref] South-South collaboration on HIV/AIDS prevention and treatment research: when birds of a..., Bdepfe [/bib_ref] [bib_ref] International collaboration in scientific publishing, Mêgnigbêto [/bib_ref] [bib_ref] Knowledge production through collaborative research in sub-Saharan Africa: how much do countries..., Onyancha [/bib_ref] we found that intra-African collaborations produced a small fraction of the health research in Africa, and of these, South African researchers were the most common collaborative partner. Beyond the benefit of local representation in first and last authorship positions, these collaborations likely focus on research priorities of the region. [bib_ref] The influence of collaboration in research priorities: the SADC case, Pouris [/bib_ref] However, there are distinct consequences, as papers resulting from intra-African collaborations are less likely to be cited [bib_ref] Knowledge production through collaborative research in sub-Saharan Africa: how much do countries..., Onyancha [/bib_ref] and established collaborations outside of Africa is a significant predictor of an African scientist's productivity. [bib_ref] The characteristics of highly cited researchers in, Confraria [/bib_ref] This paper is subject to several limitations that should be noted when interpreting the results. First, we recognise that the data extraction process, the application of inclusion/exclusion criteria, and the process of determining where an author is from, are all subject to errors. The extraction process excluded articles published in non-indexed journals, which includes many national or regional African journals; this likely decreases the overall representation of local authors and papers collaborating with researchers from other Africa countries. However, given that collaborators from Africa were significantly less likely to publish in these types of journals, the overall effect of collaborations with researchers outside of Africa on inclusion of local authors is not likely to be affected. Further, the requirement of the word 'health' likely excluded non-English articles. We note that local and non-English journals provide important forums for publishing local research [bib_ref] Quest for publication metrics undermines regional research, Neff [/bib_ref] and that local authors are more likely to be engaged in these publications. We did validate all major steps of exclusion of this process, and the error rates were low. Further, we conducted a sensitivity analysis changing how author's country was classified based on affiliation and all trends persisted.
A final limitation is the focus on top US universities instead of premier institutions in other HICs and how the top US universities were identified. We chose to focus on the US institutions because of the noted 'publish or perish' culture, but recognise that the issues elucidated in this paper may present at top institutions in Canada, Europe and elsewhere. Further, other university ranking systems could have been used to identify authors affiliated with a top US university; the use of the US News and World report list provides at minimum a proxy for general trends.
While authorship alone does not ensure that collaborations are inclusive and equitable, authorship is an important indicator for who is benefiting from research endeavours. Under the current system, researchers from sub-Saharan countries are underrepresented in prime authorship positions in health-related literature about their own countries and our data suggest that collaborations between researchers in sub-Saharan Africa and academics from HICs may perpetuate this situation, particularly when the HIC collaborators hold posts at top US universities. Our data also show that a significant fraction of papers about an African country have no coauthors from that country.
Additional scholarship is needed to identify the specific features of North-South collaborative research that could BMJ Global Health be blocking skilled African researchers from authorship entirely or locking them into middle authorship. Further, this scholarship must be coupled with open debates about how to shift these dynamics, including open discussions on the presence and consequence of inequities in global health collaborations. [bib_ref] Developing world: discuss inequality, Geissler [/bib_ref] [bib_ref] Academic Collaborations: Do's and Don'ts, Kyamanywa [/bib_ref] [bib_ref] The way the country has been carved up by researchers": ethics and..., Walsh [/bib_ref] Only by fully understanding the basis for these inequities can the various parties involved-sponsors, journals, HIC and LMIC researchers and their institutions and governments-take the necessary steps to raise the stature of research collaborators in LMICs to true partners.
[fig] ►: Global health research often relies on collaborations between institutions. Previous studies have noted underrepresentation of local coauthors in collaborative research, particularly in the prime first and last author positions. [/fig]
[fig] Figure 1: Process of identifying eligible papers and author affiliations for bibliometric analysis of African health research, 2014-2016. [/fig]
[fig] Figure 2: Definition of collaboration types considered in bibliometric analyses of authorship on publications about health in Africa, 2014-2016. [/fig]
[fig] Figure 3: Geographic distribution of African health research papers and local authors between 2014-2016. (A) shows distribution of number of papers, (B) shows per cent of papers with at least one coauthor from the country of the paper's focus and (C) shows per cent of papers with first and/or last author from the country of the paper's focus. [/fig]
[fig] Figure 4: Per cent of authors from the country of the paper's focus, stratified by authorship position and collaboration type. The top figure includes results based on primary classification, where HIC and African affiliation defaults to author being from HIC. The bottom has results from the sensitivity analysis, where HIC and African affiliation defaults to author being from Africa. HIC, high-income country. from the country of focus is higher for any of the five collaborative types studied: for papers including top US university authors only, 16.8% (n=139 out of 830); USA/ Canada authors (excluding top US universities) only, 21.4% (n=196 out of 916); European authors only, 18.9% (n=315 out of 1665); other African authors only, 23.1% (n=70 out of 303) or authors from multiple locales, 17.6% (n=199 out of 1132). The per cent of papers without a local author differed significantly across collaboration types (p=0.027).Excluding single author papers (n=281, 4.0% of all papers) since these often represent correspondence or invited editorials, 11.9% (n=812 out of 6819) of papers had no representation from the country of focus; this includes 15.3% (n=125 out of 816) of papers with authors from a top US university only, 16.6% (n=143 out of 863) with authors from the USA/Canada (excluding top US universities), 15.8% (n=253 out of 1603) with authors from Europe, 20.2% (n=59 out of 292) with authors from other African countries and 17.5% (n=198 out of 1131) with author from multiple locales. The difference by collaboration types was no longer significant (p=0.272). [/fig]
[table] Table 1: Description of papers included in systematic review of health research in Africa, 2014-2016 [/table]
[table] Table 2: Association of paper characteristics with collaborative groupings on health research from Africa, 2014-2016 For papers with collaborators outside of country of focus, authors include individuals from: [/table]
|
Radiation-Induced Sarcomas of the Head and Neck: A Systematic Review
Introduction: As a result of the increased use of radiotherapy (RT) and improved long-term overall survival of patients with cancers of the head and neck (HN), the frequency of radiationinduced sarcomas of the head and neck (RISHN) may be increasing. The main objective of this systematic review was to determine the existing evidence on the frequency, treatment, and outcome of RISHN. Methods: Using PRISMA guidelines we conducted a systematic review of the literature
# Introduction
Radiation-induced sarcomas of the head and neck (RISHNs) pose a challenging entity in head and neck (HN) oncology. This tumor appears in irradiated tissues of the HN without any subsite predilection. The median latency after radiotherapy (RT) is reported to be 10-12 years. The clinical features of RISHNs can be similar to other more common RT complications, like radionecrosis, which makes their diagnosis difficult in early stages, but sometimes symptoms are related to the region where the tumor appears: diplopia, neck mass, jaw numbness, or asymmetry in the HN region. The diagnostic criteria originally proposed by Cahan and Woodardand modified by Murray et al.are still frequently used and include (1) the tumor arises in a field that has been previously irradiated,there is evidence that the first tumor differs histologically from the subsequent one,there was no evidence of the new tumor at the time of RT, and (4) the new tumor developed after a latency period following RT.
RT has become one of the cornerstones in the treatment of HN cancer, paralleling an increase in long-term patient survival. RISHNs are rare tumors, but the population growth, increased access to and use of RT, the increasing survival numbers, and advances in diagnostics have in turn been accompanied by an obvious increase in the frequency of RISHNs. In the study by Wei et al., the incidence of RISHN increased from 0.06% to 0.17% from 1960 to 2010 at their institution.
The most common histologic subtypes are osteosarcoma, undifferentiated pleomorphic sarcoma, and fibrosarcoma. They do not show any subsite predilection and can arise in any irradiated tissue of mesenchymal origin within the HN, with most of these tumors being of high grade.
Management of RISHNs remains controversial and all treatment modalities have been applied: surgery and chemotherapy, but also re-irradiation and a variety of combination of these modalities. The overall prognosis is still poor with a 5-year overall survival rate varying between 24.2% and 38.2%.
The main objective of this systematic review was to determine the existing evidence on the frequency, treatment, and outcome of RISHN.
# Methods
The Preferred Reporting Items for Systematic Review and Meta-Analyses (PRISMA) method was used to analyze the current literature. The search strategy aimed to include articles concerning the development of RISHNs in patients treated with RT. A PubMed search updated to July 24, 2020 was performed for publications in English between 2000 and 2020 using the following search criteria in the title or abstract: ''head and neck'' coupled with ''radiation induced sarcomas'', ''radiation associated sarcomas'', and ''radiation induced malignancy''. The search results were reviewed for potentially eligible studies. When there was any information in the abstract about the study addressing RISHNs, the full text article was searched. All review articles were also screened to identify reports on these patients. References from any full text articles were cross-checked to ensure inclusion of all relevant publications Articles included in the study (n = 64) Flowchart showing the study selection process for our systematic review following inclusion criteria: (a) patients treated with RT to their primary tumor, (b) radiationinduced sarcoma developed in the HN area, and (c) histological confirmation of the sarcoma has been obtained. Studies in which the RISHNs were analyzed together with non-HN radiationinduced sarcomas were excluded.
The statistical analysis was performed with SPSS statistical software (IBM SPSS 19.0 Statistical Software package). The Spearman's rho correlation coefficient was used for comparison between the age of the patients and the latency period to development of the sarcoma. In addition, differences between mean latency times by age group were analyzed using the analysis of variance (ANOVA) test. P values of 0.05 or less were considered statistically significant.
This article is based on previously conducted studies and does not contain any studies with human participants or animals performed by any of the authors.
# Results
Our search criteria identified 3299 papers and, after removal of duplicates and those that did not meet our inclusion criteria, 64 were selected for data review as summarized in. Most papers were excluded because they involved radiation-induced sarcomas outside the HN area. Our review includes 560 patients with 561 RISHNs (one patient had two simultaneous tumors) from 64 articles during a period of 20 years (2000-2020). Some of the parameters addressed were not available in all studies. There were 363 men and 179 women (male-to-female ratio, 2:1). The most frequently reported location of the primarily treated tumor was the nasopharynx (368 cases, or 65.7%), followed by the sinonasal region (21 cases, or 3.7%), central nervous system (19 cases, or 3.4%), eye (14 cases, or 2.5%), and larynx (13 cases, or 2.3%). The mean RT dose prescribed to the primary tumor was 62 Gy (range 36-101). RISHNs most frequently appeared in the sinonasal region (33.7%), followed by the mandible and the neck (15.9% and 13.4%, respectively). The most common histologies were osteosarcoma (35.5%) and fibrosarcoma (23%). The median age at diagnosis of RISHN was 52 years (interquartile range 21.5) and the mean latency between the initial RT treatment and the diagnosis of RISHN was 11.1 years (range 1.3-38). The mean age at the treatment of the primary tumor was 43.1 years (range 2-81), but still only 37.5% of the patients had a mean age of less than 40 years. Most of the patients (62.5%) in our review had a mean age of 40 years or older when they were treated for the primary tumor with RT. This leads us to think that RISHNs are less frequent in young patients who are treated with RT. In order to investigate if age at the time of the primary RT treatment is related to latency time, we analyzed these two groups separately, i.e., patients under 40 years old and patients older than 40 years. In the group of patients under 40 years of age, the mean age was 26 years (range 2-38.6) with a mean latency until diagnosis of the sarcoma of 14.1 years (range 2.3-38). On the other hand, in the group of patients aged older than 40 years, the mean age was 55.1 years (range 40-81) and a mean latency until appearance of the sarcoma of 9.6 years (range 1..
The Spearman's rho correlation coefficient was used for comparison between these two groups. The result was that there is an inverse correlation (q = -0.423) between age at the primary RT treatment and latency time, and this result was statistically significant (P = 0.001). So, the older the patient, the shorter the latency period. Differences between mean latency times in these two age groups were analyzed using the ANOVA test, and the result was that the latency time is greater in patients under 40 years of age and the difference is statistically significant (P = 0.01). Therefore, older patients treated with RT develop the sarcoma sooner after treatment than younger patients.
The most frequently used treatment modality was surgery (45.1%), and less commonly multimodal treatment (24.1%). Chemotherapy or RT alone were used less often (8.9% and 5.2%, respectively) and these modalities were mostly applied as palliative treatment. A total number of 228 patients died of the RISHN (40.7%) after a mean time of 13.9 months In the 2011-2020 group, the patients died of disease after a mean survival time period of 13.9 months (range 3-48), the mean follow-up time for those alive with no evidence of disease was 29.2 months (range 4-120), and for those alive with disease it was 12.2 months (range 7-17). Therefore, it seems that the mean life expectancy for those patients dying of RISHN has remained practically the same in the articles published during the past two decades. The follow-up times for the patients alive were longer in the earlier group and thus no conclusions can be drawn regarding their results. The frequency of reported RISHNs in the published series is given in. In total, there were 252 RISHNs among 168,115 patients who had been treated with RT, with a mean frequency of 0.15%. It should be noted that this frequency in the present systematic review is reported according to the date of publication of the reviewed articles, i.e., between 2000 and 2020. However, the included patient series in these reviews cover much longer period, specifically from 1964 to 2015.
# Discussion
RISHNs remain a rare late complication of RT, and typically occur between 10 and 12 years after RT. The effectiveness of RT for treatment of cancers and its increasingly widespread use have led to a parallel surge in adverse effects arising from it. This rise may be also due to the higher long-term survival rate generally observed in HN cancers (owing to early diagnosis, improved oncological and surgical techniques, and the use of systemic therapy), and the use of higher RT doses. In the present review, we found a frequency of 0.15% for RISHNs in the series published during the two past decades. The significance of this finding remains questionable, as only eight series with 252 patients were included, although the whole cohort having received RT in the included reports was rather representative with 168,115 patients.
The exact etiology of RISHNs remains unknown, but radiation likely induces their development by causing damage to DNA within the cells. There is no safe threshold in terms of RT dose, below which there is no risk of causing a second tumor. The results of some studies indicated that the risk of developing RISHN increases with accumulated radiation dosage. More specifically, in one study a total dose higher than 30 Gy increased the risk of developing RISHN. As we included only patients treated with curative intent, the mean RT dose received for treating the primary tumor was 62 Gy (range 36-101) and thus these patients were considered at risk for developing RISHN. We are aware that sarcomas sometimes appear in areas of medium dose within the primary RT field, so the dose received in the primary tumor may not be as important as the dose delivered in the area where the sarcoma appears. Unfortunately, these data are usually not available in the published articles, and instead only the mean dose is reported. The age range of patients with RISHNs varied among studies, but usually occurs between the 6th and 7th decade of life. We found that the median age at diagnosis was 52 years (interquartile range 21.5), which is in line with what has been published previously.
The mean latency period in our review was 11.1 years, but with a wide range (1.3-38 years), possibly because there are many single cases published. In the past, the latency time factor was used to differentiate radiation-induced sarcomas from de novo sarcomas, but now most studies do not take that time period into consideration and those that occur with short latency periods are included as radiation-induced sarcomas, since they are impossible to differentiate. There are studies demonstrating a shorter latency period (between 7.7 and 9.3 years) than the one reported in our review (11.1 years), but the mean age in these series was almost 50 years. This is in accordance with the latency period of 9.6 years for the patients over the age of 40 years in the present review.
The location of the RISHN depends essentially on the location of the primary tumor. Nasopharyngeal cancer irradiation was typical in the present review and has been reported earlierand consequently the most common site of RISHN being the paranasal sinuses. Sinonasal tract, central nervous system, eye, and larynx were frequent primary tumor sites as well among the present studies.
With regard to the different histological RISHN subtypes that can be observed, the following histologies have been described: fibrosarcoma, osteosarcoma, undifferentiated pleomorphic sarcoma, chondrosarcoma, angiosarcoma, leiomyosarcoma, rhabdomyosarcoma, spindle cell sarcoma, and others. Cai et al.found that the most common histologies were fibrosarcoma (44.1%) and osteosarcoma (30.5%). Zhu et al.found osteosarcoma to be the most common histology (34.1%), followed by fibrosarcoma (19.2%) and undifferentiated pleomorphic sarcoma (15.8%). In the study by Thiagarajan and Gopalakrishna Iyer, undifferentiated pleomorphic sarcoma was the most frequent histology. In our study, the most frequently encountered histologic types were osteosarcoma (35.5%) and fibrosarcoma (23%), which is in line with some of the previously published studies. Less common were undifferentiated pleomorphic sarcoma (13%) and sarcomas not otherwise specified (10.2%).
Treatment of RISHN varied widely between the studies. In our review, almost half of the patients received surgical treatment (45.1%) when the lesion was resectable. Other modalities, like RT or chemotherapy alone or multimodality treatment, were used less commonly. As RISHNs are rare tumors, with varying histologies at different locations and the available knowledge being based on retrospective studies, it is difficult to define clear recommendations for their treatment. As RISHNs most frequently occur in previously irradiated areas that received a high dose, surgery remains the preferred treatment modality in case RISHN is considered resectable. Adequate surgical margins are difficult to achieve, and this is even more difficult when neighboring vital structures and the previous RT-induced fibrous transformation of tissues with impaired blood flow further complicate surgery and increase complication rates. In unresectable cases, definitive radiotherapy alone or combined with systemic treatment may be considered but is generally associated with a high risk of severe radiation-induced side effects and, furthermore, there have been no advances in terms of effective systemic therapies. According to some authors, chemotherapy will be less effective in radiation-induced sarcomas compared with de novo sarcomas because of fibrotic changes in the previously irradiated field and, consequently, impaired vascularity thus impeding chemotherapeutic agents from reaching adequate concentrations in the target area, but the evidence for this is lacking.
Cha et al.reviewed the results obtained by surgical resection of RISHNs and found that the 1-year, 3-year, and 5-year overall survival rates were 78%, 58%, and 41%, respectively, with a median survival of 48 months. The 3-year overall survival rate reported by Xi et al.was 32.4% and the 5-year overall survival rate reported by Chan et al.was 24.2%. For Yeang et al.the 2-year and 5-year overall survival rates were 57.3% and 38.2%, respectively. Unfortunately, overall survival data were not available in most of the studies in this review, given the large number of included case reports.
# Conclusion
Our review represents the most comprehensive systematic analysis of RISHNs to date, collecting data from 560 patients. Most of the included studies were case reports or only comprised a few cases. We herein emphasize that RISHNs are a rare but lethal complication of RT. The frequency has remained rather stable. Surgery with free margins remains the treatment option of choice for RISHNs. However, given the location of the tumor, this may be difficult to achieve in some cases.
## Acknowledgements
Funding. No funding or sponsorship was received for this study or publication of this article.
Authorship. All named authors meet the International Committee of Medical Journal Editors (ICMJE) criteria for authorship for this article, take responsibility for the integrity of the work as a whole, and have given their approval for this version to be published.
Disclosures. Andrés Coca-Pelaz, Antti A. Mäkitie, Primož Strojan, June Corry, Avraham Eisbruch, Jonathan J. Beitler, Sandra Nuyts, Robert Smee, Johannes A. Langendijk, William M. Mendenhall and Cesare Piazza have nothing to disclose. Alessandra Rinaldo and Alfio Ferlito are editorial board members of Advances in Therapy but have nothing else relevant to disclose.
Compliance with Ethics Guidelines. This article is based on previously conducted studies and does not contain any studies with human participants or animals performed by any of the authors.
Data Availability. Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.
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://creativecommons.org/licenses/bync/4.0/. |
A member of the Whirly family is a multifunctional RNA- and DNA-binding protein that is essential for chloroplast biogenesis
Whirly' proteins comprise a plant-specific protein family whose members have been described as DNA-binding proteins that influence nuclear transcription and telomere maintenance, and that associate with nucleoids in chloroplasts and mitochondria. We identified the maize WHY1 ortholog among proteins that coimmunoprecipitate with CRS1, which promotes the splicing of the chloroplast atpF group II intron. ZmWHY1 localizes to the chloroplast stroma and to the thylakoid membrane, to which it is tethered by DNA. Genome-wide coimmunoprecipitation assays showed that ZmWHY1 in chloroplast extract is associated with DNA from throughout the plastid genome and with a subset of plastid RNAs that includes atpF transcripts. Furthermore, ZmWHY1 binds both RNA and DNA in vitro. A severe ZmWhy1 mutant allele conditions albino seedlings lacking plastid ribosomes; these exhibit the altered plastid RNA profile characteristic of ribosome-less plastids. Hypomorphic ZmWhy1 mutants exhibit reduced atpF intron splicing and a reduced content of plastid ribosomes; aberrant 23S rRNA metabolism in these mutants suggests that a defect in the biogenesis of the large ribosomal subunit underlies the ribosome deficiency. However, these mutants contain near normal levels of chloroplast DNA and RNAs, suggesting that ZmWHY1 is not directly required for either DNA replication or for global plastid transcription.
# Introduction
Plant mitochondrial and chloroplast genomes encode $50 and $100 products, respectively, most of which participate in basal organellar gene expression or energy transduction. Post-transcriptional events play the dominant role in dictating gene product abundance in both organelles [bib_ref] Plastid and plant mitochondrial RNA processing and RNA stability, Marchfelder [/bib_ref]. In fact, the two organelles house a similar repertoire of RNA-processing pathways that includes RNA editing, group II intron splicing and endonucleolytic processing. Genetic and bioinformatic analyses suggest that many hundreds of nuclear genes encode organellelocalized nucleic acid binding proteins and influence organellar gene expression [bib_ref] The CRM domain: an RNA binding module derived from an ancient ribosome-associated..., Barkan [/bib_ref] [bib_ref] Chloroplast RNA-binding proteins, Nickelsen [/bib_ref] [bib_ref] Comparative survey of plastid and mitochondrial targeting properties of transcription factors in..., Schwacke [/bib_ref] , but only a small fraction of such genes has been studied.
The protein that is the focus of this study, ZmWHY1, came to our attention during our characterization of the chloroplast RNA splicing machinery. Nine nucleusencoded proteins that are necessary for the splicing of various subsets of the $20 chloroplast introns in vascular plants have been reported [bib_ref] Recruitment of a peptidyl-tRNA hydrolase as a facilitator of group II intron..., Jenkins [/bib_ref] [bib_ref] A CRM domain protein functions dually in group I and group II..., Asakura [/bib_ref] [bib_ref] Arabidopsis orthologs of maize chloroplast splicing factors promote splicing of orthologous and..., Asakura [/bib_ref] [bib_ref] Group II intron splicing factors derived by diversification of an ancient RNA..., Ostheimer [/bib_ref] [bib_ref] CRS1 is a novel group II intron splicing factor that was derived..., Till [/bib_ref] [bib_ref] A ribonuclease III domain protein functions in group II intron splicing in..., Watkins [/bib_ref] [bib_ref] Two CRM protein subfamilies cooperate in the splicing of group IIB introns..., Asakura [/bib_ref] [bib_ref] The pentatricopeptide repeat gene OTP51 with two LAGLIDADG motifs is required for..., Falcon De Longevialle [/bib_ref] [bib_ref] A pentatricopeptide repeat protein facilitates the trans-splicing of the maize chloroplast rps12..., Schmitz-Linneweber [/bib_ref]. One of the first to be characterized, CRS1, is necessary for the splicing of the group II intron in the chloroplast atpF gene [bib_ref] Arabidopsis orthologs of maize chloroplast splicing factors promote splicing of orthologous and..., Asakura [/bib_ref] , and binds specifically to that intron in vivo and in vitro [bib_ref] Group II intron splicing factors derived by diversification of an ancient RNA..., Ostheimer [/bib_ref] [bib_ref] CRS1 is a novel group II intron splicing factor that was derived..., Till [/bib_ref] [bib_ref] CRS1, a chloroplast group II intron splicing factor, promotes intron folding through..., Ostersetzer [/bib_ref]. However, the large size of the particles containing CRS1 and atpF intron RNA in vivo, and the fact that CRS1 is not sufficient to promote atpF intron splicing in vitro suggested that additional proteins are involved. We therefore used mass spectrometry to identify proteins that coimmunoprecipitate with CRS1; ZmWHY1 was one such protein.
ZmWHY1 is a member of the 'Whirly' protein family, whose orthologs in potato (StWHY1) and Arabidopsis (AtWHY1) were reported to be nuclear transcription factors involved in pathogen-induced transcription [bib_ref] A ''Whirly'' transcription factor is required for salicylic aciddependent disease resistance in..., Desveaux [/bib_ref] [bib_ref] PBF-2 is a novel single-stranded DNA binding factor implicated in PR-10a gene..., Desveaux [/bib_ref]. StWHY1 and AtWHY1 bind single-stranded DNA (ssDNA) in vitro, and StWHY1 adopts a propeller-like structure from which the family acquired its name [bib_ref] A ''Whirly'' transcription factor is required for salicylic aciddependent disease resistance in..., Desveaux [/bib_ref] [bib_ref] A new family of plant transcription factors displays a novel ssDNA-binding surface, Desveaux [/bib_ref]. AtWHY1 has also been implicated in telomere binding and maintenance [bib_ref] Singlestranded DNA binding factor AtWHY1 modulates telomere length homeostasis in Arabidopsis, Yoo [/bib_ref]. Additional functions for members of the Whirly family were suggested by the fact that GFP fused to each member of the family from Arabidopsis localizes to chloroplasts or mitochondria [bib_ref] DNA-binding proteins of the Whirly family in Arabidopsis thaliana are targeted to..., Krause [/bib_ref]. The copurification of AtWHY1 with a transcriptionally active chloroplast DNA complex [bib_ref] and -12 are components of the transcriptionally active plastid chromosome that are..., Pfalz [/bib_ref] and the association of AtWHY2 with mitochondrial nucleoids [bib_ref] Overexpression of mtDNAassociated AtWhy2 compromises mitochondrial function, Marechal [/bib_ref] confirmed that these proteins have organellar functions, but the nature of these functions is not known. Results presented here show that ZmWHY1 plays an essential role in the biogenesis of chloroplasts, that it is associated with DNA from throughout the chloroplast genome and that it interacts in vivo with a subset of chloroplast RNAs that includes the atpF intron. ZmWHY1 enhances atpF intron splicing and influences the biogenesis of the large ribosomal subunit. However, chloroplast DNA and RNAs in ZmWhy1 mutants accumulate to levels similar to those in other mutants with plastid ribosome deficiencies of similar magnitude. These results argue that ZmWHY1 is required neither for chloroplast DNA replication nor directly for global chloroplast transcription.
# Materials and methods
## Purification of crs1 ribonucleoproteins and mass spectrometry
Purification of CRS1 ribonucleoprotein particles and mass spectrometry were performed as described for CAF1 and CAF2 particles in [bib_ref] A ribonuclease III domain protein functions in group II intron splicing in..., Watkins [/bib_ref]. The antibody to CRS1 was described previously [bib_ref] CRS1 is a novel group II intron splicing factor that was derived..., Till [/bib_ref].
# Plant material
Our collection of Mu transposon-induced nonphotosynthetic maize mutants (http://chloroplast.uoregon.edu/) was screened by PCR to identify insertions in ZmWHY1, using methods described in [bib_ref] A chloroplast-localized PPR protein required for plastid ribosome accumulation, Williams [/bib_ref] and a ZmWhy1-specific primer (5 0 -CGGCGGCCTTTCTGGA GGA-3 0 ) in conjunction with a Mu terminal inverted repeat primer (5 0 -GCCTCCATTTCGTCGAATCCCG-3 0 ). The alleles were tested for complementation by crossing phenotypically normal siblings (+/+ or +/À) from ears segregating each allele. Seventy-four ears were recovered, 36 of which segregated chlorophyll-deficient mutants. Other mutants used in this work include iojap [bib_ref] Nuclear gene iojap conditions a programmed change to ribosome-less plastids in Zea..., Walbot [/bib_ref] , hcf7 [bib_ref] Nuclear mutants of maize with defects in chloroplast polysome assembly have altered..., Barkan [/bib_ref] and crs1 [bib_ref] Nuclear mutations that block group II RNA splicing in maize chloroplasts reveal..., Jenkins [/bib_ref]. The inbred line B73 (Pioneer HiBred) was used as the source of wild-type tissue for coimmunoprecipitation, sucrose gradient and chloroplast fractionation experiments. Plants were grown in soil in a growth chamber (16 h light, 248C)/8 h dark, 198C). Leaf tissue was harvested $9 days after planting.
## Generation of recombinant zmwhy1 for antibody production and binding assays
ESTs representing ZmWhy1 were identified as GenBank accessions DV170433 and DV503865; the corresponding cDNAs were obtained from the maize full-length cDNA project (http://www.maizecdna.org/). The complete cDNA sequence was determined and has been entered in GenBank under Accession EU595664. A ZmWHY1 protein fragment (amino acids 86 to 258) with a C-terminal 6x-histidine tag was expressed in Escherichia coli from pET28b (Novagen), purified by nickel affinity chromatography and used for the production of polyclonal antisera in rabbits at the University of Oregon antibody facility. Full-length mature ZmWHY1 (i.e. lacking the transit peptide) for nucleic acid binding assays was generated by PCR amplification of its coding sequence from the cDNA (primers 5 0 -TATAGGATCCGCCTCCTCCCGT AAG-3 0 and 5 0 -TATAGTCGACTCACCGACGCCATT C-3 0 ), digestion of the product with BamHI and SalI, and cloning into pMAL-TEV. Subsequent steps in expressing and purifying recombinant ZmWHY1 were as described previously for RNC1 [bib_ref] A ribonuclease III domain protein functions in group II intron splicing in..., Watkins [/bib_ref].
## Chloroplast fractionation and protein analysis
Leaf protein extracts were prepared and analyzed as previously described [bib_ref] Approaches to investigating nuclear genes that function in chloroplast biogenesis in land..., Barkan [/bib_ref]. Chloroplast subfractions were those described by . For RNAse and DNAse treatment of thylakoid membranes, MgCl 2 was added to a thylakoid membrane fraction to a concentration of 15 mM. The sample was divided into three 20 ml aliquots: 1 ml RNAse-free RQ1 DNAse (1 U/ml) (Promega, Madison, WI, USA), 1 ml of RNAse A (1 mg/ml), or 1 ml water was added for the DNAse, RNAse, and mock treatments, respectively. Samples were incubated at room temperature for 30 min and then centrifuged at 48C at 15 000g for 15 min. The pellet was resuspended in 10 mM Tris-HCl pH 7.5, 2 mM EDTA, 0.2 M sucrose, to a volume equivalent to that of the supernatant. The supernatant and pellet fractions were analyzed by SDS-PAGE and immunoblotting. Sucrose gradient sedimentation of stromal extract was performed as described by Jenkins and Barkan (7); aliquots of stroma were treated with either 3 units RQ1 DNAse or 50 mg/ml RNAse A for 30 min at room temperature prior to centrifugation. Antisera to spinach chloroplast RPL2 and MDH were generously provided by A. Subramanian (University of Arizona) and Kathy Newton (University of Missouri), respectively. The other antibodies were generated by us and described previously [bib_ref] Nuclear genes required for posttranslational steps in the biogensis of the chloroplast..., Voelker [/bib_ref].
## Nucleic acid coimmunoprecipitation assays
One hundred microliter aliquots of stromal extract ($500 mg of protein) were analyzed by RIP-chip, DIPchip and slot-blot hybridization using methods described in [bib_ref] Genome-wide analysis of RNA-protein interactions in plants, Barkan [/bib_ref] , except that stroma used for RIP-chip assays was treated with DNAse prior to immunoprecipitation (10 units RQ1 DNAse at 378C for 30 min) and again after purification of nucleic acids from the immunoprecipitation. For DIP-chip assays, RNAse A (100 mg/ml final concentration) was added to stroma prior to immunoprecipitation and residual RNA was removed from the recovered nucleic acids by alkali hydrolysis in 200 mM NaOH at 708C for 30 min.
## Analysis of dna and rna
DNA extraction from leaf tissue and Southern blot analysis were performed as previously described [bib_ref] Transposondisruption of a maize nuclear gene, tha1, encoding a chloroplast SecA homolog:..., Voelker [/bib_ref]. Leaf RNA was extracted from the middle of the second leaf of 9-day old seedlings, with Tri Reagent (Molecular Research Center, Cincinnati, OH, USA). RNA gel blot hybridizations were performed as previously described [bib_ref] Approaches to investigating nuclear genes that function in chloroplast biogenesis in land..., Barkan [/bib_ref] Poisoned primer extension assays to distinguish mature from precursor RNAs were performed as previously described (9) using the following primers and dideoxynucleotide: rrn23, 5 0 -CGCAAGCCT TTCCTCTTTT-3 0 (ddTTP); rpl2, 5 0 -GGCCGTGCCTAA GGGCATATC-3 0 (ddCTP); rps12, 5 0 -GGTTTTTTGGG GTTGATAG-3 0 (ddCTP). Radioactive gels and blots were imaged with a phosphorimager and analyzed using ImageQuant software (GE Healthcare).
## Nucleic acid binding assays
Gel mobility shift assays were performed with the same substrates and procedures as described in Watkins et al. [bib_ref] A ribonuclease III domain protein functions in group II intron splicing in..., Watkins [/bib_ref] except that the binding reactions contained 150 mM NaCl, 5 mM DTT, 50 mg/ml BSA, 25 mM Tris-HCl pH 7.5, 0.1 mg/ml heparin. Filter-binding assays were based on the procedure of Wong and Lohman (31) with modifications [bib_ref] CRS1, a chloroplast group II intron splicing factor, promotes intron folding through..., Ostersetzer [/bib_ref]. The atpF intron RNA substrate for filterbinding assays was transcribed in vitro by T7 RNA from a PCR product generated with the following primers: atpF forward/T7 promoter, 5 0 -TAATACGACTCACTATAGG GATGAAAAATGTAACCCATTCTT-3 0 ; atpF reverse, 5 0 -AATGAAAGTAGATTATCTTGC-3 0 . The RNA, which included atpF exon 1 and the complete intron, was heated in TE to 908C for 2 min and then placed on ice immediately prior to its addition to binding reactions (300 mM NaCl, 5 mM DTT, 50 mg/ml BSA and 25 mM Tris pH 7.5, 308C for 30 min).
## Chloroplast run-on transcription assay
The chloroplast run-on transcription assay was performed as described by . The radiolabeled products were hybridized to the following synthetic oligonucleotides (10 pmol/slot) that had been applied with a slot-blot manifold to a nylon membrane:
[formula] rrn16 5 0 -CC CATTGTAGCACGTGTGTCGCCCAGGGCATAAG GGGCATGATGACTTGG-3 0 , rrn23 5 0 -GGACTCTTG GGGAAGATCAGCCTGTTATCCCTAGAGTAACT TTTATCCGA-3 0 , trnG 5 0 -CATCTATGTCAGCTTTTC TGTCTGAATGGAACCAAAGCTCTCCGCTTTCTA GATGC-3 0 and CFM3 5 0 -ATACTCGAGCGAAAAACA GGAGGATTAGTAATCTGGCGATCAGGGACTTC TGTTTCTCTGTACCGGGGAGTAGATTATGATGA ACC-3 0 . [/formula]
# Results
## Identification of zmwhy1 in crs1 coimmunoprecipitates
To find proteins involved in the splicing of the atpF intron we used mass spectrometry to identify proteins that coimmunoprecipitate with the atpF splicing factor CRS1.
Stromal extract was initially fractionated on a sucrose gradient, and the fractions that contained the majority of the CRS1 ribonucleoprotein particles ($600-700 kDa) were used for immunoprecipitation. The immunoprecipitated proteins were separated by SDS-PAGE, and contiguous gel slices containing proteins between $20 and $120 kDa were used for in-gel trypsin digests and tandem mass spectrometry. Among the proteins identified was a member of the Whirly protein family (Supplementary , Supplementary [fig_ref] Figure 1: Mutant alleles of ZmWhy1 [/fig_ref] [bib_ref] A ''Whirly'' transcription factor is required for salicylic aciddependent disease resistance in..., Desveaux [/bib_ref] [bib_ref] A new family of plant transcription factors displays a novel ssDNA-binding surface, Desveaux [/bib_ref]. The Whirly protein family in vascular plants includes two orthologous groups (Supplementary [fig_ref] Figure 1: Mutant alleles of ZmWhy1 [/fig_ref]. The peptides detected in the CRS1 coimmunoprecipitate identified the protein as a member of the orthologous group designated Why1.
## Recovery of zmwhy1 insertion mutants
To elucidate the function of ZmWHY1 we sought insertion mutants in a reverse-genetic screen of our collection of transposon-induced non-photosynthetic maize mutants (http://pml.uoregon.edu/). Two mutant alleles were recovered [fig_ref] Figure 1: Mutant alleles of ZmWhy1 [/fig_ref] : the Zmwhy1-1 allele has a MuDR transposon insertion 35-bp downstream of the predicted start codon and conditions an ivory leaf phenotype; the Zmwhy1-2 allele has a Mu1 or Mu1.7 insertion 38-bp upstream of the predicted start codon and conditions a pale green leaf phenotype. The heteroallelic progeny of complementation crosses (Zmwhy1-1/-2) exhibit an intermediate phenotype [fig_ref] Figure 1: Mutant alleles of ZmWhy1 [/fig_ref]. Homozygous mutant plants die after the development of three to four leaves, as is typical of non-photosynthetic maize mutants.
A polyclonal antibody was raised to a recombinant fragment of ZmWHY1. This antibody detected a leaf protein whose size is consistent with that anticipated for ZmWHY1 ($25 kDa) (data not shown) and whose abundance is reduced in ZmWhy1 mutants [fig_ref] Figure 1: Mutant alleles of ZmWhy1 [/fig_ref] , indicating that the detected protein is ZmWHY1. The ZmWHY1 antibody coimmunoprecipitated CRS1 [fig_ref] Figure 1: Mutant alleles of ZmWhy1 [/fig_ref] from chloroplast extract, confirming that CRS1 and ZmWHY1 associate with one another. This association was disrupted by treatment with ribonuclease A [fig_ref] Figure 1: Mutant alleles of ZmWhy1 [/fig_ref] , indicating it is mediated by RNA. Results described below show that atpF intron RNA, which was shown previously to associate with CRS1 in vivo (10,11), mediates the CRS1/ZmWHY1 interaction.
ZmWHY1 partitions between the chloroplast stroma and thylakoid membrane, to which it is bound in a DNA-dependent manner ZmWHY1 was initially recovered from chloroplast stroma and is predicted to localize to chloroplasts by both the TargetP [bib_ref] Prediction of organellar targeting signals, Emanuelsson [/bib_ref] and Predotar [bib_ref] Predotar: A tool for rapidly screening proteomes for N-terminal targeting sequences, Small [/bib_ref] algorithms. Immunoblot analysis of proteins from leaf, chloroplasts and mitochondria confirmed that ZmWHY1 is found in chloroplasts and that it is absent, or found at only very low levels, in mitochondria [fig_ref] Figure 2: Intracellular localization of ZmWHY1 [/fig_ref]. Analysis of chloroplast subfractions showed that ZmWHY1 is recovered in both the stromal and thylakoid membrane fractions [fig_ref] Figure 2: Intracellular localization of ZmWHY1 [/fig_ref] ; this behavior differs from that of other chloroplast gene expression factors in the same fractionated chloroplast preparation (PPR2, PPR4, RNC1, CAF1, CAF2, CFM2), all of which were found solely in the stromal fraction [bib_ref] Group II intron splicing factors derived by diversification of an ancient RNA..., Ostheimer [/bib_ref] [bib_ref] A ribonuclease III domain protein functions in group II intron splicing in..., Watkins [/bib_ref] [bib_ref] A pentatricopeptide repeat protein facilitates the trans-splicing of the maize chloroplast rps12..., Schmitz-Linneweber [/bib_ref] [bib_ref] A chloroplast-localized PPR protein required for plastid ribosome accumulation, Williams [/bib_ref].
It seemed possible that ZmWHY1 associated with the thylakoid membrane via a DNA tether because chloroplast nucleoids are membrane-associated [bib_ref] Organization, developmental dynamics, and evolution of plastid nucleoids, Sato [/bib_ref] and AtWHY1 copurified with a chloroplast chromosome preparation [bib_ref] and -12 are components of the transcriptionally active plastid chromosome that are..., Pfalz [/bib_ref]. In support of this possibility, treatment of the thylakoid membrane fraction with DNAse released a portion of the membrane-associated ZmWHY1 to the soluble fraction [fig_ref] Figure 2: Intracellular localization of ZmWHY1 [/fig_ref] , whereas RNAse treatment had no effect. These results indicate that ZmWHY1 is associated with the thylakoid membrane, at least in part, via an association with chloroplast DNA.
## Zmwhy1 is associated with large rna-and dna-containing particles
The observations that RNAse and DNAse disrupt ZmWHY1's association with CRS1 and the thylakoid The same blot stained with Ponceau S is shown below, with the large subunit of Rubisco (RbcL) marked. hcf7 and iojap are pale green and albino maize mutants with weak and severe plastid ribosome deficiencies, respectively [bib_ref] Nuclear gene iojap conditions a programmed change to ribosome-less plastids in Zea..., Walbot [/bib_ref] [bib_ref] Nuclear mutants of maize with defects in chloroplast polysome assembly have altered..., Barkan [/bib_ref]. The apparently higher levels of ZmWHY1 in Zmwhy1-1 mutants relative to Zmwhy1-2 mutants may be an artifact of the fact that samples were loaded on the basis of equal total protein: the abundant photosynthetic enzyme complexes make up the bulk of the protein in the Zmwhy1-2 extract but are missing in the Zmwhy1-1 extract, causing other proteins to appear over-represented. (D) RNAdependent coimmunoprecipitation of ZmWHY1 with CRS1. Prior to immunoprecipitation, stroma was treated with DNAse or RNAse, or incubated under similar conditions without added nuclease (Mock). The stroma was then subjected to immunoprecipitation with the antibody named at top. Presence of CRS1 in the immunoprecipitation pellets was tested by immunoblot analysis with CRS1 antibody. The samples in the chloroplast (Cp) and chloroplast subfraction lanes are derived from the same initial number of chloroplasts. The same blot was probed to detect a marker for thylakoid membranes (D1) and mitochondria (MDH). These subcellular fractions are the same as those shown previously for localization of RNC1, where a marker for the envelope membrane fraction was also presented (12). Env; envelope; Mito; mitochondria; Thy; thylakoid membranes. The blot stained with Ponceau S is shown below, with the band corresponding to RbcL marked. (B) DNAdependent association of ZmWHY1 with thylakoid membranes. The thylakoid membrane fraction was treated with DNAse, RNAse or incubated under similar conditions without added nuclease (Mock). Thylakoid membranes were then pelleted by centrifugation. Pellet (Pel) and supernatant (Sup) fractions were brought to equal volumes, and an equivalent proportion of each fraction was analyzed on an immunoblot probed with ZmWHY1 antibody. The same blot stained with Ponceau S is shown below. membrane, respectively, suggested that ZmWHY1 associates with both RNA and DNA. To further explore the nature of these interactions, the effects of RNAse or DNAse treatment on the sedimentation properties of ZmWHY1 were investigated [fig_ref] Figure 3: Sucrose-gradient sedimentation demonstrating that ZmWHY1 is associated with DNA-and RNA-containing particles in... [/fig_ref]. When untreated stroma was sedimented through a sucrose gradient, ZmWHY1 was detected in two peaks ($400-500 kDa and $600-700 kDa) and was also found in pelleted material at the bottom of the gradient. The 600-700 kDa peak coincides with the peak of CRS1 in the same gradient. Treatment of stroma with DNAse reduced the amount of ZmWHY1 in the pellet and in the $400-500 kDa peak, but did not reduce its recovery in the 600-700 kDa peak. Conversely, RNAse treatment specifically reduced the recovery of ZmWHY1 in the 600-700 kDa peak. These results together with those described above suggested that ZmWHY1 resides in two types of complexes: one that includes CRS1 and RNA, and the other that includes DNA.
Coimmunoprecipitation assays demonstrate that ZmWHY1 associates with a subset of plastid RNAs that includes the atpF intron
The RNA-dependent association between ZmWHY1 and CRS1 suggested that ZmWHY might associate with CRS1's RNA ligand, the atpF intron. However, the albino phenotype conditioned by the Zmwhy1-1 allele indicated that this could not be ZmWHY1's sole ligand, because mutations in crs1 that completely block atpF intron splicing result in a much less severe chlorophyll deficiency [bib_ref] CRS1 is a novel group II intron splicing factor that was derived..., Till [/bib_ref]. To identify RNAs that associate with ZmWHY in vivo we used a 'RIP-Chip' assay [bib_ref] RNA immunoprecipitation and microarray analysis show a chloroplast pentatricopeptide repeat protein to..., Schmitz-Linneweber [/bib_ref] as an initial screen: RNAs that coimmunoprecipitate with ZmWHY1 from stromal extract were identified by hybridization to a tiling microarray of the maize chloroplast genome. To ensure that DNA associated with ZmWHY did not contribute to the signal, the extract was treated with DNAse prior to immunoprecipitation, and the nucleic acids recovered from the immunoprecipitation pellet and supernatant were again treated with DNAse. RNAs recovered from the pellet and supernatant were then labeled with red-or green-fluorescing dye, respectively, combined, and hybridized to the microarray. Two replicate immunoprecipitations were analyzed in this manner. To highlight sequences that are enriched in the ZmWHY1 immunoprecipitations, the median enrichment ratio [red(F635)/green (F532)] was plotted according to chromosomal position, after subtracting the median enrichment ratios from control assays [fig_ref] Figure 4: Identification of chloroplast RNAs and DNAs that coimmunoprecipitate with ZmWHY1 [/fig_ref]. The results highlight the atpF intron as the major RNA ligand of ZmWHY. The results suggested, in addition, an association between ZmWHY1 and RNAs derived from several other loci (e.g. rps14, rpoC, ycf3, rps12, petD, rpl16, orf99). When the same data were analyzed by considering only the signal in the immunoprecipitation pellets, the results were similar (Supplementary [fig_ref] Figure 2: Intracellular localization of ZmWHY1 [/fig_ref].
To validate candidate RNA ligands to emerge from the RIP-chip experiment, RNAs that coimmunoprecipitate with ZmWHY1 were analyzed by slot-blot hybridization using probes corresponding to each RIP-chip peak [fig_ref] Figure 4: Identification of chloroplast RNAs and DNAs that coimmunoprecipitate with ZmWHY1 [/fig_ref]. RNAs purified from immunoprecipitations with antibodies to CRS1 and OE16 (a protein that does not bind RNA) were analyzed as controls. As for the RIPchip assays, the stromal extract was treated with DNAse prior to immunoprecipitation and the nucleic acids recovered from the immunoprecipitation were treated again with DNAse. The results largely recapitulated the RIPchip data (see lanes 'R' in [fig_ref] Figure 4: Identification of chloroplast RNAs and DNAs that coimmunoprecipitate with ZmWHY1 [/fig_ref] : atpF intron RNA was confirmed to be strongly enriched in ZmWHY1 immunoprecipitations, whereas RNAs from the psbA and petN loci, which did not appear as positives in RIPchip assays, likewise scored negative in the slot-blot hybridization assay. Coimmunoprecipitation with ZmWHY1 was also confirmed for RNAs from the rps12, ndhA, rpl16, ycf3 and rps14 loci; as predicted by the RIP-chip data, their degree of enrichment was less than that for the atpF intron. However, RNAs from the petD, orf99 and rrn5 loci, which appeared as minor peaks in the RIPchip data, did not appear to be enriched based on the slot-blot data; the orf99 transcript is of very low abundance, however, so it may be enriched in the pellet at levels that are too low to detect. These issues notwithstanding, the RIP-chip and slot-blot hybridization data together show that ZmWHY1 associates with a subset of RNAs in chloroplast extract, and that the atpF intron is its major RNA ligand.
DNA from throughout the plastid genome coimmunoprecipitates with ZmWHY1
The effects of DNAse-treatment on ZmWHY1's association with the thylakoid membrane [fig_ref] Figure 2: Intracellular localization of ZmWHY1 [/fig_ref] and on its sedimentation rate [fig_ref] Figure 3: Sucrose-gradient sedimentation demonstrating that ZmWHY1 is associated with DNA-and RNA-containing particles in... [/fig_ref] indicated that ZmWHY1 is The same data are plotted using an alternative analysis method in Supplementary [fig_ref] Figure 2: Intracellular localization of ZmWHY1 [/fig_ref] ; the atpF intron is the most prominent peak in both analyses, but the proportional sizes of other peaks vary depending on the comparison used. (B) Validation of RIP-chip and DIP-chip data by slotblot hybridization. Stroma was pretreated with DNAse or RNAse or left untreated and then subjected to immunoprecipitation with the antibodies indicated at the top. Nucleic acids purified from the pellets (Pel) and supernatants (Sup) were further treated with DNAse or alkali to remove residual DNA or RNA. The resulting total nucleic acids (T), RNA (R) or DNA (D), were applied to a nylon membrane with a slot blot manifold and hybridized with probes specific for the indicated sequences. Slots contained 1/9th or 1/27th of the nucleic acid recovered from each pellet or supernatant, respectively. (C) DIP-chip data showing genome-wide enrichment of chloroplast DNA in ZmWHY1 immunoprecipitations. Stroma was treated with RNAse prior to immunoprecipitation. Nucleic acids were extracted from the immunoprecipitation pellets and from total input stroma, and subjected to alkali hydrolysis to remove residual RNA prior to analysis by microarray hybridization. The median log2-transformed ratio of fluorescence in the pellet versus the input is plotted for replicate array fragments as a function of chromosomal position. associated with chloroplast DNA in vivo. To gain insight into which DNA sequences were involved in these interactions, we modified the RIP-chip protocol to detect coimmunoprecipitating DNA (DIP-chip): stromal extract was treated with ribonuclease prior to the immunoprecipitation, and alkali hydrolysis was used to remove residual RNA after the immunoprecipitation. A control immunoprecipitation used antibody to CAF1, a splicing factor that associates with specific chloroplast intron RNAs in vivo [bib_ref] Group II intron splicing factors derived by diversification of an ancient RNA..., Ostheimer [/bib_ref]. Both ZmWHY1 and CAF1 were efficiently immunoprecipitated (Supplementary [fig_ref] Figure 2: Intracellular localization of ZmWHY1 [/fig_ref] , but the DIP-chip data were strikingly different [fig_ref] Figure 4: Identification of chloroplast RNAs and DNAs that coimmunoprecipitate with ZmWHY1 [/fig_ref] : nearly all of the DNA in the input stromal sample coimmunoprecipitated with ZmWHY1, whereas very little DNA was recovered in CAF1 immunoprecipitations. These results confirm that ZmWHY1 is associated with chloroplast DNA and show further that ZmWHY1 either binds throughout the chloroplast genome, or binds to specific DNA regions and coimmunoprecipitates all other DNA sequences due to their linkage to ZmWHY1-binding sites. Incubation of the extract with various restriction enzymes prior to the immunoprecipitation did not reveal the specific enrichment of any DNA sequences (Supplementary [fig_ref] Figure 2: Intracellular localization of ZmWHY1 [/fig_ref] , leading us to favor the interpretation that ZmWHY1 is associated with many sites throughout the chloroplast genome. Nucleic acids recovered from the CAF1 and ZmWHY1 immunoprecipitations were also used as a direct template for PCR (Supplementary [fig_ref] Figure 2: Intracellular localization of ZmWHY1 [/fig_ref]. The results support the DIP-Chip data: PCR product was obtained using a variety of chloroplast genome primers from the ZmWHY1 coimmunoprecipitation and not from the CAF1 coimmunoprecipitation.
The enrichment of DNA sequences in ZmWHY1 immunoprecipitations was further confirmed by slot-blot hybridization [fig_ref] Figure 4: Identification of chloroplast RNAs and DNAs that coimmunoprecipitate with ZmWHY1 [/fig_ref]. As for the DIP-chip assays, stroma was treated with RNAse prior to the immunoprecipitation, and residual RNA was removed by alkali hydrolysis after the immunoprecipitation [fig_ref] Figure 4: Identification of chloroplast RNAs and DNAs that coimmunoprecipitate with ZmWHY1 [/fig_ref] , lanes 'D'). Antibody to ZmWHY1 coimmunoprecipitated DNA from all sequences tested, whereas DNA was not detected in either the CRS1 or OE16 immunoprecipitations. The DIP-chip, PCR and slot-blot hybridization data provide strong evidence that ZmWHY1 is associated with chloroplast DNA in vivo and that it has many binding sites throughout the genome.
Zm Why1 mutants are deficient for plastid ribosomes A role for WHY1 in chloroplast gene expression was suggested by the coimmunoprecipitation of ZmWHY1 with CRS1, RNA and DNA, and by the copurification of AtWHY1 with the plastid transcriptionally active chromosome [bib_ref] and -12 are components of the transcriptionally active plastid chromosome that are..., Pfalz [/bib_ref]. In support of this possibility, core subunits of the chloroplast ATP synthase, photosystem II, photosystem I, the cytochrome b 6 f complex and Rubisco accumulate to reduced levels in ZmWhy1 mutants . The protein deficiencies conditioned by the weak allele combinations (Zmwhy1-2/-2 and Zmwhy1-1/-2) resemble those in hcf7mutants, which have a reduced content of chloroplast ribosomes [bib_ref] Nuclear mutants of maize with defects in chloroplast polysome assembly have altered..., Barkan [/bib_ref]. These proteins were not detectable in Zmwhy1-1 homozygotes, as in albino iojap mutants which lack plastid ribosomes .
The global loss of photosynthetic enzyme complexes in ZmWhy1 mutants suggested an underlying loss of plastid ribosomes. This possibility was confirmed by RNA gel blot hybridizations, which showed a loss of mature 23S, 4.5S and 16S rRNAs in hypomorphic ZmWhy1 mutants, and an increased accumulation of rRNA precursors . Chloroplast rRNAs were not detectable in plants homozygous for the null Zmwhy1-1 allele, as in albino iojap leaves. Whereas hcf7 mutants show a more severe loss of 16S rRNA than 23S and 4.5S rRNAs, the reverse is true for hypomorphic ZmWhy1 mutants. A dramatic increase in the ratio of 23S rRNA precursors to mature 23S rRNA in these mutants was confirmed with a poisoned-primer extension assay (Supplementary [fig_ref] Figure 3: Sucrose-gradient sedimentation demonstrating that ZmWHY1 is associated with DNA-and RNA-containing particles in... [/fig_ref].
Some steps in rRNA processing are dependent upon ribosome assembly in chloroplasts, as in bacteria [bib_ref] A pentatricopeptide repeat protein facilitates the trans-splicing of the maize chloroplast rps12..., Schmitz-Linneweber [/bib_ref] [bib_ref] Nuclear mutants of maize with defects in chloroplast polysome assembly have altered..., Barkan [/bib_ref] [bib_ref] The pentatricopeptide repeat protein PPR5 stabilizes a specific tRNA precursor in maize..., Beick [/bib_ref]. The aberrant 23S and 4.5S rRNA processing in ZmWhy1 mutants suggested therefore that ZmWHY1 might promote the expression of a gene needed for the assembly of the large ribosomal subunit (an rRNA or ribosomal protein), with loss of the small ribosomal subunit being a secondary effect. It seemed plausible, for example, that ZmWHY1 might promote processive transcription through the chloroplast rrn operon; this would differentially affect the large ribosomal subunit due to the distal position of the genes encoding its rRNA components (23S, 4.5S and 5S rRNA) in the operon (see map in . However, the results of chloroplast run-on transcriptions assays argue against this possibility : the ratio of polymerase transit through the 23S gene in comparison to the 16S rRNA gene, and the ratio of rrn operon transcription in comparison to transcription from a different chloroplast locus (trnG-UCC) were similar in wild-type and Zmwhy1-1/-2 mutant chloroplasts. Furthermore, the rRNA components of the large ribosomal subunit were not reproducibly enriched in ZmWHY coimmunoprecipitates [fig_ref] Figure 4: Identification of chloroplast RNAs and DNAs that coimmunoprecipitate with ZmWHY1 [/fig_ref] , Supplementary [fig_ref] Figure 2: Intracellular localization of ZmWHY1 [/fig_ref] ; this suggests that ZmWHY1 does not interact directly with rRNAs or 50S ribosomal subunits, although such interactions cannot be eliminated based on these negative results. Taken together, these results argue that ZmWHY1 directly impacts the expression of a gene encoding a component of the large ribosomal subunit and/or promotes ribosome assembly.
Elucidation of its precise role in this process will require further study.
## Zmwhy1 promotes atpf intron splicing
The coimmunoprecipitation of ZmWHY1 with the atpF splicing factor CRS1 and with RNA from the atpF locus suggested that ZmWHY1 might be involved in the splicing of atpF pre-mRNA. To test this possibility, atpF RNA from Zmwhy1 mutants was analyzed by RNA gel blot hybridization . To control for pleiotropic effects of weak and severe plastid ribosome deficiencies, RNAs in pale green (hypomorphic) Zmwhy1-2 and Zmwhy1-2/-1 mutants were compared to those in hcf7 mutants, and RNAs in albino (null) Zmwhy1-1 mutants were compared to those in iojap mutants. These comparisons were important because the complete absence of plastid ribosomes results in the failure to splice all chloroplast subgroup IIA introns, including the atpF intron [bib_ref] Inefficient rpl2 splicing in barley mutants with ribosome-deficient plastids, Hess [/bib_ref] [bib_ref] Comparative analysis of splicing of the complete set of chloroplast group II..., Vogel [/bib_ref].
The results in show that the ratio of spliced (S) to unspliced (U) atpF transcripts is reduced in hypomorphic ZmWhy1 mutants in comparison to wild-type and hcf7 plants, albeit not as severely as in crs1 mutants. The ratio of excised intron (asterisks) to unspliced RNA is also reduced, supporting the interpretation that ZmWHY1 promotes atpF splicing rather than stabilizing the spliced product. The normal splicing of the atpF intron in hcf7 mutants argues that the partial plastid ribosome deficiency in hypomorphic ZmWhy1 mutants cannot account for their reduced atpF splicing. Furthermore, a different subgroup IIA intron, the rpl2 intron, is spliced normally in the same plants (Supplementary [fig_ref] Figure 3: Sucrose-gradient sedimentation demonstrating that ZmWHY1 is associated with DNA-and RNA-containing particles in... [/fig_ref] , showing that not all subgroup IIA introns are affected in the hypomorphic ZmWhy1 mutants. These results provide strong evidence that ZmWHY1's association with atpF RNA enhances the splicing of the atpF intron.
The coimmunoprecipitation data demonstrated an association between ZmWHY1 and RNAs from several loci other than atpF. However, RNA gel blot hybridizations showed that the transcripts from all such genes were . Plastid ribosome deficiency in ZmWhy1 mutants. (A) Total seedling leaf RNA (0.5 mg) was analyzed by RNA gel blot hybridization using probes for the RNAs indicated at the bottom. A map of the plastid rRNA operon is shown below. A cDNA probe was used to detect mature trnA; this lacks intron sequences and therefore hybridizes poorly to unspliced precursor. The probe for 23S rRNA is derived from the 5 0 portion of the rrn23 gene and detects just one of the two 23S rRNA fragments found in ribosomes in vivo. The leaf pigmentation conditioned by each mutant allele is indicated: iv: ivory leaves; pg: pale green leaves. The blot used to detect 16S rRNA is shown after staining with methylene blue to illustrate equal loading of cytosolic rRNAs (18S, 28S). Mature RNA forms are indicated with asterisks. (B) Reduced accumulation of photosynthetic enzyme complexes in ZmWhy1 mutants. Immunoblots of leaf extract (5 mg protein or the indicated dilutions) were probed with antibodies to core subunits of photosynthetic enzyme complexes: AtpA (ATP synthase), D1 (photosystem II), PsaD (photosystem I) and PetD (cytochrome b 6 f complex). The same blot stained with Ponceau S is shown below to illustrate sample loading and the abundance of RbcL. (C) Plastid runon transcription. Chloroplasts prepared from Zmwhy1-1/-2 heteroallelic mutants or their normal siblings (wt) were used for runon transcription assays. RNAs purified from the reactions were hybridized to slot blots harboring oligonucleotides corresponding to the genes indicated at the top. Each probe was present in duplicate. cfm3, a nuclear gene, served as a negative control. The results were quantified with a phosphorimager and plotted on the bar graph below.
qualitatively similar in ZmWhy1 mutants and in the relevant control mutant . The coimmunoprecipitation of ZmWHY1 with RNAs from both loci encoding the trans-spliced group II intron in rps12 was intriguing [fig_ref] Figure 4: Identification of chloroplast RNAs and DNAs that coimmunoprecipitate with ZmWHY1 [/fig_ref] , but splicing of this RNA is not disrupted in ZmWhy1 mutants (Supplementary [fig_ref] Figure 3: Sucrose-gradient sedimentation demonstrating that ZmWHY1 is associated with DNA-and RNA-containing particles in... [/fig_ref]. These results show that ZmWHY1 is not necessary for the normal processing of most chloroplast transcripts.
A structural homolog of ZmWHY1 in Trypanosoma brucei is required for mitochondrial RNA editing [bib_ref] Crystal structures of T. brucei MRP1/MRP2 guide-RNA binding complex reveal RNA matchmaking..., Schumacher [/bib_ref]. Several plastid RNAs that are known to be substrates for RNA editing were represented among the RNAs that coimmunoprecipitate with ZmWHY1. Direct sequencing of RT-PCR products demonstrated, however, that the editing of the known edited sites in the petB, rpl20, ycf3 and rps14 transcripts is not disrupted in Zmwhy1-1 and Zmwhy1-2/-1 mutants (data not shown), suggesting that ZmWHY1 is not required for RNA editing.
## Zmwhy1 is required neither for chloroplast dna replication nor for global plastid transcription
The association of ZmWHY1 with plastid DNA suggested that it might be involved in chloroplast transcription or DNA replication. However, Southern blot analysis of total leaf DNA showed that plastid DNA levels in ZmWhy1 mutants, although somewhat variable from sample to sample, were generally similar to those in normal and control mutant plants [fig_ref] Figure 8: Chloroplast DNA levels in ZmWhy1 mutants [/fig_ref]. In addition to the plastid transcripts shown in , a variety of other transcripts were examined by RNA gel blot hybridization (Supplementary [fig_ref] Figure 3: Sucrose-gradient sedimentation demonstrating that ZmWHY1 is associated with DNA-and RNA-containing particles in... [/fig_ref]. In no case was a significant reduction in transcript level detected, indicating that ZmWHY1 is not necessary for global plastid transcription. In fact, a trend is apparent toward increased transcript abundance in ZmWhy1 mutants, but these changes are rather subtle and indirect effects on RNA abundance cannot be excluded.
## Recombinant zmwhy1 binds single-stranded rna and dna in vitro
To determine whether ZmWHY1 can directly bind both RNA and DNA, recombinant ZmWHY1 (rWHY1) was generated by expression as a maltose-binding protein (MBP) fusion. rWHY1 was released from the MBP moiety by protease cleavage and further purified on a gel filtration column [fig_ref] Figure 9: Recombinant ZmWHY1 binds ssRNA and DNA [/fig_ref]. rWHY1 eluted from the sizing column at a position corresponding to a globular protein of $100 kDa, consistent with the report that StWHY1 forms a homotetramer [bib_ref] A new family of plant transcription factors displays a novel ssDNA-binding surface, Desveaux [/bib_ref]. Filter-binding assays showed that rWHY1 binds to unspliced atpF RNA in vitro [fig_ref] Figure 9: Recombinant ZmWHY1 binds ssRNA and DNA [/fig_ref] , but it did not show specificity for this RNA relative to other RNAs of similar size under the conditions tested (data not shown).
To compare the affinity of ZmWHY1 for singlestranded and double-stranded RNA and DNA, gel mobility shift assays were used to detect binding to a synthetic 31-mer oligonucleotide in the context of ssDNA, ssRNA, ds DNA or dsRNA [fig_ref] Figure 9: Recombinant ZmWHY1 binds ssRNA and DNA [/fig_ref]. ZmWHY1 bound rather weakly to these short oligonucleotides but the results showed, nonetheless, that rWHY1 binds both ssDNA and ssRNA, and binds poorly to dsRNA and dsDNA.
# Discussion
Previous reports have attributed diverse functions and intracellular locations to WHY1. WHY1 in dicots has been reported to be a ssDNA-binding protein that functions in the nucleus as both a transcription factor [bib_ref] A ''Whirly'' transcription factor is required for salicylic aciddependent disease resistance in..., Desveaux [/bib_ref] [bib_ref] A new family of plant transcription factors displays a novel ssDNA-binding surface, Desveaux [/bib_ref] and as a negative regulator of telomere length [bib_ref] Singlestranded DNA binding factor AtWHY1 modulates telomere length homeostasis in Arabidopsis, Yoo [/bib_ref]. Arabidopsis WHY1 copurified with the 'transcriptionally active chromosome' from chloroplasts [bib_ref] and -12 are components of the transcriptionally active plastid chromosome that are..., Pfalz [/bib_ref]. Our results add another layer to this complex picture. We demonstrate that ZmWHY1 is essential for chloroplast biogenesis, and that it localizes to the chloroplast where it plays multiple roles in gene expression. We also add RNA binding to WHY1's repertoire of biochemical activities and demonstrate that ZmWHY1 is bound to a subset of chloroplast RNAs in chloroplast extract.
## Multiple roles of zmwhy1 in chloroplast biogenesis
ZmWHY was identified among proteins that coimmunoprecipitate with CRS1, which is required for the splicing of the group II intron in the chloroplast atpF pre-mRNA. . Reduced atpF intron splicing in ZmWhy1 mutants. (A) RNA gel blot analysis of atpF splicing. Total seedling leaf RNA (5 mg) was analyzed by RNA gel blot analysis using a probe including atpF exon 2 and a portion of the atpF intron (atpF int/ex2), or with an intronspecific probe (atpFint). The atpF gene is part of a polycistronic transcription unit that gives rise to a previously characterized population of RNAs [bib_ref] Tissue-dependent plastid RNA splicing in maize: Transcripts from four plastid genes are..., Barkan [/bib_ref] [bib_ref] Co-transcription pattern of an introgressed operon in the maize chloroplast genome comprising..., Stahl [/bib_ref]. Spliced (S) and unspliced (U) transcripts are indicated. Asterisks mark bands that we believe correspond to the excised intron and its degradation products. The ratio of spliced to unspliced transcripts was quantified with a phosphorimager, normalized to the wildtype ratio and plotted below using arbitrary units.
We showed that ZmWHY1 is associated with atpF intron RNA in vivo and that the coimmunoprecipitation of ZmWHY1 and CRS1 is disrupted by RNAse, indicating that they coimmunoprecipitate due to their association with the same RNA molecule. ZmWHY1's association with atpF RNA is functionally significant, as atpF intron splicing is disrupted in ZmWhy1 mutants. However, the splicing of this intron is more sensitive to a partial loss of CRS1 than to a partial loss of ZmWHY1, suggesting that ZmWHY1 plays an accessory function in atpF splicing but may not be absolutely required.
The atpF splicing defect in ZmWhy1 mutants cannot account for their loss of plastid ribosomes, as the more severe atpF splicing defect in crs1-1 mutants is not accompanied by a substantial plastid ribosome deficiency [bib_ref] CRS1 is a novel group II intron splicing factor that was derived..., Till [/bib_ref]. The specific role of ZmWHY1 in promoting the biogenesis of the plastid translation machinery remains unclear.
Although several RNAs with translation-related functions are among the RNAs that coimmunoprecipitate with ZmWHY1, the abundance and processing of these RNAs are similar in ZmWhy1 mutants and in control mutants that exhibit a ribosome-deficiency of similar severity. The specific rRNA deficiencies in ZmWhy1 mutants do suggest, however, that ZmWHY1 is most directly involved in the biogenesis of the large ribosomal subunit: the accumulation and processing of the 23S and 4.5S rRNAs are more sensitive to the partial loss of ZmWhy1 function than are those of 16S rRNA, whereas the reverse is true for hcf7 mutants. Furthermore, in ppr5 mutants, whose primary defect is in the maturation of a specific plastid tRNA, the rRNAs from the two ribosomal subunits are impacted to a similar extent [bib_ref] The pentatricopeptide repeat protein PPR5 stabilizes a specific tRNA precursor in maize..., Beick [/bib_ref]. Thus, our results point to the biogenesis of the plastid large ribosomal subunit as one function of ZmWHY1 but definition of . Accumulation of plastid RNAs in ZmWhy1 mutants. Total seedling leaf RNA (5 mg) was analyzed by RNA gel blot hybridization using probes specific for the RNAs indicated at bottom. The rps12 probe was a cDNA probe containing exons 1 and 2. The leaf pigmentation conditioned by each mutant allele is indicated: iv: ivory; pg: pale green. The methylene blue-stained blots are shown below, with rRNAs marked. Additional RNAs that were analyzed analogously are shown in Supplementary [fig_ref] Figure 3: Sucrose-gradient sedimentation demonstrating that ZmWHY1 is associated with DNA-and RNA-containing particles in... [/fig_ref] its precise role in this process will require additional study. The strong defect in the processing step that separates 23S rRNA from 4.5S rRNA in hypomorphic ZmWhy1 mutants is reminiscent of defects reported for mutations in the DCL, DAL and RNR1 genes in dicots [bib_ref] RNR1, a 3 0 -5 0 exoribonuclease belonging to the RNR superfamily,..., Bollenbach [/bib_ref] [bib_ref] DCL is a plantspecific protein required for plastid ribosomal RNA processing and..., Bellaoui [/bib_ref] [bib_ref] Altered expression of the Arabidopsis ortholog of DCL affects normal plant development, Bellaoui [/bib_ref] [bib_ref] The Arabidopsis nuclear DAL gene encodes a chloroplast protein which is required..., Bisanz [/bib_ref]. Although it is unclear whether any of these genes function directly in 23S/4.5S rRNA processing, it is possible that WHY1 acts in concert with one or more of these proteins.
## Zmwhy1 binds both rna and dna in vitro and in vivo
We show here that chloroplast DNA coimmunoprecipitates with ZmWHY1 from plastid extract, that a fraction of ZmWHY1 is tethered to the thylakoid membrane in a DNA-dependent fashion, that a fraction of stromal ZmWHY1 is found in DNA-containing particles of $400 kDa, and that ZmWHY1 binds ssDNA in vitro. These results are consistent with previous reports that dicot WHY1 binds ssDNA [bib_ref] A new family of plant transcription factors displays a novel ssDNA-binding surface, Desveaux [/bib_ref] [bib_ref] Singlestranded DNA binding factor AtWHY1 modulates telomere length homeostasis in Arabidopsis, Yoo [/bib_ref] and that it copurifies with a chloroplast 'transcriptionally active chromosome' [bib_ref] and -12 are components of the transcriptionally active plastid chromosome that are..., Pfalz [/bib_ref]. Our findings suggest that ZmWHY1 either binds DNA in a sequence non-specific fashion or that it has many binding sites distributed throughout the plastid genome, because DNA sequences from throughout the plastid genome coimmunoprecipitated to a similar extent with ZmWHY1. It remains possible, however, that ZmWHY1 associates with specific DNA regions in vivo, but that these associations were disrupted during lysate preparation. A DNA immunoprecipitation experiment was recently reported for AtWHY2, a mitochondriallocalized Whirly protein (23), with analogous results: DNA sequences from a variety of regions throughout the mitochondrial genome coimmunoprecipitated with AtWHY2, when assayed by PCR.
We demonstrate here that ZmWHY1 interacts not only with DNA, as anticipated by previous reports, but that it also binds RNA in vivo and in vitro. That ZmWHY1 interacts with RNA is, perhaps, not surprising given that a structural homolog of ZmWHY1 has been shown to bind RNAs involved in kinetoplastid RNA editing [bib_ref] Crystal structures of T. brucei MRP1/MRP2 guide-RNA binding complex reveal RNA matchmaking..., Schumacher [/bib_ref] , and that many proteins that bind ssDNA also bind RNA. The atpF intron RNA was the major RNA ligand of ZmWHY1 detected in the RNA coimmunoprecipitation assays. This RNA is not particularly abundant in vivo so its enrichment in ZmWHY1 immunoprecipitations likely reflects a specific interaction in vivo. Although intrinsic specificity for this RNA did not emerge from in vitro binding assays using the entire intron, a high-affinity site within a large RNA such as the atpF intron ($800 nt) can be masked in vitro due to the overwhelming number of nonspecific sites available for protein binding. Therefore, more detailed studies involving smaller RNA ligands will be required to determine whether ZmWHY1 binds RNA with sequence-specificity or whether it is Protein-RNA complexes were captured on the nitrocellulose (bound); unbound RNA was captured on the nylon membrane below. (C) Gel mobility shift assay showing rWHY1's relative affinity for double-and single-stranded RNA and DNA. A 31-mer oligonucleotide in RNA or DNA form was radiolabeled, heated and, either snap cooled (ssRNA, ssDNA) or cooled slowly in the presence of monovalent salts and a two-fold excess of its complement (dsRNA, dsDNA). The substrate (40 pM) was mixed with increasing concentrations of ZmWHY1 (17, 50 and 150 nM). Protein binding is illustrated by the appearance of an upper band and retention at the top of the gel, and by the disappearance of unbound substrate. recruited to the atpF intron via protein-protein interactions.
What is WHY1's DNA-related function in the chloroplast?
The association of ZmyWHY1 with DNA sequences from throughout the chloroplast genome suggests that it participates in transcription and/or DNA metabolism. However, our results argue against a general role in transcription, as all plastid mRNAs examined accumulate in hypomorphic Zmwhy1 mutants to levels that are comparable to those in the relevant control mutants. The results of chloroplast transcription runon experiments argue that the preferential loss of 23S rRNA in these mutants is due to aberrant ribosome assembly rather than to reduced rRNA transcription rates. It remains possible, however, that ZmWHY1 does play a role in chloroplast transcription but that another gene with a partially redundant function serves this purpose in ZmWhy1 mutants.
It is intriguing that ZmWHY1 binds preferentially to DNA in single stranded form because opportunities to interact with ssDNA in vivo are expected to be limited. DNA replication, recombination and repair involve the transient occurrence of ssDNA, and torsional stress can induce DNA unwinding. The Southern blot data showing that plastid DNA levels are no more than minimally decreased in ZmWhy1 null mutants argue against a central role for ZmWHY1 in DNA replication; however participation of ZmWHY1 in DNA recombination or repair remains possible. In fact, the participation of an unrelated ssDNA-binding protein, OSB1, in plant mitochondrial DNA recombination was reported recently [bib_ref] The plant-specific ssDNA binding protein OSB1 is involved in the stoichiometric transmission..., Zaegel [/bib_ref].
There are several parallels between our findings with ZmWHY1 and the activities reported for the bacterial protein HU. HU is associated with the bacterial nucleoid, binds preferentially to DNA with irregular structural features (e.g. single stranded gaps and bulges), and is involved in DNA recombination and repair [bib_ref] Regulation of gene expression by histone-like proteins in bacteria, Dorman [/bib_ref] [bib_ref] HU binds and folds single-stranded DNA, Kamashev [/bib_ref]. Despite its high conservation in bacteria and the presence of an HU homolog in a plastid genome in red algae [bib_ref] Detection and localization of a chloroplast-encoded HU-like protein that organizes chloroplast nucleoids, Kobayashi [/bib_ref] , HU homologs are not encoded in the nuclear or plastid genomes of vascular plants [bib_ref] Detection and localization of a chloroplast-encoded HU-like protein that organizes chloroplast nucleoids, Kobayashi [/bib_ref] [bib_ref] Was the evolution of plastid genetic machinery discontinuous?, Sato [/bib_ref]. Thus, alternative proteins have presumably been recruited in vascular plants to fulfill the functions performed by HU in the chloroplast's cyanobacterial ancestor. The nucleoid-associated protein sulfite reductase has been suggested to be one such protein [bib_ref] Was the evolution of plastid genetic machinery discontinuous?, Sato [/bib_ref] [bib_ref] The 70-kDa major DNA-compacting protein of the chloroplast nucleoid is sulfite reductase, Sato [/bib_ref] [bib_ref] DNA binding and partial nucleoid localization of the chloroplast stromal enzyme ferredoxin:sulfite..., Sekine [/bib_ref] , and perhaps WHY1 is another. HU influences global transcription patterns through its effect on nucleoid architecture, and mediates the formation of DNA loops that repress transcription from specific genes [bib_ref] Supercoiling and denaturation in Gal repressor/heat unstable nucleoid protein (HU)-mediated DNA looping, Lia [/bib_ref] [bib_ref] Nucleoid remodeling by an altered HU protein: reorganization of the transcription program, Kar [/bib_ref] [bib_ref] Role of HU and DNA supercoiling in transcription repression: specialized nucleoprotein repression..., Lewis [/bib_ref]. HU is also an RNA-binding protein, and functions in vivo to repress the translation of the E. coli rpoS mRNA [bib_ref] The bacterial histone-like protein HU specifically recognizes similar structures in all nucleic..., Balandina [/bib_ref] [bib_ref] The Escherichia coli histone-like protein HU regulates rpoS translation, Balandina [/bib_ref]. Like HU, ZmWHY1 interacts globally with plastid DNA, but specifically with certain plastid RNAs, and binds preferentially to nucleic acids with single-stranded character. The abundance of several chloroplast mRNAs is increased in ZmWhy1 mutants, consistent with a global repressive role for ZmWHY1 in transcription. This possibility is in accord with the recent report that over-expression of AtWHY2 in Arabidopsis causes a reduction in the levels of several mitochondrial RNAs [bib_ref] Overexpression of mtDNAassociated AtWhy2 compromises mitochondrial function, Marechal [/bib_ref]. Although its role in DNA metabolism remains uncertain, our results demonstrate that description of WHY1 as a chloroplast transcription factor is, at best, an over-simplification of the complex roles played by this interesting protein.
[fig] Figure 1: Mutant alleles of ZmWhy1. (A) Positions of Mu transposon insertions in the ZmWhy1 gene. Protein coding regions are indicated by rectangles, untranslated regions and introns by lines and Mu transposon insertions by triangles. The sequence of each insertion site is shown below, with the nine nucleotides that were duplicated during insertion underlined. The identity of the member of the Mu family is shown for each insertion (why1-2: Mu1/1.7; why1-1: MuDR), and was inferred from polymorphisms in the terminal inverted repeats. (B) Phenotypes of ZmWhy1 mutant seedlings grown for nine days in soil. Seedlings shown are homozygous for either the Zmwhy1-1 or Zmwhy1-2 allele, or are the heteroallelic progeny of a complementation cross. (C) Immunoblot showing loss of ZmWHY1 in mutant leaf tissue. Total leaf extract (10 mg protein, or dilutions as indicated) were analyzed. [/fig]
[fig] Figure 2: Intracellular localization of ZmWHY1. (A) Immunoblots of extracts from leaf and subcellular fractions. [/fig]
[fig] Figure 3: Sucrose-gradient sedimentation demonstrating that ZmWHY1 is associated with DNA-and RNA-containing particles in chloroplast stroma. Stromal extract was treated with DNAse or RNAse, or incubated under similar conditions without nuclease (Mock), and then sedimented through a sucrose gradient. An equal volume of each gradient fraction was analyzed by probing immunoblots with the antibodies indicated to the left. RPL2, a protein in the large ribosomal subunit, marks the position of ribosomes. Shown below is the blot of the mock-treated fractions stained with Ponceau S, with the RbcL band marked to illustrate the position of Rubisco. The Ponceau S stained blots of experiments involving the DNAseand RNAse-treated extracts looked similar (data not shown). [/fig]
[fig] Figure 4: Identification of chloroplast RNAs and DNAs that coimmunoprecipitate with ZmWHY1. (A) RIP-chip data showing coimmunoprecipitation of specific chloroplast RNAs with ZmWHY1. The ratio of signal in the pellet versus the supernatant (F635/F532) for each array fragment is plotted according to chromosomal position. The plot shows the median values for replicate spots across two replicate ZmWHY1 immunoprecipitations after subtracting the corresponding values for two negative control immunoprecipitations (one with OE16 antibody and one without antibody). [/fig]
[fig] Figure 9: Recombinant ZmWHY1 binds ssRNA and DNA. (A) Elution of recombinant ZmWHY1 from a gel filtration column. MBP-WHY1 was purified by amylose affinity chromatography, cleaved with TEV protease to separate the WHY1 and MBP moieties, and applied to a Superdex 200 column. Column fractions were analyzed by SDS-PAGE and staining with Coomassie blue. The elution position of size markers (alcohol dehydrogenase, 150 kDa; BSA, 67 kDa, MBP, 42 kDa) is shown. The peak WHY1 fractions were pooled and used for in vitro assays. (B) Filter binding assay showing RNA binding activity of ZmWHY1. Assays containing 10 pM radiolabeled atpF intron RNA and increasing ZmWHY1 concentrations (50 nM maximum) were filtered through sandwiched nitrocellulose and nylon membranes. [/fig]
[fig] Figure 8: Chloroplast DNA levels in ZmWhy1 mutants. Seedling leaf DNA (5 mg) was digested with EcoRI (left), or PvuII (right) and analyzed by DNA gel blot hybridization using a probe from the chloroplast rrn23 gene (top left), or orf99 (top right). The same gels stained with ethidium bromide are shown below. The small fluctuations in relative band intensity may result from small differences in sample loading. [/fig]
|
A case of recalcitrant silicone granuloma treated with adalimumab: A case report
Liquid silicone is a relatively inexpensive injectable used for soft tissue augmentation. Injection of silicone is associated with a risk of delayed granuloma formation associated with elevated levels of tumour necrosis factor-α. We report a case of recalcitrant silicone granulomas following facial injections of silicone successfully treated with tumour necrosis factor-α blockade. Our case, as well as previous reports, demonstrates the effectiveness of this therapy for the treatment of foreign body granulomas from due to silicone.
# Introduction
Subcutaneous liquid silicone injections are a relatively inexpensive, minimally invasive approach to cosmetic alteration of facial appearance used by licenced and non-licenced practitioners. [bib_ref] Adverse granulomatous reaction after cosmetic dermal silicone injection, Bigatà [/bib_ref] Complications include localized bruising, erythema, edema, allergic response, cellulitis, abscesses, necrosis, scarring, contractures, and granulomas. [bib_ref] Etanercept therapy for silicone granuloma, Desai [/bib_ref] Systemic adverse events range from acute pneumonitis and hepatic granulomas to intravascular embolization leading to sudden death. [bib_ref] Injectable fluid silicone therapy: human morbidity and mortality, Ellenbogen [/bib_ref] [bib_ref] Silicone pneumonitis after a cosmetic augmentation procedure, Gurvits [/bib_ref] Formation of silicone granulomas was first described in 1964 by Winer et al. [bib_ref] Tissue reactions to injected silicone liquids: a report of three cases, Winer [/bib_ref] as an immunologic response. They may present locally at the site of injection, but can migrate elsewhere, and usually form within 12 months.Clinically, they appear as tender nodules with induration, ulceration, and surrounding lymphadenopathy. 7 Treatment can be challenging and without a clear consensus. Options include local and systemic steroids, tetracycline antibiotics, hydroxychloroquine, 5-fluorouracil, and isotretinoin. [bib_ref] Characteristics and treatment of silicone granulomas: a retrospective multicenter cohort of 21..., Kassamali [/bib_ref] Surgical management is difficult and not recommended due to the tendency of silicone to migrate and cause scarring which may be more distressing than original presenting concern of granulomatous inflammation. [bib_ref] Silicone granulomas treated with etanercept, Pasternack [/bib_ref] The interferon-γ modulator, imiquimod, has been utilized, as well as tumour necrosis factor-α (TNFα) inhibitors etanercept and adalimumab. [bib_ref] Etanercept therapy for silicone granuloma, Desai [/bib_ref] [bib_ref] Silicone granulomas treated with etanercept, Pasternack [/bib_ref] [bib_ref] Granuloma formation secondary to surreptitiously placed silicone, Guhan [/bib_ref] Case report A 78-year-old woman with a past medical history of depression and breast cancer presented with erythematous, tender plaques on her cheeks, and forehead several weeks after a dental crown repair [fig_ref] Figure 1: Before and after treatment of recalcitrant silicone granuloma with adalimumab [/fig_ref] and (b)). Eighteen years prior, she had received silicone injections in her cheeks and glabella. Magnetic resonance imaging demonstrated findings consistent with plastic and inflammation in the areas which had been injected. She was successfully treated with a 10-day course of oral prednisone; however, she developed a flare during the taper. A skin biopsy demonstrated a granulomatous inflammation consistent with a foreign body reaction, and chest X-ray was non-contributory. She was treated with intralesional and intramuscular triamcinolone acetonide and transitioned to oral doxycycline 100 mg twice daily for 3 months, with minimal response.
She was subsequently started on oral cyclosporine, and then mycophenolate mofetil (MMF), but was unable to tolerate either medication. Repeated attempts to wean her from cyclosporine and MMF resulted in four flares requiring extended use of prednisone, cyclosporine, and MMF.
Four years after her initial presentation, the patient was started on adalimumab 40 mg subcutaneously every 2 weeks. She was weaned from prednisone, cyclosporine, and MMF and achieved a complete response. After 8 months of biologics, she developed a new exacerbation with increased facial swelling and was treated with a 2-week course of 15 mg PO daily prednisone followed by a 6-week taper. At her latest follow-up 14 months after starting adalimumab, she has not experienced any further recurrences of her symptoms [fig_ref] Figure 1: Before and after treatment of recalcitrant silicone granuloma with adalimumab [/fig_ref]. The patient will be continued on adalimumab indefinitely.
# Discussion
The etiology of silicone granulomas is not well-understood and the variation in silicone formulations and methods of injection lead to highly heterogeneous reactions. However, it is hypothesized that T-cell activation may be initiated by the silicone itself or adulterants in the injection diluent. These granulomatous reactions are associated with elevations in the pro-inflammatory cytokine, TNF-α. [bib_ref] Silicone granulomas treated with etanercept, Pasternack [/bib_ref] Pasternack et al. [bib_ref] Silicone granulomas treated with etanercept, Pasternack [/bib_ref] described treatment of two patients with etanercept twice weekly. Similarly, Desai et al. [bib_ref] Etanercept therapy for silicone granuloma, Desai [/bib_ref] reported a patient treated with etanercept twice weekly resulting in significant improvement in symptoms within 1 month. Our case adds to the growing evidence that TNF-α inhibition may be an important and effective modality of treatment for the otherwise recalcitrant silicone granulomas. [bib_ref] Etanercept therapy for silicone granuloma, Desai [/bib_ref] [bib_ref] Characteristics and treatment of silicone granulomas: a retrospective multicenter cohort of 21..., Kassamali [/bib_ref] [bib_ref] Silicone granulomas treated with etanercept, Pasternack [/bib_ref] [bib_ref] Granuloma formation secondary to surreptitiously placed silicone, Guhan [/bib_ref] As in the case of our patient, TNF-α inhibitors may be safer and more effective therapies than either cyclosporine or MMF. Thus, consideration should be given to initiating a TNF-α inhibitor for recalcitrant silicone granulomas before other immunomodulatory therapies.
## 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.
## Informed consent
The patients in this article have given written informed consent to publication of their case details.
## Orcid id
Benoit M Cyrenne https://orcid.org/0000-0001-7220-4983
[fig] Figure 1: Before and after treatment of recalcitrant silicone granuloma with adalimumab. Patient on initial presentation of erythematous, tender plaques to cheeks and forehead (a), (b). Subsequent images of patient following treatment with adalimumab (c), (d). [/fig]
|
Online community engagement in response to COVID‐19 pandemic
| 729LETTER TO THE EDITOR were used to mitigate these issues. We recognize that community engagement is difficult during crises due to changing lifestyles and commitments with some CFs volunteering to deliver necessities in their community. Whilst not observed, participants may be reluctant to attend online meetings if they feel they lack privacy in their home setting. Whilst not always practically possible, we encouraged CFs to undertake meetings in a private room in their house. Other limitations and/or risks of virtual meetings include: (a) loss of body language and non-verbal cues, particularly if individuals switch off their cameras and (b) risk worsening health inequities by reliance on digital methods as the highest risk, hardest to reach populations are likely those who experience greater digital poverty. 16 After conducting online parent/carer workshops to gather feedback around the COVID-19 experience, the G2KCP project team is undertaking qualitative focus group discussions (FGD) and semi-structured interviews online. The team have incorporated solutions on how to conduct online qualitative research from Tuttas et al (2014) such as reducing FGDs sizes so that they are more manageable online. 17 In addition, they are exploring the feasibility of digital implementation of the programme. Collaborators in Rwanda and Uganda are similarly currently exploring the digital adaptation of the ABaANA programme. To advance patient and public involvement and community engagement, an online Facebook group is being set up as a virtual support group for East London parents during the pandemic. Furthermore, this virtual community can act as a space to disseminate information and promote shared learning.NEON and community parent groups for neurodisability represent two exemplar community participatory research projects adapting to COVID-19 by implementing online research, meetings and intervention implementation as a successful alternative to faceto-face meetings. There may still be inevitable digital divide in the community, and evaluation of these is planned and required to ascertain impact on equity and efficacy. We hope that many generalizable lessons will be learned from our experiences that may be applied to the conduct of participatory and qualitative research and to the implementation of community-based interventions in the UK and beyond.K E Y WO R DSchildren, community engagement, community-based intervention, infant, online interventions CO N
## Online community engagement in response to covid-19 pandemic
Community participatory research is a method of active engagement between community members to develop interventions that catalyse social change and improve health outcomes.Increasingly, particularly in low-middle-income countries (LMICs), participatory research is used to co-develop low-cost community interventions.
UK health policy and providers are also placing increasing emphasis on strengthening community services to equitably optimize health outcomes. South-to-north translation of community-based participatory interventions provides enormous potential to address these aims. In this article, we provide two examples of interventions which have been successfully implemented in LMICs 2-4 and are currently undergoing adaptation and testing in the UK, driven by the emphasis of NHS policy shifting towards community strengthening.The first is community mobilization through participatory learning and action (PLA) cycles with women's groups and the second community-based parent group programme for children with complex neurodisability (Ubuntu). [bib_ref] Evaluating the impact of a community-based parent training programme for children with..., Zuurmond [/bib_ref] [bib_ref] Long-term impact of community-based participatory women's groups on child and maternal mortality..., Heys [/bib_ref] Here we describe how these programmes are being further adapted to run partially or fully online in response to COVID-19, and lessons can be learnt from these two examples, which have accelerated digital innovation in the delivery of health services to enable non-contact approaches.
Women's groups PLA cycle is a WHO-recommended intervention to produce sustained improvements in maternal and newborn outcomes in rural LMICs. PLA cycles are iterative processes led by community facilitators (CFs) guiding participants through four-stage cycles of identifying and prioritizing contextual health issues, designing strategies to address these issues and evaluating post-implementation. The Nurture Early for Optimal Nutrition (NEON) programme 9 aims to co-adapt PLA from LMIC to theUK to improve infant feeding, care and dental hygiene practices for South Asian infants <2 years in East London. Since COVID-19, there has been an increase in anxiety, depression and other mental health morbidity, which may have impacted on CF's proactiveness to participate online. Participation promotion techniques, for example inviting participants by name to speak in turn, splitting the group into smaller discussion groups, |
Systemic Immunosuppression in High-Risk Penetrating Keratoplasty: A Systematic Review
Cornea transplantation has a high success rate and typically only requires topical immunomodulation. However, in high-risk cases, systemic immunosuppression can be used. We conducted a systematic review on the efficacy and side effects of systemic immunosuppression for high-risk cornea transplantation. The study population was 18 years old or older with a high-risk transplant (two or more clock hours of cornea vascularization or a previous failed graft or a graft needed because of herpes simplex keratitis). A comprehensive search strategy was performed with the help of an information specialist and content experts from ophthalmology. All study designs were accepted for assessment. Level 1 and level 2 screening was performed by two reviewers followed by data abstraction. Forest plots were created whenever possible to synthesize treatment effects. Quality assessment was done with a Downs and Blacks score. From 1,150 articles, 29 were ultimately used for data abstraction. The odds ratios (ORs) for clear graft survival in cyclosporine and controls were 2.43 (95% CI: 1.00 -5.88) and 3.64 (95% CI: 1.48 -8.91) for rejection free episodes. Mycophenolate mofetil (MMF) significantly improved the rejection free graft survival rates at 1 year (OR: 4.05, 95% CI: 1.83 -8.96). The overall results suggested that both systemic cyclosporine and MMF improved 1-year rejection free graft survival in high-risk keratoplasty. Cyclosporine also significantly improved clear graft survival rates at 1 year; however, there were insufficient data to analyze the same in the MMF group. Higher quality studies are needed to understand this issue better.
# Introduction
Corneal transplantation is one of the most commonly performed human transplantation surgeries. The overall 10-year survival rates of corneal grafts range between 75% and 80% [bib_ref] Long-term graft survival after penetrating keratoplasty, Thompson [/bib_ref] [bib_ref] Penetrating keratoplasty in Asian eyes: the Singapore Corneal Transplant Study, Tan [/bib_ref]. However, in the presence of "high-risk" conditions, the survival rate drops to 30-50% at 3-to 5-year follow-up [bib_ref] Longterm outcome in high-risk corneal transplantation and the influence of HLA-A and..., Bartels [/bib_ref] [bib_ref] Corneal graft survival and visual outcome. A multicenter Study. Corneal Transplant Follow-up..., Vail [/bib_ref] [bib_ref] Outcome of optical penetrating keratoplasties at a tertiary care eye institute in..., Joshi [/bib_ref]. The Singapore Corneal Transplant Study revealed that while the 5-year transplant survival rate was as high as 100% for keratoconic cases, only 18% of corneas with regrafts survived at 5-year followup [bib_ref] Penetrating keratoplasty in Asian eyes: the Singapore Corneal Transplant Study, Tan [/bib_ref]. Bersudsky et al, in their study on repeat corneal grafts, noted that only 28% of the 78 first regrafts remained clear at the end of follow-up of 54 months [bib_ref] The profile of repeated corneal transplantation, Bersudsky [/bib_ref]. The number dropped to 20% for subsequent regrafting [bib_ref] The profile of repeated corneal transplantation, Bersudsky [/bib_ref]. The most common reason responsible for graft failure is immunologic rejection that accounts for almost one-third of the failure rates [bib_ref] Penetrating keratoplasty in Asian eyes: the Singapore Corneal Transplant Study, Tan [/bib_ref] [bib_ref] The profile of repeated corneal transplantation, Bersudsky [/bib_ref] [bib_ref] The impact of corneal allograft rejection on the long-term outcome of corneal..., Coster [/bib_ref].
Ocular immune privilege has been described by Niederkorn as a "three legged stool" comprising an afferent arm blockade, deviation of the immune response to a state of immune tolerance and blockade of efferent arm [bib_ref] Immune privilege of corneal allografts, Niederkorn [/bib_ref] [bib_ref] Graft failure IV. Immunologic mechanisms of corneal transplant rejection, Chong [/bib_ref]. The afferent arm blockade results from avascularity, lack of lymphatics, low major histocompatibility complex expression and presence of native immunosuppression molecules. The clonal deletion, anergy and immune deviation serve as the central arm and protective molecules FasL and PD-L1 act to block the efferent arm of the immune cascade. A breach in any of the above increases the immunogenicity of the corneal transplant and places that cornea at "high risk" for immunologic rejection. The Collaborative Cornea Transplantation Study has described high-risk conditions as presence of more than two quadrants of corneal neovascularization and sensitization due to a previous graft [bib_ref] High risk corneal grafting, Hill [/bib_ref]. The other conditions that may place the cornea at a higher risk of rejection are position of the graft close to limbus, simultaneous limbo-keratoplasty, severe atopic dermatitisand herpes simplex keratitis (HSV) [bib_ref] Penetrating keratoplasty for herpes simplex keratitis and keratoconus. Allograft rejection and survival, Epstein [/bib_ref] [bib_ref] Efficacy of postoperative immunosuppression after keratoplasty in herpetic keratitis, Maier [/bib_ref].
Reports from experimental models suggest that transported corneal alloantigens lead to clonal expansion of T cells in regional lymph nodes and spleen [bib_ref] Involvement of CD8+ RT1.B+ and CD4+ RT1. B+ cells of cervical lymph..., Okada [/bib_ref] [bib_ref] The critical role of lymph nodes in corneal alloimmunization and graft rejection, Yamagami [/bib_ref] , and have justified the use of systemic immunosuppression in high-risk cases in an attempt to prolong corneal graft survival. The common drugs included in these protocols are calcineurin inhibitors, including cyclosporine and tacrolimus, and antimetabolites, including MMF and azathioprine [bib_ref] Systemic cyclosporine in high-risk keratoplasty. Short-versus long-term therapy, Hill [/bib_ref] [bib_ref] Mycophenolate mofetil (MMF) following penetrating high-risk keratoplasty: long-term results of a prospective,..., Birnbaum [/bib_ref] [bib_ref] Tacrolimus immunosuppression in high-risk corneal grafts, Joseph [/bib_ref] [bib_ref] Sa-Immunosuppression in Penetrating Keratoplasty, Sanchez [/bib_ref]. The calcineurin inhibitors bind to a specific cytosol protein and inhibit calcineurin-calm- A number of studies have been published to report the efficacy of these immunosuppressive agents as prophylaxis against corneal graft rejection [bib_ref] Systemic cyclosporine in high-risk keratoplasty. Short-versus long-term therapy, Hill [/bib_ref] [bib_ref] Mycophenolate mofetil (MMF) following penetrating high-risk keratoplasty: long-term results of a prospective,..., Birnbaum [/bib_ref] [bib_ref] Tacrolimus immunosuppression in high-risk corneal grafts, Joseph [/bib_ref] [bib_ref] Sa-Immunosuppression in Penetrating Keratoplasty, Sanchez [/bib_ref]. However, the results have been inconsistent and there is lack of evidence-based guidelines about the use of these agents for immunoprophylaxis in high-risk corneal transplantation.
The purpose of this study was to conduct a systematic review of the published literature and examine the published evidence on efficacy of systemic immunosuppressive agents as prophylaxis against corneal graft rejection in cases at "high risk" of immunologic rejection after corneal transplantation.
# Methods
## Inclusion criteria
Any study investigating the role of systemic immunosuppression in high-risk penetrating keratoplasty was included in our review to maximize interpretability and generalizability. Patients had to be adults over 18 years of age. High risk was defined as two or more vascular quadrants in the cornea, surgery for a previously failed graft or a history of HSV keratitis in the operated eye.
## Study identification and databases
The search strategy for this project was comprehensive and was tailored to achieve the highest possible recall of relevant studies. An electronic search strategy was developed by an information specialist in consultation with two clinical content experts in corneal transplantation. The following bibliographic databases were searched: PubMed; Ovid's Medline in-Process & Other Non-Indexed Citations, Medline, EMBASE and CINAHL; Thomson's Social Sciences Citation Index and Biosis Previews.
Searches were not restricted by publication type, or study design. The search had no initial time restriction. The final search time was March 2014 but updates continued monthly until August 2014.
## Study design
Published and unpublished reports of any design were included for our systematic review. The review included studies with a comparison group (randomized controlled trial (RCT)), noncontrolled and observational studies (pre-versus post-, prospective cohort, and case series designs).
## Study selection
We drafted specific screening questions for all levels of rele-vance assessment (level 1: title and abstract screening; level 2: full text relevance screening) and performed a calibration exercise involving questions developed specifically for this review. All records were uploaded into an internet-based, secured, software program (evidence for policy and practice information (EPPI)) to help manage the review. All records retrieved through searches were initially screened broadly (level 1) using titles, and abstracts and done by two reviewers. All records that were tagged at this level as a review article, report, or statement were screened for relevance for our review for reference matching. Reference lists of reviews that were thought to be relevant were screened for potentially relevant publications using reference list checking. Discrepancies were addressed by consensus between the two reviewers and when this was not possible, an adjudicator (WGH) resolved the conflict.
All studies identified as potentially relevant were retrieved in full-text format, and screened independently (level 2), again by two reviewers. The same method to address reliability and discrepancies was used as in level 1. All studies excluded at this level were placed in an excluded database, and exclusionary reasons were noted, and used to create a PRISMA (preferred reporting items for systematic reviews and meta-analysis) checklist.
# Data abstraction
Following a calibration exercise involving two studies, data abstraction was performed using an electronic data abstraction form developed and tailored specifically for this review. Study and outcome characteristics to be extracted included report (e.g., publication type, location, year of publication), study (e.g., sample size, research design, number of study arms/groups, cohorts, or phases), population (e.g., age, gender, diagnosis description, high-risk characteristics, duration of follow-up), treatment under study outcomes (e.g. type, dosage and duration of immunosuppressive agent, clear graft survival at 1 and 3 years, rejection free graft survival at 1 and 3 years), and adverse events (e.g., side effects, post-keratoplasty complications, systemic illnesses associated with the medication, withdrawals or termination of treatment). Data abstraction occurred with two reviewers, a primary reviewer and a secondary reviewer who verified the work.
## Study quality
We used the instrument generated by Downs and Black for appraising quality [bib_ref] The feasibility of creating a checklist for the assessment of the methodological..., Downs [/bib_ref]. The Downs and Black checklist for quality assessment was selected as it has been developed to use with both randomized and non-randomized studies and is recommended as being suitable for use in systematic reviews [bib_ref] The inclusion of reports of randomised trials published in languages other than..., Moher [/bib_ref] [bib_ref] Public health interventions in midwifery: a systematic review of systematic reviews, Mcneill [/bib_ref].
## Summarizing the evidence
Qualitative data synthesis was used when outcomes were not amenable to quantitative synthesis. When quantitative synthesis was available but too heterogenous for meta-analysis, the data were summarized but not in a forest plot. Quantitative data synthesis studies of the association between treatment option and outcome were considered for meta-analysis. The measure of association used was the odds ratio (OR) adjusted for possible confounding factors and was combined across studies where possible. For forest plots presented, only studies with a comparison group were used (i.e. case series were not used). [fig_ref] Figure 1: PRISMA [/fig_ref] shows the PRISMA diagram for this review. A total of 1,150 studies of potential interest were identified by the original literature search. Two hundred sixty-eight articles were duplications and were eliminated. After screening the titles and abstracts, 735 further articles were excluded. For the remaining 147 articles, the full texts were retrieved and screened. A further 78 articles were excluded from the review based on this screening. Of the remaining 69 articles, 40 were excluded either because the full texts articles were inaccessible (n = 4), or the articles were not in English (n = 19), or were not relevant to the review (n = 17). Thus, a total of 29 papers were selected for the final analysis.
# Results
Of the 29 articles selected, eight were RCTs, 10 were cohort studies and 11 were case series. Sixteen studies evaluated the results of systemic immunosuppression cyclosporine A (CsA) alone, of which three were RCTs [bib_ref] Systemic cyclosporine in high-risk keratoplasty. Short-versus long-term therapy, Hill [/bib_ref] [bib_ref] Sa-Immunosuppression in Penetrating Keratoplasty, Sanchez [/bib_ref] [bib_ref] Systemic Cyclosporine in High-Risk Corneal Transplants, Akata [/bib_ref] [bib_ref] Randomized Study on Efficacy of Systemic Cyclosporine a in High-Risk Corneal Transplantation, Den [/bib_ref] [bib_ref] The use of systemic cyclosporin a in human corneal transplantation: a preliminary..., Hill [/bib_ref] [bib_ref] The use of cyclosporine in high-risk keratoplasty, Hill [/bib_ref] [bib_ref] Long-term outcome of systemic cyclosporine treatment following penetrating keratoplasty, Inoue [/bib_ref] [bib_ref] Successful engraftment of high-risk corneal allografts with short-term immunosuppression with cyclosporine, Miller [/bib_ref] [bib_ref] Systemic cyclosporin A in high risk penetrating keratoplasties: a case-control study, Poon [/bib_ref] [bib_ref] Systemic ciclosporin A in high-risk keratoplasties, Reinhard [/bib_ref] [bib_ref] Systemic cyclosporin A in high failure risk, repeated corneal transplantation, Rumelt [/bib_ref] [bib_ref] Prospective, randomized study of the efficacy of systemic cyclosporine in high-risk corneal..., Shimazaki [/bib_ref] [bib_ref] Six years' experience with systemic cyclosporin A prophylaxis in highrisk perforating keratoplasty..., Sundmacher [/bib_ref] [bib_ref] Systemic cyclosporin treatment for high-risk corneal transplantation, Uusitalo [/bib_ref] [bib_ref] Cyclosporin A in the treatment of penetrating keratoplasty, Vlkova [/bib_ref]. Four studies investigated the outcomes of high-risk grafts after immunosuppression with MMF alone and involved three RCTs [bib_ref] Mycophenolate mofetil (MMF) following penetrating high-risk keratoplasty: long-term results of a prospective,..., Birnbaum [/bib_ref] [bib_ref] Synergistic antiherpetic effect of acyclovir and mycophenolate mofetil following keratoplasty in patients..., Mayer [/bib_ref] [bib_ref] Systemic mycophenolate mofetil avoids immune reactions in penetrating high-risk keratoplasty: preliminary results..., Reinhard [/bib_ref] [bib_ref] Long-term results of immunosuppressive therapy with mycophenolate mofetil (CellCept(R)) for keratoplasty in..., Soares-Wulf [/bib_ref]. Four studies conducted a comparative evaluation of CsA and MMF and two of these studies were RCTs [bib_ref] Efficacy of postoperative immunosuppression after keratoplasty in herpetic keratitis, Maier [/bib_ref] [bib_ref] Immunosuppression with cyclosporine A and mycophenolate mofetil after penetrating highrisk keratoplasty: a..., Birnbaum [/bib_ref] [bib_ref] Systemic mycophenolate mofetil in comparison with systemic cyclosporin A in high-risk keratoplasty..., Reinhard [/bib_ref] [bib_ref] Mycophenolate mofetil versus cyclosporin A in high risk keratoplasty patients: a prospectively..., Reis [/bib_ref]. The immunosuppression regimens used in other studies were rapamycin alone (n = 1) [bib_ref] An open prospective pilot study on the use of rapamycin after penetrating..., Birnbaum [/bib_ref] or in combination with MMF (n = 1) [bib_ref] Sirolimus and mycophenolate as combination prophylaxis in corneal transplant recipients at high..., Chatel [/bib_ref] , tacrolimus (n = 4) [bib_ref] Tacrolimus immunosuppression in high-risk corneal grafts, Joseph [/bib_ref] [bib_ref] Efficacy of tacrolimus in the management of high-risk corneal grafts, Joseph [/bib_ref] [bib_ref] Tacrolimus in the management of high-risk corneal grafts, Joseph [/bib_ref] [bib_ref] Tacrolimus (FK506) in the management of high-risk corneal and limbal grafts, Sloper [/bib_ref] , and azathioprine (n = 1) [bib_ref] Sa-Immunosuppression in Penetrating Keratoplasty, Sanchez [/bib_ref]. The quality assessment of the articles was done using the Downs and Blacks score [bib_ref] The feasibility of creating a checklist for the assessment of the methodological..., Downs [/bib_ref]. The median score for all studies was 18, with a range from 6 to 24. Overall, the mean age of study participants was 41.73 ± 22.74 years. The mean percent of females for all the studies was 45.3%. The average followup period for all the studies was 25.45 ± 11.16 months.
At level 1 screening where 882 articles were screened, there were 25 disagreements for a kappa score of 0.90. At level 2 screening, there were 147 articles screened and there were 15 disagreements for a kappa score of 0.71.
## Csa
Sixteen studies evaluated the results of systemic immunosuppression in high-risk keratoplasty with CsA alone, of which eight were cohort studies, five were case series and three were RCTs.
The total number of subjects enrolled in all studies using cyclosporine was 817, of which 35.7% were females. Five hundred eighteen patients received systemic cyclosporine for post-keratoplasty immunosuppression (there were 299 controls) with the average blood CsA concentration ranging from 210 to 395 ng/mL. The mean age of patients included for studies was 49.5 ± 12.4 years. The mean post-operative followup period was 26.5 ± 12.9 months. The mean rejection free and clear graft survival rates at 1 year were 80.5±12.1% and 85.3±14.4%, respectively. Only six studies evaluated the longterm Kaplan-Meir survival rates at 3 years. In these studies, the mean rejection free and clear graft survival rates at 3 years were 67.6±16.4% and 54±26.8%, respectively. In patients on systemic CsA, 66.4% of the rejection episodes were successfully reversed with medical management. On the other hand, among control subjects, the average rate of rejection reversal was 27.8% (P = 0.02).
We found three RCTs comparing the efficacy of systemic cyclosporine against controls for immunosuppression following high-risk keratoplasty. Den et al evaluated 38 patients with high-risk keratoplasty, of which 57.9% were females [bib_ref] Randomized Study on Efficacy of Systemic Cyclosporine a in High-Risk Corneal Transplantation, Den [/bib_ref]. Systemic CsA was used at a serum concentration of 500 -800 ng/mL for at least 12 months. They did not find any significant difference in the clear graft survival rates between the cases and controls. Sanchez et al conducted an RCT wherein four patients received oral prednisone and CsA each for a period of 6 months following high-risk keratoplasty [bib_ref] Sa-Immunosuppression in Penetrating Keratoplasty, Sanchez [/bib_ref]. At a follow-up of 12 months, 100% grafts in the CsA group were rejection free, whereas in the control group, the rejection free graft survival was 0%. In another study, Shimazaki et al enrolled 39 patients in an RCT, 20 of which received systemic CsA (Serum CsA concentration: 600-1000 ng/mL) for prophylactic immunosuppression [bib_ref] Prospective, randomized study of the efficacy of systemic cyclosporine in high-risk corneal..., Shimazaki [/bib_ref]. It was found that cyclosporine failed to show a positive effect in preventing rejection in high-risk corneal transplantation (1-year graft survival rates of 94.7% for CsA group and 94.1% for the control group). [fig_ref] Figure 2: One-year clear grafts OR for cyclosporine vs [/fig_ref] show the forest plots for 1-year clear We performed a meta-regression to see if any covariates significantly affected this relationship and studied age, gender, diagnosis, topical steroid duration, systemic steroid duration, study quality, and study design. No variable had a significant effect in the meta-regression. The Horbold-Egger bias was used to detect funnel plot asymmetry and publication bias was not found (P = 0.41).
## Mmf
Four studies evaluated the results of systemic MMF on graft survival rates for patients undergoing high-risk corneal transplantation, and three of the studies were RCTs. The total number of subjects enrolled in the four studies was 242, of which 147 received systemic MMF for prophylactic immunosuppression (95 were control subjects). MMF was used at a dosage of 2 g/day for 6 months in three studies, whereas one study used it for 12 months. The mean age of the participants was 56.95 ± 2.26 years and the mean follow-up duration was 25.9 ± 12.32 months. Overall, the mean Kaplan-Meir rejection free graft survival rate was 89.05% at 1 year. Only two studies reported the 3-year graft survival rate, with a mean of 76.5%. In patients on systemic MMF, the rate of reversibility of rejection episodes was 91.7%. This was higher as compared to the control group where only 52.05% rejection episodes were reversible (P = 0.01).
Three of these studies were RCTs. Birnbaum et al enrolled 98 of 140 scheduled patients after an interim evaluation showed a statistically significant result [bib_ref] Mycophenolate mofetil (MMF) following penetrating high-risk keratoplasty: long-term results of a prospective,..., Birnbaum [/bib_ref]. They noted that 83% of patients in the MMF group and 64.5% in the control group were free of immunologic rejection at a follow-up of 39.5 months (P = 0.044). Mayer et al evaluated 30 patients who underwent penetrating keratoplasty for herpetic eye disease [bib_ref] Synergistic antiherpetic effect of acyclovir and mycophenolate mofetil following keratoplasty in patients..., Mayer [/bib_ref]. Ten of these patients received MMF for 1 year for prophylaxis against immunologic reaction. They found that 90% of patients on MMF and 80% of controls were free of rejection at 36 months follow-up. This difference was found to be statistically significant. Reinhard et al conducted a randomized multicenter trial and published preliminary results of 86 patients [bib_ref] Systemic mycophenolate mofetil avoids immune reactions in penetrating high-risk keratoplasty: preliminary results..., Reinhard [/bib_ref]. Forty-eight of these patients received MMF for immunoprophylaxis and 89% of these grafts had no immune reaction at 1-year follow-up. This was noted to be significantly higher than the controls, where the 1-year rejection free graft survival was 67%.
Forest plot results [fig_ref] Figure 4: One-year rejection free OR for MMF vs [/fig_ref] showed that systemic MMF significantly improved the rejection free graft survival rates at 1 year in high-risk keratoplasty (OR: 4.05, 95% CI: 1.83 -8.96, I 2 = 0%) There were not enough data to analyze the clear graft survival rates in this group. Meta-regression failed to show significant effect of any of the variables on graft survival or immunologic rejection outcomes. Studied covariates included age, gender, topical steroid duration, systemic steroid duration, study quality or study design. The Horbold-Egger bias was used to detect funnel plot asymmetry and publication bias was not found (P = 0.31).
## Cyclosporine versus mmf
Four studies compared the efficacy of CsA and MMF in promoting graft survival in high-risk penetrating keratoplasty. The total number of patients enrolled in these studies was 598. Of these, 304 received CsA, 278 received MMF and 16 patients received combined CsA and MMF. The mean age of study participants was 57.5 ± 4.02 years and the mean follow-up was 24.3 ± 12.7 months. In the cyclosporine group, at 1 year, the mean rejection free and clear graft survival rates were 88.8% and 88.6%, respectively. The respective mean rates for MMF patients were 88.6% and 97.3%. Overall, 88% of immune reactions in the CsA group and 81.3% in the MMF group were successfully reversed. Two of these studies were RCTs. In both the trials, no significant difference was noted in the efficacy of CsA and MMF in prevention of graft rejection in high-risk penetrating keratoplasty [bib_ref] Systemic mycophenolate mofetil in comparison with systemic cyclosporin A in high-risk keratoplasty..., Reinhard [/bib_ref] [bib_ref] Mycophenolate mofetil versus cyclosporin A in high risk keratoplasty patients: a prospectively..., Reis [/bib_ref].
## Safety
Overall, nephrotoxicity was the most common adverse effect reported with the use of CsA and was noted in an average of 5.5% patients. The other reported significant side effects were gastrointestinal (4.3%), elevated blood pressure (4.02%), neu-rotoxicity (1.4%), skin rash (0.7%), and systemic infection (0.5%). For all studies, in 6.3% patients, CsA had to be withdrawn as a result of significant toxicity.
The most common adverse effect reported with MMF was gastrointestinal, reported in 13.6% of individuals. The other significant reported side effects were elevated blood pressure (9.2%), systemic infection (6.5%), hyperlipidemia (5.2%), hepatotoxicity (2.2%), tachycardia (2.1%), lymphoma (1.8%), and arthralgia/myalgia (1.2%). Overall, MMF had to be withdrawn in 10.5% of patients as a result of these adverse effects.
# Discussion
The overall results suggest that the use of both systemic cyclosporine and MMF improves 1-year rejection free graft survival in high-risk keratoplasty. CsA also significantly improved clear graft survival rates at 1 year; however, there were insufficient data to analyze the same in the MMF group. The use of both drugs was associated with higher reversibility of rejection episodes as compared with controls. The other drugs used in different studies were rapamycin, tacrolimus and azathioprine, but there were insufficient data to perform an accurate synthesis.
The safety profile of the systemic drugs was acceptable. The most common toxicity noted with CSA was nephrotoxicity followed by gastrointestinal and hypertension. MMF use was most commonly associated with gastrointestinal upset and hypertension. Overall, the medications had to be withdrawn in less than about 11% of patients due to toxicity.
The meta-analysis performed on MMF rejection free efficacy showed no heterogeneity among the small number of studies found. There was moderate heterogeneity for the metaanalysis performed on clear graft survival for cyclosporine (I 2 = 37.9%). There was large heterogeneity found in the meta- analysis for cyclosporine rejection free survival (I 2 = 64.8%). This was likely from two studies with very high ORs. Removing these studies from the meta-analysis produced results with a smaller OR but in the same direction as the full meta-analysis. Overall, there were some lacunae in the data available pertaining to this research question. Firstly, the overall quality assessment of the studies included was average.
Of the 29 articles selected, only eight were RCTs. Eleven studies were case series and hence did not have a control group. The mean quality assessment score of the studies using Downs and Blacks instrument was 18, out of a possible total of 30. Secondly, there were insufficient follow-up data and results to conduct a meta-analysis on the effect of the drugs on longterm graft survival.
Also noteworthy is that while cyclosporine showed an overall beneficial effect on rejection and survival in analytic studies, it had a surprisingly null effect on both outcomes in the two larger RCTs [bib_ref] Randomized Study on Efficacy of Systemic Cyclosporine a in High-Risk Corneal Transplantation, Den [/bib_ref] [bib_ref] Prospective, randomized study of the efficacy of systemic cyclosporine in high-risk corneal..., Shimazaki [/bib_ref] except for one very small RCT [bib_ref] Sa-Immunosuppression in Penetrating Keratoplasty, Sanchez [/bib_ref]. This is an important point in that RCTs represent our highest level of evidence and the discrepancy between the RCT results and overall results indicate that these questions need to be studied further.
To conclude, the results of this systematic review demonstrate that both CSA and MMF have a clinical benefit in improving 1-year graft survival in patients undergoing high-risk keratoplasty. The choice between the two may be dictated by cost and availability, presence of co-morbidities and the clinical experience of the prescribing physician. More high quality studies are needed to understand this important issue better.
# Funding support
The review was supported by AMOSO Innovation Fund (INN12-010).
# Disclosure
There was no commercial interest with any of the authors.
[fig] Figure 1: PRISMA [/fig]
[fig] Figure 2: One-year clear grafts OR for cyclosporine vs. controls. and 1-year rejection free episodes amongst all of the analytic studies (i.e. RCTs plus observational studies). Although there were 11 analytic studies in total, only 10 had 1-year data or longer. Six of the 10 studies reported both clear graft survival and rejection free episodes and were included in both forest plots, whereas four studies only reported one of the outcomes and hence were only found in one or the other forest plot. The OR for clear graft survival in cyclosporine vs. controls was 2.43 (95% CI: 1.00 -5.88, I 2 = 37.9%) and for rejection free episodes was 3.64 (95% CI: 1.48 -8.91, I 2 = 64.8%). [/fig]
[fig] Figure 3: One-year rejection free OR for cyclosporine vs. controls. [/fig]
[fig] Figure 4: One-year rejection free OR for MMF vs. controls. [/fig]
|
Point-of-Care Virologic Testing to Improve Outcomes of HIV-Infected Children in Zambia: A Clinical Trial Protocol
# Introduction
Despite recent advances in prevention science, pediatric AIDS remains an important public health problem in many countries with high HIV burden. In 2014, an estimated 170,000 children were newly infected with HIV.Motherto-child transmission is responsible for the majority of these pediatric infections and can be prevented with antiretroviral therapy (ART). Yet, in many settings, the coverage of prevention interventions remains suboptimal. [bib_ref] Performance of HIV prevention of mother-to-child transmission programs in sub-Saharan Africa: longitudinal..., Ladner [/bib_ref] [bib_ref] Missed opportunities to prevent mother-to-child-transmission: systematic review and meta-analysis, Wettstein [/bib_ref] [bib_ref] Coverage of nevirapinebased services to prevent mother-to-child HIV transmission in 4 African..., Stringer [/bib_ref] [bib_ref] Effectiveness of a city-wide program to prevent mother-to-child HIV transmission in Lusaka,..., Stringer [/bib_ref] When ART or antiretroviral prophylaxis is not provided (or fails), early infant diagnosis (EID) and immediate initiation of ART become critical because HIV infection progresses rapidly in infants. [bib_ref] Early antiretroviral therapy and mortality among HIV-infected infants, Violari [/bib_ref] It is well established that early ART has a profound survival effect, [bib_ref] Early antiretroviral therapy and mortality among HIV-infected infants, Violari [/bib_ref] [bib_ref] Optimization of antiretroviral therapy in HIV-infected children under 3 years of age:..., Penazzato [/bib_ref] [bib_ref] The effect of early initiation of antiretroviral treatment in infants on pediatric..., Johnson [/bib_ref] and WHO guidelines recommend that ART be initiated in HIV-infected infants as soon as possible after diagnosis.Despite this global recommendation, however, the majority of HIV-infected children in sub-Saharan Africa (SSA) never start treatment.An important reason for this failure lies in the complexity of EID. Most national programs use HIV DNA polymerase chain reaction (PCR) assays for infant diagnosis, a sophisticated technology that is best suited for high-volume, centralized laboratories. In a typical program, dried blood spot (DBS) specimens are obtained and transported from peripheral locations to a centralized laboratory. This logistically challenging arrangement can result in delays of many weeks or months before results are returned to patients and can also contribute to results being lost during sample and data transfer. In the meantime, children may become lost to follow-up (LTFU). In Malawi, for example, delays between collection of the blood sample and parents' receipt of their infant's HIV test results were identified as a major cause of pediatric HIV program attrition. [bib_ref] Inadequate coordination of maternal and infant HIV services detrimentally affects early infant..., Braun [/bib_ref] The challenges are similar in Zambia, where only 4 public-sector laboratories currently perform HIV DNA PCR testing for EID.
Alere q Detect Qualitative Test for EID is a point-ofcare (POC) nucleic acid testing platform that is suitable for placement in low-volume and remote settings. [bib_ref] Point-of-care tests poised to alter course of HIV treatment, Laursen [/bib_ref] Results are available within 1 hour, which allows reporting of results on the same day. Clinical performance of the Alere q was recently validated in Mozambique in a study that included heel-stick samples from 827 HIV-exposed infants. The prevalence of HIV among infants in this study was 7.9%. Gold standard testing was performed using the Roche COBAS Ampliprep/Taqman system. The sensitivity of Alere q was reported as 98.5% [95% confidence interval (CI): 91.7% to 99.9%] and the specificity as 99.9% (95% CI: 99.3% to 100%). The positive and negative predictive values for the test were also high [98.5 (95% CI: 96.3% to 100%) and 99.9% (95% CI: 99.7% to 100%), respectively]. [bib_ref] Accurate early infant HIV diagnosis in primary health clinics using a point-of-care..., Jani [/bib_ref] In this article, we present the study protocol for randomized, clinical trial that tests the field effectiveness of Alere q as a tool to improve outcomes of HIV-infected children in Zambia. HIV-exposed infants are randomized to receive their diagnosis either using Alere q or using the usual standard of off-site PCR. Our primary hypothesis is that HIVinfected children who receive their diagnosis using the Alere q POC test will be more likely to start ART and to remain alive, in care, with suppressed HIV plasma viral load 12 months after their diagnostic blood draw, compared with those who are randomized to receive a diagnosis using the current standard for EID.
# Methods and design
## Study design
Our study is an unmasked, randomized, controlled trial. Participating HIV-exposed infants are randomly assigned to receive either (1) standard, off-site EID through HIV DNA PCR testing of DBS samples or (2) POC EID using the Alere q HIV Detect test. Following the principles of a "pragmatic trial," our intervention is overlaid on the background of routine clinical practice, with minimal protocol-related constraints, liberal inclusion criteria, and few exclusion criteria. [bib_ref] What are pragmatic trials?, Roland [/bib_ref] [bib_ref] Why do we need some large, simple randomized trials, Yusuf [/bib_ref] The study is being conducted in prevention of mother-to-child transmission and pediatric ART clinics at 6 participating public health facilities in Lusaka, Zambia. These study sites were selected because the volume of HIV-exposed infants receiving EID services at each site is .100 infants per year.
## Study participants
To be eligible for inclusion, infants are between the ages of 4 and 12 weeks of life and have HIV exposure documented by a seropositive maternal or infant HIV antibody test. To align the trial population with what would be expected in a real-world implementation of this technology, we have elected to exclude infants with major congenital anomalies or other medical conditions that require management at a referral facility. We also excluded those whose parents/guardians are unwilling to provide written consent for study participation.
## Intervention
The trial includes 2 study arms: a standard of care (SOC) or control arm and an intervention arm. In both study arms, EID is provided at ;6 weeks of life. Infants randomized to the SOC arm receive EID through the extant public-sector EID program in Lusaka, with DBS samples sent to an off-site laboratory for DNA PCR testing. Infants randomized to the intervention arm receive POC Alere q testing along with a DBS drawn for confirmatory off-site DNA PCR testing of Alere q positive results.
## Primary objective
The main objective of our trial is to determine whether POC virologic testing improves clinical outcomes of HIVinfected children in Zambia. We define an ideal outcome as being alive, retained in care, receiving ART, and having a suppressed viral load, and we have designed the study to measure this composite.
## Study procedures
In Zambia, EID is provided at 6 weeks of life and corresponds with the infant's initial vaccination visit. During this routine visit, study personnel approach parents/guardians of potential participants to determine eligibility for enrollment. Research personnel describe the study and obtain written informed consent for study enrollment from caregivers of eligible participants.
Eligible infants are randomly allocated to 1 of 2 study arms in a 1:1 ratio. Infants randomly assigned to the SOC arm have a blood specimen drawn for DNA PCR testing and are asked to return to clinic within 4 weeks to receive the result. Positive DBS results are confirmed with a second test, in keeping with national guidelines. Infants randomly assigned to the intervention arm receive POC EID testing using the Alere q HIV Detect test. Those who are Alere q positive are immediately enrolled into the ART program with ART initiated as soon as possible. Once again, all positive tests are confirmed with a second test.
All HIV-infected infants are initiated on a lopinavir/ ritonavir-based regimen [bib_ref] Antiretroviral treatment for children with peripartum nevirapine exposure, Palumbo [/bib_ref] and followed in the government clinics with monthly follow-up visits during the first year on ART. Their care is provided by clinic staff and includes physical examinations, laboratory monitoring, and basic adherence counseling. ART is provided through the existing government treatment programs. Current Zambian guidelines also recommend viral load monitoring 3 times during the first year of treatment (at 6, 9, and 12 months) and annually thereafter, although this schedule is not always followed in real practice.
After ART initiation, 3 additional follow-up study visits are scheduled for the infants at 3, 6, and 12 months [fig_ref] TABLE 1: Schedule of Events in the Control Arm [/fig_ref]. Per study protocol, all infants undergo plasma HIV viral load testing at 12 months to measure the primary outcome. If at any time during follow-up there is clinical or laboratory evidence of virologic failure, participants are referred for intensive adherence counseling with clinical changes made in accordance with local guidelines for management of virologic failure in pediatric populations.
## Laboratory testing
For HIV DNA PCR testing, five 75-100 mL DBS aliquots are prepared on a Whatman 903 collection card (GE Healthcare Bio-Sciences Corporation, Piscataway, NJ) using a heel-stick, whole-blood sample. Appropriately labeled DBS cards are then transported to the central laboratory in a sealed bag containing desiccant. DBS samples are tested using the Roche COBAS Ampliprep/Taqman Qualitative HIV-1 Test v2.0 (Roche Molecular Diagnostics, Pleasanton, CA) in accordance with the manufacturer's recommendations. We designed the study to ensure that all infants receive their initial HIV test result within 4 weeks of their blood being drawn, and use a commercial laboratory if DBS test results are not available from the routine central laboratory within 3 weeks.
The Alere q Detect Qualitative Test (Alere Technologies, Jena, Germany) is performed using 25 mL of whole blood, collected through heel-stick and performed according to the manufacturer's instructions. The test has been optimized for HIV groups prevalent in SSA, including HIV-1 groups M, N, and O and HIV-2. All test reagents are contained within a single-use cartridge, and no pretreatment (or other handling) of the sample is required. PCR amplification and real-time fluorescence detection are performed using a small, bench-top analyzer. [bib_ref] Point-of-care tests poised to alter course of HIV treatment, Laursen [/bib_ref] The turnaround time is approximately 1 hour, and the test is performed in the study clinic.
As noted above, positive DBS tests are repeated and positive Alere q tests are confirmed through DBS testing. Although we expect discrepancies between the Alere q and Roche EID tests to be rare, [bib_ref] Accurate early infant HIV diagnosis in primary health clinics using a point-of-care..., Jani [/bib_ref] we have developed a 2-step algorithm to resolve any such discrepancies. For discordant results, we repeat both the Alere q and Roche EID tests using fresh samples, and obtain 3 additional 75-100 mL DBS aliquots. If the discrepancy persists, we perform Abbott m2000 HIV Qualitative PCR testing (using remaining DBS aliquots), which serves as a "tie-breaker" between the Alere q and Roche tests. This algorithm ensures that the HIV status of every randomized infant is determined definitively.
HIV-1 plasma viral load testing is also performed using the Roche Cobas Ampliprep/Taqman platform (Roche Molecular Diagnostics). Plasma is isolated from whole-blood samples, and then processed and tested in a central laboratory according to the manufacturers' recommendations.
## Primary endpoint
The primary endpoint is the proportion of HIV-infected children who initiate ART and remain in care with suppressed plasma viral load (#200 copies/mL) 12 months after their diagnostic blood draw. This composite outcome is designed to measure the overall effectiveness of the intervention.
## Sample size
We postulate that the availability of same-day POC EID will improve ART initiation from 40% to 80% and improve viral suppression at 12 months from 80% to 90%. We also postulate equally in both study arms that 72% of children will be alive and in care by 12 months, with 22% LTFU and 6% deceased by 12 months. [bib_ref] Clinical outcomes and CD4 cell response in children receiving antiretroviral therapy at..., Bolton-Moore [/bib_ref] Applying these estimates as conditional probabilities, we anticipate a success rate of 0.23 in the SOC arm and 0.52 in the intervention arm. Assuming 12-month success probabilities of 23% in the SOC arm and 52% in the intervention arm, a sample size of 86 HIV-infected children (43 per arm) provides 80% power to detect a difference in success probability between arms using a x 2 test with 2-sided alpha of 0.05 (SAS power procedure, SAS version 9.3, Cary, NC). Given an expected mother-tochild transmission rate of 3%, 1 we aim to enroll 2867 HIVexposed infants to identify 86 infected infants for follow-up. Children who are confirmed uninfected are exited from the study and not followed for the 12 months.
# Statistical analysis
Descriptive statistics will be used to delineate the overall study population and assess for clinically meaningful imbalances between the intervention and SOC arms. The primary endpoint (proportion who are in care and virally suppressed at 12 months) will be evaluated among HIVinfected participants and compared between the intervention arm and SOC arm using an estimated risk ratio and corresponding 95% Wald confidence interval. HIV infection status at the time of randomization will be ascertained for all infants such that the subset of HIV-infected infants can be fully identified.
A sensitivity analysis will be conducted using a logbinomial model [bib_ref] Categorical Data Analysis, Agresti [/bib_ref] to estimate a risk ratio adjusted for prespecified possible confounders that, if unbalanced in the 2 arms, could potentially bias the estimate of effect. Prespecified possible confounders include infant birth weight, CD4 + percentage at ART commencement, presence of clinical AIDS, and maternal vital status.
In a secondary analysis among all enrolled infants (anticipated n = 2867), the HIV transmission proportion in this study population will be estimated with a corresponding Clopper and Pearson 95% CI. All statistical analyses will be conducted using a 2-sided 0.05 significance level.
# Ethical considerations
Participation in the trial is voluntary, and caregivers of study participants provide written, informed consent before study enrollment. All procedures are conducted in accordance with local SOC. The study protocol has been reviewed and approved by the University of Zambia's Biomedical Research Ethics Committee, the Zambian Ministry of Health, and the University of North Carolina at Chapel Hill's Institutional Review Board. All staff who have contact with participants receive training on the protection of human research participants before conducting any study activities and every 2 years thereafter. Key staff also complete Good Clinical Practice or Good Clinical Laboratory Practice training.
Risks to participants in this study are minimal and do not differ significantly from the risks inherent in the local SOC for HIV-infected infants, with the exception of the additional risks to participant confidentiality and data privacy. Study participation may also cause stigmatization or discrimination against participants and their families. The confidentiality of all study records is safeguarded to the extent legally possible, and all laboratory specimens, reports, study data, and administrative forms are identified by a coded number only. Databases are also secured with password-protected access systems.
Participants in the intervention arm may benefit from POC HIV diagnosis and more timely initiation of ART. The knowledge generated from this study regarding the potential benefits of the Alere q platform and the feasibility of its implementation will enable the Zambian Ministry of Health to make evidence-based decisions regarding POC assays for EID.
# Discussion
Unlike HIV testing in adults, diagnosis in infants requires specialized laboratory equipment, traditionally operated by personnel available only in central laboratories. As a result, many infants either have no access to HIV diagnosis at all or must wait for weeks or months to receive their test result. In Zambia-as in many countries in the region-treatment is never started for more than half of newborns who are tested and found to be infected. 1 A POC assay for early infant HIV diagnosis may offer a distinct advantage over the current standard clinical practice of off-site PCR in SSA. The POC platform, Alere q, was designed specifically for near-patient use in resource-poor settings, as the instrument is portable, battery-operated, and ruggedized. The Alere q test has also been optimized for HIV groups prevalent in SSA.
This trial will be the first of its kind in Zambia, and our composite primary outcome is designed to measure the incremental effectiveness of the intervention when compared with the usual standard, thus providing key estimates for policy makers who must decide whether or not to make substantial countrywide investments in POC technologies in the future. The composite outcome measure accounts for the immediate effect of the POC technology on ART initiation and any potential effect on remaining in care with virologic suppression. We have chosen to link these outcomes into a composite because we believe that they are inextricably linked in real practice. The POC technology may very well increase the proportion of infants who start ART immediately, but this early "win" could be diluted by higher rates of LTFU.
[table] TABLE 1: Schedule of Events in the Control Arm [/table]
[table] TABLE 2: Schedule of Events in the Intervention Arm [/table]
|
Polylactide-co-glycolide nanoparticles for controlled delivery of anticancer agents
# Introduction
Recent developments in nanotechnology have enabled new research strategies to flourish in the field of drug delivery. There has been considerable interest in developing nanoparticles as effective drug carriers. [bib_ref] Drug-loaded nanoparticles -preparation methods and drug targeting issues, Allémann [/bib_ref] For pharmaceutical purposes, nanoparticles are defined as solid colloidal particles ranging in size from 10 nm to 400 nm. They consist of macromolecular materials in which the active agent (drug or biologically active material) is dissolved, entrapped, and/or encapsulated, or to which the active agent is adsorbed or attached. [bib_ref] Targeted nanomedicines: Effective treatment modalities for cancer, AIDS and brain disorders, Muthu [/bib_ref] Drug delivery has been improved with the translation of several nanoscale drug delivery systems into the clinic, although the full potential of these systems is only now starting to be explored. Nanoscale drug delivery systems have shown the ability to encapsulate a variety of therapeutic agents, such as small molecules (hydrophilic and/or hydrophobic), peptide protein-based drugs, and nucleic acids. By encapsulating these molecules inside a nanocarrier, the solubility and stability of drugs can be improved, providing an opportunity to re-evaluate the therapeutic potential of drugs previously discounted because of poor pharmacokinetics. [bib_ref] Drug delivery and targeting, Langer [/bib_ref] The diversity of drug delivery systems allows nanoparticles to be developed with a diverse array of shapes, sizes, and components, enabling them to be tailored for specific applications. However, the primary consideration when designing any drug delivery system is to control the drug concentration in the therapeutic window, while reducing side effects and improving patient compliance. This allows effective treatment cycles to be maintained, and at the same time reduces damage to healthy cells and minimizes the recovery period. [bib_ref] Nanoparticle therapeutics: An emerging treatment modality for cancer, Davis [/bib_ref] [bib_ref] Tumour-targeted nanomedicines: Principles and practice, Lammers [/bib_ref] [bib_ref] Nanocarriers as an emerging platform for cancer therapy, Peer [/bib_ref] Langer and Folkman were the first to demonstrate the controlled release of macromolecules using polymers, which enabled the development of antiangiogenic drug delivery systems for cancer therapy and opened up new areas for the delivery of macromolecules. [bib_ref] Polymers for the sustained release of proteins and other macromolecules, Langer [/bib_ref] Polymeric nanoparticles provide significant flexibility in design because different polymers from synthetic or natural sources can be used. Polymeric nanoparticles may represent the most effective nanocarriers for targeted drug delivery. Some common polymers used for nanoparticle formation include polylactide-co-glycolide (PLGA), polylactic acid, dextran, and chitosan. Biodegradable polymers are typically degraded into oligomers and individual monomers, which are metabolized and removed from the body via normal pathways. [bib_ref] Polymer conjugates as anticancer nanomedicines, Duncan [/bib_ref] [bib_ref] Polymeric nanomedicines for image-guided drug delivery and tumortargeted combination therapy, Lammers [/bib_ref] [bib_ref] Polymers in drug delivery -state of the art and future trends, Grund [/bib_ref] Degradation and drug release kinetics can be precisely controlled by the physicochemical properties of the polymer, such as molecular weight, polydispersity index, hydrophobicity, and crystallinity. In general, drugs can be released in a controlled manner following Fickian kinetics due to drug diffusion through the polymeric matrix, or be triggered in response to environmental stimuli or released in the course of chemical degradation. The nanoparticle surface may be sterically stabilized by grafting, conjugating, or adsorbing hydrophilic polymers, such as polyethylene glycol (PEG), to its surface, which can reduce hepatic uptake and improve the circulation half-life of the nanoparticles. [bib_ref] Stealth'corona-core nanoparticles surface modified by polyethylene glycol (PEG): Influences of the corona..., Gref [/bib_ref] [bib_ref] Visualization of in vitro protein-rejecting properties of PEGylated Stealth ® polycyanoacrylate nanoparticles, Peracchia [/bib_ref] PLGA is one of the most commonly used degradable polymers and is discussed extensively here.
## Properties of plga
PLGA has generated tremendous interest due to its excellent biocompatibility, biodegradability, and mechanical strength [fig_ref] Figure 1: Chemical structure of polylactide-co-glycolide [/fig_ref]. [bib_ref] Sterilization, toxicity, biocompatibility and clinical applications of polylactic acid/polyglycolic acid copolymers, Athanasiou [/bib_ref] PLGA can be synthesized by a polycondensation reaction or via ring-opening polymerization of cyclic diesters. Ring-opening polymerization is currently the preferred method for the synthesis for PLGA and polylactic acid due to shorter reaction times and higher monomer conversion rates. [bib_ref] Poly(lactide-co-glycolide)-based micro and nanoparticles for the controlled drug delivery of vitamins, Stevanovic [/bib_ref] The discovery and synthetic work on low molecular weight oligomeric forms of lactide and/or glycolide polymers were first carried out several decades ago. [bib_ref] The manufacturing techniques of various drug loaded biodegradable poly (lactide-co-glycolide)(PLGA) devices, Jain [/bib_ref] A method to synthesize high molecular weight forms of such polymers was first reported by Lowe.During the late 1960s and early 1970s, a number of groups published their pioneering work on the utility of these polymers in the manufacture of sutures and fibers. [bib_ref] The manufacturing techniques of various drug loaded biodegradable poly (lactide-co-glycolide)(PLGA) devices, Jain [/bib_ref] Since then, these polymers have been used in a broad range of pharmaceutical and biomedical applications. Various polymeric devices, such as microspheres, microcapsules, nanoparticles, pellets, implants, and films have been fabricated using these polymers. They are also easy to formulate into various delivery systems for carrying a variety of active agents, such as vaccines, peptides, proteins, and micromolecules, some of which are now approved by the Food and Drug Administration for use in drug delivery. [bib_ref] Poly(lactide-co-glycolide)-based micro and nanoparticles for the controlled drug delivery of vitamins, Stevanovic [/bib_ref] [bib_ref] The manufacturing techniques of various drug loaded biodegradable poly (lactide-co-glycolide)(PLGA) devices, Jain [/bib_ref] [bib_ref] Biodegradable polymers for use in surgerypolyglycolic/poly (actic acid) homo-and copolymers: 1, Gilding [/bib_ref] [bib_ref] PLGA nanoparticles in drug delivery: The state of the art, Bala [/bib_ref] In order to improve the formulation of controlled drug delivery devices, an understanding of the physical, chemical, and biological properties of polymers is helpful. The polylactic acid polymer can exist in an optically active stereoregular form (L-polylactic acid) and in an optically inactive racemic form (D, L-polylactic acid). L-polylactic acid is semicrystalline in nature due to the high regularity of its polymer chain structure, while D, L-polylactic acid is an amorphous polymer because of irregularities in its polymer chain structure. Polyglycolide is highly crystalline because it lacks the methyl side groups of polylactic acid. [bib_ref] The manufacturing techniques of various drug loaded biodegradable poly (lactide-co-glycolide)(PLGA) devices, Jain [/bib_ref] PLGA copolymers prepared from L-polylactic acid and polyglycolide are crystalline, while those from D, L-polylactic acid and polyglycolide are amorphous in nature. It has been found that PLGAs containing less than 70% glycolide are amorphous. The degree of crystallinity and the melting point of the polymers are directly related to their molecular weight. The mechanical strength, swelling behavior, capacity to undergo hydrolysis, and, subsequently, the biodegradation rate, are directly influenced by the crystallinity of the PLGA polymer, which depends on the type and molar ratio of the individual monomer components (lactide and glycolide) in the copolymer chain. [bib_ref] The manufacturing techniques of various drug loaded biodegradable poly (lactide-co-glycolide)(PLGA) devices, Jain [/bib_ref] Lactic acid is more hydrophobic than glycolic acid and, therefore, lactide-rich PLGA copolymers are less hydrophilic, absorb less water, and subsequently, degrade more slowly. [bib_ref] Degradation of poly (D, L-lactic acid) microspheres: Effect of molecular weight, Park [/bib_ref] The glass transition temperatures of PLGA copolymers are above the physiological temperature of 37°C, and hence they are normally glassy in nature, with sufficient strength to be formulated as a drug delivery system. The glass transition temperatureof a PLGA copolymer decreases with reducing molecular weight and lactide content. 15
## Mechanism of biodegradation of plga
A fundamental understanding of the in vivo phenomenon of PLGA biodegradation is important because this determines the rate and mechanism of the release of therapeutic agents. The therapeutic agent is either dispersed throughout the polymeric matrix or encapsulated in the hydrophobic nanoparticle core. The release of a therapeutic agent from nanoparticles has usually been shown to be biphasic, initially by diffusion through the polymer matrix and later bydiffusion of the therapeutic agent and degradation of the polymer matrix itself. [bib_ref] Polymer degradation and in vitro release of a model protein from poly..., Panyam [/bib_ref] [bib_ref] Therapeutic applications of colloidal drug carriers, Barratt [/bib_ref] PLGA copolymers are degraded in the body by hydrolytic cleavage of the ester linkage to lactic and glycolic acid . These monomers are easily metabolized in the body via the Krebs cycle and eliminated as carbon dioxide and water. [bib_ref] The manufacturing techniques of various drug loaded biodegradable poly (lactide-co-glycolide)(PLGA) devices, Jain [/bib_ref] [bib_ref] Rapid endolysosomal escape of poly(DL-lactide-co-glycolide) nanoparticles: Implications for drug and gene delivery, Panyam [/bib_ref] The polymer degradation process both in vitro and in vivo is affected by several factors, including the method of preparation, the presence of low molecular weight compounds (monomers, oligomers, catalysts), size, shape and morphology, the intrinsic properties of the polymer (molecular weight, chemical structure, hydrophobicity, crystallinity, and glass transition temperature), [bib_ref] The manufacturing techniques of various drug loaded biodegradable poly (lactide-co-glycolide)(PLGA) devices, Jain [/bib_ref] physicochemical parameters (pH, temperature, and ionic strength of the environment), site of implantation, and mechanism of hydrolysis. In general, the degradation time will be shorter for low molecular weight, more hydrophilic, and more amorphous polymers, and for copolymers with a higher glycolide content.
Although PLGA copolymers can undergo surface erosion in some conditions, bulk erosion is the main degradation pathway [fig_ref] Figure 3: as shown by Konan et al who A B Degradation mechanisms of... [/fig_ref]. This occurs by random scission of ester bonds in the polymer backbone occurring homogeneously throughout the device. [bib_ref] PLGA nanoparticles in drug delivery: The state of the art, Bala [/bib_ref] A three-phase mechanism for PLGA biodegradation has been proposed. Initially, a significant decrease in the molecular weight of the polymer is observed, with no appreciable weight loss and no soluble monomer products formed after random chain scission. This phase is followed by a decrease in molecular weight, with rapid loss of mass and formation of soluble monomeric and oligomeric products. Finally, soluble monomer products are formed from soluble oligomeric fragments, resulting in complete polymer degradation. 15
## Methods for preparation of plga nanoparticles
Several methods for polymeric nanoparticle production have been developed by researchers. These methods generally include two main steps. The first step is to prepare an emulsified system, and this is common to all the methods used. The nanoparticles are formed during the second step, which varies according to the method used. In general, the principle of this second step gives its name to the method. Some methods do not require the preparation of an emulsion prior to obtaining the nanoparticles, and are based on spontaneous precipitation of a polymer or through self assembly of macromolecules. [bib_ref] Methods for the preparation and manufacture of polymeric nanoparticles, Vauthier [/bib_ref] The commonly used methods for preparation of PLGA nanoparticles are briefly described.
## Emulsification solvent evaporation
Emulsification solvent evaporation is one of the most frequently used methods. The polymer and the drug are first dissolved in a water-immiscible volatile solvent, such as dichloromethane or chloroform, which is then emulsified in an aqueous solution containing a stabilizer. The emulsification is brought about by subsequent exposure to a high-energy shearing source, such as an ultrasonic device or homogenizer. The organic phase is evaporated under reduced pressure or vacuum, resulting in a fine aqueous dispersion of nanoparticles. The nanoparticles are collected by ultracentrifugation and washed with distilled water to remove stabilizer residues or any free drug, and then lyophilized for storage. [bib_ref] Drug-loaded nanoparticles -preparation methods and drug targeting issues, Allémann [/bib_ref] [bib_ref] Design and production of nanoparticles formulated from nano-emulsion templates -a review, Anton [/bib_ref] [bib_ref] Synthesis and characterization of PLGA nanoparticles, Astete [/bib_ref] The emulsion evaporation method can be used for the preparation of particles with sizes varying from a few
## 880
Dinarvand et al nanometers to micrometers by controlling the stirring rates and conditions, and shows high efficiency forthe incorporation of lipophilic drugs. [bib_ref] A novel controlled release formulation for the anticancer drug paclitaxel (Taxol ®..., Mu [/bib_ref] [bib_ref] Susceptibility of nanoparticleencapsulated paclitaxel to P-glycoprotein-mediated drug efflux, Chavanpatil [/bib_ref] [bib_ref] Nanoparticles of biodegradable polymers for new-concept chemotherapy, Feng [/bib_ref] To entrap hydrophilic drugs, the double-emulsion technique is employed, which involves the addition of aqueous drug solution to organic polymer solution under vigorous stirring to form a water-in-oil emulsion. This water-in-oil emulsion is added into a second aqueous phase containing a stabilizer with stirring to form the water-in-oilin-water emulsion. The emulsion is then subjected to solvent removal by evaporation. [bib_ref] Influences of process parameters on nanoparticle preparation performed by a double emulsion..., Lamprecht [/bib_ref] [bib_ref] Mucoadhesive nanoparticulate systems for peptide drug delivery, Takeuchi [/bib_ref] Emulsification solvent diffusion
This method is also known as emulsification and solvent displacement. The solvent used to prepare the emulsion needs to be partly soluble in water. [bib_ref] Nanoparticles for drug delivery: Review of the formulation and process difficulties illustrated..., Moinard-Chécot [/bib_ref] The polymeric solution is added to an aqueous solution, containing stabilizer, under vigorous stirring. [bib_ref] PLGA nanoparticles in drug delivery: The state of the art, Bala [/bib_ref] Once the oil-in-water emulsion is obtained, it is diluted with a large quantity of pure water. As a result of this dilution, additional organic solvent from the organic phase contained in the dispersed droplets can diffuse out of the droplets, leading to precipitation of the polymer. Suitable solvents include benzyl alcohol, propylene carbonate, ethyl acetate, isopropyl acetate, methyl acetate, methyl ethyl ketone, benzyl alcohol, butyl lactate, and isovaleric acid. This method has been used for PLGA nanoparticle preparation in many studies. [bib_ref] Preparation of poly (D, L-lactide-co-glycolide) nanoparticles by modified spontaneous emulsification solvent diffusion..., Murakami [/bib_ref] [bib_ref] Preparation of PLGA nanoparticles containing estrogen by emulsification-diffusion method*1, Kwon [/bib_ref] [bib_ref] In vitro and in vivo anti-tumor activities of nanoparticles based on doxorubicin-PLGA..., Yoo [/bib_ref] [bib_ref] Preparation and characterization of sterile sub-200 nm meso-tetra (4-hydroxylphenyl) porphyrin-loaded nanoparticles for..., Konan [/bib_ref] It should be noted that the solvent evaporation process is similar to this method, in the sense that the solvent must first diffuse out into the external aqueous dispersion medium before it can be removed from the system by evaporation. [bib_ref] Methods for the preparation and manufacture of polymeric nanoparticles, Vauthier [/bib_ref] Emulsification reverse salting-out
The emulsification reverse salting-out technique involves the addition of polymer and drug solution to a water-miscible solvent, such as acetone, and to an aqueous solution containing the salting-out agent, such as magnesium chloride, calcium chloride, and a colloidal stabilizer, such as polyvinyl pyrrolidone, under vigorous mechanical stirring. When this oil-in-water emulsion is diluted with a sufficient volume of water, it induces the formation of nanoparticles by enhancing the diffusion of acetone into the aqueous phase. The dilution produces a sudden decrease in the salt concentration in the continuous phase of the emulsion, inducing the polymer solvent to migrate out of the emulsion droplets. The remaining solvent and salting-out agent are eliminated by cross-flow filtration. [bib_ref] Drug-loaded nanoparticles -preparation methods and drug targeting issues, Allémann [/bib_ref] [bib_ref] Aqueous nanodispersions prepared by a salting-out process, Ibrahim [/bib_ref] [bib_ref] Preparation and characterization of sterile and freeze-dried sub-200 nm nanoparticles, Konan [/bib_ref] Although the emulsification-diffusion method is a modification of the salting-out procedure, it has the advantage of avoiding the use of salts and thus eliminates the need for intensive purification steps. 18
## Nanoprecipitation
The nanoprecipitation method is a one-step procedure, also known as the solvent displacement method. [bib_ref] Nanocapsule formation by interfacial polymer deposition following solvent displacement, Fessi [/bib_ref] It is usually employed to incorporate lipophilic drugs into the carriers based on the interfacial deposition of a polymer. [bib_ref] PLGA nanoparticles prepared by nanoprecipitation: Drug loading and release studies of a..., Govender [/bib_ref] [bib_ref] Doxorubicin-loaded PLGA nanoparticles by nanoprecipitation: Preparation, characterization and in vitro evaluation, Betancourt [/bib_ref] Nanoprecipitation is performed using systems containing three basic ingredients, ie, the polymer, the polymer solvent, and the nonsolvent of the polymer. The solvent should be organic, miscible in water, and easily removed by evaporation. For this reason, acetone is the most frequently used solvent with this method. [bib_ref] Preparation of pseudolatex by nanoprecipitation: Influence of the solvent nature on intrinsic..., Thioune [/bib_ref] Sometimes it exists as a binary blend of solvents, as acetone with a small amount of water, or as a blend of ethanol and acetone. Polymer, drug, and lipophilic surfactant (eg, phospholipids) are dissolved in a semipolar watermiscible solvent, such as acetone or ethanol. The solution is then poured or injected into an aqueous solution containing stabilizer under magnetic stirring. Nanoparticles are formed immediately by rapid solvent diffusion. The solvent is then removed from the suspension under reduced pressure. [bib_ref] The manufacturing techniques of various drug loaded biodegradable poly (lactide-co-glycolide)(PLGA) devices, Jain [/bib_ref] [bib_ref] PLGA nanoparticles in drug delivery: The state of the art, Bala [/bib_ref]
## Nanoparticle characterization techniques
Characterization of nanoparticles is essential for a thorough understanding of their properties before developing them further for pharmaceutical application. Nanoparticle size is critical, not only in determining the release profile and degradation behavior, but also in determining the efficacy of the therapeutic agent in terms of tissue penetration and cellular uptake. [bib_ref] Nanoparticles for drug delivery: The need for precision in reporting particle size..., Gaumet [/bib_ref] The molecular weight of the polymer influences the nanoparticle size, encapsulation efficiency, and degradation rate of the polymer, hence affecting the release rate of the therapeutic agent, [bib_ref] Influence of particle size and dissolution conditions on the degradation properties of..., Dunne [/bib_ref] , and 213 kDa) by the emulsion-diffusionevaporation method, and showed that entrapment efficiency did not follow a regular pattern. First, it decreased from 51.3% to 34.5% as molecular weight was increased from about 14.5 kDa to 85 kDa, and, thereafter, a significant increase (P , 0.05) to 67.8% was observed as molecular weight was increased to 213 kDa. Entrapment efficiency of particles using the oil-in-water method depends mainly on the drug partition coefficient in the internal and external phases. It is possible that an increase in viscosity on increasing the molecular weight might have decreased the diffusion rate of the solvent into the external aqueous phase. The polymer precipitated slowly because of a slow rate of solvent removal, giving the drug molecules more time to come into the aqueous phase, resulting in low entrapment efficiency. However, an increase inentrapment efficiency was observed for particles with molecular weights of 137 kDa and 213 kDa, which could be due to strong hydrophobic interaction between the molecular chains of the polymer and the drug. [bib_ref] Estradiol loaded PLGA nanoparticles for oral administration: Effect of polymer molecular weight..., Mittal [/bib_ref] An increase in molecular weight from 14.5 kDa to 213 kDa was associated with a significant decrease in the rate of release of estradiol. Molecular weight is indicative of polymer chain length, and the higher the molecular weight, the longer the chain length. Furthermore, chain length reflects the hydrophilicity or lipophilicity of the polymer. An increase in chain length increases the lipophilicity and decreases the degradation rate of the polymer. Therefore, by varying the molecular weight, the degradation rate of the polymer and release kinetics of the drug can be controlled. [bib_ref] The manufacturing techniques of various drug loaded biodegradable poly (lactide-co-glycolide)(PLGA) devices, Jain [/bib_ref] [bib_ref] Estradiol loaded PLGA nanoparticles for oral administration: Effect of polymer molecular weight..., Mittal [/bib_ref] The physical state of both the drug and the polymer need to be determined because this will have an influence on the in vitro and in vivo drug release characteristics. The zeta potential can influence particle stability and mucoadhesion, as well as intracellular trafficking of nanoparticles as a function of pH. Hydrophobicity determines the distribution of nanoparticles in the body after administration. Hydrophilic particles tend to remain in the blood for a longer time. [bib_ref] PLGA nanoparticles in drug delivery: The state of the art, Bala [/bib_ref] [bib_ref] Synthesis and characterization of PLGA nanoparticles, Astete [/bib_ref] [bib_ref] Biodegradable polymeric nanoparticles as drug delivery devices, Soppimath [/bib_ref] There are many sensitive techniques for characterizing nanoparticles, depending upon the parameter being investigated. These techniques are summarized in .
## Plga nanoparticles for drug delivery to tumors
Cancer is a worldwide public health problem, and tens of millions of people presently suffer from this deadly disease. [bib_ref] Novel drug delivery systems for sustained and targeted delivery of anti-cancer drugs:..., Utreja [/bib_ref] Cancer research involves intensive scientific efforts to identify the causes of cancer and to develop specific strategies for its prevention, diagnosis, treatment, and cure. Despite considerable progress in its early diagnosis, but progress concerning its treatment has been less so. In current anticancer therapy, drugs are administered via the intravenous and/or oral route using conventional formulations, including injections, tablets, and capsules. Controlled and targeted delivery of an anticancer agent at the site of action is necessary to maximize the killing effect during the tumor growth phase and to avoid drug exposure to healthy adjacent cells, thereby reducing drug toxicity. It is also desirable to maintain a steady rate of infusion of the drug into the tumor to maximize exposure to dividing cells, resulting in tumor regression. [bib_ref] Poly (ethylene oxide)-modified poly (-amino ester) nanoparticles as a pH-sensitive system for..., Shenoy [/bib_ref] Development of novel systems for delivery of anticancer drugs is a recent topic of research. Abraxane ® , an albumin-based formulation of paclitaxel from Abraxis Oncology, Nanoxel ® , a nanoliposome containing paclitaxel from DaburPharma, and Doxil ® , a nanoliposome containing doxorubicin from Ortho Biotech, are now well-known commercial products. 82,84-89 Techniques for polylactide-co-glycolide nanoparticle characterization Particle size, size distribution, morphology Dynamic light scattering or photon correlation spectroscopy [bib_ref] PLGA nanoparticles prepared by nanoprecipitation: Drug loading and release studies of a..., Govender [/bib_ref] [bib_ref] Stabilizer-free poly (lactide-co-glycolide) nanoparticles for multimodal biomedical probes, Cheng [/bib_ref] [bib_ref] Paclitaxel-loaded PLGA nanoparticles: Preparation, physicochemical characterization and in vitro anti-tumoral activity, Fonseca [/bib_ref] Scanning electron microscopy [bib_ref] A novel controlled release formulation for the anticancer drug paclitaxel (Taxol ®..., Mu [/bib_ref] [bib_ref] Folate-receptor-targeted delivery of docetaxel nanoparticles prepared by PLGA-PEG-folate conjugate, Esmaeili [/bib_ref] [bib_ref] Zinc (II) phthalocyanine loaded PLGA nanoparticles for photodynamic therapy use, Ricci-Junior [/bib_ref] Transmission electron microscopy [bib_ref] Preparation and in vitro anticancer activity of wheat germ agglutinin (WGA)-conjugated PLGA..., Mo [/bib_ref] [bib_ref] Fluorescence and electron microscopy probes for cellular and tissue uptake of poly..., Panyam [/bib_ref] [bib_ref] Tumor necrosis factor alpha blocking peptide loaded PEG-PLGA nanoparticles: Preparation and in..., Yang [/bib_ref] Atomic force microscopy [bib_ref] Poly (D, L-lactide-co-glycolide)/montmorillonite nanoparticles for oral delivery of anticancer drugs, Dong [/bib_ref] [bib_ref] Preparation and characterization of cationic PLGA nanospheres as DNA carriers, Kumar [/bib_ref] [bib_ref] The effect of type of organic phase solvents on the particle size..., Song [/bib_ref]
## Molecular weight
Size exclusion chromatography 58-60 Surface chemistry analysis X-ray photoelectron spectroscopy [bib_ref] A novel controlled release formulation for the anticancer drug paclitaxel (Taxol ®..., Mu [/bib_ref] [bib_ref] Target-specific cellular uptake of PLGA nanoparticles coated with poly (l-lysine)-poly (ethylene glycol)-folate..., Kim [/bib_ref] [bib_ref] Nanoparticles of biodegradable polymers for clinical administration of paclitaxel, Si-Shen [/bib_ref] Fourier transform infrared spectroscopy [bib_ref] Tumor necrosis factor alpha blocking peptide loaded PEG-PLGA nanoparticles: Preparation and in..., Yang [/bib_ref] [bib_ref] Design of surface-modified poly (D, L-lactideco-glycolide) nanoparticles for targeted drug delivery to..., Choi [/bib_ref] [bib_ref] PEGylated PLGA nanoparticles as protein carriers: Synthesis, preparation and biodistribution in rats, Li [/bib_ref] Nuclear magnetic resonance spectroscopy [bib_ref] PEGylated PLGA nanoparticles as protein carriers: Synthesis, preparation and biodistribution in rats, Li [/bib_ref] Nanoparticle-based drug delivery systems have many advantages for anticancer drug delivery, including an ability to pass through the smallest capillary vessels, because of their very small volume, and being able to avoid rapid clearance by phagocytes, so that their presence in the blood stream is greatly prolonged. [bib_ref] Biodegradable nanoparticles for oral delivery of peptides: Is there a role for..., Jung [/bib_ref] Nanoparticles can also penetrate cells and gaps in tissue to arrive at target organs, including the liver, spleen, lung, spinal cord, and lymph. They may have controlled-release properties due to their biodegradability, pH, ions, and/or temperature sensitivity. All these properties can improve the utility of anticancer drugs and reduce their toxic side effects.
PLGA nanoparticles linked to targeting ligands are used to target malignant tumors with high affinity. PLGA nanoparticles also have large surface areas and functional groups for conjugating to multiple diagnostic (eg, optical, radioisotopic, or magnetic) agents. [bib_ref] Current advances in research and clinical applications of PLGA-based nanotechnology, Lü [/bib_ref] Nanoparticle carriers have high stability in biological fluids, and are more able to avoid enzymatic metabolism than other colloidal carriers, such as liposomes or lipid vesicles. [bib_ref] Surface functionalized biocompatible magnetic nanospheres for cancer hyperthermia, Liu [/bib_ref] Most anticancer drugs that have been investigated in PLGA nanoparticle preparations are discussed below. They are also summarized in [fig_ref] Table 2: Summary of polylactide-co-glycolide nanoparticles [/fig_ref].
## Paclitaxel
Many anticancer drugs can be used clinically to treat various cancers, but have limited efficacy due to poor cell penetration. For example, paclitaxel, a mitotic inhibitor, has had limited clinical application because of its low therapeutic index, and its low solubility in water and many other pharmaceutical solvents acceptable for intravascular administration. Incorporation of paclitaxel into PLGA nanoparticles strongly enhances its antitumoral efficacy compared with the free drug (Taxol ® ), with this effect being more relevant after more prolonged incubation with cells. Based on these results, it can be concluded that the formulations developed so far may be considered promising systems for in vivo paclitaxel delivery. [bib_ref] Paclitaxel-loaded PLGA nanoparticles: Preparation, physicochemical characterization and in vitro anti-tumoral activity, Fonseca [/bib_ref] In an animal model, Van Vlerken et al encapsulated paclitaxel and the apoptotic signaling molecule, C6-ceramide, into a PLGA/poly (β-amino ester)-blended polymer. When this nanoparticle formulation was administered intravenously to MCF7 and MCF7 TR -tumor-bearing mice, higher concentrations of paclitaxel were found in the blood due to a longer retention time and enhanced tumoral accumulation compared with the free drug. In addition, the PLGA/poly (β-amino ester)-blended nanoparticles were effective in enhancing the residence time of both drugs at the tumor site by reducing systemic clearance. [bib_ref] Biodistribution and pharmacokinetic analysis of paclitaxel and ceramide administered in multifunctional polymer-blend..., Van Vlerken [/bib_ref] In another study, Feng et al developed paclitaxel-loaded nanoparticles to achieve better therapeutic effects with minimum side effects. In this investigation, phospholipids, cholesterol, and vitamins were used to replace traditional chemical emulsifiers to achieve high encapsulation efficacy and the desired drug release rate. [bib_ref] Polymeric nanospheres fabricated with natural emulsifiers for clinical administration of an anticancer..., Feng [/bib_ref] The methodology and experimental parameters used for nanoparticle preparation can impact the physicochemical properties of the resulting formulations. Danhier et al have reported significantly higher encapsulation efficacies for paclitaxel loaded into PLGA nanoparticles using the nanoprecipitation method (70%) compared with the emulsion/ solvent evaporation technique (40%). [bib_ref] Paclitaxel-loaded PEGylated PLGA-based nanoparticles: In vitro and in vivo evaluation, Danhier [/bib_ref] Elsewhere, it was shown that an increase in the oil-to-water phase ratio 49 and the polymer concentration of the organic phase 95 could enhance the entrapment efficacy of paclitaxel within polyester-based nanoparticles produced by the nanoprecipitation technique. Alternatively, the organic solvent can be dialyzed against water to obtain polyester-based nanoparticles incorporating paclitaxel. [bib_ref] Hydrophobically modified glycol chitosan nanoparticles as carriers for paclitaxel, Kim [/bib_ref] [bib_ref] Self-assembled biodegradable nanoparticles developed by direct dialysis for the delivery of paclitaxel, Xie [/bib_ref] [bib_ref] Taxol-loaded block copolymer nanospheres composed of methoxy poly (ethylene glycol) and poly..., Kim [/bib_ref] [bib_ref] In vitro and in vivo investigation on PLA-TPGS nanoparticles for controlled and..., Zhang [/bib_ref] Mu and Feng used α-tocopheryl polyethylene glycol l000 succinate (vitamin E TPGS) as well as a matrix material with other biodegradable polymers for the fabrication of a nanoparticle formulation of paclitaxel. They concluded that vitamin E TPGS was advantageous either as an emulsifier or as matrix material blended with PLGA for the manufacture of nanoparticles enabling controlled release of paclitaxel. [bib_ref] A novel controlled release formulation for the anticancer drug paclitaxel (Taxol ®..., Mu [/bib_ref] Surfactants and stabilizers are used to increase the physical stability of nanoparticles. Reports of the positive surface charge of a quaternary ammonium salt, didodecyl dimethyl ammonium bromide (DMAB), provided the incentive to aid the delivery of paclitaxel, because it was expected to ensure better interaction with the negatively charged cell membrane. This could result in increased retention time at the cell surface, thus increasing particle uptake. [bib_ref] Arterial uptake of biodegradable nanoparticles for intravascular local drug delivery: Results with..., Song [/bib_ref] In another study, the safety and utility of DMAB for stabilizing PLGA nanoparticles was studied. The preliminary data from this study provide proof-of-concept of improved efficacy and safety of oral paclitaxel chemotherapy. [bib_ref] PLGA nanoparticles stabilized with cationic surfactant: Safety studies and application in oral..., Bhardwaj [/bib_ref] Docetaxel PLGA nanoparticles containing docetaxel with the desired size and drug-loading characteristics suitable for intravenous administration can be prepared without using Tween ® 80. Esmaeili et al showed that the cellular cytotoxicity of the nanoparticles was higher than for the free drug. Docetaxel-loaded nanoparticles reached good plasma levels in vivo in comparison with a conventional formulation of docetaxel (Taxotere ® ). [bib_ref] Cellular cytotoxicity and in-vivo biodistribution of docetaxel poly (lactide-co-glycolide) nanoparticles, Esmaeili [/bib_ref] The nanoprecipitation process has been applied for the formation of docetaxel-loaded nanoparticles. [bib_ref] Formulation of functionalized PLGA-PEG nanoparticles for in vivo targeted drug delivery, Cheng [/bib_ref] [bib_ref] PLA/PLGA nanoparticles for sustained release of docetaxel, Musumeci [/bib_ref] Cheng et al showed that limiting drugloading to 1% (w/w) minimized particle aggregation and yielded docetaxel-loaded PLGA nanoparticles with narrower size distributions. 66
## Cisplatin
Mattheolabakis et al prepared cisplatin nanoparticles with an average size of 150-160 nm and an approximately 2% w/w cisplatin content using a modified emulsification and solvent evaporation method. The cisplatin-loaded PLGAmonomethoxy (m)PEG nanoparticles appeared to be effective in delaying tumor growth in HT29 tumor-bearing mice with severe combined immune deficiency. The group of mice treated with cisplatin-loaded nanoparticles had a higher survival rate compared with the free cisplatin group. [bib_ref] In vivo investigation of tolerance and antitumor activity of cisplatin-loaded PLGA-mPEG nanoparticles, Mattheolabakis [/bib_ref] Cisplatin-loaded PLGA-mPEG nanoparticles also resulted in prolonged cisplatin residence time in the systemic circulation when used in mice with prostate cancer. 104
## Doxorubicin
Betancourt et al formulated nanoparticles by nanoprecipitation of acid-ended PLGA to control the release of doxorubicin in a pH-dependent manner and deliver high loads of active drug to an MDA-MB-231 breast cancer cell line. The pH-dependent release behavior could be a result of accelerated degradation of the polymer and decreasing ionic interaction between the drug and the polymer at an acidic pH. [bib_ref] Doxorubicin-loaded PLGA nanoparticles by nanoprecipitation: Preparation, characterization and in vitro evaluation, Betancourt [/bib_ref] Another approach to improve the efficacy and selectivity of cancer treatment is the application of hyperthermia in combination with traditional cancer therapeutics, such as radiation therapy and chemotherapy. [bib_ref] Hyperthermia in combined treatment of cancer, Wust [/bib_ref] [bib_ref] Heating the patient: A promising approach?, Van Der Zee [/bib_ref] Hyperthermia makes some cancer cells more sensitive to radiation and can also enhance the effect of certain anticancer drugs, [bib_ref] Heating the patient: A promising approach?, Van Der Zee [/bib_ref] thus allowing the use of decreased chemotherapy doses.
Indocyanine green is an optical tracer that can generate heat by absorbing near-infrared light. The significance of the Manchanda et al study is the synthesis of multifunctional PLGA nanoparticles and the incorporation of drugs with different physical properties (indocyanine green being amphiphilic and doxorubicin being hydrophobic). These indocyanine green-doxorubicin nanoparticles have potential applications as drug delivery systems for combined chemotherapy and localized hyperthermia. [bib_ref] Preparation and characterization of a polymeric (PLGA) nanoparticulate drug delivery system with..., Manchanda [/bib_ref] Curcumin Curcumin has been used in traditional medicine for many centuries in India and China. [bib_ref] Curcumin: Getting back to the roots, Shishodia [/bib_ref] It is chemically diferuloylmethane, a yellow polyphenol extracted from the rhizomes of turmeric (Curcuma longa). The only factor that limits the use of free curcumin for cancer therapy is its poor solubility in water, which in turn limits its systemic bioavailability when administered orally. Mukerjee and Vishwanatha formulated curcumin-loaded PLGA nanoparticles, and suggested that a nanoparticle-based formulation of curcumin has high potential as adjuvant therapy in prostate cancer. [bib_ref] Formulation, characterization and evaluation of curcumin-loaded PLGA nanospheres for cancer therapy, Mukerjee [/bib_ref] Another study demonstrated that curcumin encapsulation in PLGA nanoparticles employing a nanoprecipitation approach in the presence of polyvinyl alcohol and poly L-lysine stabilizers not only produced a very stable nanoformulation but also enhanced cellular drug uptake and retention, as well as sustained release of curcumin. The optimized nanoparticle formulation has shown a greater inhibitory effect on the growth of metastatic cancer (A2780CP and MDA-MB-231) cells than free curcumin. 110
## Multidrug resistance
A major barrier to successful cancer treatment is multidrug resistance. It has been reported that most patients, even those who are initially responsive, acquire a multidrug-resistant phenotype, and some patients show multidrug-resistance even at their first treatment. In metastatic breast cancer, the development of a multidrug-resistant phenotype is primarily responsible for insensitivity to a new drug. [bib_ref] How cancer cells evade chemotherapy: Sixteenth Richard and Hinda Rosenthal Foundation award..., Gottesman [/bib_ref] Therefore, resistance to chemotherapeutic agents is the major challenge in the treatment of breast cancer. It has been shown that nanoparticles can reduce the multidrug-resistance characterizing many anticancer drugs via a drug internalization mechanism mediated by P-glycoprotein, thereby reducing its efflux from cells. [bib_ref] Characterization of nanoparticle uptake by endothelial cells, Davda [/bib_ref] [bib_ref] Nanoparticles in cancer therapy and diagnosis, Brigger [/bib_ref] Yan et al showed that the cytotoxicity of a PLGA-poloxamer188 nanoparticle blend containing docetaxel against MCF-7 TAX30 cells was higher than that of the free drug, indicating that poloxamer188 could enhance the ability of PLGA nanoparticles to overcome multidrug resistance. A docetaxel-loaded PLGA-poloxamer188 nanoparticle formulation has been developed to overcome multidrug resistance in a docetaxel-resistant human breast cancer cell line (MCF-7 TAX30). [bib_ref] The effect of poloxamer 188 on nanoparticle morphology, size, cancer cell uptake,..., Yan [/bib_ref] Multidrug resistance may be treated using a combination of entrapped cytotoxic drugs and chemosensitizers. To optimize the effectiveness of this approach, Song et al prepared PLGA nanoparticle formulations capable of delivering vincristine (a cytotoxic drug) and verapamil (a chemosensitizer), or a combination of these agents by combining the oil-in-water emulsion solvent evaporation and salting-out methods. PLGA nanoparticles showed moderate multidrug-resistance reversal in MCF-7/ADR cells resistant to vincristine. Coencapsulation of an anticancer drug and chemosensitizer may cause lower drug toxicity and fewer drug-drug interactions. Therefore, PLGA nanoparticles simultaneously loaded with an anticancer drug and a chemosensitizer may potentially be a very promising formulation for treatment of drug-resistant cancers in vivo. [bib_ref] Reversion of multidrug resistance by co-encapsulation of vincristine and verapamil in PLGA..., Song [/bib_ref]
## Photodynamic therapy
Photodynamic therapy combines a photosensitizing agent with a specific type of light to kill cancer cells. This technique is minimally invasive, with great potential in both malignant and premalignant conditions. Photodynamic therapy requires the presence of a photosensitizing agent, oxygen, and light of a specific wavelength matching the absorption characteristics of the photosensitizer agent. Administration of the photosensitizer is followed by illumination of the tumor with visible light in a wavelength range matching the absorption spectrum of the photosensitizer agent.
Hypericin, a natural photosensitizer extracted from Hypericum perforatum, is a potential tool for the detection and treatment of ovarian and other cancers. Due to its hydrophobicity, systemic administration of hypericin is problematic. A photodynamic approach has been suggested by Zeisser-Labou et al to improve the diagnosis and treatment of ovarian cancer. [bib_ref] Hypericin-loaded nanoparticles for the photodynamic treatment of ovarian cancer, Zeisser-Labouèbe [/bib_ref] This group used polymeric nanoparticles of polylactic acid or PLGA as a drug delivery system, and compared the in vitro photoactivity of the nanoparticles and that of the free drug using the NuTu-19 ovarian cancer cell model derived from Fischer 344 rats. Their studies showed that hypericin-loaded nanoparticles exhibited higher photoactivity than did the free drug, and increasing light dose and cell incubation time enhanced their activity.
Ricci-Junior and Marchetti have prepared, characterized, and assayed the photocytotoxicity of PLGA nanoparticles Dovepress Dovepress 885 PLGA nanoparticles for cancer therapy containing zinc (II) phthalocyanine for use in photodynamic therapy. These nanoparticles maintained their photophysical behavior after the encapsulation process. 51
## Surface modification of plga nanoparticles
Polymeric nanoparticles have been characterized by their morphology and polymer composition. The drug molecule is either conjugated to the surface of the nanoparticles or entrapped and protected inside the core. The unique sizes of these nanoparticles are amenable to surface functionalization or modification to achieve any desired characteristics. This has been achieved by various methods to increase drug retention time in blood, to reduce nonspecific distribution, and to target tissues or specific cell surface antigens with targeting ligands, such as peptides, aptamers, antibodies, and small molecules. [bib_ref] Factors affecting the clearance and biodistribution of polymeric nanoparticles, Pridgen [/bib_ref] Opson in proteins present in the bloodstream quickly bind to conventional nonstealth nanoparticles, allowing macrophages of the molecular phagocytic system to recognize and remove these drug delivery devices before they can exert their therapeutic effects. [bib_ref] Opsonization, biodistribution, and pharmacokinetics of polymeric nanoparticles, Owens [/bib_ref] Opsonization of injected particles by antibodies in the circulation, attachment of opsonized particles to macrophages, and subsequent internalization by phagocytosis are important steps in the clearance of particles by the molecular phagocytic system. Different materials are used for the preparation of nanoparticles, leading to distinct surface properties. Surface modification of nanoparticles is important for escaping the body's natural defense systems when transporting drugs to the bloodstream. [bib_ref] Surface modification of nanoparticles to oppose uptake by the mononuclear phagocyte system, Storm [/bib_ref] Unless nanoparticles are modeled to escape recognition by the molecular phagocytic system, they will be eliminated from the body. A long circulation time increases the probability that the nanoparticles will reach their target. Nanostructures smaller than 100 nm with a hydrophilic surface have the greatest ability to evade the molecular phagocytic system. [bib_ref] Nanoparticles of biodegradable polymers for new-concept chemotherapy, Feng [/bib_ref] To increase circulation time, the particles can be coated with molecules that provide a hydrophilic protective layer, such as PEG, polyvinylpyrrolidone, human serum albumin, poloxamers, polysorbate 80, polysorbate 20, vitamin E TPGS, polysaccharides (eg, dextran) and different types of copolymers. [bib_ref] Which polymers can make nanoparticulate drug carriers long-circulating?, Torchilin [/bib_ref] Here we discuss three of the more important materials used for surface modification of PLGA nanoparticles.
## Polyethylene glycol
Adsorption or grafting of PEG to the surface of nanoparticles is the preferred method for shielding nanoparticles from the molecular phagocytic system. [bib_ref] Paclitaxel-loaded PEGylated PLGA-based nanoparticles: In vitro and in vivo evaluation, Danhier [/bib_ref] [bib_ref] In vivo investigation of tolerance and antitumor activity of cisplatin-loaded PLGA-mPEG nanoparticles, Mattheolabakis [/bib_ref] [bib_ref] Anticancer activity of cisplatin-loaded PLGA-mPEG nanoparticles on LNCaP prostate cancer cells, Gryparis [/bib_ref] Addition of PEG or PEG-containing copolymers to the nanoparticle surface results in an increase of half-life in the blood circulation by several orders of magnitude. This method creates a protective hydrophilic layer around the nanoparticles that is able to repel the adsorption of proteins via steric repulsion forces, thereby blocking and delaying the first step in the opsonization process. [bib_ref] Opsonization, biodistribution, and pharmacokinetics of polymeric nanoparticles, Owens [/bib_ref] PEG is a hydrophilic nonionic polymer that has been shown to have excellent biocompatibility. PEG coating on the surface of polymers reduces interaction between the nanoparticles and digestive enzymes, and increases uptake of the encapsulated drug in the bloodstream and lymphatic tissue. [bib_ref] The role of PEG on the stability in digestive fluids and in..., Tobio [/bib_ref] PLGA nanoparticles have been surface-modified with PEG in an attempt to protect them from the reticuloendothelial system. [bib_ref] Paclitaxel-loaded PEGylated PLGA-based nanoparticles: In vitro and in vivo evaluation, Danhier [/bib_ref] [bib_ref] In vivo investigation of tolerance and antitumor activity of cisplatin-loaded PLGA-mPEG nanoparticles, Mattheolabakis [/bib_ref] [bib_ref] Anticancer activity of cisplatin-loaded PLGA-mPEG nanoparticles on LNCaP prostate cancer cells, Gryparis [/bib_ref] Avgoustakis et al considered the feasibility of using long circulating PLGA-mPEG nanoparticles as carriers for passive targeting of cisplatin to tumors. They showed that intravenous administration of PLGA-mPEG nanoparticles loaded with cisplatin in mice resulted in a prolonged residence time of cisplatin in the systemic circulation. [bib_ref] PLGA-mPEG nanoparticles of cisplatin: In vitro nanoparticle degradation, in vitro drug release..., Avgoustakis [/bib_ref] vitamin e TPGS TPGS is a widely used form of vitamin E that has been used as a solubilizer, an emulsifier, and a vehicle in drug delivery formulations. TPGS has been used as an emulsifier for producing nanoparticles containing hydrophobic drugs and for improving encapsulation efficiency, drug loading, and the release profile of nanoparticles. [bib_ref] Preparation of PLGA nanoparticles using TPGS in the spontaneous emulsification solvent diffusion..., Esmaeili [/bib_ref] Surface modification by TPGS increases adhesion of the nanoparticles to the tumor cell surface. [bib_ref] Surface modification of nanoparticles to oppose uptake by the mononuclear phagocyte system, Storm [/bib_ref] [bib_ref] Long circulating microparticulate drug carriers, Stolnik [/bib_ref] TPGS has also been blended with PLGA for the preparation of nanoparticles. [bib_ref] A novel controlled release formulation for the anticancer drug paclitaxel (Taxol ®..., Mu [/bib_ref] TPGS increases adhesion of PLGA nanoparticles to cells and the hemodynamic properties of the nanoparticles.
Ma et al synthesized PLGA-TPGS copolymers as nanoparticle carriers for small molecular weight anticancer drugs, using docetaxel as a model drug. In vitro cellular uptake of these nanoparticles was investigated by confocal laser scanning microscopy, which demonstrated that fluorescent PLGA-TPGS nanoparticles could be internalized by HeLa, a human cervical carcinoma cell line. Random PLGA-TPGS copolymers could act as a novel and potential biocompatible polymeric matrix material applicable to nanoparticle-based drug delivery systems for cancer chemotherapy. [bib_ref] Nanoparticles of poly(lactideco-glycolide)-d-a-tocopheryl polyethylene glycol 1000 succinate random copolymer for cancer treatment, Ma [/bib_ref] Pluronic ® Pluronic block copolymers consist of ethylene oxide and propylene oxide blocks arranged in a basic A-B-A structure, ie, ethylene oxide-propylene oxide-ethylene oxide. The incorporation of drugs into the core of the micelles formed by Pluronic results in increased solubility, metabolic stability, and a longer circulation time for the drug. [bib_ref] Pluronic ® block copolymers as novel polymer therapeutics for drug and gene..., Kabanov [/bib_ref] Kabanov et al demonstrated that Pluronics P85, F68, and L61 reduced the respiration rate of both drug-sensitive cells and the molecular phagocytic system, resulted in hypersensitivity to chemotherapy. Cells often achieve their drug resistance via the efflux of drugs through energy-dependent transporters, so reduction in respiration impairs this mechanism. [bib_ref] Pluronic ® block copolymers as novel polymer therapeutics for drug and gene..., Kabanov [/bib_ref] Shah et al formulated paclitaxel-PLGA nanoparticles surface-modified with Pluronic P85 for inhibition of P-glycoprotein. This study demonstrated the feasibility of targeting nanoparticles to specific cancer tumor cells, cell internalization, and successful overcoming of the P-glycoprotein-mediated paclitaxel efflux mechanism, and demonstrated a significant increase in uptake of Pluronic P85-coated nanoparticles. 124
## Plga nanoparticle targeting strategies
In the process of killing cancer cells, chemotherapeutic agents also damage healthy tissues, leading to systemic toxicity and adverse side effects. Thus, an important step in improving treatment regimes is achieving more effective targeting of anticancer drugs to tumor tissues. Therefore, the need to develop novel cancer therapies and drug delivery strategies specifically targeted to tumor cells continues to be a major focus for scientists. Nanoparticle systems offer major improvements in therapeutics through site specificity, an ability to evade multidrug resistance, and efficient delivery of anticancer agents. [bib_ref] Polymeric nanoparticles for cancer therapy, Parveen [/bib_ref] Ideally, drug delivery systems should improve the stability, absorption, and therapeutic action of the drug within the target tissue and permit controlled release of the drug. Targeting cancer cells using nanoparticles loaded with anticancer agents is a promising strategy that could overcome these challenges. Drug targeting can be achieved by taking advantage of the distinct pathophysiological features of tumor tissue (ie, passive targeting) or by actively targeting the drug carrier using target-specific ligands (ie, active targeting). [bib_ref] Active targeting schemes for nanoparticle systems in cancer therapeutics, Byrne [/bib_ref] [bib_ref] Nanoparticle and targeted systems for cancer therapy, Brannon-Peppas [/bib_ref] Passive targeting Passive targeting exploits the anatomical differences between normal tissue and tumor tissue to deliver drugs to the desired site. The vasculature in tumor tissues is very different from that in normal tissues. It is more heterogeneous in distribution, larger in size, has higher vascular density, and is more permeable and leaky, with gap sizes of 100 nm to 2 µm depending on the tumor type, unlike the tight endothelium of normal blood vessels. In addition, extensive production of vascular mediators, including bradykinins, nitric oxide, vascular endothelial growth factor, and prostaglandins, facilitates extravasation. This, coupled with impaired lymphatic drainage of macromolecules in solid tumors, allows enhanced accumulation and retention of high molecular weight drugs in solid tumors. This is popularly known as the "enhanced permeation and retention effect", that allows extravasation of circulating polymeric nanoparticles within the tumor interstitium and also increases concentrations of the chemotherapeutic agent within the tumor tissue . [bib_ref] Polymeric nanoparticles for cancer therapy, Parveen [/bib_ref] [bib_ref] Active targeting schemes for nanoparticle systems in cancer therapeutics, Byrne [/bib_ref] [bib_ref] Nanoparticle and targeted systems for cancer therapy, Brannon-Peppas [/bib_ref] It should be mentioned that nanoparticles within the size range of 60-400 nm are effective for this type of targeting. [bib_ref] Photo-thermal tumor ablation in mice using near infrared-absorbing nanoparticles, O'neal [/bib_ref] [bib_ref] Size-dependent extravasation and interstitial localization of polyethyleneglycol liposomes in solid tumor-bearing mice, Ishida [/bib_ref] These are three in vivo studies based on the permeation and retention effect. Bhardwaj et al formulated PLGApaclitaxel nanoparticles stabilized with DMAB and administered orally to female Sprague Dawley rats. [bib_ref] PLGA nanoparticles stabilized with cationic surfactant: Safety studies and application in oral..., Bhardwaj [/bib_ref] Paclitaxel nanoparticles administered orally were as effective as paclitaxel given intravenously with Cremophor ® EL at a 50% reduced dose, and was significantly better than oral paclitaxel in the Cremophor EL group. This in vivo proofof-concept study provides encouraging evidence of the value of nanoparticulate anticancer formulations, and highlights the anticancer efficacy of paclitaxel when incorporated into nanoparticles. The nanoparticle formulation was as effective as the drug used alone, but at half the dose. The average tumor weight measured after paclitaxel orally administered in Cremophor was roughly three-fold higher than after the nanoparticulate formulation. [bib_ref] PLGA nanoparticles stabilized with cationic surfactant: Safety studies and application in oral..., Bhardwaj [/bib_ref] Danhier et al formulated Cremophor EL-free paclitaxelloaded PEGylated PLGA-based nanoparticles by a nanoprecipitation method. In vivo tumor growth inhibition by the paclitaxel-loaded nanoparticles was then investigated in transplantable liver tumor-bearing mice. Paclitaxel was shown to reach the tumor site via the enhanced permeation and retention effect and maintain an effective therapeutic concentration. [bib_ref] Paclitaxel-loaded PEGylated PLGA-based nanoparticles: In vitro and in vivo evaluation, Danhier [/bib_ref] In another study, Mattheolabakis et al investigated tolerance of BALB/c mice to different doses of blank and cisplatinloaded PLGA-mPEG nanoparticles, and the in vivo anticancer activity of cisplatin-loaded PLGA-mPEG nanoparticles in mice with severe combined immune deficiency bearing HT29 colon adenocarcinoma. By in vivo antitumor activity assay, the PLGA-mPEG-cisplatin nanoparticles appeared to reduce tumor growth in mice with severe combined immune deficiency and HT29 xenografts, and these mice had higher survival rates than those treated with free cisplatin. [bib_ref] In vivo investigation of tolerance and antitumor activity of cisplatin-loaded PLGA-mPEG nanoparticles, Mattheolabakis [/bib_ref] International Journal of Nanomedicine 2011:6 submit your manuscript | www.dovepress.com
## Dovepress
## Dovepress
## 887
PLGA nanoparticles for cancer therapy Active targeting Active targeting involves the use of peripherally attached targeting moieties for enhanced delivery of nanoparticle systems . Active targeting to the tumor can be achieved by molecular recognition of cancer cells either via ligand-receptor or antigen-antibody interactions, or by targeting through aptamers, that allow preferential accumulation of the drug in tumor tissue, within individual cancer cells or specific molecules in cancer cells. Long circulation times will allow for effective transport of the nanoparticles to the tumor site through the enhanced permeation and retention effect, and the targeting molecule can increase endocytosis of the nanoparticles. The success of drug targeting depends on the selection of the targeting agent, which should be abundant, have high affinity and specificity of binding to cell surface receptors, and should be well-suited to chemical modification by conjugation.
The receptors and their surface-bound antigens may be expressed only in diseased cells or may exhibit a differentially higher expression in diseased cells as compared with normal cells. Thus, the tumor endothelium provides many targets for cancer therapy . [bib_ref] Nanoparticle and targeted systems for cancer therapy, Brannon-Peppas [/bib_ref] [bib_ref] Antibody targeting of longcirculating lipidic nanoparticles does not increase tumor localization but..., Kirpotin [/bib_ref] Here we discuss the three main categories of targeting moieties used for PLGA nanoparticles. These targeted nanoparticles are also shown in [fig_ref] Table 3: Summary of targeted polylactide-co-glycolide nanoparticles [/fig_ref].
## Ligand-receptor interaction
The lectin-carbohydrate interaction is highly specific, and can be exploited for the development of nanoparticles containing carbohydrate moieties that are directed to certain lectins, or vice versa. Several lectins have been found to possess anticancer properties, and are used as therapeutic agents, preferentially binding to cancer cell membranes or their receptors, causing cytotoxicity, apoptosis, and inhibition of tumor growth. 125 isopropyl myristate or wheat germ agglutinin, or paclitaxelloaded nanoparticles with only isopropyl myristate or wheat germ agglutinin. Thus, their studies showed that these nanoparticles exhibited a stronger cytotoxic effect because of more efficient cellular uptake via wheat germ agglutinin receptormediated endocytosis and isopropyl myristate-facilitated release of paclitaxel from the nanoparticles. [bib_ref] Preparation and in vitro anticancer activity of wheat germ agglutinin (WGA)-conjugated PLGA..., Mo [/bib_ref] There is broad interest in the development of nanoparticles carrying carbohydrates, such as sialic acids, on their surface. Macromolecular compounds containing these carbohydrates show an antirecognition effect, exert an antiviral effect, and are also able to be recognized by the cell surface of some cancer cell types. [bib_ref] PLGA nanoparticles surface decorated with the sialic acid, N-acetylneuraminic acid, Bondioli [/bib_ref] Bondioli et al used two different approaches to obtain polymeric PLGA nanoparticles surface-decorated with sialic acid N-acetylneuraminic acid (Neu5Ac). The first strategy used is based on derivatization of PLGA with the thioderivative of Neu5Ac as the starting material for the preparation of nanoparticles, and the second strategy is based on the synthesis of compounds potentially able to insert their lipophilic moiety into nonderivatized PLGA nanoparticles during their preparation and display their hydrophilic moiety (Neu5Ac) on their surface. 131
## Dovepress
## Dovepress
## 889
PLGA nanoparticles for cancer therapy Transferrin receptors are overexpressed by 2-10-fold in most tumor cells, and thus, transferrin and/or transferrin antibodies may be used for targeting drugs to tumor cells. [bib_ref] Polymeric nanoparticles for cancer therapy, Parveen [/bib_ref] Zheng et al synthesized transferrin-conjugated lipid-coated PLGA nanoparticles carrying the aromatase inhibitor, 7-(4-amino) phenylthio-1, 4-androstadiene-3, 17-dione. The aromatase inhibition activity of the nanoparticles was evaluated in a SKBR-3 breast cancer cell line. The IC 50 value of the nanoparticles incorporating transferrin ranged from 0.77 to 1.21 nM, and the IC 50 value of the nanoparticles ranged from 1.90 to 3.41 nM (n = 3). These results suggest that the aromatase inhibition activity of the transferrin nanoparticles was enhanced relative to that of the nontargeted nanoparticles, which was attributable to transferrin receptor-mediated uptake. [bib_ref] Transferrin-conjugated lipidcoated PLGA nanoparticles for targeted delivery of aromatase inhibitor 7α-APTADD to..., Zheng [/bib_ref] The α v β 3 integrin is an endothelial cell receptor for extracellular matrix proteins harboring the arginine-glycine-aspartic acid (RGD) sequence. [bib_ref] Active targeting schemes for nanoparticle systems in cancer therapeutics, Byrne [/bib_ref] Wang et al conjugated doxorubicin to PLGA, and the nanoparticle surfaces were then linked with PEG and the RGD peptides to achieve both passive and active targeting functions. The nanoparticle targeting ability was enhanced via strong affinity to various integrin-expressing cancer cells, and much less affinity to low integrin-expressing cancer cells. [bib_ref] Design of a multifunctional PLGA nanoparticulate drug delivery system: Evaluation of its..., Wang [/bib_ref] Danhier et al showed that PEGylated PLGAbased nanoparticles grafted with the RGD peptide or an RGD peptidomimetic, targeted the tumor endothelium and would enhance the antitumor efficacy of paclitaxel. They observed that RGD-grafted nanoparticles were more associated with human umbilical vein endothelial cells in vitro by binding to α v β 3 integrin than were nontargeted nanoparticles, and they also demonstrated the targeting of RGD and RGD peptidomimetic-grafted nanoparticles to tumor vessels, as well as effective retardation of transplantable liver tumor growth and prolonged survival times in mice treated by paclitaxel-loaded RGD nanoparticles when compared with nontargeted nanoparticles. [bib_ref] Targeting of tumor endothelium by RGD-grafted PLGA-nanoparticles loaded with paclitaxel, Danhier [/bib_ref] The folate receptor is a highly specific tumor marker that is overexpressed in many human cancers. In addition, this receptor is absent in most normal tissues, so is frequently exploited for drug-targeting purposes. With the proper design, folate-drug conjugates display high-affinity properties which enable them to bind rapidly to the folate receptor and become internalized via an endocytic process. [bib_ref] Polymeric nanoparticles for cancer therapy, Parveen [/bib_ref] Yoo and Park prepared biodegradable polymeric micelles self-assembled from a diblock copolymer of PLGA and PEG to achieve delivery of doxorubicin to the folate receptor. They showed higher cytotoxicity to the folate-conjugated mixed micelles than for free doxorubicin, suggesting that folate receptor-mediated endocytosis of micelles plays an important role in transporting an increased amount of doxorubicin into cancer cells. In addition, in vivo animal experiments using a nude mouse xenograft model demonstrated that, when systemically administered, tumor volume was significantly reduced. [bib_ref] Folate receptor targeted biodegradable polymeric doxorubicin micelles, Yoo [/bib_ref] In another study, Zhang et al synthesized doxorubicinloaded folate-decorated PLGA-vitamin E TPGS nanoparticles for targeted chemotherapy to folate receptor-rich tumors. Their research showed that the TPGS-folate uptake was 1.5 and 1.7 times higher in MCF-7 and C6 cells, respectively, as compared with nanoparticles with no TPGS-folate component after 30 minutes of incubation. [bib_ref] Folate-decorated poly(lactideco-glycolide)-vitamin E TPGS nanoparticles for targeted drug delivery, Zhang [/bib_ref] The authors of the present study produced docetaxel nanoparticles by using PLGA-PEG-folate conjugates for folate receptor-targeted anticancer delivery. The folate receptor-targeted nanoparticles showed greater intracellular uptake in folate receptor-positive cancer cells (SKOV3) in comparison with the nontargeted nanoparticles, indicating that a folate receptor-mediated endocytosis mechanism could have a role in the cellular uptake of nanoparticles. [bib_ref] Folate-receptor-targeted delivery of docetaxel nanoparticles prepared by PLGA-PEG-folate conjugate, Esmaeili [/bib_ref] In another study, the same research group studied the use of PLGA-PEG-folate nanoparticles containing SN-38 (7-ethyl-10-hydroxy camptothecin). These folate receptortargeted nanoparticles showed greater cytotoxicity against cancer cells than nontargeted SN-38 nanoparticles.
Biotin is an essential micronutrient for normal cellular functions (eg, biosynthesis of fatty acids, gluconeogenesis), growth, and development. Humans and other mammals cannot synthesize biotin and thus must obtain it from exogenous sources via intestinal absorption. Rapidly dividing cells, such as cancer cells, have a voracious appetite for certain vitamins, including biotin, vitamin B12, and folate. Biotin levels have been found to be significantly higher in some cancer cells compared with normal tissue. [bib_ref] Targeting cancer cells with biotin-dendrimer conjugates, Yang [/bib_ref] Interestingly, tumor cell lines, including ovarian and colorectal, which overexpress receptors involved in folate or vitamin B12 uptake also show overexpression of biotin receptors. [bib_ref] Vitamin-mediated targeting as a potential mechanism to increase drug uptake by tumours, Mctavish [/bib_ref] Accordingly, several research groups have tested different biotinylated chemotherapeutic agents for cancer cell-specific drug delivery. [bib_ref] Biotinylated poly (amido) amine (pamam) dendrimers as carriers for drug delivery to..., Yellepeddi [/bib_ref] [bib_ref] Biotin-pyrene conjugates with poly (ethylene glycol) spacers are convenient fluorescent probes for..., Marek [/bib_ref] [bib_ref] A novel biotinylated degradable polymer for cell-interactive applications, Cannizzaro [/bib_ref] [bib_ref] Enhancing the anticancer efficacy of camptothecin using biotinylated poly (ethyleneglycol) conjugates in..., Minko [/bib_ref] Patil et al investigated simultaneous targeted delivery of paclitaxel with tariquidar (a third-generation P-glycoprotein modulator) using PLGA nanoparticles to overcome tumor drug resistance. Nanoparticles were surface-functionalized with biotin for active tumor targeting. [bib_ref] Nanoparticle-mediated simultaneous and targeted delivery of paclitaxel and tariquidar overcomes tumor drug..., Patil [/bib_ref] In the same study, this group used small interfering RNA instead of tariquidar to silence expression of the P-glycoprotein efflux transporter in Balb/c mice bearing JC (mammary adenocarcinoma) tumors. Growth in tumor volume and survival were monitored on a regular basis. Biotin-conjugated nanoparticles demonstrated more tumor growth inhibition than the nontargeted nanoparticles. [bib_ref] The use of nanoparticle-mediated targeted gene silencing and drug delivery to overcome..., Patil [/bib_ref]
## Antigen-antibody interaction
In recent years, overexpressing cancer-specific antigens have become an important tool in developing different delivery technologies for cancer treatment. The advent of monoclonal antibody technology in the 1970s and the development of genetically engineered derivatives in the 1980s, along with technological advances in the bulk production of monoclonal antibodies, have led to a number of clinical studies evaluating the efficacy of cancer-specific monoclonal antibodies in targeted drug delivery systems. The overexpression of receptors and antigens in human cancers also helps efficient uptake via receptor-mediated endocytosis. Moreover, due to their transformed nature, tumor cells overexpress many new proteins in comparison with normal cells, and these markers may be exploited for active drug targeting. [bib_ref] Polymeric nanoparticles for cancer therapy, Parveen [/bib_ref] Kou et al developed paclitaxel-loaded PLGA nanoparticles coated with cationic SM5-1 single-chain antibody. Nonradioactive cell proliferation assay demonstrated that targeted nanoparticles had significantly better in vitro cytotoxicity against Ch-hep-3 human hepatocellular cancer cells than nontargeted paclitaxel-loaded PLGAnanoparticles. [bib_ref] Preparation and characterization of paclitaxel-loaded PLGA nanoparticles coated with cationic SM5-1 single-chain..., Kou [/bib_ref] Kocbek et al prepared PLGA immunonanoparticles for targeting invasive epithelial breast tumor cells. The monoclonal antibody was prepared against soluble membrane proteins of MCF-7 human invasive ductal breast carcinoma and was attached to the nanoparticle surface either covalently or noncovalently. These nanoparticles were more likely to attach to the targeted cells than the noncoated nanoparticles. In coculture of MCF-10A neoT and Caco-2 cells, immunonanoparticles entered only MCF-10A neoT cells, while noncoated nanoparticles were taken up by both cell types, indicating specific targeting by the immunonanoparticles. [bib_ref] Targeting cancer cells using PLGA nanoparticles surface modified with monoclonal antibody, Kocbek [/bib_ref] The human epidermal receptor (HER) family of receptor tyrosine kinases offers two highly upregulated targets on tumor cell surfaces. These receptor tyrosine kinases, epidermal growth factor receptor and human epidermal receptor-2 (HER2), are known to mediate a cell signaling pathway for growth and proliferation in response to binding of the growth factor ligand. [bib_ref] Polymeric nanoparticles for cancer therapy, Parveen [/bib_ref] Achraya et al prepared and characterized rapamycin-loaded PLGA nanoparticles so that their surfaces were modified with antibodies to the epidermal growth factor receptor, highly expressed on breast cancer cells, and improved the cytotoxicity of nanoparticles in a malignant MCF7 breast cancer cell line. IC 50 doses, determined by MTT assay, showed superior antiproliferative activity of antibodyconjugated rapamycin-loaded nanoparticles compared with the unconjugated nanoparticles and native rapamycin due to higher cellular uptake by malignant breast cancer cells. The molecular basis of apoptosis, studied by Western blotting, revealed the involvement of a cytoplasmic protein in activating the programmed cell death pathway. Thus, it was concluded that epidermal growth factor receptor antibodyconjugated rapamycin-loaded nanoparticles provide efficient and targeted delivery of anticancer drugs. 147
## Dovepress
## Dovepress
## 891
PLGA nanoparticles for cancer therapy Trastuzumab, approved by the Food and Drug Administration, is amonoclonal antibody designed to antagonize HER2 function, and has been used as a targeting moiety in various nanoparticle systems. [bib_ref] Active targeting schemes for nanoparticle systems in cancer therapeutics, Byrne [/bib_ref] Sun et al investigated a nanoparticle formulation containing PLGA-montmorillonite and trastuzumab for targeted breast cancer chemotherapy, with paclitaxel as a model anticancer drug. The in vitro drug release from this nanoparticle formulation showed a biphasic drug-release pattern, with a moderate initial burst release followed by a sustained release profile. [bib_ref] Multifunctional poly (D, L-lactideco-glycolide)/montmorillonite (PLGA/MMT) nanoparticles decorated by trastuzumab for targeted chemotherapy..., Sun [/bib_ref] Gao et al developed Pseudomonas exotoxin (PE38KDEL)-loaded PLGA nanoparticles conjugated with Fab' fragments (rhuMAbHER2) from a humanized anti-HER2 monoclonal antibody (PE-NP-HER). Compared with nontargeted nanoparticles lacking anti-HER2 Fab', PE-NP-HER specifically bound to and was sequentially internalized into HER2-overexpressing breast cancer cells, which resulted in significant cytotoxicity in vitro. Notably, PE-NP-HER was of low immunogenicity in the development of anti-PE38KDEL-neutralizing antibodies, and was less susceptible to inactivation by anti-PE38KDEL antibodies compared with PE-HER. [bib_ref] PE38 KDEL-loaded anti-HER2 nanoparticles inhibit breast tumor progression with reduced toxicity and..., Gao [/bib_ref]
## Aptamer-mediated targeting
Aptamers are DNA or RNA oligonucleotides (short DNA or RNA oligonucleotide ligands) capable of binding to target antigens with high affinity and specificity, and are analogous to antibodies. [bib_ref] Polymeric nanoparticles for cancer therapy, Parveen [/bib_ref] The group led by [bib_ref] Targeted delivery of cisplatin to prostate cancer cells by aptamer functionalized Pt(IV)..., Dhar [/bib_ref] investigated the targeting potential of aptamers specific to the extracellular domain of the prostate-specific membrane antigen (PSMA). Farokhzad et al developed 180 nm docetaxel-encapsulated nanoparticles using a PLGAblock-PEG copolymer surface-functionalized with the A10 20-fluoropyrimidine RNA aptamers that recognize PSMA. Their results showed significant enhancement of the cellular toxicity of the functionalized nanoparticles in vitro as compared with the nontargeted nanoparticles lacking the PSMA aptamer. In addition, after a single intratumoral injection of docetaxel-aptamer bioconjugate nanoparticles, complete tumor reduction was observed in five of seven LNCaP xenograft nude mice, and 100% survival was observed as compared with mice treated with docetaxel nanoparticles alone. They also studied the biodistribution of the nanoparticles in aLNCaP (PSMA+) xenograft mouse model of prostate cancer. Surface functionalization of the nanoparticles with A10 PSMA aptamer significantly enhanced (3.77-fold increase at 24 hours in nanoparticle aptamer of injected dose per gram of tissue) the delivery of nanoparticles to tumors versus equivalent nanoparticles lacking the A10 PSMA aptamer. [bib_ref] Targeted nanoparticleaptamer bioconjugates for cancer chemotherapy in vivo, Farokhzad [/bib_ref] In another study, Dhar et al also used PLGA-b-PEG nanoparticles with PMSA-targeting aptamers on the surface as a vehicle for the platinum (IV) compound, c, t, c-[Pt(NH3) (2)(O2CCH2CH2CH2CH2CH3)(2)Cl-2], as a strategy to deliver cisplatin to prostate cancer cells. A comparison between the cytotoxic activities of Pt(IV)-encapsulated PLGA-b-PEG nanoparticles with the PSMA aptamer on the surface (Pt-NP-Apt), cisplatin, and the nontargeted Pt(IV)encapsulated nanoparticles against human PSMA overexpressing LNCaP and PSMA(−) PC3 cancer cells revealed significant differences. The effectiveness of PSMA-targeted Pt-NP-Apt nanoparticles against the PSMA (+) LNCaP cells was significantly greater than that of free cisplatin. 151
# Conclusion
Current polymeric nanocarrier technologies have demonstrated remarkable advantages for cancer therapy when compared with conventional drugs. Among the polymers utilized to date, PLGA is very promising for the preparation of novel anticancer drug delivery systems due to its desirable characteristics, including good biodegradability and biocompatibility. PLGA nanoparticles can achieve tumortargeted drug delivery via passive targeting based on the enhanced permeation and retention effect, or active targeting by an appropriate ligand, which improves antitumor efficacy and reduces toxicity on healthy tissues. For further advancement, it will be necessary to focus more research attention on the pharmacokinetics, biodistribution, and safety of these novel drug delivery systems.
Development of multifunctional PLGA nanoparticles containing specific ligands for active targeting, and multiple drugs for synergistic anticancer effects and overcoming drug resistance, will provide a versatile and straightforward approach to improving chemotherapy. Combined therapeutic and diagnostic (theranostic) nanoparticles will be another potential future direction. In this case, theranostic PLGA nanoparticles will be detected in tumors by using a diagnostic agent. These nanoparticles can modulate their responses based on changes in the environment. When imaging shows maximum nanoparticle accumulation in tumor tissue, physical sources, such as light or heat, can be applied to modify nanoparticle drug-releasing activity.
Lastly, despite significant progress so far, a large gap between the cost of preparing PLGA nanoparticles and that of conventional delivery systems is seen as an impediment to their commercial application. Until now, most of the methods reported for preparation and surface modification of PLGA nanoparticles have involved small batches. Scale-up to large production volumes will certainly introduce additional challenges. As a result, the preparation process for PLGA nanoparticles needs to be further developed to achieve the reproducibility and scalability necessary in the marketplace.
# Disclosure
The authors report no conflicts of interest in this work. submit your manuscript | www.dovepress.com Dovepress Dovepress
[fig] Figure 1: Chemical structure of polylactide-co-glycolide. Abbreviations: m, number of units of lactide acid, n, number of units of glycolic acid. [/fig]
[fig] Figure 3: as shown by Konan et al who A B Degradation mechanisms of biodegradable polymeric nanoparticles: A) bulk erosion, B) surface erosion. [/fig]
[fig] Figure 4, Figure 5, Figure 6: enhanced permeability and retention effect. Passive tissue targeting is achieved by extravasation of nanoparticles through increased permeability of the tumor vasculature and ineffective lymphatic drainage. Abbreviations: PLGA, polylactide-co-glycolide; NP, nanoparticles. submit your manuscript | www.dovepress.Lym developed novel lectin-conjugated isopropyl myristate-incorporating PLGA nanoparticles for local delivery of paclitaxel to the lungs. These nanoparticles had superior in vitro cytotoxicity against A549 and H1299 cells as compared with paclitaxel-Targeted polylactide-co-glycolide nanoparticle carrying the chemotherapeutic drug. Internalization of targeted polylactide-co-glycolide nanoparticles via receptor-mediated endocytosis. Since specific receptors are overexpressed on tumor cells, the nanoparticles are selectively uptaken by the tumor cells via receptor-ligand interaction. Abbreviations: PLGA, polylactide-co-glycolide; NP, nanoparticles. [/fig]
[table] Table 2: Summary of polylactide-co-glycolide nanoparticles [/table]
[table] Table 3: Summary of targeted polylactide-co-glycolide nanoparticles [/table]
|
Disposal of iNKT cell deficiency and an increase in expression of SLAM signaling factors characterizes sarcoidosis remission: a 4-year longitudinal study
Background: Invariant NKT (iNKT) cells are regulatory lymphocytes that may be important in disorders with increased Th1 responses. We utilized a 4-year longitudinal observational study of iNKT cells and SLAM signaling pathway factors, which are important for iNKT development in patients with newly diagnosed sarcoidosis. Methods: Detailed clinical, functional, and radiographic evaluation and determination of iNKT peripheral blood cell counts and expression of SLAM signaling factors was carried out at presentation and after 3 months, 1 year, and 4 years of disease follow-up in 29 patients with pulmonary sarcoidosis. At presentation, we also evaluated the frequencies of pulmonary BALF iNKT cells. We also included 37 control subjects.Results: We demonstrated a marked deficiency of blood and lung iNKT cells and decreased expression of SLAM signaling factors in patients with newly diagnosed sarcoidosis. During 4 years of disease follow-up, there was a significant increase in blood iNKT cell numbers and in expression of SLAM signaling factors, mainly SLAMF1, SLAMF6, and FYN. This increase clearly correlated with improvement in patients' clinical symptoms. At the 4-year endpoint, the disease had gone into remission in the great majority of patients and thus also iNKT cell deficiency. Moreover, at the 4-year endpoint iNKT level reached the iNKT level of the control subjects.Conclusions: Our longitudinal study showed that a disposal of iNKT deficiency in parallel with an increase in expression of SLAM signaling factors characterizes the clinical remission of sarcoidosis.
# Introduction
In contrast to both conventional T lymphocytes and other types of Tregs, invariant natural killer T (iNKT) cells comprise a unique subgroup of immunoregulatory T lymphocytes that are restricted by the non-classical MHC class I molecule CD1d and expression of the invariant Vα24-Jα18 T cell receptor (TCR) preferably paired with Vβ11 [bib_ref] Cytotoxic gammadelta or alphabeta T cells with a natural killer cell marker,..., Satoh [/bib_ref] [bib_ref] Unique subpopulations of CD56+ NK and NK-T peripheral blood lymphocytes identified by..., Campbell [/bib_ref] [bib_ref] Natural killer T cells reactive to a single glycolipid exhibit a highly..., Matsuda [/bib_ref] [bib_ref] Van Kaer L: NKT cells: what's in a name?, Godfrey [/bib_ref] [bib_ref] Raising the NKT cell family, Godfrey [/bib_ref] [bib_ref] Natural killer T cells in pulmonary disorders, Rijavec [/bib_ref]. Furthermore, iNKT cells can either upregulate or downregulate cell-mediated immunity and thus these cells could have diverse influences in various disease models [bib_ref] Raising the NKT cell family, Godfrey [/bib_ref] [bib_ref] Natural killer T cells in pulmonary disorders, Rijavec [/bib_ref] [bib_ref] The unconventional lifestyle of NKT cells, Kronenberg [/bib_ref] [bib_ref] Ex vivo expansion of CD8 + CD56+ and CD8 + CD56-natural killer..., Wajchman [/bib_ref].
Sarcoidosis is the most common interstitial lung disease in the western world. It frequently involves organs other than the lungs. This multisystem disorder features the presence of CD4 + T cells that differentiate into Th1-like cells and amplify the local immune response [bib_ref] Statement on sarcoidosis, Ats/Ers/Wasog Committee [/bib_ref]. Despite much research over the past decade, the cause of the Th1-biased hyperactive response remains unknown. Ho et al. [bib_ref] Deficiency of a subset of T-cells with immunoregulatory properties in sarcoidosis, Ho [/bib_ref] first hypothesized that the loss of immunoregulation by iNKT cells could explain amplified and persistent T cell activity. This hypothesis was based on NOD mice that are deficient for iNKT cells. These mice are generally healthy, but they are predisposed to disorders with increased Th1 responses [bib_ref] Presumed guilty: natural killer T cell defects and human disease, Berzins [/bib_ref]. Ho reported a significant decrease in iNKT cell frequency in the peripheral blood and BALF of sarcoidosis patients. We recently confirmed the deficiency of iNKT cells in BALF of corticosteroid-naive sarcoidosis patients [bib_ref] Deficiency of pulmonary Valpha24 Vbeta11 natural killer T cells in corticosteroid-naïve sarcoidosis..., Korosec [/bib_ref]. Nevertheless, these association studies examined only one sample from each patient and, despite the fact that the iNKT cell defect was evident at some disease stages, there are currently no data about the nature of this defect or whether iNKT cells could be important for the long-term course of the disease. A more definitive study would have to closely follow the progress of iNKT cells over a longer period. Another question that arises is whether the iNKT cell defect is also affected by changes in iNKT cell development, which is highly dependent on signaling via the SLAM family of surface receptors [bib_ref] Raising the NKT cell family, Godfrey [/bib_ref] [bib_ref] Signaling lymphocytic activation molecule-associated protein controls NKT cell functions, Chung [/bib_ref] [bib_ref] Regulation of NKT cell development by SAP, the protein defective in XLP, Nichols [/bib_ref] [bib_ref] Defective NKT cell development in mice and humans lacking the adapter SAP,..., Pasquier [/bib_ref]. For this reason, we followed up blood iNKT cells and expression of SLAM signaling factors together with detailed clinical data in 29 patients with newly diagnosed pulmonary sarcoidosis over 4 prospective years.
# Methods
## Study subjects
Twenty-nine patients (mean age 38 years; range 25-74 years; 14 women, Caucasians) with newly diagnosed, histologically confirmed pulmonary sarcoidosis were recruited in a prospective manner from the Golnik University Clinic and first examined at diagnosis and then compared at 3 months, 1 year, and 4 years (mean after 4.2 years; range 3.5-4.6 years). Five subjects had Löfgren syndrome . For iNKT and SLAM analysis, four patients were missing at 3 months and five at 1 year [fig_ref] Figure 1: The number of patients included with sarcoidosis at each longitudinal time point... [/fig_ref]. We also included 28 healthy control subjects (mean age 39 years; range 20-62 years; 15 women); 15 control subjects were paired-sampled in a 3-month interval. The analysis involving BAL included an additional nine control subjects without any pulmonary morbidity (mean age 47 years; range 27-84 years; five women) and normal lung function tests. Some of these bronchoalveolar lavage (BAL) controls have already been published [bib_ref] Deficiency of pulmonary Valpha24 Vbeta11 natural killer T cells in corticosteroid-naïve sarcoidosis..., Korosec [/bib_ref]. All subjects gave their written consent and the study was approved by the Slovenian National Medical Ethics Committee.
## Chest radiographs, functional and clinical data
Pulmonary disease was evaluated at presentation and at the follow-up by chest radiography and pulmonary function testing in order to determine the disease course. FEV1, FVC, and DLCO were used to assess the presence of lung function impairment. All patients were also assessed for extrapulmonary disease and serum angiotensin-converting enzyme (sACE; .
## Bronchoalveolar lavage
BAL differential cell count and BAL TBNK immunophenotyping were performed at diagnosis, as previously described [bib_ref] Deficiency of pulmonary Valpha24 Vbeta11 natural killer T cells in corticosteroid-naïve sarcoidosis..., Korosec [/bib_ref] [bib_ref] Expansion of pulmonary CD8+ CD56+ natural killer T-cells in hypersensitivity pneumonitis, Korosec [/bib_ref]. In brief: we instilled seven aliquots of 20 ml and one of 10 ml, up to a total volume of 150 ml of saline. After each instillation, the aliquot was immediately recovered, pooled, and thereafter processed for cytological and flow cytometry analysis. Cytospin preparations were stained according to the May-Grünwald-Giemsa and Papanicolaou methods and differential cell count was performed by counting 200 cells.
For TBNK immunophenotyping, the BALF was strained through a 70 mm cell strainer (BD Biosciences), centrifuged and resuspended in Haemaccel (Behringwerke, Marburg, Germany), and incubated with the respective mAbs (CD3, CD4, CD8, CD14, CD19, CD45, CD16/56, BD Biosciences, San Jose, CA, USA) for 15 min, followed by lysing, washing, Results of chest radiographic stages, serum angiotensin-converting enzyme (sACE), lung function testing, and organ involvement in patients with newly diagnosed pulmonary sarcoidosis iNKT cells
Invariant NKT (CD3 + Vα24-Jα18 + Vβ11+) were identified by using mAb against CD3, Vα24-Jα18 (clone 6B11) T cell receptor (TCR; all from BD Biosciences), and Vβ11 TCR (Immunotech, Marseille, France) [bib_ref] Deficiency of pulmonary Valpha24 Vbeta11 natural killer T cells in corticosteroid-naïve sarcoidosis..., Korosec [/bib_ref] [bib_ref] Expansion of pulmonary CD8+ CD56+ natural killer T-cells in hypersensitivity pneumonitis, Korosec [/bib_ref]. Absolute iNKT counts were performed in combination with routine TruCOUNT tubes and TriTEST CD3/CD4/ CD45 staining (all from BD Biosciences). The frequencies of BALF iNKT cells were established at diagnosis and the cells were prepared alongside cells for TBNK immunophenotyping. First, the BALF cells were incubated with the respective iNKT mAbs for 15 min, followed by lysing, washing, and fixation, and then analyzed using FACSCalibur and CellQuest software (BD Biosciences). The BALF iNKT cells in patients were expressed as % of BALF T lymphocytes and were also compared with frequencies of BALF iNKT cells in control subjects.
Invariant NKT cells in peripheral blood were first examined at diagnosis and then compared at 3 months, 1 year, and 4 years for each patient. They were also compared with control subjects. Whole blood with EDTA anticoagulant was first incubated with iNKT mAbs for 15 min, followed by lysing, washing, and fixation, and analyzed using FACSCalibur and CellQuest software (BD Biosciences). The frequencies of blood iNKT cells were expressed as % of blood T lymphocytes. The parallel blood sample for absolute count was incubated with standardized TriTEST CD3/CD4/CD45 mAbs in TruCOUNT tubes followed by lysing and analyzing of T lymphocyte absolute numbers by FACSCalibur and MultiSET software (BD Biosciences). The absolute number of iNKT cells in whole blood was calculated by multiplying the frequencies of blood iNKT cells expressed as the proportion of blood T lymphocytes with the absolute number of blood T lymphocytes, and this was expressed as the number of iNKT cells per μl of whole blood.
## Slam signaling factors gene expression
Whole blood was collected in PAXgene Blood RNA Tubes (PreAnalytiX GmbH, Hombrechtikon, Switzerland) at diagnosis and then at 3 months, 1 year, and 4 years for each patient and also in control subjects and stored at −30°C until further processing. Total RNA was isolated from blood samples using the PAXgene Blood RNA Kit (PreAnalytiX GmbH) and quantified by a Qubit fluorometer using the Quant-iT™ RNA Assay Kit (Invitrogen Corporation, Carlsbad, CA, USA) according to the manufacturer's instructions. A reverse transcription reaction was performed using a High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA, USA).
## Real-time pcr was performed on an abi prism 7500
Real-Time PCR System (Sequence Detection System instrument equipped with SDS version 1.3.0 software; Applied Biosystems).
The human GAPD (GAPDH) Endogenous Control (Applied Biosystems) with the MGB probe labeled with FAM at the 5′ end was used for normalization of the target genes. Primers and probes for the genes of interest (SLAMF1, SLAMF6, FYN, and SLAM adaptor protein (SAP) (SH2D1A)) were also supplied by Applied Biosystems as TaqMan® Gene Expression Assays with the MGB probes labeled with FAM at the 5′ end (the assay IDs are Hs00158978_m1, Hs00941607_m1, Hs00234150_m1, and Hs00372941_m1, respectively). PCR reactions were set up in separate tubes with TaqMan Universal PCR Master Mix (Applied Biosystems) at default thermal conditions. All measurements were performed in triplicate for each time point and relative expression was analyzed using the ΔΔCt method.
# Statistical analysis
Data distribution was evaluated using the D' Agostino-Pearson test. Because the majority of data were not normally distributed, we performed a Wilcoxon or Mann-Whitney test. A chi-square test was used on binominal data. Data were expressed as median and range, if not otherwise specified. Probability values of P < 0.05 were accepted as significant. Analyses were performed using GraphPad Prism (GraphPad Software, La Jolla, CA, USA).
# Results
## Monitoring of pulmonary and extrapulmonary disease and treatment
At diagnosis, nine patients showed chest radiographic stage I, 19 showed stage II, and two showed stage III . At 3 months, the stages were highly comparable. However, after 1 year there was a significant improvement in radiographic chest stages (six showed stage 0), and even more after 4 years, when 17 patients showed stage 0, two stage I, nine stage II, and one stage III [fig_ref] Figure 2: Radiographic, functional, clinical, and treatment data in newly diagnosed patients with sarcoidosis... [/fig_ref]. The pulmonary function data also significantly improved, with a median increase of predicted FEV1 from 92 to 97%, FVC from 94 to 103%, and DLCO from 92 to 100% after 4 years of disease follow-up [fig_ref] Figure 2: Radiographic, functional, clinical, and treatment data in newly diagnosed patients with sarcoidosis... [/fig_ref]. There was also a significant decrease of sACE (from 1 to 0.6 μkat/l; [fig_ref] Figure 2: Radiographic, functional, clinical, and treatment data in newly diagnosed patients with sarcoidosis... [/fig_ref] after 4 years of disease follow-up.
At presentation, 16 patients showed extrapulmonary disease . At follow-up there was a decrease to 12 patients at 3 months, to eight at 1 year, and finally after 4 years to six patients (four with skin involvement and two with liver involvement) [fig_ref] Figure 2: Radiographic, functional, clinical, and treatment data in newly diagnosed patients with sarcoidosis... [/fig_ref].
Among the 29 patients, only five needed initial treatment with oral corticosteroids. At 1 year only one patient remained on treatment; later, in the 3rd year, the disease again progressed in another patient, and so finally after 4 years two patients that seemed to have chronic disease needed corticosteroid treatment [fig_ref] Figure 2: Radiographic, functional, clinical, and treatment data in newly diagnosed patients with sarcoidosis... [/fig_ref].
## Monitoring of inkt cells
The frequencies of BALF iNKT cells were determined only at the diagnosis and they were compared with the BALF of control subjects. The median (range) BALF iNKT cell frequency was 0.11% (0-2.55), whereas in controls the iNKT frequencies were about 10 times higher (0.96%; 0.2-2.51; [fig_ref] Figure 3: Frequencies [/fig_ref]. The accompanying analysis of BALF cell counts and TBNK immunophenotyping are presented in [fig_ref] Table 2: Bronchoalveolar lavage fluid [/fig_ref]. BALF of sarcoidosis patients showed a higher lymphocyte total cell count and CD4/CD8 T ratio with higher CD4 and lower CD8 T lymphocyte subset frequencies, but comparable CD3 T and B lymphocyte, NK cell, and nonbiased NKT cell frequencies.
In peripheral blood, iNKT cells were closely monitored from the diagnosis of sarcoidosis over 4 prospective years. These data were also compared with iNKT cells in control subjects; in some of the controls iNKT cells were also monitored over 3 months. The median (range) frequency of blood iNKT cells at diagnosis was 0.12% (0.02-0.58), at 3 months 0.15% (0.01-0.76), at 1 year 0.15% (0.02-0.79), and after 4 years 0.21% (0.01-1.3) [fig_ref] Figure 3: Frequencies [/fig_ref]. The median (range) absolute numbers of blood iNKT cells at diagnosis were 0.94 cells/μl (0.12-7.47), at 3 months 1.16 cells/μl (0.00-10.11), at 1 year 1.45 cells/μl (0.26-11.2), and after 4 years 1.87 cells/μl (0.14-15.7; [fig_ref] Figure 3: Frequencies [/fig_ref]. Thus after 4 years iNKT cell frequencies and absolute count significantly increased, and became comparable with control subjects (0.2%; 0.03-1.4 and 2.29 cells/μl; 0.41-12.3). Representative dot plots of iNKT cell analysis over 4 years of disease follow-up show a sarcoidosis patient with disease remission [fig_ref] Figure 4: Representative flow cytometry dot plots of iNKT cell analyzes over 4 years... [/fig_ref] and a sarcoidosis patient with chronic disease [fig_ref] Figure 4: Representative flow cytometry dot plots of iNKT cell analyzes over 4 years... [/fig_ref].
Monitoring of blood iNKT cells in 15 controls at the 3month interval showed no differences, either in frequencies (0.25% (0.03-1.44) vs. 0.20% (0.03-1.42)) or in absolute counts (4.7 cells/μl (0.41-12.3) vs. 5 cells/μl (0.44-12.30)). Similarly, when we followed the frequencies and absolute count of blood T lymphocytes in sarcoidosis patients, we did not find any significant differences (at diagnosis vs.
## Monitoring the expression of slam signaling factors
The expression of SLAM signaling factors SLAMF1, SLAMF6, FYN, and SAP was monitored in patients' whole blood at the same time points as iNKT cells. SLAMF1 expression remained similarly low after 3 months and 1 year of disease follow-up, but after 4 years it significantly increased and became comparable with the level in control subjects [fig_ref] Figure 5: Relative expression levels of genes [/fig_ref]. SLAMF6 expression demonstrated a moderate increase through 3 months and significantly through 1 to 4 years of follow-up, and it was already comparable to controls after 1 year [fig_ref] Figure 5: Relative expression levels of genes [/fig_ref]. A very similar significant increase was also shown for FYN expression [fig_ref] Figure 5: Relative expression levels of genes [/fig_ref]. SAP expression varied through 3 months and 1 year of disease follow-up, and after 4 years it remained at the same level as at diagnosis and still lower than in control subjects [fig_ref] Figure 5: Relative expression levels of genes [/fig_ref].
# Discussion
This study confirmed a marked blood and lung deficiency of invariant NKT cells in newly diagnosed patients with pulmonary sarcoidosis. It also showed that this deficiency is associated with decreased expression of the SLAM signaling factors that are important for development of these immunoregulatory T cells. Next, we closely followed the progress of blood iNKT cells and the expression of SLAM factors in sarcoidosis patients over a 4-year period. During longitudinal disease follow-up, there was a significant and gradual increase in iNKT cell numbers and expression of SLAM signaling factors. This increase correlated with improvement in patients' overall radiographic, functional, and clinical symptoms, and finally at the 4-year endpoint the disease had gone into remission in the great majority of patients and thus also iNKT cell deficiency.
A causative association between iNKT cells and Th1biased disorders such as sarcoidosis is based on the hypothesis that an iNKT cell deficiency affects their Th1related immunoregulatory role. The most direct evidence for this hypothesis comes from animal models, such as NOD mice with lower iNKT cell numbers, which are characterized by a Th1-biased hyperactive response, and in which the adoptive transfer of iNKT cells has been shown to prevent the development of Th1-biased disorders [bib_ref] Overexpression of natural killer T cells protects Valpha14-Jalpha281 transgenic nonobese diabetic mice..., Lehuen [/bib_ref] [bib_ref] Activation of natural killer T cells by alpha-galactosylceramide treatment prevents the onset..., Sharif [/bib_ref] [bib_ref] CD1-restricted NK T cells protect nonobese diabetic mice from developing diabetes, Wang [/bib_ref] [bib_ref] The natural killer T lymphocyte: a player in the complex regulation of..., Cardell [/bib_ref]. Unfortunately, in the case of sarcoidosis, no reliable animal model is available; therefore only clinical analyses of iNKT cells can be studied. One of the key questions that arose with the first evidence of iNKT cell deficiency in sarcoidosis [bib_ref] Deficiency of a subset of T-cells with immunoregulatory properties in sarcoidosis, Ho [/bib_ref] [bib_ref] Deficiency of pulmonary Valpha24 Vbeta11 natural killer T cells in corticosteroid-naïve sarcoidosis..., Korosec [/bib_ref] was whether the peripheral blood deficiency reflects the systemic state or is simply a consequence of iNKT compartmentalization; that is, the lack of blood iNKT cells might be due to sequestration of these cells in the compartments of disease activity. It seems that compartmentalization is very unlikely because the deficiency of blood iNKT cells was clearly evident alongside the iNKT deficiency in airways [bib_ref] Deficiency of a subset of T-cells with immunoregulatory properties in sarcoidosis, Ho [/bib_ref] [bib_ref] Deficiency of pulmonary Valpha24 Vbeta11 natural killer T cells in corticosteroid-naïve sarcoidosis..., Korosec [/bib_ref] , and current report] as well as in the sarcoid mediastinal lymph nodes [bib_ref] Distinct functional lineages of human V(alpha)24 natural killer T cells, Lee [/bib_ref] and in granulomatous lesions of cutaneous sarcoidosis [bib_ref] Comparison of the T cell patterns in leprous and cutaneous sarcoid granulomas...., Mempel [/bib_ref]. In contrast, only one report demonstrated increased Vα24 TCR expression by RT-PCR in the lymph nodes of sarcoidosis patients in comparison to the lymph nodes of lung-cancer patients [bib_ref] Impaired IFN-gamma production of Valpha24 NKT cells in non-remitting sarcoidosis, Kobayashi [/bib_ref]. Nevertheless, this observation was not confirmed by methodologically more stringent cellular approaches, which was the case for other reports [bib_ref] Deficiency of a subset of T-cells with immunoregulatory properties in sarcoidosis, Ho [/bib_ref] [bib_ref] Deficiency of pulmonary Valpha24 Vbeta11 natural killer T cells in corticosteroid-naïve sarcoidosis..., Korosec [/bib_ref] [bib_ref] Comparison of the T cell patterns in leprous and cutaneous sarcoid granulomas...., Mempel [/bib_ref].
In contrast to previous studies [bib_ref] Deficiency of a subset of T-cells with immunoregulatory properties in sarcoidosis, Ho [/bib_ref] [bib_ref] Deficiency of pulmonary Valpha24 Vbeta11 natural killer T cells in corticosteroid-naïve sarcoidosis..., Korosec [/bib_ref] [bib_ref] Comparison of the T cell patterns in leprous and cutaneous sarcoid granulomas...., Mempel [/bib_ref] , this study benefited from long-term longitudinal sampling and detailed patient follow-up. Importantly, we closely followed the longitudinal progress of blood iNKT cells and showed that the increase in iNKT cells clearly correlated with patients' clinical improvement. Moreover, after 4 years, when the disease resolved in the great majority of patients, blood iNKT levels reached the iNKT levels of the control subjects. Unfortunately, BAL sampling was carried out only at diagnosis and, because the disease was mostly resolved after 4 years, it did not seem ethical to perform another bronchoscopy and thus obtain a paired pulmonary sample. The observations that the restoration of iNKT cells could contribute to disease resolution also support the idea of possible therapeutic targeting of iNKT cells. Namely, clinical trials have shown that administering iNKT cell mitogens is well tolerated by patients and this can result in the expansion of residual iNKT cell populations [bib_ref] A phase I-II study of alpha-galactosylceramide-pulsed IL-2/GM-CSF-cultured peripheral blood mononuclear cells in..., Motohashi [/bib_ref]. In addition, this study did not show any differences concerning the nonbiased CD3 + CD16/56+ NKT cells in sarcoidosis patients, which is similar to previous reports [bib_ref] Deficiency of pulmonary Valpha24 Vbeta11 natural killer T cells in corticosteroid-naïve sarcoidosis..., Korosec [/bib_ref] [bib_ref] Expansion of pulmonary CD8+ CD56+ natural killer T-cells in hypersensitivity pneumonitis, Korosec [/bib_ref]. It is not known what causes iNKT cell defects in sarcoidosis. In particular, iNKT cell development is highly dependent on SLAM signaling factors. SLAMF1 and its homologue SLAMF6 are type 1 plasma membranes and, in response to ligation, Src tyrosine kinase FYN is recruited to SAP [bib_ref] Homotypic interactions mediated by Slamf1 and Slamf6 receptors control NKT cell lineage..., Griewank [/bib_ref]. In NOD mice a defect in SLAMF1 expression is responsible for the lower iNKT cell numbers [bib_ref] Slamf1, the NKT cell control gene Nkt1, Jordan [/bib_ref]. Furthermore, mice lacking SAP completely lack iNKT cells, yet conventional T cells are present in normal numbers. In patients with X-linked lymphoproliferative disease (defect in the SH2D1A gene, which encodes the SAP molecule) development of iNKT cell is also absent [bib_ref] Defective NKT cell development in mice and humans lacking the adapter SAP,..., Pasquier [/bib_ref]. However, a significant difference between human and mouse iNKT cells development is related to their place of maturation. In mice, mature iNKT cells develop in the thymus. In contrast, the equivalent development steps in humans are not completed in the thymus, but are completed in the periphery-or, at least, mature NKT cells are extremely rare in the human thymus, in contrast to their frequency in blood [bib_ref] Distinct homeostatic requirements of CD4+ and CD4-subsets of Vα24-invariant natural killer T..., Baev [/bib_ref] [bib_ref] Limited correlation between human thymus and blood NKT cell content revealed by..., Berzins [/bib_ref]. Due to the importance of the periphery for iNKT development in humans, it was reasonable to monitor the expression of SLAM signaling factors in the periphery of sarcoidosis patients. We showed that their expression upon disease presentation was much lower than in the control subjects, suggesting that impaired iNKT development might be the cause for the iNKT cell deficiency. Furthermore, during the course of the disease the expression of SLAMF1, SLAMF6, and FYN factors increased significantly, and this increase correlated with the increase in the iNKT cell numbers and the clinical remission of the disease.
The limitation of expression analysis was that SLAM family receptors are also expressed on T cells, B cells, and dendritic cells and SLAM adaptor protein (SAP) on T and NK cells [bib_ref] Role of SLAM-associated protein in the pathogenesis of autoimmune diseases and immunological..., Furukawa [/bib_ref]. However, only development of iNKT cells, but not other hematopoietic cells, is absent in patients with X-linked lymphoproliferative disease (SAP defect) [bib_ref] Defective NKT cell development in mice and humans lacking the adapter SAP,..., Pasquier [/bib_ref] and SAP-deficient mice [bib_ref] Regulation of NKT cell development by SAP, the protein defective in XLP, Nichols [/bib_ref] [bib_ref] Defective NKT cell development in mice and humans lacking the adapter SAP,..., Pasquier [/bib_ref] or reduced in NOD [bib_ref] Overexpression of natural killer T cells protects Valpha14-Jalpha281 transgenic nonobese diabetic mice..., Lehuen [/bib_ref] [bib_ref] Activation of natural killer T cells by alpha-galactosylceramide treatment prevents the onset..., Sharif [/bib_ref] [bib_ref] CD1-restricted NK T cells protect nonobese diabetic mice from developing diabetes, Wang [/bib_ref] [bib_ref] The natural killer T lymphocyte: a player in the complex regulation of..., Cardell [/bib_ref] , SLAMF1/SLAMF6 double mutant chimeras [bib_ref] Homotypic interactions mediated by Slamf1 and Slamf6 receptors control NKT cell lineage..., Griewank [/bib_ref] and in Fyn deficient mice [bib_ref] Cutting edge: NKT cell development is selectively impaired in Fyn-deficient mice, Eberl [/bib_ref] [bib_ref] The Src family tyrosine kinase Fyn regulates natural killer T cell development, Gadue [/bib_ref]. Because of that specific SLAM cellular development effect we believe that the measurement and quantitative follow-up of SLAM expression in whole blood sample is specific enough for their association with iNKT cells and also a possible reason for iNKT cell deficiency.
The heterogeneity among patients with sarcoidosis is well known [bib_ref] Statement on sarcoidosis, Ats/Ers/Wasog Committee [/bib_ref]. However, the majority of patients with sarcoidosis have a remission within 5 years after diagnosis, with few or no consequences [bib_ref] Statement on sarcoidosis, Ats/Ers/Wasog Committee [/bib_ref]. Furthermore, a recurrence after 1 or more years of remission is uncommon and affects less than 5% of patients, despite the fact that recurrent disease may develop at any age and in any organ. Conversely, up to one-third of patients show unrelenting chronic disease that can lead to clinically significant organ impairment. The incidence of chronic disease varies throughout the world, with the highest peak and severity in black Americans. In our study, only two patients seem to have had chronic disease, and in all other patients the remission was clearly evident within 4 years. Furthermore, only five patients needed corticosteroid treatment up to the 1st year, and after 1 year only one chronic patient remained continually on treatment and another started again in the 3rd year. Löfgren syndrome-an acute presentation consisting of arthralgia, erythema nodosum, and bilateral hilar adenopathy-was exhibited in five patients. Obviously, our study group represents patients with pulmonary sarcoidosis that have a good prognosis. Therefore, the iNKT cellular and SLAM expression results and implication should be predisposing only for patients with such a reversible clinical phenotype. We also did not separately analyze chronic or Löfgren syndrome patients because those subgroups were too small.
# Conclusion
In conclusion, we showed that the deficiency of iNKT cells and decreased expression of SLAM signaling factors constitute a significant immunoregulatory T cell abnormality in sarcoidosis, and further that the longitudinal increase in iNKT cell numbers and the changes in expression of SLAM signaling factors characterize the clinical remission of sarcoidosis. These observations support the hypothesis that the loss of immunoregulatory iNKT cells might contribute to the clinical appearance of sarcoidosis and, vice versa, that recovery of immunoregulatory iNKT cells might contribute to an improvement in symptoms and thus to the clinical remission of sarcoidosis. Because most of the data were based on analysis of patients that had a reversible clinical phenotype and in whom remission occurred within 4 years, further prospective studies on iNKT cells should obviously focus on patients with unrelenting chronic disease that leads to clinically significant organ impairment. Furthermore, further studies are also needed to better understand the mechanistic and possible Th1related role of iNKT cells in sarcoidosis. Only such data could confirm the biological and functional importance of this specific population of immunoregulatory T cells in immunopathogenesis of sarcoidosis.
## Competing interests
The authors declare that they have no competing interests.
[fig] Figure 1: The number of patients included with sarcoidosis at each longitudinal time point for cellular iNKT and SLAM expression analysis. and fixation. Labeled cells were analyzed by FACSCalibur and SimulSET software (BD Biosciences). [/fig]
[fig] Figure 2: Radiographic, functional, clinical, and treatment data in newly diagnosed patients with sarcoidosis and then over 4 years of disease follow-up: (A) chest radiographic stages, (B, C, D) lung functions, (E) serum ACE levels, (F) manifestations of granuloma in organs other than the lungs, and (G) systemic corticosteroid treatment. The data measured are presented with the median and interquartile range. [/fig]
[fig] 4: years: 67% (39-86) vs. 70% (42-89) and 860 cells/μl (479-1832) vs. 1019 cells/μl (311-2416)). [/fig]
[fig] Figure 3: Frequencies (A) of BALF iNKT (CD3 + Vα24-Jα18-Vβ11) cells in newly diagnosed sarcoidosis patients and in comparison to BALF of the control subjects; Frequencies (B) and absolute counts of iNKT cells in whole blood (C) of the same patients over 4 years of disease follow-up and in comparison to the healthy control subjects. Data are presented with the median and interquartile range. [/fig]
[fig] Figure 5: Relative expression levels of genes (A) SLAMF1, (B) SLAMF6, (C) FYN, and (D) SAP in whole blood of patients with sarcoidosis over 4 years of disease follow-up and in comparison to healthy control subjects. Data are presented with the median and interquartile range. [/fig]
[table] Table 2: Bronchoalveolar lavage fluid (BALF) cell counts and T lymphocytes, B lymphocytes, NK, and NKT-like cells in BALF from patients with sarcoidosis sampled upon diagnosis and control subjects [/table]
|
High prevalence of unawareness of HCV infection status among both HCV-seronegative and seropositive people living with human immunodeficiency virus in Taiwan
ObjectivesHCV infection status awareness is crucial in the HCV care continuum for both HCV-seropositive (HCV-positive status awareness) and seronegative (HCV-negative status awareness) populations. However, trends in the unawareness of HCV infection status (UoHCV) remain unknown in HIV-positive patients. This study investigated UoHCV prevalence, the associated factors of UoHCV, and its association with HCV-related knowledge in HIV-positive patients.MethodsFor this cross-sectional, multicenter, questionnaire-based study, 844 HIV-infected participants were recruited from three hospitals in Taiwan from June 2018 to March 2020. Participants were grouped by HCV serostatus (HCV-seronegative [n = 734] and HCV-seropositive [n = 110]) and categorized by their HIV diagnosis date (before 2008, 2008-2013, and 2014-2020). Exploratory factor analysis was used to categorize the 15 items of HCV-related knowledge into three domains: route of HCV transmission, HCV course and complications, and HCV treatment.
# Introduction
The global seroprevalence of the hepatitis C virus (HCV) is approximately 2.5%. Although treatment with direct-acting antivirals (DAAs) can lead to the elimination of HCV viremia and a curative outcome in more than 90% of patients with chronic HCV infection, several barriers to eradicating HCV infection still exist, including the high costs of drugs, frequent loss to follow-up after diagnosis, high rate of early HCV reinfection among patients who have recently received drug injections, and ongoing high-risk behaviors associated with HCV infection (even after clearance of HCV infection). The identification of undiagnosed patients, timely provision of DAAs to HCV-seropositive populations, and the prevention of transmission among at-risk HCV-seronegative populations must be prioritized to eliminate HCV infection.
Awareness of HCV infection status is crucial for both HCV-seropositive and HCV-seronegative populations. Although the short-and long-term impacts of HCV-positive status awareness among HCV-seropositive patients on their risk behavior remain matters of debate, HCV-positive status awareness is essential in the HCV care continuum in terms of treatment eligibility and taking medical advice on viral transmission. However, unawareness of HCV infection status (UoHCV) remains common in the HCV-seropositive population, with prevalence rates of 20.5% in Italy, 16.6%-35.1% in Taiwan, and 14%-51% in the United States. In the HCV-seronegative population at risk of contracting HCV, people who inject drugs (PWID) may engage in high-risk behaviors (e.g., sharing a syringe or injecting themselves with drugs) less frequently if they are aware of their HCV infection status (i.e., HCV-negative status awareness). However, a nationwide screening program in Taiwan revealed a 33% prevalence of UoHCV in the HCV-seronegative population.
People living with human immunodeficiency virus (HIV) infection (PLWH) are at risk of HCV infection because the transmission routes of HCV infection, such as unprotected sex and drug injection, are similar to those of HIV infection. Moreover, individuals coinfected with HCV and HIV are less likely to seek HCV care, which may contribute to a significantly decreased quality of life and quicker progression of liver disease, especially in those who are homeless or marginally housed. Additionally, patients with HCV/HIV coinfection have higher rates of death and disease progression, including the progression of histological fibrosis/cirrhosis and decompensated liver disease, than do patients with HCV monoinfection. Therefore, HCV screening, treatment, and prevention strategies should be strictly implemented among PLWH. However, although strategies have been implemented worldwide to combat the spread of HCV, no study has explored the prevalence or associated factors of UoHCV among PLWH. Serostatus awareness facilitates the next step in the continuum of HCV care, namely providing affected patients with access to health care, relevant consultation, and potential treatment, which are necessary to eradicate HCV. Furthermore, knowledge regarding HCV infection is essential for the further utilization of HCV treatment. In one study, the majority of participants had limited knowledge regarding the complications of chronic HCV infection despite being aware of the high prevalence of HCV among men who have sex with men (MSM). However, little is known regarding the association of UoHCV with knowledge of HCV among PLWH. We hypothesized that UoHCV is negatively associated with an individual's HCV-related knowledge, regardless of their HCV serostatus.
We conducted a cross-sectional questionnaire-based study at three HIV referral centers in Taiwan from June 2018 to March 2020. We evaluated the prevalence of UoHCV, explored the determinants of UoHCV, and evaluated the associations of UoHCV with different domains of HCV-related knowledge (i.e., route of HCV transmission, HCV course and complications, and HCV treatment) among a sample of PLWH stratified by HCV serostatus.
# Materials and methods
## Study design and setting
This cross-sectional, multicenter, questionnaire-based study was conducted from June 1, 2018, to March 31, 2020, at Kaohsiung Municipal Siaogang Hospital and Kaohsiung Municipal Ta-Tung Hospital, which are regional hospitals in southern Taiwan, and at Kaohsiung Medical University Hospital, which is the largest referral center for PLWH in southern Taiwan. The HCV seropositivity in southern Taiwan is 8.6%, which is higher than that in northern Taiwan (1.2%-2.7%).
## Development of the study questionnaire
A questionnaire was designed to investigate participants' awareness of HCV infection status, knowledge of different aspects of the disease, perceived risk of HCV infection, and assessment of potential exposure to HCV.
An expert group comprising an HIV case manager, HIV specialists, hepatologists, and researchers developed the preliminary questionnaire used in this study. The preliminary questionnaire was then modified based on feedback from 30 PLWH after they had completed a pretest. The questionnaire was tested again with 20 PLWH and further modified. Because of the lack of a standardized scoring system for evaluating different categories of HCV-related knowledge, the questionnaire items were modified after a review of pertinent studies. The section on the perceived risk of HCV infection was also modified.
The final questionnaire comprised variables in the following five categories: sociodemographic characteristics, awareness of HCV infection status, knowledge of HCV infection, perceived risk of HCV infection, and assessment of potential exposure to HCV (S1 File). Participants were instructed to answer the 15 HCV-related knowledge items by providing one of the following responses: "yes," "no," or "I do not know". One point was awarded for each correct response, and no point was awarded for incorrect or "I do not know" responses. Therefore, the mean scale scores ranged from 0 to 1, with higher scores indicating greater HCV-related knowledge. Variables correlated with the respondents' perceived risk of HCV infection were measured on a 5-point Likert scale (strongly disagree, disagree, neither agree nor disagree, agree, and strongly agree).
## Participants and study procedure
Two trained investigators screened PLWH by reviewing their available medical records at the participating hospitals for the period from January 1, 2000, to March 31, 2020. Patients who were less than 20 years old during the screening period (June 1, 2018, to March 31, 2020), had not undergone HCV antibody tests within 1 year before enrollment, or were lost to follow-up during the screening period were excluded. The participants completed the questionnaires on Google Forms.
The participants were classified into one of two groups according to their HCV serostatus: a HCV-seronegative group and a HCV-seropositive group. They were then stratified according to their awareness of their HCV infection status. Finally, each of the two groups was further divided into two subgroups: the unawareness/HCV-seronegative (subgroup 1), awareness/ HCV-seronegative (subgroup 2), unawareness/HCV-seropositive (subgroup 3), and awareness/HCV-seropositive (subgroup 4).
The study was approved by the Institutional Review Board of Kaohsiung Medical University Hospital (KMUHIRB-SV(I)-20180024) and adhered to the principles of the Declaration of Helsinki. The investigators obtained signed informed consent forms from all patients before enrollment.
## Definitions
The participants were categorized by three distinct periods based on the calendar year of their first confirmed HIV diagnosis: before 2008 (period 1, before the remission of the HIV epidemic among PWID), 2008-2013 (period 2, remission of the HIV epidemic among PWID but before the introduction of oral DAAs), and 2014-2020 (period 3, after the introduction of oral DAAs).
In this study, awareness of HCV infection status was defined as participants' self-reported recognition of their HCV infection status at the time of enrollment in the study (i.e., HCVseropositive patients' awareness of their HCV-positive status and HCV-seronegative patients' awareness of their HCV-negative status), whereas UoHCV was defined as participants' selfreported unawareness of their HCV infection status.
The behavioral indicators of a high risk of exposure to HCV infection were modified from other studies and included using any intravenous recreational drugs, engaging in chemosexual behaviors within the preceding 6 months, having a sexual partner within the preceding 6 months (assessment options were no sexual partners, one regular sexual partner, no regular sexual partners/less than five partners, and no regular sexual partner/more than five partners), and engaging in other activities involving sexual contact within the preceding 6 months.
## Outcomes of interest
The primary outcome of interest was the prevalence of UoHCV across the three study periods, stratified by the participants' HCV serostatus. Secondary outcomes were factors associated with UoHCV and the associations of UoHCV with the mean scores for three domains of HCV knowledge (route of HCV transmission, HCV course and complications, and HCV treatment) among the participants stratified by their HCV serostatus.
# Statistical analysis
Descriptive analyses were performed on the characteristics of the participants in the four subgroups. The categorical and continuous variables in each subgroup were compared through χ 2 tests (or Fisher's exact tests) and independent t tests, respectively. The prevalence of UoHCV was defined as the proportion of participants with UoHCV at the time of questionnaire completion. The trend analyses of the prevalence of UoHCV stratified by HCV serostatus in the three periods were performed using a Cochran-Armitage trend test with modified ridit scores.
Backward stepwise binary logistic regressions were performed to calculate odds ratios and evaluate associations in the bivariate and multivariable analyses between surveyed factors and UoHCV among all the participants and among those in the two HCV serostatus groups. To simultaneously consider the effects of all variables in the multivariable model, we adopted a backward approach.
To determine the validity of the 15 items used to measure the participants' knowledge regarding HCV infection, an item analysis was performed for the assessment of item discrimination. We also performed exploratory factor analysis by using principal axis factoring with varimax rotation to investigate the structural domain of the 15 items, and three domains were finally categorized: route of HCV transmission (domain 1), HCV course and complications (domain 2), and HCV treatment (domain 3). Cronbach's α was used to measure the internal consistency of the items in each structural domain, where α represented the function of the number of items in a test. Cronbach's α � 0.7 indicates high reliability.
Finally, to determine the association of UoHCV with the means of the total and domainspecific scores of HCV-related knowledge, we employed a multilinear regression model with a backward approach. β along with 95% confidence intervals were calculated to estimate the effects of UoHCV and directions of all associations. A backward approach was also adopted to enable the effects of all the variables to be simultaneously considered in the multivariable model.
All tests were two-tailed, and p < 0.05 was considered significant. Statistical analyses were performed using SPSS Statistics version 25.0 (IBM, Armonk, NY, USA).
# Results
## Participants
The study flowchart is displayed inAmong the 1448 eligible PLWH, 525 were unwilling to participate in the study, and 79 were excluded for incomplete responses to questionnaire items. A total of 844 PLWH were included in the final analysis. They were divided into HCVseronegative (n = 734) and HCV-seropositive (n = 110) groups. The two groups were further divided into subgroups 1 (unawareness/HCV-seronegative; n = 448), 2 (awareness/HCV-seronegative; n = 286), 3 (unawareness/HCV-seropositive; n = 22), and 4 (awareness/HCV-seropositive; n = 88).presents the sociodemographic characteristics, laboratory profiles, perceived risk of HCV infection, and high-risk behaviors for HCV infection of the participants in each subgroup. The mean (standard deviation) age at enrollment among all the participants was 36.6 (±9.8) years, and 98.1% of the participants were men. The routes of HIV transmission included men who have sex with men (MSM) (73.5%), bisexual contact (13.0%), heterosexual contact (8.3%), and drug injection (5.2%). Approximately 21.0%, 30.2%, and 48.8% of the participants were diagnosed as having HIV in periods 1, 2, and 3, respectively.
## Characteristics of the study participants
## Trend analysis of the prevalence of uohcv across the three periods stratified by participants' hcv serostatus
The prevalence rate of UoHCV was 46.9% in period 1, 54.5% in period 2, and 60.2% in period 3 (p for trend = 0.003;. The overall prevalence of UoHCV was 58.7%-62.6% in the HCV-seronegative group (p for trend = 0.497) and 15.1%-31.3% in the HCV-seropositive group (p for trend = 0.086). Participants with a history of treatment for HCV were excluded from the HCV-seropositive group because all of these patients were aware of their HCV infection status and were thus not part of the target population; this exclusion increased the prevalence of UoHCV to 33.8% (ranging from 30.8% in period 1 to 37.0% in period 3, p for trend = 0.632).
## Factors associated with uohcv among plwh stratified by hcv serostatus
After stratification by HCV serostatus, the two groups differed in terms of factors associated with UoHCV in a binary logistic regression. In the HCV-seropositive group, the proportion of participants who believed that intravenous injection was a requirement for contracting HCV was greater among individuals with UoHCV than among those who were aware of their HCV status. Furthermore, these individuals with UoHCV were less likely to have received an HIV diagnosis in period 3 (vs. period 1), be PWID (vs. MSM), and have hepatitis A virus (HAV) seropositivity, compared with individuals with awareness of their HCV status. In the HCV-seronegative group, compared with the participants who were aware of their HCV status, those with UoHCV were more likely to have a history of sexually transmitted diseases within the preceding 6 months and believe that intravenous injection was a requirement for contracting HCV. Moreover, they were less likely to have received higher education, previously heard of HCV infection, and received information on HCV infection from clinicians.
## Development of the hcv knowledge scale among plwh
The item analysis initially indicated that the original 15 items for measuring HCV knowledge exhibited satisfactory discriminant validity. Exploratory factor analysis finally identified three domains of the 15 items: route of HCV transmission (domain 1), HCV course and complications (domain 2), and HCV treatment (domain 3). Cronbach's α revealed suitable reliability in the three domains, which were thus further analyzed (S1 .. Multilinear regression analyses revealed that UoHCV was associated with lower mean scores overall and for each domain of HCV-related knowledge compared with HCV status awareness, both in the HCV-seronegative group and the HCV-seropositive group.
## Association between uohcv and hcv-related knowledge scores stratified by hcv serostatus
## Comparison of correct responses to structural questions on hcv knowledge between participants with and without awareness of their hcv status, stratified by hcv serostatus.
## Correct answer
## Hcv-seronegative group p-value hcv-seropositive group p-value
# Discussion
To the best of our knowledge, ours is the first study to reveal differences in the prevalence rates of UoHCV among PLWH; these rates were 61.0% in the HCV-seronegative group and 20.0% in the HCV-seropositive group. The prevalence rate in the HCV-seropositive group increased to 33.8% after the exclusion of participants with a history of HCV treatment, all of whom knew their HCV status. After stratification by HCV serostatus, the two groups differed in terms of their sociodemographic characteristics and laboratory variables associated with UoHCV, indicating the need for strategies to be tailored according to HCV serostatus when attempting to reduce UoHCV among PLWH.
Our results revealed a higher prevalence of UoHCV among MSM than among PWID (25.9% vs.7.3%, p = 0.037) in the HCV seropositive population. Since the 1990s, several measures have been implemented to promote HCV testing and to enhance the awareness of HCV infection status among PWID. These measures have achieved significant effects in increasing HCV status awareness. The risk of sexual transmission of HCV was low in the 1990s. However, since the mid-2000s, sexually transmitted HCV has been increasingly detected among sexually active MSM. The high prevalence of UoHCV among MSM may be attributable to their unawareness of their HCV seropositivity status due to the fundamental misunderstanding that HCV can only be contracted from intravenous exposure to infected blood (e.g., through the use of unsterile injection equipment and contaminated blood products). This explanation is supported by the finding that individuals in this study who believed that intravenous injection was a requirement for contracting HCV were more likely to have UoHCV. Furthermore, an increasing trend of MSM was observed in the HCV-seropositive PLWH group (28.3%, 68.0%, and 81.3% in periods 1, 2, and 3, respectively [p for trend < 0.001]; S2 , which is consistent with findings of another study conducted in Taiwan. Therefore, substantial efforts are urgently required to implement interventions that reduce the prevalence rate of UoHCV among HIV-positive individuals with sexually transmitted HCV. The lower prevalence of UoHCV among HCV-seropositive PLWH diagnosed in 2014-2020 compared with the prevalence among those diagnosed before 2008 may be attributable to the results of recent efforts to fight the spread of HCV. These efforts include the increase in publicity on HCV worldwide, newly available DAAs, and implementation of measures to ensure the affordability of oral DAA agents in Taiwan; such measures include national programs for using DAA agents to treat patients with advanced HCV (implemented in 2017-2018) and for providing treatment support to all eligible patients enrolled in Taiwan's National Health Insurance program (2019-present). This means that compared with before 2008, there is now a higher likelihood of primary physicians testing for HCV and identifying HCV seropositivity in people recently diagnosed as having HIV and informing them of their condition. Moreover, primary physicians may be more likely to inform individuals of their HCV-seropositive status when diagnosing them as having a coinfection with sexually transmitted HAV because of the emerging trend of concomitant sexually transmitted HAV and HCV among PLWH in Taiwan.
Reducing the engagement in high-risk behavior is critical for reducing rates of HCV infection. People who have received treatment for HCV who have ongoing exposure to HCV are at risk of reinfection. Limiting such exposure is necessary for making continual progress toward eradication of the disease. Studies have revealed that awareness of HCV infection status reduces engagement in high-risk behaviors among HCV-seropositive PWID. However, a significant association was not observed between UoHCV and high-risk behaviors related to HCV infection in the present study, in which the majority (94.8%) of participants had sexually transmitted HIV infection. This finding has two possible explanations. First, individuals who were aware of their HCV status may not have known its transmission route (e.g., sexual transmission). Second, individuals may have known the transmission route but been unwilling or unable to alter their behavior accordingly. Our findings support the second explanation because the participants with awareness of their HCV-seropositive status had higher scores in knowledge domain 1 (route of HCV transmission) than those without awareness of their HCV-seropositive status, indicating that they did not alter their behavior despite knowing the HCV transmission route. The results of a cross-sectional online survey of 48 Australian MSM further support our findings. The survey revealed that most participants knew that HCV infection can be sexually transmitted between men. However, participants generally did not know how to prevent the sexual transmission of HCV. Although notifying PWID of their HCV status can effectively reduce their engagement in high-risk behaviors, effective behavior modification approaches for reducing the sexual transmission of HCV in at-risk populations have yet to be identified. Moreover, interventions to reduce modifiable behavioral risk factors, such as condom distribution, the promotion of abstinence from illicit drugs, and advocacy for safe sex, may attract serious criticism because of the stigma associated with sex and drug use. Therefore, to optimize the awareness of the HCV care cascade among HCV-seropositive PLWH, further research is necessary to identify effective interventions for modifiable behavioral risk factors and settings during the process of notifying patients of their HCV status to promote long-term reductions in behaviors that involve a high-risk of exposure to HCV.
In the HCV-seronegative population, the prevalence of UoHCV (61.0%) among PLWH reported in the present study is substantially higher than the prevalence of general population reported by a nationwide screening program in Taiwan (33.0%), possibly because of the high proportion of MSM (76.6%) who believe that intravenous injection is a requirement for contracting HCV.
In the present study, only 288 (39.2%) of the 734 HCV-seronegative PLWH responded "yes" to the question "has your doctor ever provided information regarding HCV infection?" This number was unexpectedly low considering that these PLWH were undergoing follow-up every 1-3 months at HIV referral hospitals in Taiwan. According to Taiwan's guidelines for HIV and acquired immunodeficiency syndrome, HCV-seronegative PLWH should undergo annual screening for HCV antibodies. HIV experts may overlook the importance of informing PLWH of their HCV-seronegative status. However, our data revealed that UoHCV is associated with lower mean scores for total and domain-specific HCV-related knowledge compared with HCV status awareness among PLWH with HCVseronegative status. Furthermore, young, sexually active PLWH may primarily consult and receive serostatus notifications of HCV infection from HIV experts because of the stigma and discrimination that PLWH may experience when consulting with non-HIV experts. Therefore, HIV experts should notify sexually active PLWH of their HCV-seronegative status. This is especially crucial for individuals who believe that HCV can only be transmitted between PWID or who have been diagnosed as having a sexually transmitted disease within the preceding 6 months.
Overall, our study revealed that the correct response rate for questions on HCV-related knowledge varied considerably (16.8% to 71.1%). This finding provides a basis for targeting gaps in patient HCV knowledge during counseling. In our study, UoHCV was determined to be negatively associated with different aspects of HCV-related knowledge, which is considered critical for initiating treatment for HCV infection. The findings of the present study also indicate that interventions tailored to each patient's HCV serostatus should be actively enforced to reduce the prevalence of UoHCV.
The present study has several strengths. This is the first study to analyze UoHCV among PLWH and identify different determinants associated with UoHCV among PLWH according to their HCV serostatus. The study results also aid in the customization of strategies according to their HCV serostatus for reducing the prevalence of UoHCV among PLWH. Second, no standardized, validated HCV knowledge scoring system for PLWH is currently available. This study provides clinically relevant, structural, and valid measurements of HCV-related knowledge among PLWH. This structural measure of HCV-related knowledge can be applied to assess HCV-related knowledge before and after HCV-related educational interventions. These findings can also be used to longitudinally assess the influence of knowledge regarding HCV transmission on patients' engagement in high-risk behaviors associated with HCV infection after they have completed treatment interventions. However, this study also has several limitations. First, 525 patients were unwilling to participate, which might have caused selection bias. However, the demographic characteristics, HCV serostatus, and HIV-related risks did not differ between the enrolled and unwilling patients. Second, although the trained investigators instructed participants to recall their experiences to answer certain questions, inaccurate recall was unavoidable. Third, although the current study revealed no association between high-risk behaviors for HCV infection and UoHCV among HCV-seropositive and HCV-seronegative PLWH, this study had a cross-sectional design. Therefore, additional prospective cohort studies are warranted to further clarify changes in individuals' engagement in high-risk behaviors after they have been notified of their HCV status. Finally, the participants in the present study were PLWH, most of whom were MSMs. Therefore, the findings may not be generalizable to all PLWH populations.
# Conclusion
With the current availability of all-oral, interferon-free DAAs, which achieve a curative outcome in more than 90% of patients with HCV and require only 12 weeks of treatment, the eradication of HCV in prevalent areas is now possible; however, our study revealed that the majority of PLWH in Taiwan were unaware of their HCV serostatus. In the present study, UoHCV was associated with lower scores in different domains of HCV-related knowledge, which is considered critical for initiating treatment for HCV infection and reducing the engagement in risky behaviors among affected individuals. This finding suggests that interventions should be implemented to reduce the prevalence of UoHCV, regardless of patients' HCV serostatus. Because the factors associated with UoHCV differed between HCV-seronegative and HCV-seropositive PLWH, strategies for reducing UoHCV may be more effective if they target PLWH according to their HCV serostastus.
Supporting information S1 File. Questionnaire on HCV infection status and related knowledge among patients living with HIV. (DOCX) S1 |
Timing of Histologic Progression from Chorio-Deciduitis to Chorio-Deciduo-Amnionitis in the Setting of Preterm Labor and Preterm Premature Rupture of Membranes with Sterile Amniotic Fluid
BackgroundHistologic chorio-deciduitis and chorio-deciduo-amnionitis (amnionitis) in extra-placental membranes are known to represent the early and advanced stages of ascending intrauterine infection. However, there are no data in humans about the time required for chorio-deciduitis to develop and for chorio-deciduitis without amnionitis to progress to chorio-deciduitis with amnionitis, and the effect of prolongation of pregnancy on the development of chorio-deciduitis and amnionitis in patients with preterm labor and intact membranes (PTL) and preterm premature rupture of membranes (preterm-PROM). We examined these issues in this study.MethodsThe study population consisted of 289 women who delivered preterm (133 cases with PTL, and 156 cases with preterm-PROM) and who had sterile amniotic fluid (AF) defined as a negative AF culture and the absence of inflammation as evidenced by a matrix metalloproteinase-8 (MMP-8) level <23 ng/ml. We examined the association between amniocentesisto-delivery interval and inflammatory status in the extra-placental membranes (i.e., inflammation-free extra-placental membranes, choroi-deciduitis only, and chorio-deciduitis with amnionitis) in patients with PTL and preterm-PROM.ResultsAmniocentesis-to-delivery interval was longer in cases of chorio-deciduitis with amnionitis than in cases of chorio-deciduitis only in both PTL(median [interquartile-range (IQR)]; PLOS ONE | Competing Interests: The authors have declared that no competing interests exist. 645.4 [319.5] vs. 113.9 [526.9] hours; P = 0.005) and preterm-PROM (131.3 [135.4] vs. 95.2 [140.5] hours; P<0.05). Amniocentesis-to-delivery interval was an independent predictor of the development of both chorio-deciduitis and amnionitis after correction for confounding variables such as gestational age at delivery in the setting of PTL, but not preterm-PROM.ConclusionsThese data confirm for the first time that, in cases of both PTL and preterm-PROM with sterile AF, more time is required to develop chorio-deciduitis with amnionitis than chorio-deciduitis alone in extra-placental membranes. Moreover, prolongation of pregnancy is an independent predictor of the development of both chorio-deciduitis and amnionitis in cases of PTL with sterile AF.Progression from Chorio-Deciduitis to Chorio-Deciduo-Amnionitis PLOS ONE | 2 / 15 R Of 133 cases, 102 cases were only included in the analysis, because the test of umbilical arterial pH was not performed in 31 cases due to the failure in extraction of umbilical arterial blood at birth. ‡ Of 133 patients, the information about cervical-dilatation at amniocentesis was not available in 2 cases. SD, standard deviation; GA, gestational age; NS, not significant.
# Introduction
Ascending intra-uterine infection is the most common cause of spontaneous preterm birth [bib_ref] Intrauterine infection and preterm delivery, Goldenberg [/bib_ref] [bib_ref] Infection and preterm labor, Romero [/bib_ref] [bib_ref] Intrauterine infection and prematurity, Goncalves [/bib_ref]. It is hypothesized that infectious organisms from vagina gain access to the uterus through the cervix and then proceed to invade the chorio-decidua and finally the amnion in the extraplacental membranes [bib_ref] Intrauterine infection and preterm delivery, Goldenberg [/bib_ref] [bib_ref] Infection and preterm labor, Romero [/bib_ref] [bib_ref] Intrauterine infection and prematurity, Goncalves [/bib_ref]. Indeed, levels of inflammatory biomarkers in the amniotic fluid (AF) are increased in the presence of histologic chorio-deciduitis compared with inflammation-free placentas [bib_ref] Fetal, amniotic and maternal inflammatory responses in early stage of ascending intrauterine..., Park [/bib_ref] , and are increased even further in the presence of amnionitis compared with chorio-deciduitis alone [bib_ref] The involvement of human amnion in histologic chorioamnionitis is an indicator that..., Park [/bib_ref]. However, there is no data in humans about the time required for ascending intra-uterine infection to invade the extra-placental membranes leading to histologic chorio-deciduitis and amnionitis in the setting of preterm labor and intact membranes (PTL) and preterm premature rupture of membranes (preterm-PROM). We hypothesized that more time would be required to develop histologic chorio-deciduitis with amnionitis than chorio-deciduitis without amnionitis in cases of both PTL and preterm-PROM with sterile AF, and prolongation of pregnancy would be an independent risk factor of chorio-deciduitis and amnionitis in the setting of PTL and preterm-PROM with sterile AF.
# Methods
## Study design and data collection
A retrospective cohort study was performed of 289 patients who delivered preterm (defined as delivery prior to 37 weeks of gestation) due to PTL (133 cases) or preterm-PROM (156 cases) from the year 1993 to 2007. Patients were eligible for inclusion in the analysis if they presented with a singleton pregnancy and PTL or preterm-PROM and also had: (1) an amniocentesis between 23.5 weeks and 36.0 weeks of gestation; (2) "sterile" AF as evidenced by a negative AF culture and the absence of AF inflammation defined as an AF matrix metalloproteinase-8 (MMP-8) level <23 ng/mL; and (3) documented results of placental pathology. On histologic examination of the placenta and extra-placental fetal membranes, inflammation was examined in choriodecidua, amnion, umbilical-cord, and chorionic-plate. The association between histologic findings and amniocentesis-to-delivery interval was then examined. Patients were divided into 3 groups according to the amniocentesis-to-delivery interval (2 days, 2-7 days, and >7 days). Transabdominal amniocentesis and placental pathology are routinely offered to patients admitted with a diagnosis of PTL or preterm-PROM at our institution. PTL was defined as the presence of regular phasic uterine contractions with a frequency of at least two contractions every 10 minutes and associated cervical change. Rupture of membranes was diagnosed at the time of sterile speculum examination by confirming the presence of pooling of AF in the vagina, a positive nitrazine pH test, and a positive ferning test result as previously reported [bib_ref] Maternal blood C-reactive protein, white blood cell count, and temperature in preterm..., Yoon [/bib_ref] [bib_ref] Serum C-reactive protein, white blood cell count, and amniotic fluid white blood..., Yoon [/bib_ref]. AF was routinely analyzed for intraamniotic infection and inflammation. We obtained a written informed consent in all subjects included in this study. The collection or use of these samples and information for research purposes were approved by the Institutional Review Board of Seoul National University Hospital. The Institutional Review Board of Seoul National University Hospital specifically approved this study.
The demographic and clinical characteristics of mothers and their neonates were abstracted from the medical records. Data abstracted included maternal age, parity, gestational age (GA) at amniocentesis, cervical dilatation at amniocentesis, GA at delivery, vaginal delivery or Cesarean section after trial of labor, birth weight, gender of newborn, 1 min and 5 min Apgar scores, umbilical cord arterial pH at birth (if collected), antenatal use of antibiotics, corticosteroids and tocolytic agents. Clinical chorioamnionitis was diagnosed when maternal temperature was elevated to 37.8°C and 2 of the following criteria were present: uterine tenderness, malodorous vaginal discharge, maternal leukocytosis (>15,000 cells/mm 3 ), maternal tachycardia (>100 beats/min), and fetal tachycardia (>160 beats/min).
## Placental pathologic examination
Placental tissue samples included choriodecidua, amnion, umbilical-cord, and chorionic-plate. Samples were fixed in 10% neutral buffered formalin and embedded in paraffin. Hematoxylin / eosin staining was performed in the sections of prepared tissue-blocks. Clinical information was not disclosed to pathologists. Chorio-deciduitis, amnionitis, funisitis and chorionic plate inflammation were diagnosed according to previously published criteria [bib_ref] Amniotic fluid interleukin-6: a sensitive test for antenatal diagnosis of acute inflammatory..., Yoon [/bib_ref] , which included the following: (1) for chorio-deciduitis, at least one focus of more than five neutrophils in the choriodecidua; [bib_ref] Infection and preterm labor, Romero [/bib_ref] for amnionitis, at least one focus of more than five neutrophils in the amnion; (3) for funisitis, neutrophil infiltration into the umbilical vessel walls or Wharton's jelly; and (4) for chorionic-plate inflammation, more than one focus of at least ten neutrophilic-collections or diffuse inflammation in subchorionic-fibrin, or diffuse / dense inflammation, neutrophilic-infiltration into connective-tissue of placental-plate, or placental-vasculitis. For the purposes of the analysis, the histologic inflammatory status of the extra-placental membranes was classified into three groups: inflammation-free extra-placental membranes, choriodeciduitis only, or chorio-deciduitis with amnionitis.
## Amniotic fluid testing
AF was cultured for aerobic / anaerobic bacteria and genital-mycoplasmas (e.g., U. urealyticum, and M. hominis) with the use of the methods previously reported [bib_ref] Maternal blood C-reactive protein, white blood cell count, and temperature in preterm..., Yoon [/bib_ref] [bib_ref] Serum C-reactive protein, white blood cell count, and amniotic fluid white blood..., Yoon [/bib_ref]. We centrifuged and stored the remaining-fluid in polypropylene-tubes at -70°C. We measured MMP-8 concentrations in stored-AF as previously reported [bib_ref] Clinical significance of intra-amniotic inflammation in patients with preterm premature rupture of..., Shim [/bib_ref] [bib_ref] The relationship between amniotic fluid matrix metalloproteinase-8 and funisitis, Park [/bib_ref]. The sensitivity of the test was <0.3 ng/mL. Both intra-and inter-assay coefficients of variation were <10%. AF inflammation was defined as an elevated AF MMP-8 concentration (>23 ng/mL) as previously reported [bib_ref] Clinical significance of intra-amniotic inflammation in patients with preterm premature rupture of..., Shim [/bib_ref]. AF was regarded as sterile if the AF culture was negative and there was no AF inflammation (i.e., AF MMP-8 levels were <23 ng/mL).
# Statistical analysis
The comparisons of continuous variables for clinical characteristics and pregnancy outcomes were performed using Kruskal-Wallis test and the comparisons of proportions were made using Pearson's chi-square test. Spearman rank correlation test was used to investigate the relationship between amniocentesis-to-delivery interval and inflammatory status in the extra-placental membranes (i.e., inflammation-free extra-placental membranes, chorio-deciduitis only, and chorio-deciduitis with amnionitis). Mann-Whitney U test was used for the comparison of amniocentesis-to-delivery interval between each inflammatory status group in extra-placental membranes. The linear-by-linear association test was used to investigate for trend. Logistic regression analysis was used to examine the relationship between chorio-deciduitis with or without amnionitis, and amniocentesis-to-delivery interval controlling for the effect of any other potential confounding variables. Data were analyzed using SPSS Statistics 20.0. Statistical significance was defined as P < .05.
# Results
The clinical characteristics and pregnancy outcomes in cases with PTL and in those with preterm-PROM are shown in and , respectively. There was a significant difference in GA at amniocentesis, GA at delivery, vaginal delivery or Cesarean section after trial of labor, and antenatal corticosteroids use between the groups according to amniocentesis-to-delivery interval in both PTL and preterm-PROM (P<0.05 for each, and . There was also a significant difference in antenatal use of antibiotics between the three amniocentesis-todelivery interval groups in preterm-PROM, but not PTL (see and . . Clinical characteristics and pregnancy outcomes according to amniocentesis-to-delivery interval (i.e., 2 days, 2-7 days, and >7 days) in 133 patients with preterm labor and intact membranes (PTL) and sterile amniotic fluid. Inflammation in each of the placental compartments (i.e., chorio-deciduitis, amnionitis, funisitis, and chorionic plate inflammation) increased with increasing amniocentesis-to-delivery interval in both PTL and preterm-PROM (P<0.05 for each using linear-by-linear association analysis; [fig_ref] Fig 1: Frequency of inflammation in placental compartments according to amniocentesis-to-delivery interval [/fig_ref]. Chorio-deciduitis developed within 2 days of the documentation of sterile AF in 10.4% (8/77) cases of PTL and 17.8% (13/73) cases of preterm-PROM [fig_ref] Fig 1: Frequency of inflammation in placental compartments according to amniocentesis-to-delivery interval [/fig_ref]. However, amnionitis did not appear until 2 days of the documentation of sterile AF in the preterm-PROM cohort and until 7 days in the PTL cohort [fig_ref] Fig 1: Frequency of inflammation in placental compartments according to amniocentesis-to-delivery interval [/fig_ref]. Amnionitis was first documented on day 6 after documentation of sterile AF in the preterm-PROM cohort and on day 14 in the PTL cohort [fig_ref] Fig 2: Amniocentesis-to-delivery interval in accordance with histologic status of the extra-placental membranes [/fig_ref]. Of note, histologic amnionitis was never present without chorio-deciduitis in both PTL and preterm-PROM cohorts (data not shown). [fig_ref] Fig 2: Amniocentesis-to-delivery interval in accordance with histologic status of the extra-placental membranes [/fig_ref] amniocentesis-to-delivery interval according to the inflammatory status in the extra-placental membranes. There was a strong correlation between amniocentesis-to-delivery interval and inflammation in the extra-placental membranes in both PTL (r = 0.322; P<0.0005; Spearman rank correlation test) and preterm-PROM (r = 0.290; P<0.0005; Spearman rank correlation test). Moreover, median [interquartile range (IQR)] amniocentesis-todelivery interval was longer in cases of chorio-deciduitis with amnionitis than in cases of . Clinical characteristics and pregnancy outcomes according to amniocentesis-to-delivery interval (i.e., 2 days, 2-7 days, and >7 days) in 156 patients with preterm premature rupture of membranes (preterm-PROM) and sterile amniotic fluid. [fig_ref] Fig 2: Amniocentesis-to-delivery interval in accordance with histologic status of the extra-placental membranes [/fig_ref]. Of note, chorio-deciuditis with amnionitis developed later in PTL than in preterm-PROM (P<0.005) [fig_ref] Fig 2: Amniocentesis-to-delivery interval in accordance with histologic status of the extra-placental membranes [/fig_ref]. To determine the relative value of clinical parameters in the prediction of chorio-deciduitis with or without amnionitis, we conducted multiple logistic regression analysis with variables commonly regarded as confounding factors for the development of chorio-deciduitis and amnionitis. Of all these independent variables, amniocentesis-to-delivery interval retained statistical significance in the prediction of both chorio-deciduitis and amnionitis after adjusting for confounding variables such as GA at delivery in PTL [fig_ref] Table 3: Relationship of various independent variables with the development of chorio-deciduitis and amnionitis... [/fig_ref] while early GA at delivery was an independent risk factor for amnionitis after adjusting for the contribution of potential confounding variables in preterm-PROM [fig_ref] Table 4: Relationship of various independent variables with the development of chorio-deciduitis and amnionitis... [/fig_ref].
# Discussion
Principal findings of this study are that, in cases with confirmed sterile AF, longer time is required to develop histologic chorio-deciduitis with amnionitis than chorio-deciduitis alone in the setting of both PTL and preterm-PROM and prolongation of pregnancy is an independent predictor of both chorio-deciduitis and amnionitis in PTL.
The chorio-decidua, the outermost layer of the extra-placental membranes, is exposed to the vagina via the cervical canal and may be readily accessible by ascending intra-uterine infection. In contrast, the amnion, the innermost layer of the extra-placental membranes, is more protected and is not readily accessible by ascending intra-uterine infection [bib_ref] Intrauterine infection and preterm delivery, Goldenberg [/bib_ref] [bib_ref] Infection and preterm labor, Romero [/bib_ref] [bib_ref] Intrauterine infection and prematurity, Goncalves [/bib_ref]. We hypothesize, therefore, that more time is required for ascending intra-uterine infection to cause choriodeciduitis with amnionitis than chorio-deciduitis alone. Unfortunately, no study has yet been reported in humans to test this hypothesis, and there are no good animal models of ascending infection in the setting of PTL or preterm-PROM. To test this hypothesis, we therefore chose a large cohort of women with PTL or preterm-PROM, who underwent amniocentesis to exclude intra-uterine infection, and who subsequently delivered preterm and had their placentas examined by placental pathologists. In order to avoid the possibility of trans-placental infection and to enrich the population with cases that were likely to have a prolonged amniocentesis-todelivery interval, we chose to exclude patients who had preexisting intra-amniotic infection and/or inflammation (as evidenced by a positive AF culture or elevated AF MMP-8 concentration).
Our data demonstrated a strong correlation between amniocentesis-to-delivery interval and inflammatory status in the extra-placental membranes [fig_ref] Fig 2: Amniocentesis-to-delivery interval in accordance with histologic status of the extra-placental membranes [/fig_ref]. Moreover, the median amniocentesis-to-delivery interval was longer in cases of chorio-deciduitis with amnionitis than in cases of chorio-deciduitis alone in both the PTL and preterm-PROM cohorts [fig_ref] Fig 2: Amniocentesis-to-delivery interval in accordance with histologic status of the extra-placental membranes [/fig_ref]. It should be noted that, while histologic chorio-deciduitis was evident as early as one day after documentation of sterile AF in both the PTL and preterm-PROM groups, histologic evidence of amnionitis first appeared on day 6 in the preterm-PROM cohort and day 14 in the PTL cohort [fig_ref] Fig 2: Amniocentesis-to-delivery interval in accordance with histologic status of the extra-placental membranes [/fig_ref]. In this study, amnionitis developed later in the PTL group than in the preterm-PROM group [fig_ref] Fig 2: Amniocentesis-to-delivery interval in accordance with histologic status of the extra-placental membranes [/fig_ref]. Given that the amnion lines the amniotic cavity, in patients with PTL, microorganisms or neutrophils originating within the maternal decidua must migrate through the chorion and chorio-amniotic interface before reaching the amnion [bib_ref] Intrauterine infection and preterm delivery, Goldenberg [/bib_ref] [bib_ref] Infection and preterm labor, Romero [/bib_ref] [bib_ref] Intrauterine infection and prematurity, Goncalves [/bib_ref]. However, in the context of ruptured membranes, microorganisms from the vagina may directly invade the amnion by entering the intra-uterine cavity through site of membranes rupture [bib_ref] Intrauterine infection and preterm delivery, Goldenberg [/bib_ref] [bib_ref] Infection and preterm labor, Romero [/bib_ref] [bib_ref] Intrauterine infection and prematurity, Goncalves [/bib_ref]. This may explain why PTL is likely to need more time to manifest amnionitis than preterm-PROM.
There is no information about the relationship between prolongation of pregnancy and the development of acute histologic chorioamnionitis in cases of PTL. Moreover, most data about the latency after preterm-PROM analyzed clinical, but not histologic, chorioamnionitis [bib_ref] Intentional delivery versus expectant management with preterm ruptured membranes at 30-34 weeks'..., Cox [/bib_ref] [bib_ref] Premature rupture of membranes at 34 to 37 weeks' gestation: aggressive versus..., Naef Rw 3rd [/bib_ref] [bib_ref] Chorioamnionitis increases neonatal morbidity in pregnancies complicated by preterm premature rupture of..., Ramsey [/bib_ref] [bib_ref] Factors and outcomes associated with longer latency in preterm premature rupture of..., Aziz [/bib_ref] [bib_ref] Neonatal outcomes are associated with latency after preterm premature rupture of membranes, Nayot [/bib_ref] [bib_ref] Risk factors and perinatal outcomes associated with latency in preterm premature rupture..., Ekin [/bib_ref] [bib_ref] Induction versus expectant management in premature rupture of the membranes with mature..., Mercer [/bib_ref] , and very little data regarding the length of the latency period and histologic chorio-deciduitis / amnionitis in the setting of preterm-PROM is controversial [bib_ref] Prolonged latency after preterm premature rupture of membranes: an evaluation of histologic..., Mcelrath [/bib_ref] [bib_ref] Relationship between endomyometritis and the duration of premature membrane rupture, Üstün [/bib_ref] [bib_ref] Lack of relationship between histologic chorioamnionitis and duration of the latency period..., Ghidini [/bib_ref]. While some studies have reported an association between the development of acute chorioamnionitis and prolonged rupture of membranes [bib_ref] Prolonged latency after preterm premature rupture of membranes: an evaluation of histologic..., Mcelrath [/bib_ref] [bib_ref] Relationship between endomyometritis and the duration of premature membrane rupture, Üstün [/bib_ref] , other study has not [bib_ref] Lack of relationship between histologic chorioamnionitis and duration of the latency period..., Ghidini [/bib_ref] (see [fig_ref] Table 5: Relationship between the latency after ROM and chorioamnionitis in previous studies [/fig_ref]. However, all of these studies have one major limitation, which is that they failed to perform amniocentesis and confirm the absence of pre-existing intra-amniotic infection/inflammation, thereby failing to exclude a major source of bias leading to a shorter amniocentesis-to-delivery interval in cases of intra-amniotic infection/inflammation [bib_ref] Clinical significance of intra-amniotic inflammation in patients with preterm premature rupture of..., Shim [/bib_ref] [bib_ref] Clinical significance of intra-amniotic inflammation in patients with preterm labor and intact..., Yoon [/bib_ref]. Progression from Chorio-Deciduitis to Chorio-Deciduo-Amnionitis
Our results also show that prolongation of pregnancy is an independent predictor of both chorio-deciduitis and amnionitis after adjusting for confounding variables (i.e., GA at delivery) in patients with PTL [fig_ref] Table 3: Relationship of various independent variables with the development of chorio-deciduitis and amnionitis... [/fig_ref]. These results are consistent with that of Lee and colleagues' [bib_ref] The frequency and risk factors of funisitis and histologic chorioamnionitis in pregnant..., Lee [/bib_ref] who reported that a longer duration of labor was associated with a higher frequency of acute-chorioamnionitis and funisitis in pregnant women in labor at term with intact membranes, although it should be noted that this study did not include women with PTL and there was no information on infection or inflammation in the AF. How then is PTL with sterile AF causally related to histologic chorio-deciduitis with increasing amniocentesis-to-delivery interval? It has been proposed that, in addition to causing the cervix to efface and dilate, uterine contractions have a "suction-like effect" that draws vaginal fluid with a microbial load into the uterine cavity [bib_ref] Physiology of upward transport in the human female genital tract, Zervomanolakis [/bib_ref] [bib_ref] The frequency of microbial invasion of the amniotic cavity and histologic chorioamnionitis..., Seong [/bib_ref]. Moreover, vaginal delivery or Cesarean section after trial of labor was an independent risk factor for the development of chorio-deciduitis in patients with PTL [fig_ref] Table 3: Relationship of various independent variables with the development of chorio-deciduitis and amnionitis... [/fig_ref]. It is plausible that delivery mode influence the histologic finding of the extra-placental membranes because the patient who delivered vaginally or tried to deliver vaginally experienced more times of uterine contractions and probably received more times of digital examinations. Indeed, this result is similar to the previous report that acute-histologic chorioamnionitis was significantly more common in women with regular uterine contractions with cervical dilatation 2cm than in those without regular uterine contraction [bib_ref] The risk of intra-amniotic infection, inflammation and histologic chorioamnionitis in term pregnant..., Lee [/bib_ref].
The major strengths of the study are that it included a large cohort of singleton preterm pregnancies with AF biomarkers and placental histology (n = 289) and that we specifically excluded patients with preexisting intra-amniotic infection/inflammation, which is known to have a strong effect on both the amniocentesis-to-delivery interval and the incidence of acute chorioamnionitis [bib_ref] The involvement of human amnion in histologic chorioamnionitis is an indicator that..., Park [/bib_ref] [bib_ref] Clinical significance of intra-amniotic inflammation in patients with preterm premature rupture of..., Shim [/bib_ref] [bib_ref] Clinical significance of intra-amniotic inflammation in patients with preterm labor and intact..., Yoon [/bib_ref] [bib_ref] The frequency and clinical significance of intra-uterine infection and inflammation in patients..., Park [/bib_ref]. The potential weaknesses of this study are as follows. Firstly, we measured latency as amniocentesis-to-delivery interval, since it is difficult to define the precise time at which PTL or preterm-PROM developed. However, we do not believe this is a major source of bias, since our prior study has shown that there is no relationship between intraamniotic inflammation and the time elapsed between membranes rupture and amniocentesis after the first 6 hours of membranes rupture [bib_ref] Is there a relationship between the interval between rupture of membranes and..., Kim [/bib_ref]. Secondly, we used "the absence of pre-existing intra-amniotic inflammation" instead of "the absence of pre-existing uterine cervicitis" as a starting point of ascending intra-uterine infection, although the first step of ascending intrauterine infection is the bacterial infection to the uterine cervix in theory. However, all the previous studies about the uterine cervicitis (defined as an elevated cervical fluid inflammatory cytokine concentration) and acute histologic chorioamnionitis consistently reported that the presence or absence of uterine cervicitis did not discriminate between the presence and absence of acute histologic chorioamnionitis [bib_ref] Increased interleukin-6 concentrations in cervical secretions are associated with preterm delivery, Lockwood [/bib_ref] [bib_ref] Expression of cytokines and chemokines in cervical and amniotic fluid: relationship to..., Holst [/bib_ref] [bib_ref] Cervical and vaginal fluid soluble Toll-like receptor 2 in pregnancies complicated by..., Kacerovsky [/bib_ref] [bib_ref] Cervical microbiota in women with preterm prelabor rupture of membranes, Kacerovsky [/bib_ref]. Indeed, we found "the absence of intra-amniotic inflammation"(defined as AF MMP-8 < 23 ng/ml) is superior to "the absence of uterine cervicitis" (defined as cervical fluid IL-6 < 350 pg/ml) in the identification of inflammation-free placenta in another retrospective cohort of our database which consisted of 38 singleton pregnant women who underwent amniocentesis and cervical fluid collection before 37 weeks of gestation and delivered within 7 days of amniocentesis, and we could predict "inflammation-free placenta" only with the use of "the absence of intra-amniotic inflammation" (sensitivity, 89.5% . Thirdly, we regarded "5 min Apgar score < 7" instead of umbilical cord arterial pH at birth and abnormal fetal heart rate pattern (FHR) as the representative of fetal distress in the overall logistic regression analysis for the relationship between various independent variables and the development of chorio-deciduitis and amnionitis, although the Apgar scoring system remains as relevant for the prediction of neonatal survival in preterm and term neonates as it was almost 50 years ago [bib_ref] The continuing value of the Apgar score for the assessment of newborn..., Casey [/bib_ref] [bib_ref] Low 5-minute Apgar score: a population-based register study of 1 million term..., Thorngren-Jerneck [/bib_ref] and the validity of the relationship between certain FHR patterns (i.e., degree of FHR variability and depth of decelerations) and fetal academia and/or 5-minute Apgar score 7 was demonstrated [bib_ref] Fetal acidemia and electronic fetal heart rate patterns: is there evidence of..., Parer [/bib_ref]. However, we did not have umbilical cord arterial pH result at birth in 31 patients (23.3%) among 133 cases with PTL and in 21 patients (13.5%) among 156 cases with preterm-PROM due to the failure in the extraction of umbilical cord arterial blood at birth. Moreover, it would seem that it is inadequate to accept abnormal FHR pattern as an objective indicator of fetal distress because there is high inter-observer and intra-observer variability in the interpretation of FHR tracings [bib_ref] Intrapartum fetal heart rate monitoring: nomenclature, interpretation, and general management principles, College [/bib_ref] [bib_ref] Intra-and inter-observer variability in the assessment of intrapartum cardiotocograms, Nielsen [/bib_ref] [bib_ref] The reproducibility of intrapartum cardiotocogram assessments, Beaulieu [/bib_ref]. Finally, we did not quantitatively evaluate the frequency of uterine contraction as an independent risk factor of chorio-amnionitis. However, it would seem that we should consider the amplitude or intensity in addition to the frequency of PTL for the exact quantitation of uterine contractility. Unfortunately, external uterine tocodynamometry is unlikely to represent the exact intensity of uterine contraction. Indeed, Paul MJ et al. demonstrated that external tocodynamometry reliably distinguishes between uterine contractions and uterine quiescence in preterm pregnancies but does not adequately measure the intensity of contractions [bib_ref] Relationship of measured external tocodynamometry with measured internal uterine activity, Paul [/bib_ref]. However, we did not routinely apply the intrauterine pressure monitor for the exact measurement of the intensity of uterine contractions in patients with PTL, and moreover, we did not find any study reporting the quantitation of internal uterine activity with the use of intrauterine pressure monitor as an objective quantitation method of uterine contractility in patients with PTL. For these reasons, we did not consider the markers related to the measurement of uterine contraction as an independent risk factor of chorioamnionitis. However, we already performed the overall logistic regression analysis for the relationship between various independent variables and the development of chorio-deciduitis and amnionitis separating patients with PTL from those with preterm-PROM, and moreover, we already included "cervix dilatation at amniocentesis" in the overall logistic regression analysis, because cervix dilatation was likely to represent the result of uterine contractions and a previous report demonstrated that the more advanced the cervical dilatation, the higher the risk of acute-histologic chorioamnionitis [bib_ref] The risk of intra-amniotic infection, inflammation and histologic chorioamnionitis in term pregnant..., Lee [/bib_ref].
To our knowledge, this is the first study in humans showing that more time is required to develop histologic chorio-deciduitis with amnionitis than chorio-deciduitis alone in the setting of both PTL and preterm-PROM after documentation of sterile AF. We also show that prolongation of pregnancy is an independent predictor of both chorio-deciduitis and amnionitis in the setting of PTL. Further research is needed to better understand the spatio-temporal migration of ascending intra-uterine infection in the extra-placental membranes, and to determine the effect of pregnancy prolongation on neonatal outcome.
[fig] Fig 1: Frequency of inflammation in placental compartments according to amniocentesis-to-delivery interval. The frequency of chorio-deciduitis (A), amnionitis (B), funisitis (C), and chorionic plate inflammation (D) according to amniocentesis-to-delivery interval (i.e., 2 days, 2-7 days, and >7 days) is shown in patients with PTL and preterm-PROM. Frequency and P-values are shown. doi:10.1371/journal.pone.0143023.g001 [/fig]
[fig] Fig 2: Amniocentesis-to-delivery interval in accordance with histologic status of the extra-placental membranes. Amniocentesis-to-delivery interval is shown according to histologic status of the extra-placental membranes (i.e., inflammation-free, chorio-deciduitis only, and chorio-deciduitis with amnionitis) in patients with PTL (median, range; 20.6 hours [0.01-1794.8 hours] vs. 113.9 hours [0.10-1093.1 hours] vs. 654.5 hours [312.2-718.2 hours]) and preterm-PROM (median, range; 42.9 hours [0.01-1267.2 hours] vs. 95.2 hours [0.01-1833.9 hours] vs. 131.3 hours [90.2-655.0 hours]). P-values and Spearman rank correlation coefficients are shown. doi:10.1371/journal.pone.0143023.g002 [/fig]
[fig] Fig 3: Histopathology of extra-placental membranes (chorio-decidua and amnion) (These images are based on the magnification setting X200). [/fig]
[table] Table 3: Relationship of various independent variables with the development of chorio-deciduitis and amnionitis analyzed by overall logistic regression analysis in PTL with sterile amniotic fluid. [/table]
[table] Table 4: Relationship of various independent variables with the development of chorio-deciduitis and amnionitis analyzed by overall logistic regression analysis in preterm-PROM with sterile amniotic fluid. [/table]
[table] Table 5: Relationship between the latency after ROM and chorioamnionitis in previous studies. [/table]
|
Increasing secular trends in height and obesity in children with type 1 diabetes: JSGIT cohort
BackgroundRecently, anthropometric indices in children with type 1 diabetes mellitus (T1DM) have begun to change.ObjectiveTo examine secular trends in patients' anthropometric indices.SubjectsJapanese children with T1DM from the 1995, 2000, 2008 and 2013 cohorts of The Japanese Study Group of Insulin Therapy for Childhood and Adolescent Diabetes.MethodsWe analysed serum haemoglobin A1c (HbA1c) levels, the incidence of severe hypoglycaemic events, the types and doses of insulin, height standard deviation scores (SDS), body mass index (BMI) percentiles compared with healthy Japanese children and obesity prevalence over time. We also stratified the patients according to glycaemic control levels of <58 mmol/mol (optimal), 58-75 mmol/mol (suboptimal) and �75 mmol/mol (high-risk).ResultsData for 513-978 patients from each of the cohorts were analysed. The incidence of severe hypoglycaemic events decreased over time (from 21 to 4.8/100 patient-years), while the PLOS ONE PLOS ONE | https://doi.
[formula] a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 [/formula]
# Introduction
Recently, clinicians have noticed improvements in glycaemic control in patients with type 1 diabetes mellitus (T1DM). We previously reported a secular trend in glycaemic control improvement in Japanese patients. Although glycaemic control varies among countries, age groups and patients [bib_ref] Glycaemic control of Type 1 diabetes in clinical practice early in the..., Mcknight [/bib_ref] , recent studies have reported a decreased mean haemoglobin A1c (HbA1c) level in Japanese patients with T1DM of between 7.7% and 8.2% [bib_ref] Serum glucose, cholesterol and blood pressure levels in Japanese type 1 and..., Yokomichi [/bib_ref]. The 2007 International Society for Pediatric and Adolescent Diabetes (ISPAD) consensus guidelines management objectives indicated that unless there was severe hypoglycaemia and repeated hypoglycaemia, the target HbA1c for all children under the age of 18 was < 7.5%. The Japanese Society for Pediatric Endocrinology (JSPE) translated these guidelines into Japanese in 2008 [bib_ref] The International Society for Pediatric and Adolescent Diabetes, Rewers [/bib_ref]. The JSPE and the Japan Diabetes Society (JDS) recommended an HbA1c levels of < 7.5% for optimal glycaemic control since 2011, consistent with the ISPAD recommendations. The Japanese Study Group of Insulin Therapy for Childhood and Adolescent Diabetes (JSGIT) uses this target in Japanese children and adolescents with T1DM.
Given these new, lower target HbA1c levels, the major focus of T1DM treatment has shifted from decreasing the incidence of serum hypoglycaemic levels to preventing diabetic complications. These complications include hypoglycaemia, vascular complications and obesity. Disrupted growth and maturation are also critical complications for paediatric patients. Given the improvement in glycaemic control over time, paediatric diabetologists should be aware of secular changes in the growth and maturation of their patients.
In 2010, researchers in the USA reported temporal patterns in overweight and obesity in adult patients with T1DM, and suggested that the prevalence of overweight was increasing in this population [bib_ref] Temporal patterns in overweight and obesity in Type 1 diabetes, Conway [/bib_ref]. However, little information is available regarding anthropometric measurements in paediatric patients, and, in particular, in Japanese paediatric patients with T1DM. In this study, we analysed changes in height and body mass index (BMI) of cohorts of Japanese paediatric patients with T1DM. We also assessed changes in the prevalence of overweight and obesity, which could increase the risk of future cardiovascular complications.
# Materials and methods
## Participants
Data for paediatric patients with T1DM were obtained from the JSGIT cohorts [bib_ref] The Japanese Study Group of Insulin Therapy for Childhood and Adolescent Diabetes..., Matsuura [/bib_ref]. The JSGIT is the only multi-institutional joint research group in Japan, and the group was established in 1995 for patients with childhood-onset type 1 diabetes [bib_ref] The Japanese Study Group of Insulin Therapy for Childhood and Adolescent Diabetes..., Matsuura [/bib_ref]. The first cohort study began in 1995 and was designed to improve the treatment and glycaemic control of diabetic children in Japan. The number of participating institutions and registered patients is increasing cohort by cohort. The current study was a cross-sectional study performed using data from the entry year for the four cohort studies. We analysed patients' baseline data from cohorts recruited , and we excluded data from patients who had T1DM for less than 6 months. Patients were diagnosed with T1DM by their paediatricians according to the criteria established by the JDS and the American Diabetes Association [bib_ref] Report of the committee on the classification and diagnostic criteria of diabetes..., Seino [/bib_ref]. Participants' clinical characteristic are shown in [fig_ref] Table 1: Participants' characteristics [/fig_ref].
## Anthropometric indices
We recorded the height, weight and HbA1c levels for all of the patients, and we analysed data collected between July and October for each cohort. BMI was calculated as weight in kilograms divided by the square of the height in metres. We defined overweight and obesity as BMI in the 85th-94th percentile and BMI � 95th percentile [bib_ref] Prevalence of overweight and obesity in the United States, Ogden [/bib_ref] , respectively, according to the 2000 JSPE data for each age and sex. We also evaluated BMI standard deviation scores (SDSs) and height SDSs to compare the data from patients across all age groups. SDS values for controls were calculated using data from healthy Japanese children recorded in the National Survey in 2000.
## Other parameters
According to the ISPAD guidelines, we categorised patients into three groups according to their level of glycaemic control as follows: optimal: < 58 mmol/mol using Independent Federal Constitution Commission units (IFCC) (< 7.5% in the National Glycohemoglobin Standardization Program (NGSP) guidelines); suboptimal: 58-75 mmol/mol (7.5%-9.0%); and high-risk: � 75 mmol/mol (� 9.0%). We also recorded the incidence of severe hypoglycaemic events, as well as the types and doses of insulin formulations used. Severe hypoglycaemia was defined as an event associated with severe cognitive impairment, including coma and convulsions, requiring external assistance by another person [bib_ref] Assessment and management of hypoglycemia in children and adolescents with diabetes, Clarke [/bib_ref]. We categorised patients into four age groups; 5-6-year-old group, yearold group, according to sex. Pubertal stage was classified as follows: in boys, the 5-6-year-old and 10-11-year-old groups indicated the pre-pubertal stage; the 12-13-year-old group indicated the early pubertal stage and the 15-16-year-old group indicted the middle to the end of the pubertal stage. In girls, the 5-6-year-old group indicated the pre-pubertal stage; the 10-11-year-old group indicated the early pubertal stage; the 12-13-year-old group indicated the middle pubertal stage and the 15-16-year-old group indicated the end of the pubertal stage [bib_ref] Fifty years of child height and weight in Japan and South Korea:..., Cole [/bib_ref]. The data are presented as mean (standard deviation) or number (%). � P-values for the trends as determined by Cochran-Armitage or Jonckheere-Terpstra tests.
https://doi.org/10.1371/journal.pone.0242259.t001
# Statistical analysis
We described glycaemic control, the incidence of severe hypoglycaemic events, types and doses of insulin, the proportion of patients using bolus insulin at tea-time, height SDSs, BMI and percentile and the incidence of obesity and overweight over time. We also analysed secular change in these indices over time according to glycaemic control groups. Secular change in height SDSs and BMI percentile among boys and girls were analysed over time according to the four age groups: 5-6-year-old group, 10-11-year-old group, 12-13-year-old group and 15-16-year-old group. We used the Mann-Whitney U test to compare skewed data and the chi-square test to compare categorical data. We statistically evaluated secular trends by the Jonckheere-Terpstra test for continuous variables and the Cochran-Armitage test for categorical variables. We reported descriptive statistics as means and standard deviations (SDs). Two-sided probability values less than 0.05 were considered statistically significant. All statistical tests and descriptive analyses were performed using JMP software (version 13.0; SAS Institute Inc., Cary, NC, USA) and SAS statistical software (version 9.3; SAS Institute Inc., Cary, NC, USA).
# Ethics approval statement
This study was approved by the institutional review boards of the ethics committee of Saitama Medical University (no: 12-077) and School of Medicine, University of Yamanashi (no: 1817), and the study was conducted in accordance with the provisions of the Declaration of Helsinki. Written consent was obtained from all patients or their caregivers.
# Results
In total, 513, 685, 734 and 978 patients from 1995, 2000, 2008 and 2013, respectively, were included in the study. The percentage of boys in these cohorts was 41.7%, 37.4%, 39.9% and 41.3%, respectively, and the mean disease duration at baseline was 5-6 years [fig_ref] Table 1: Participants' characteristics [/fig_ref]. [fig_ref] Table 2: Secular changes in height SDSs and the prevalence of overweight and obesity... [/fig_ref] shows the mean HbA1c levels, dose and type of insulin, incidence of severe hypoglycaemic events, SDSs for height and BMI, BMI percentile and prevalence of obesity in the patients with T1DM selected for inclusion [fig_ref] Fig 1: Prevalence of obesity and overweight in each cohort [/fig_ref]. Over time, there were statistically significant decreases in HbA1c levels and the incidence of severe hypoglycaemic events and statistically significant increases in the proportion of patients who took bolus insulin at tea-time, as well as in height SDSs, BMI SDSs and the prevalence of overweight and obesity. [fig_ref] Table 3: Secular changes in height SDSs and BMI percentile according to the level... [/fig_ref] and [fig_ref] Fig 2: Secular changes in total daily insulin dose, proportion of obesity, height SDSs... [/fig_ref] show the secular changes in the means of height SDS, BMI percentile and the prevalence of obesity compared with the glycaemic control group. In all glycaemic control groups, height SDSs showed an increasing trend over time. In the suboptimal and high-risk glycaemic control groups, the BMI percentile and prevalence of obesity showed increasing trends. Total daily insulin dose showed an increasing trend in the order of high-risk control group, suboptimal control group and optimal control group [fig_ref] Fig 2: Secular changes in total daily insulin dose, proportion of obesity, height SDSs... [/fig_ref]. Height SDSs, BMI percentile and proportion of obesity between T1DM with or without using bolus insulin at tea-time did not differ significantly (−0.20 The proportions of girls were higher in each cohort, but there were no significant differences between the cohorts. A dominant proportion of girls in a cohort is a common characteristic in Japanese patients with T1DM.
According to the age groups, height SDSs and BMI percentile showed significant increasing trends in the 15-16-year-old group (P = 0.0003 and P = 0.014 in boys and P<0.0001 and P = 0.0055 in girls, respectively) [fig_ref] Fig 3: Secular changes in height SDSs and BMI percentile in the four age... [/fig_ref].
# Discussion
Our data suggest that height SDSs, BMI percentile and the prevalence of obesity have increased in Japanese paediatric patients with T1DM from 1995 to 2013. The height SDSs of these patients are currently similar to those of healthy Japanese children. These trends were accompanied by an increase in the use of insulin analogues and the use of bolus insulin at tea-time. A trend toward increased BMI percentile and obesity prevalence was observed in the suboptimal and high-risk glycaemic control groups.
Paediatric patients with T1DM have been reported to be shorter than healthy children even after initiating insulin therapy, and this phenomenon is particularly pronounced in patients with poor glycaemic control [bib_ref] Factors influencing height and weight development in children with diabetes: results of..., Danne [/bib_ref]. In patients with T1DM, peripheral insulin administration results in lower levels of portal insulin, which leads to higher concentrations of growth hormone binding proteins (GHBPs) than in healthy individuals. Higher levels of GHBPs, in turn, lower free insulin-like growth hormone-1 (IGF-1) concentrations. Thus, the impaired growth seen in children with T1DM may be due to low portal insulin levels. In our study, we observed increased height SDSs over time in all of the cohorts, as shown in [fig_ref] Table 2: Secular changes in height SDSs and the prevalence of overweight and obesity... [/fig_ref]. This suggests that, especially in the high-risk glycaemic control group, height can be increased by peripherally administering larger doses of insulin, which lead to higher levels of portal insulin.
## Comparison of the trends in the cohorts with those in the general population
The proportion of overweight children (weight > 20% over the standard weight adjusted for sex and height) in the Japanese general population increased from 1995 to 2003, but did not change from 2003 to 2013. The proportion of overweight and obese children with T1DM in this study's cohorts increased. The difference in the proportion of obesity from 2008 to 2013 might be associated with having diabetes. Considering puberty, the onset of puberty among Japanese children advanced from 1950 to 2010, but stabilised by 1970 [bib_ref] Fifty years of child height and weight in Japan and South Korea:..., Cole [/bib_ref]. Height SDSs, BMI percentile and the proportion of overweight and obese children with T1DM could not be evaluated using Japanese data for the year 2000.
Height SDSs and BMI percentile showed a significant increasing trend in the 15-16-yearold group for both boys and girls in the cohorts, which reflect increased height SDSs and BMI percentiles at the end of the pubertal stage.
## Shift to analogue insulins and increased total daily dose
The proportion of obese patients increased in the suboptimal and high-risk glycaemic control groups over time. Previous studies also reported an increase in BMI in paediatric patients with T1DM compared with healthy children [bib_ref] Temporal patterns in overweight and obesity in Type 1 diabetes, Conway [/bib_ref]. The increased use of insulin analogues, and a shift from NPH insulin to long-acting insulin and from regular insulin to rapid-acting insulin analogues, which is associated with a decreased risk of hypoglycaemic events, may account for the increase in BMI that we observed in our study. Although data regarding dietary intake were not available for the cohorts analysed in our study, we suspect that the high proportion of obesity observed in the suboptimal and high-risk glycaemic control groups may be due to excessive food intake and inappropriate insulin administration. Increased body fat in these The black lines and bars indicate the optimal control group, the grey lines and bars indicate the sub-optimal control group and the broken lines and dotted bars indicate the high-risk control group. BMI, body mass index; SDS, standard deviation score.
https://doi.org/10.1371/journal.pone.0242259.g002 children could increase their risk of cardiovascular disease in the future, and obesity in patients with T1DM increases their risk of developing metabolic syndrome and microvascular complications [bib_ref] Relationship between overweight and obesity with hospitalization for heart failure in 20,985..., Vestberg [/bib_ref] [bib_ref] Obesity is associated with retinopathy and macrovascular disease in type 1 diabetes, Price [/bib_ref] ; therefore, paediatricians need to pay attention to obesity risk in patients receiving insulin therapy.
The spread of analogue insulins may influence improved glycaemic control, reducing severe hypoglycaemia and improving anthropometric indices. However, in 2008 and 2013, almost all patients used analogue insulins, so the effects of these insulins could not be evaluated individually. Analogue insulin therapy was introduced in Japan in 2000, and increased patients' dietary options . The JSGIT intensively introduced basal bolus insulin therapy for patients with T1DM in 1993 [bib_ref] The Japanese Study Group of Insulin Therapy for Childhood and Adolescent Diabetes..., Matsuura [/bib_ref]. In Japan, rapid-acting insulin analogues were introduced in 2000, and long-acting insulin analogues were introduced in 2003. Furthermore, the pharmacokinetics of insulin analogues were suitable for basal bolus therapy in childhood patients. Rapid-acting insulin analogue (e.g., lispro) blood concentrations rise sharply and drop precipitously [28] without increasing the risk of severe hypoglycaemic events . The long-acting insulin analogues (e.g., glargine and detemir) enable diabetic patients to maintain peak-less basal insulin levels with fewer hypoglycaemic episodes . Combining glargine and detemir also enables patients to receive insulin while eating snacks. Although we did not assess patients' diets in our study, the increase that we observed in the use of bolus insulin at tea-time might suggest that a subset of the patients had excess caloric intake. In this study, although there was no information regarding the time of onset of severe hypoglycaemia, previous studies stated that hypoglycaemia in children occurred mostly as nocturnal hypoglycaemia, and that this was more common at a younger age . In addition, the incidence of severe hypoglycaemia has been reduced by continuous subcutaneous insulin infusion (CSII) therapy . As in our study, Yamamoto et al reported that the use of CSII increased in all age groups, especially in the 0-5-year age group . This might be due to a decrease in nocturnal hypoglycaemia because of the increased use of CSII in younger children and the shift from conventional insulin to insulin analogues. It is possible that positive attitudes towards treatment, including insulin therapy, progressed because of a reduced fear of hypoglycaemia.
Anthropometric indices between T1DM with or without using bolus insulin at tea-time did not differ significantly. This lack of difference might be because data for bolus insulin at teatime indicated only whether this occurred; no data were available for consistency of use and frequency during the day and dose.
There are limitations to this study. First, we were unable to track the outcomes of individual children longitudinally; however, we were able to identify secular trends in height SDSs, BMI percentile and the prevalence of obesity in the sample population. Furthermore, determining BMI percentiles enabled us to evaluate overall secular trends in addition to age-specific changes. Because T1DM is rare in Japan, it was difficult to compare increasing or decreasing trends in obesity with healthy children by age group. Second, we analysed data collected only during the summer; however, this approach enabled us to eliminate seasonal influences on the outcomes . Third, we were unable to assess the influence of puberty on patient outcomes. The onset of puberty was not assessed because we had no data to evaluate. However, height SDSs and BMI percentile indices were based on data for Japanese children with no information on the onset of puberty, and the indices included variations in the onset of puberty within the group. Therefore, these indices could be used without considering the evaluation of individual adolescence levels. Fourth, the possibility of selection bias cannot be ruled out. The number of participants increased, as did the number of institutes, over the course of the study period. Additionally, age and the duration of diabetes in the participants differed between the four cohorts. This may have contributed to the difference in age and duration of diabetes indicated found in participants in 1995 who were older than 6 years of age, to avoid including patients with monogenic diabetes [bib_ref] The Japanese Study Group of Insulin Therapy for Childhood and Adolescent Diabetes..., Matsuura [/bib_ref]. As the original institutes involved in the 1995 cohort were the core centres for paediatric diabetes, it might be undeniable that participants in those institutes were more difficult cases.
Our results suggest that the increased use of insulin analogues has reduced the risk of severe hypoglycaemic events and improved glycaemic control. However, the use of these insulins may also have resulted in excess food intake. Patients with T1DM, as well as type 2 diabetes, need to be educated about appropriate caloric intake, nutritional composition and BMI.
# Conclusions
Japanese paediatric patients with T1DM exhibited increasing trends in height SDSs and BMI percentile from 1995 to 2013. These results could be due to improvements in patients' conditions because of the availability of new treatment options. While glycaemic control has improved and the incidence of severe hypoglycaemia has decreased, the proportion of obesity has increased. Supporting information S1 File. STROBE statement-checklist of items that should be included in reports of observational studies.
[fig] Fig 1: Prevalence of obesity and overweight in each cohort. The black bars indicate the prevalence of obesity (� 95 th BMI percentile), and the white bars indicate overweight (85 th -95 th BMI percentile). https://doi.org/10.1371/journal.pone.0242259.g001 [/fig]
[fig] Fig 2: Secular changes in total daily insulin dose, proportion of obesity, height SDSs and BMI percentile according to the level of glycaemic control in Japanese children with type 1 diabetes mellitus. a) Mean total daily insulin dose (units/kg/day), b) Prevalence of obesity (%), c) Mean height SDS and d) Mean BMI percentile. [/fig]
[fig] Fig 3: Secular changes in height SDSs and BMI percentile in the four age groups. a) Mean height SDS in boys, b) Mean BMI percentile in boys, c) Mean height SDSs in girls and d) Mean BMI percentile in girls. SDS, standard deviation score; n.s., not significant; BMI, body mass index; P-value for the trend as determined by the Jonckheere-Terpstra test. a, P = 0.0499; b, P = 0.0003; c, P = 0.014; d, P<0.0001, e, P = 0.0055. https://doi.org/10.1371/journal.pone.0242259.g003 [/fig]
[table] Table 1: Participants' characteristics. [/table]
[table] Table 2: Secular changes in height SDSs and the prevalence of overweight and obesity in Japanese children with type 1 diabetes mellitus. P<0.001 vs. 1995 as determined by the Mann-Whitney U test. SD, standard deviation; 95% CI, 95% confidence interval; BMI, body mass index; SDS, standard deviation score; obese, BMI �95th percentile; overweight, BMI in the 85th-94th percentile; 0, no patients used insulin analogues or insulin analogues as a bolus or as basal therapy.https://doi.org/10.1371/journal.pone.0242259.t002 [/table]
[table] Table 3: Secular changes in height SDSs and BMI percentile according to the level of glycaemic control in Japanese children with type 1 diabetes mellitus. [/table]
[table] 27: Sä mann A, Lehmann T, Kloos C, Braun A, Hunger-Dathe W, Wolf G, et al. Flexible, intensive insulin therapy and dietary freedom in adolescents and young adults with Type 1 diabetes: a prospective implementation study. Diabetic medicine. 2008; 25(5):592-6. https://doi.org/10.1111/j.1464-5491.2008. 02406.x PMID: 18445173 28. Bolli GB, Di Marchi RD, Park GD, Pramming S, Koivisto VA. Insulin analogues and their potential in the management of diabetes mellitus. Diabetologia. 1999; 42(10):1151-67. https://doi.org/10.1007/ s001250051286 PMID: 10525654 29. Chase HP, Lockspeiser T, Peery B, Shepherd M, MacKenzie T, Anderson J, et al. The impact of the diabetes control and complications trial and humalog insulin on glycohemoglobin levels and severe hypoglycemia in type 1 diabetes. Diabetes Care. 2001; 24(3):430-4. https://doi.org/10.2337/diacare.24.3. 430 PMID: 11289463 [/table]
|
The Inhaler Technique Questionnaire (InTeQ): Development and Validation of a Brief Patient-Reported Measure
Citation: Lizano-Barrantes, C.; Garin, O.; Dima, A.L.; van Ganse, E.; de Bruin, M.; Belhassen, M.; Mayoral, K.; Pont, À.; Ferrer, M. The Inhaler Technique Questionnaire (InTeQ): Development and Validation of a Brief Patient-Reported Measure. Int. J.
# Introduction
Asthma is a common chronic respiratory conditionwhich causes a substantial burden of disease. Inhaled medications are the cornerstone of asthma management [bib_ref] A brief history of inhaled asthma therapy over the last fifty years, Crompton [/bib_ref] [bib_ref] Prescription of inhalers in asthma and COPD: Towards a rational, rapid and..., Dekhuijzen [/bib_ref] [bib_ref] Asthma phenotypes: The evolution from clinical to molecular approaches, Wenzel [/bib_ref] , but clinical evidence suggests that asthma control is often not achieved in practice. Poor inhaler technique is one of the main contributing factors, as it can result in suboptimal drug delivery and compromise treatment effectiveness [bib_ref] Device errors in asthma and COPD: Systematic literature review and meta-analysis, Chrystyn [/bib_ref] [bib_ref] Urging Europe to put non-adherence to inhaled respiratory medication higher on the..., Van Boven [/bib_ref] [bib_ref] Systematic review of association between critical errors in inhalation and health outcomes..., Kocks [/bib_ref] [bib_ref] Critical inhaler errors in asthma and COPD: A systematic review of impact..., Usmani [/bib_ref]. Besides its difficult acquisition, inhaler technique deteriorates over time [bib_ref] Repeated instruction on inhalation technique improves adherence to the therapeutic regimen in..., Takemura [/bib_ref] , with approximately 50% of patients failing to maintain it appropriately [bib_ref] Practice makes perfect: Self-reported adherence a positive marker of inhaler technique maintenance, Azzi [/bib_ref]. Therefore, it has to be assessed and improved repeatedly, particularly before considering any step-up in treatment when asthma remains uncontrolled.
Although the assessment of inhalation technique at every opportunity is recommended [bib_ref] Identifying Critical Errors: Addressing Inhaler Technique in the Context of Asthma Management, Bosnic-Anticevich [/bib_ref] , it is rarely performed systematically with patients after their first prescription of inhaled medication [bib_ref] Metered-dose inhaler technique: The effect of two educational interventions delivered in community..., Bosnic-Anticevich [/bib_ref]. Checklists are the most common, feasible, and accessible method with which to assess inhalation technique [bib_ref] Systematic review of association between critical errors in inhalation and health outcomes..., Kocks [/bib_ref] [bib_ref] Checklists for powder inhaler technique: A review and recommendations, Basheti [/bib_ref]. Most are intended for assessments performed by healthcare professionals [bib_ref] A systematic review of instruments aimed at evaluating metered-dose inhaler administration technique..., Rodríguez-Martínez [/bib_ref] with considerable variation [bib_ref] Systematic review of association between critical errors in inhalation and health outcomes..., Kocks [/bib_ref] [bib_ref] Checklists for powder inhaler technique: A review and recommendations, Basheti [/bib_ref] [bib_ref] Controller Inhalers: Overview of Devices, Instructions for Use, Errors, and Interventions to..., Gleeson [/bib_ref] in sources [bib_ref] Controller Inhalers: Overview of Devices, Instructions for Use, Errors, and Interventions to..., Gleeson [/bib_ref] , steps included (from 3 to [bib_ref] A multinational observational study identifying primary care patients at risk of overestimation..., Kritikos [/bib_ref] , and scoring systems [bib_ref] Systematic review of association between critical errors in inhalation and health outcomes..., Kocks [/bib_ref] [bib_ref] Checklists for powder inhaler technique: A review and recommendations, Basheti [/bib_ref].
We have found only two patient-reported questionnaires developed for inhaler technique assessment [bib_ref] Assessing Metered-Dose Inhaler Technique: Comparison of Observation vs, Erickson [/bib_ref] [bib_ref] Asthma and COPD Patients' Perception of Appropriate Metered-Dose Inhaler Technique, Ramadan [/bib_ref] , both specifically for metered-dose inhalers. One focuses mainly on the steps for device use, with nine questions [bib_ref] Assessing Metered-Dose Inhaler Technique: Comparison of Observation vs, Erickson [/bib_ref] , and the other comprises steps for device preparation and use and for spacer use, with twenty questions [bib_ref] Asthma and COPD Patients' Perception of Appropriate Metered-Dose Inhaler Technique, Ramadan [/bib_ref]. However, most studies ask patients to self-assess their inhaler technique or how confident they are about appropriate use with one [bib_ref] A multinational observational study identifying primary care patients at risk of overestimation..., Kritikos [/bib_ref] [bib_ref] Confidence in correct inhaler technique and its association with treatment adherence and..., Amin [/bib_ref] [bib_ref] Inhaler use and education characteristics among English and non-English speaking patients: A..., Kher [/bib_ref] [bib_ref] Inhaler devices in asthma and COPD-An assessment of inhaler technique and patient..., Chorão [/bib_ref] [bib_ref] Identifying patients at risk of poor asthma outcomes associated with making inhaler..., Barbara [/bib_ref] to three questions [bib_ref] Association between inhaler technique and confidence among hospitalized children with asthma, Litt [/bib_ref]. To our knowledge, there are no instruments designed to self-report the performance of the inhaler technique steps common to any type of device.
Therefore, we aimed to develop a new patient-reported tool for inhaler technique assessment, the InTeQ (Inhaler Technique Questionnaire), which could also serve as a simple training and self-monitoring resource for any type of inhaler device. We aimed to evaluate the feasibility, validity, and reliability of the InTeQ in adults with asthma.
# Materials and methods
## Development of the inhaler technique questionnaire (inteq)
Our target was to design a brief assessment questionnaire for patients with asthma to measure the quality of their inhaler technique in their daily living. The development was based on a literature review of existing instruments and pilot testing with clinicians and patients. It was developed in English and translated in parallel into French by a specialized company using forward translation, independent back translation, review with an investigator's input, and independent proofreading. The resulting English and French versions were pretested with 5 adult patients and 4 caregivers through cognitive interviews.
Although the ideal checklists are those standardized across devices as much as possible [bib_ref] Systematic review of association between critical errors in inhalation and health outcomes..., Kocks [/bib_ref] , and the metered-dose inhalers (MDI) have relatively similar recommendations for administration, the dry powder inhalers (DPI) have several designs and mechanisms for preparing the dose that slightly modify the administration requirements. Few common steps apply to both MDIs and DPIs, and there are some contrasting recommendations (e.g., shaking vigorously is important for MDIs, but detrimental for some DPIs). Checklists may include steps related to 3 overall processes: the preparation of the device/loading of the dose, the delivery of the dose, and the preparation of the device for storage [bib_ref] Checklists for powder inhaler technique: A review and recommendations, Basheti [/bib_ref]. For this questionnaire, we focused on key steps of the delivery of the dose process common across inhaler types.
Previous questionnaires [bib_ref] Assessing Metered-Dose Inhaler Technique: Comparison of Observation vs, Erickson [/bib_ref] [bib_ref] Asthma and COPD Patients' Perception of Appropriate Metered-Dose Inhaler Technique, Ramadan [/bib_ref] require patients to select the correct response option for each inhalation technique step. Choosing between focusing on correctness or on frequency was a major decision in the InTeQ design, which was finally based on information obtained in cognitive interviews. [fig_ref] Table 1: InTeQ items and answer options [/fig_ref] shows the description of a patient-interviewee's experience with using the inhalers, indicating that, even for a person who has received training, inhaler technique may vary. The patient has an overall perception that, at times, they may not be performing all steps at the optimum level. This suggests that inquiring about the frequency with which a patient performs the key steps of the inhaler technique (as opposed to whether they know or perform the various steps correctly or not) may be a useful approach. We conceptualized the inhaler technique in this questionnaire as the frequency with which a patient appropriately performs key steps when administering inhaled medication. This approach is consistent with new electronic devices for measuring the timing of administration and the quality of technique at each inhalation, which have shown association with positive clinical outcomes [bib_ref] A Method to Calculate Adherence to Inhaled Therapy that Reflects the Changes..., Sulaiman [/bib_ref] [bib_ref] A novel statistical method for assessing effective adherence to medication and calculating..., Greene [/bib_ref]. Moreover, our approach is in line with methodological recommendations for health behavior change assessment and intervention [bib_ref] Focusing on fidelity: Recommendations for improving intervention fidelity within trials of health..., Toomey [/bib_ref].
In the final version of the InTeQ, respondents were asked to report on how often they performed five key steps when using their inhaler on a 5-point Likert scale from "Never" to "Always", and a response option to indicate uncertainty ("Don't know"; see [fig_ref] Table 1: InTeQ items and answer options [/fig_ref]. The InTeQ was revised and modified iteratively within the ASTRO-LAB project [bib_ref] Assessment of the safety of long-acting β2 -agonists in routine asthma care:..., Van Ganse [/bib_ref] , with extensive pretesting until achieving this final version.
## Study design and participants
The ASTRO-LAB project has been described elsewhere [bib_ref] Assessment of the safety of long-acting β2 -agonists in routine asthma care:..., Van Ganse [/bib_ref] [bib_ref] Does adherence to inhaled corticosteroids predict asthma-related outcomes over time? A cohort..., Dima [/bib_ref] [bib_ref] Impact of asthma on women and men: Comparison with the general population..., Hernandez [/bib_ref]. Briefly, it was a prospective observational study conducted in France and the United Kingdom, designed to provide new evidence about the safety of long-acting β-agonists in routine clinical care. Inclusion criteria were: aged 6-40 years, ≥6 months of prescribed use of controller inhalers during the 12 months before enrolment, no history of omalizumab therapy and/or any other concomitant respiratory disease, no chronic oral corticosteroid use during the 3 months before enrolment, and no asthma exacerbations 2 months before enrolment.
We analyzed data from adult patients (18-40 years old) from the ASTRO-LAB project, which was approved by the ethics and regulatory boards of the participating centers and was conducted following the Declaration of the World Medical Association. Written informed consent was obtained from all participants before inclusion.
## Study variables
Participants were followed through online surveys at 12-month intervals, with computerassisted telephone interviews every 4 months and monthly text messages to detect severe asthma exacerbations. Demographic information and asthma severity markers were collected from the primary care records.
The yearly online survey included the InTeQ (inquiring about the previous 4 months), a question about the use of a spacer ("During the last 4 months, have you used your inhaler(s) with a spacer: always, often, sometimes, rarely, never or don't know?"), and two questions related to the support that patients received from healthcare professionals: "Looking at the care you have received in the last year, please indicate how much this topic has been discussed with you: (A) Make together a concrete plan of where, when, and how I need to use my inhalers, and (B) Teach me how to use my inhaler." These were answered with a numeric scale from 1 ('not at all') to 7 ('in a lot of detail').
Telephone interviews were used to collect patient-reported data on asthma treatment use and adherence with the Medication Intake Survey-Asthma, among others. The type of device for controller treatment was classified from manufacturers' technical information into DPIs, MDIs, and breath-actuated MDIs (BA-MDIs).
## Analytical strategy
The information analyzed for this study was from the first online survey, except for the assessment of reproducibility, in which follow-up data until twelve months were also used. Patients' characteristics were described by calculating percentages or means and standard deviations. The distribution of the InTeQ items was examined according to the reported use of reliever treatment, types of inhaler devices, and countries, and feasibility was evaluated through completion rate and data completeness.
We examined the InTeQ measurement model using Mokken scaling analysis (MSA), which assesses whether an item set orders respondents accurately on a continuum representing a latent trait [bib_ref] lavaan: An R Package for Structural Equation Modeling, Rosseel [/bib_ref]. We estimated the statistics of central tendency and dispersion of the items [bib_ref] Scale validation in applied health research: Tutorial for a 6-step R-based psychometrics..., Dima [/bib_ref] , inter-item correlations to identify any negative correlations, and multivariate outliers (Mahalanobis D 2 p < 0.001). Finally, unidimensionality was tested by examining homogeneity and by performing an automated selection procedure [bib_ref] Selection of Unidimensional Scales from a Multidimensional Item Bank in the Polytomous..., Hemker [/bib_ref] of items dichotomized into "Always" vs. the rest, due to their skewed distribution. Loevinger's homogeneity coefficient (H i between each item and the item set, and H between all items) thresholds are: 0.3-0.4 (weak), 0.4-0.5 (medium), and 0.5-1 (good). After confirming unidimensionality, a global score was defined as a variable that counts the number of the InTeQ items answered as "Always", ranging from 0 to 5 (the best inhaler technique). This global score orders patients along a continuum from systematically performing all 5 steps through to less frequent rigorous inhaler technique.
We evaluated construct validity by assessing the ability of the InTeQ items and its global score to discriminate among known groups defined by the support received from healthcare practitioners (inhaler technique "Not discussed/only in general" vs. "Discussed in detail") and by spacer use with MDI ("Always-Often-Sometimes" vs. "Rarely-Never"). The hypotheses raised a priori were: patients discussing in detail with healthcare professionals and those with MDI using a spacer "Rarely-Never" might have better inhaler technique. To assess the discrimination capacity of the InTeQ items, the five response options were collapsed into three categories (Always, Often-Sometimes, and Rarely-Never); and the InTeQ global score was categorized into: 4-5 (Good inhaler technique), 3 (Fair), and 1-2 (Poor). Differences were tested using Chi-square.
Internal consistency was estimated in the total sample with Cronbach's alpha coefficient. To evaluate reproducibility, we estimated the agreement between baseline and the 12-month follow-up in the subsample of stable participants (those reporting no exacerbations between both responses) with the kappa coefficient at item level and with the intraclass correlation coefficient (ICC) for the InTeQ global score. The difference between baseline and 12-month follow-up on the InTeQ global score was tested using paired Wilcoxon signed ranks test.
The prevalence of correct inhaler technique was displayed using an alluvial plot, an infographic that allows the representation of multiple pathways.
Of the 388 participants answering the online survey, 361 (93%) completed the InTeQ. Considering these and a Type I error of 0.05, the statistical power was 0.8to detect differences between known groups of 15% in the InTeQ items, with a kappa of 0.5 (SE 0.077) and a 95%CI of +/−0.15 in the subsample of stable participants.
# Results
Around 60% of the participants were women, and 75% were living in France [fig_ref] Table 2: Patients' characteristics at baseline [/fig_ref]. Almost all declared no hospitalizations in the 12 months before (96.6%), 71% used inhaled corticosteroids and long-acting β-agonists in fixed-dose combinations, and 64% used DPIs. Most participants (81.6%) never used a spacer, and very few (14%) reported having made a detailed inhaler plan. Data are presented as n (%) or mean (±SD); n = 361. [fig_ref] Figure 1: Distribution of InTeQ items [/fig_ref] shows that the frequency of items responded with "Always" in patients with frequent and non-frequent reliever treatment use (FRTU and non-FRTU) varied from 30% and 26% ('Breathe out slowly') to 72% and 86% ('Breathe in deeply'). For all items, less than 3% of participants answered "Don't know", and the proportion of missing items was less than 1%. The comparison of response distributions showed no statistically significant differences between FRTU and non-FRTU. French patients answered "Always" more frequently [fig_ref] Figure 2: Comparison of inhalation technique between French and British participants [/fig_ref] , with statistically significant differences for 'Breathe out fully before' and 'Hold breath after' (p = 0.006 and 0.001, respectively). The comparison of response distributions among inhaler device types [fig_ref] Figure 3: Comparison of inhalation technique between inhaler device types [/fig_ref] showed no statistically significant differences. [fig_ref] Table 3: InTeQ items descriptive statistics, inter-item correlations, and Loevinger's scalability coefficients [/fig_ref] shows that the five-point Likert response scale of the InTeQ items was skewed, with means in the range of 0. [fig_ref] Table 4: Validity of the InTeQ items and global score comparing known groups defined... [/fig_ref] shows the results of the InTeQ's validity based on the known groups. Statistically significant differences between the groups defined according to "Make an inhaler use plan together" were observed in only one InTeQ item (p = 0.049) and in the global score (p = 0.023). Patients that reported having "Discussed in detail" their inhaler technique with healthcare professionals answered more frequently having followed the steps "Always" than those who reported "Not discussed/only in general". Additionally, one InTeQ item presented a statistically significant difference on spacer use among patients with MDI (p = 0.035). The Cronbach's alpha coefficient estimated in the total sample was 0.716. [fig_ref] Table 5: Reproducibility of the InTeQ evaluated among stable participants at baseline and 12... [/fig_ref] shows the test-retest reproducibility results in the stable subsample. In total, 105 participants answered the InTeQ twice (at baseline and 12 months) and, after excluding 24 who suffered exacerbations, 81 patients were included in this subsample. Agreement ranged from 75.7% to 81.1%, and the kappa coefficient was moderate for four of the items (0.460-0.549). Global score change was not statistically significant and the intraclass correlation coefficient was 0.775. [fig_ref] Figure 4: Alluvial plot showing the prevalence of correct inhaler technique [/fig_ref] shows the prevalence of correct inhaler technique: 19.4% of patients reported "Always" performing the five steps and 10.9% did not report "Always" in any step. White vertical bars correspond to each InTeQ item; "A" indicates patients responding "Always"; "O" indicates patients responding other options ("Often-Sometime-Rarely-Never"). Gray bars on the right correspond to the number of items to which patients responded "Always": the lowest horizontal band is patients that did not respond "Always" to any item, while the top, dark-blue band is patients who responded "Always" to every item. Width of the horizontal bands is proportional to the flow of patients' responses.
# Discussion
To the best of our knowledge, this is the first study describing the development and metric properties of an instrument for self-reporting the frequency of adequately performing the key common steps of inhaler technique for any type of device. With the InTeQ, it was possible to estimate the actual prevalence of correct inhaler technique in real life. The InTeQ showed good feasibility, based on the high response rate and negligible missing data; good unidimensionality, which allows us to calculate a global score; good construct validity, based on the capacity of discriminating among known groups; and good reliability, based on internal consistency and test-retest reproducibility.
The selection of items for the InTeQ considered the five most important steps in the dose delivery process of the inhaler technique that were generic across devices. Four of the steps composing the InTeQ were among the five most common errors in inhaler use identified in a systematic review, and there is scientific evidence [bib_ref] Misuse of corticosteroid metered-dose inhaler is associated with decreased asthma stability, Giraud [/bib_ref] [bib_ref] Training issues in the use of inhalers, Duerden [/bib_ref] [bib_ref] Effect of incorrect use of dry powder inhalers on management of patients..., Lavorini [/bib_ref] supporting that the five steps selected for this review are the most common errors in inhaler use. The abovementioned systematic review reports 'no post-inhalation breath-hold' as the most frequent error for both types of devices, and 31% for overall prevalence of poor inhaler technique, both results that are consistent with our findings. Furthermore, the three InTeQ steps where more patients did not answer "Always" ('Breathe out slowly', 'Hold breath after', and 'Breathe out fully before') were among the 12 steps generic for all devices in the CRITIKAL study [bib_ref] Inhaler Errors in the CRITIKAL Study: Type, Frequency, and Association with Asthma..., Price [/bib_ref].
The CRITIKAL study showed association with uncontrolled asthma for four and three steps of the InTeQ in DPIs and MDIs, respectively [bib_ref] Inhaler Errors in the CRITIKAL Study: Type, Frequency, and Association with Asthma..., Price [/bib_ref] , since 'Breathe in deeply' did not in MDIs. Nonetheless, this step has been linked to an improvement in quality of life [bib_ref] Inadequate literacy is a barrier to asthma knowledge and self-care, Williams [/bib_ref]. It has to be noted that 'Breathe in deeply' involves insufficient respiratory effort for DPIs, and slow inhalation for MDIs. Finally, even though 'Breathe out after' was not related to uncontrolled asthma or exacerbations, it was among the most frequent errors in all devices [bib_ref] Inhaler Errors in the CRITIKAL Study: Type, Frequency, and Association with Asthma..., Price [/bib_ref]. Other inhaler technique errors associated with asthma outcomes that were not included in the InTeQ because they were specific for MDIs are: 'incorrect second dose preparation', 'poor coordination between the start of the inhalation and the actuation of the dose', 'exhaling into the mouthpiece', and 'not holding the inhaler upright'. Two generic errors associated with asthma outcomes in the CRITIKAL study which were not included in the InTeQ, 'not having the head tilted (chin slightly upward)' and 'not removing the cap', merit further research.
In the InTeQ, we chose the approach of patients rating the frequency of correct technique in each step because we prioritized this over the most usual option of measuring whether the patient knows the correct answer or not [bib_ref] Assessing Metered-Dose Inhaler Technique: Comparison of Observation vs, Erickson [/bib_ref] [bib_ref] Asthma and COPD Patients' Perception of Appropriate Metered-Dose Inhaler Technique, Ramadan [/bib_ref]. While being a prerequisite for a good daily inhaler technique, knowing the technique does not necessarily correspond with what patients actually do regularly. Other instruments measure patients' confidence about their inhaler technique, without considering the actual steps [bib_ref] A multinational observational study identifying primary care patients at risk of overestimation..., Kritikos [/bib_ref] [bib_ref] Confidence in correct inhaler technique and its association with treatment adherence and..., Amin [/bib_ref] [bib_ref] Inhaler use and education characteristics among English and non-English speaking patients: A..., Kher [/bib_ref] [bib_ref] Inhaler devices in asthma and COPD-An assessment of inhaler technique and patient..., Chorão [/bib_ref] [bib_ref] Identifying patients at risk of poor asthma outcomes associated with making inhaler..., Barbara [/bib_ref] [bib_ref] Association between inhaler technique and confidence among hospitalized children with asthma, Litt [/bib_ref]. Patients' confidence in the quality of their inhaler technique, in general, may not be a reliable indicator of their actual performance [bib_ref] Inhaler Technique in Asthma: How Does It Relate to Patients' Preferences and..., Jahedi [/bib_ref] because patients can overestimate their inhaler skills [bib_ref] Inhaler Technique in Asthma: How Does It Relate to Patients' Preferences and..., Jahedi [/bib_ref] [bib_ref] What can be done to impact respiratory inhaler misuse: Exploring the problem,..., Volerman [/bib_ref]. Thus, the focus of the InTeQ on the frequency of performing specific steps may make it a more suitable tool for training and self-monitoring. Repeated instructions on inhaler technique have proven to help achieve effective inhalation skillsand improve adherence to therapy and asthma outcomes [bib_ref] Repeated instruction on inhalation technique improves adherence to the therapeutic regimen in..., Takemura [/bib_ref] [bib_ref] Effect of novel inhaler technique reminder labels on the retention of inhaler..., Basheti [/bib_ref].
The Mokken analyses showed that the InTeQ can be considered to be unidimensional and that all the items measure a single underlying construct with good homogeneity. This implies that items can be used to order respondents according to their latent frequency of performing each step, thus justifying the construction of the global score. The high response rate and low proportion of missing values suggest an easy completion for a wide range of patients, thus indicating the feasibility of the InTeQ.
The order of InTeQ items according to the percentage of participants that answered "Always" did not vary among devices in our study, but patients using MDIs answered "Always" to a lower number of steps than patients using other devices. Although the dif-ference was not statistically significant in our study, this pattern is consistent with previous findings [bib_ref] Device errors in asthma and COPD: Systematic literature review and meta-analysis, Chrystyn [/bib_ref]. In a study that applied the self-reported inhaler technique questionnaire specific for MDIs with 20 questions [bib_ref] Asthma and COPD Patients' Perception of Appropriate Metered-Dose Inhaler Technique, Ramadan [/bib_ref] , 68% of the patients reported "exhaling before using their inhaler", and 77% reported "inhaling slowly and deeply". In our study, 49% and 75% of the patients using MDIs reported always performing these steps, respectively; the former is probably lower in the InTeQ because of the focus on frequency, which reduces the skewness.
A systematic review of pediatric inhaler technique instruments [bib_ref] A systematic review of instruments aimed at evaluating metered-dose inhaler administration technique..., Rodríguez-Martínez [/bib_ref] highlighted the lack of a construct validity assessment (1/24). None of the two prior patient-reported inhaler technique questionnaires assessed this [bib_ref] Assessing Metered-Dose Inhaler Technique: Comparison of Observation vs, Erickson [/bib_ref] [bib_ref] Asthma and COPD Patients' Perception of Appropriate Metered-Dose Inhaler Technique, Ramadan [/bib_ref]. The InTeQ was able to discriminate in the hypothesized direction among known groups based on the support received from healthcare professionals for inhaler technique, indicating the adequate construct validity of the questionnaire. There is evidence supporting the association between a lack of prior inhaler training and poor inhaler technique [bib_ref] Inhaler mishandling remains common in real life and is associated with reduced..., Melani [/bib_ref] [bib_ref] Predictors of Incorrect Inhalation Technique in Patients with Asthma or COPD: A..., Rootmensen [/bib_ref]. Although there are several factors associated with poor inhaler technique, such as old age [bib_ref] Inhalation technique and variables associated with misuse of conventional metered-dose inhalers and..., Melani [/bib_ref] , gender [bib_ref] Comparison of elderly people's technique in using two dry powder inhalers to..., Diggory [/bib_ref] , a low education level [bib_ref] Tailored Education May Reduce Health Literacy Disparities in Asthma Self-Management, Paasche-Orlow [/bib_ref] , disease-specific knowledge [bib_ref] Inadequate literacy is a barrier to asthma knowledge and self-care, Williams [/bib_ref] , and limited access to primary care [bib_ref] Inadequate outpatient medical therapy for patients with asthma admitted to two urban..., Hartert [/bib_ref] , we selected inhaler technique training as the most proximally related.
The InTeQ global score can be used for comparing groups, as both Cronbach's alpha and intraclass correlation coefficients were above the standard of 0.7 [bib_ref] Assessing health status and quality-of-life instruments: Attributes and review criteria, Lohr [/bib_ref] [bib_ref] A note on the use of the intraclass correlation coefficient in the..., Bland [/bib_ref] and reproducibility at item level was moderate [bib_ref] The Measurement of Observer Agreement for Categorical Data, Landis [/bib_ref]. The aforementioned patient-reported questionnaire with 20 items [bib_ref] Asthma and COPD Patients' Perception of Appropriate Metered-Dose Inhaler Technique, Ramadan [/bib_ref] also showed good reliability: Cronbach's alpha was 0.86 for the total survey score, and 0.94, 0.82, and 0.75 for the subscales. The abovementioned systematic review of pediatric inhaler technique instruments [bib_ref] A systematic review of instruments aimed at evaluating metered-dose inhaler administration technique..., Rodríguez-Martínez [/bib_ref] highlighted that reliability was only evaluated in a few studies (4/24), obtaining kappa coefficients for inter-rater agreement from moderate [bib_ref] Evaluation of inhaler device technique in caregivers of young children with asthma, Welch [/bib_ref] to almost perfect [bib_ref] Development of a scale to measure children's metered-dose inhaler and spacer technique, Boccuti [/bib_ref] [bib_ref] Provider demonstration and assessment of child device technique during pediatric asthma visits, Sleath [/bib_ref] and Cronbach's alpha coefficients from acceptable [bib_ref] An exploration of the skills needed for inhalation therapy in schoolchildren with..., Chen [/bib_ref] to excellent [bib_ref] Development of a scale to measure children's metered-dose inhaler and spacer technique, Boccuti [/bib_ref].
Some potential limitations of this study need to be considered. First, participants may have been affected by social desirability bias when answering the InTeQ. We cannot evaluate criterion validity because the ASTRO-LAB cohort did not include a gold standard measure for inhaler technique performance to limit participant burden. Probably the most suitable gold standard measure for the frequency of correct inhaler technique would be electronic devices added onto the inhalers, but this was found to be unfeasible in this project, and the cost effectiveness of devices is yet to be determined in routine use. Further research should include evaluating the concordance between the InTeQ and objective measurement using electronic devices. Second, since this cohort only included patients up to 40 years old, the results may not be generalized to older adults with asthma. Finally, the items included in the InTeQ are clinically relevant; even so, further work should focus on including all critical steps across inhaler device types, such as the abovementioned two generic errors associated with asthma outcomes in the CRITIKAL study [bib_ref] Inhaler Errors in the CRITIKAL Study: Type, Frequency, and Association with Asthma..., Price [/bib_ref] , which are not measured in the InTeQ.
# Conclusions
In conclusion, the InTeQ is a feasible, valid, and reliable instrument for self-reporting inhaler technique in any type of device. Patients' self-monitoring using the InTeQ has the potential to extend the benefit of prior training through the repetition of critical steps, and to help them become aware of under-performing some steps or not performing them as frequently as desired. It can be used by patients to self-monitor their inhaler technique between visits to the healthcare professional, by healthcare professionals to teach patients or other professionals, and by researchers to assess inhaler technique with an easy and accessible tool, which could allow further comparisons among studies. Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.
# Data availability statement:
The data presented in this study are available on request from the corresponding author.
[fig] Figure 1: Distribution of InTeQ items (frequency of performing step during the last 4 months). FRTU: frequent reliever treatment use; non-FRTU: non-frequent reliever treatment use. [/fig]
[fig] Figure 2: Comparison of inhalation technique between French and British participants. *: significant at p < 0.05 (0.008 and 0.001, respectively). [/fig]
[fig] Figure 3: Comparison of inhalation technique between inhaler device types. DPI: dry powder inhaler; MDI: metered-dose inhaler; BA-MDI: breath-actuated metered-dose inhaler; DPI + MDI: more than one type of device. [/fig]
[fig] Figure 4: Alluvial plot showing the prevalence of correct inhaler technique. [/fig]
[fig] Author: Contributions: All the authors have been actively involved in the study in different capacities. Conceptualization, C.L.-B., A.L.D., M.d.B. and E.v.G.; methodology, C.L.-B., O.G., A.L.D., M.d.B. and M.F.; validation, C.L.-B., O.G., A.L.D. and M.F.; formal analysis, À.P. and O.G.; investigation, O.G., M.F., A.L.D., E.v.G., M.d.B. and M.B.; resources, E.v.G. and M.F.; data curation, À.P. and O.G.; writing-original draft preparation, C.L.-B. and O.G.; writing-review and editing, C.L.-B., O.G., M.F., A.L.D., E.v.G., M.d.B., M.B. and K.M.; visualization, C.L.-B., O.G., M.F. and A.L.D.; supervision, M.F. and A.L.D.; project administration, E.v.G., A.L.D. and M.d.B.; funding acquisition, E.v.G., M.F. and C.L.-B. All authors have read and agreed to the published version of the manuscript. Funding: The research leading to these results has received funding from the European Community 7th Framework (FP7/2007-2013) under grant agreement number 282593, and the H2020 programme (MSCA-IF) under grant agreement number 706028 for A.L.D. during manuscript preparation. The following researchers have worked on this manuscript while funded by grants: C.L.-B. (University of Costa Rica OAICE-85-2019), K.M. (Instituto de Salud Carlos III FEDER: Fondo Europeo de Desarrollo Regional FI16/00071) and A.L.D. (IDEXLYON 16-IDEX-0005). The Health Services Research Group, IMIM, received funding from Generalitat de Catalunya (2017 SGR 452). Funding information for this article has been deposited with the Crossref Funder Registry. The funding agreements ensure the authors' independence in designing the study, interpreting the data, and writing and publishing the report. Institutional Review Board Statement: The study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics and Regulatory Boards in France (Comité consultatif sur le traitement de l'information en matière de recherche dans le domaine dela santé, Dossier N • 12702; and Commission Nationale d'Informatique et LiberteÂ, DR-2013-264) and United Kingdom (NRES Committee London-West London REC, reference 12/LO/2039). [/fig]
[table] Table 1: InTeQ items and answer options. Verbatim response regarding options, emerged during a cognitive interview "I think I know how to do it; the doctor asks me to do it in front of them every time I have an asthma clinic. Sometimes I do feel like it is not doing anything, so I do wonder. I take the blue inhaler with me and take a couple of puffs before exercising, and this really worked [ . . . ] because of actually paying attention on [ . . . ] how I use it [ . . . ].[But] for the morning and evening one, I take it rather quickly, and sometimes I wonder if I shouldn't actually focus more on how I do it, because sometimes I just take a tiny breath in, and sometimes I am tired or I just woke up and I don't have any breath capacity, so I do wonder if I do it properly. It's not because I don't know, it's just that it becomes more usual that I use it in a really quick way." [/table]
[table] Table 2: Patients' characteristics at baseline. [/table]
[table] Table 3: InTeQ items descriptive statistics, inter-item correlations, and Loevinger's scalability coefficients. [/table]
[table] Table 4: Validity of the InTeQ items and global score comparing known groups defined by the support received from healthcare practitioners and use of spacer. [/table]
[table] Table 5: Reproducibility of the InTeQ evaluated among stable participants at baseline and 12 months. [/table]
|
Prediction of organic homolytic bond dissociation enthalpies at near chemical accuracy with sub-second computational cost
Bond dissociation enthalpies (BDEs) of organic molecules play a fundamental role in determining chemical reactivity and selectivity. However, BDE computations at sufficiently high levels of quantum mechanical theory require substantial computing resources. In this paper, we develop a machine learning model capable of accurately predicting BDEs for organic molecules in a fraction of a second. We perform automated density functional theory (DFT) calculations at the M06-2X/def2-TZVP level of theory for 42,577 small organic molecules, resulting in 290,664 BDEs. A graph neural network trained on a subset of these results achieves a mean absolute error of 0.58 kcal mol −1 (vs DFT) for BDEs of unseen molecules. We further demonstrate the model on two applications: first, we rapidly and accurately predict major sites of hydrogen abstraction in the metabolism of drug-like molecules, and second, we determine the dominant molecular fragmentation pathways during soot formation.
N early all chemical reactions of organic compounds involve the breaking and formation of covalent bonds. Unsurprisingly, bond energies feature as an essential ingredient in many predictive models of chemical reactivity. Homolytic bond dissociation enthalpies (BDEs) are defined by the enthalpy change for the gas-phase reaction at 298 K:
[formula] A À B ! A Á þBÁ ð 1Þ [/formula]
The cumulative difference between BDE values of all bonds broken and formed in a chemical reaction thus provides an estimate of the overall reaction enthalpy. BDE values are thermodynamic quantities but they are also used widely to predict reaction kinetics. For example, BDE values are used to predict relative reaction rates using well-established Evans-Polanyi-type correlations with bond strengths in radical hydrogen atom abstractions 2 . BDEs also provide insight into thermodynamically accessible reaction mechanisms for a given compound, and their calculation is often the first step in characterizing dominant pathways in combustion 3 , polymer synthesis 4 and thermal stability, lignin depolymerization 7 , drug metabolism 8-10 , explosives, organic synthesis planning, and other applications to energy-related materials.
The accurate measurement and calculation of BDEs underlies numerous applications in organic chemistry. Experimental measurement of BDEs for polyatomic molecules are difficult, but a variety of techniques existwith a typical uncertainty of ±1-2 kcal mol −116 . Calculation of BDEs with ab initio quantum chemistry methods is possible, however, the choice of method is known to greatly affect the resulting computational accuracy. Despite this, density functional theory (DFT) computations using M06-2X and M05-2X functionals have been shown to achieve accuracies comparable to the uncertainties of the underlying experimental measurements. As a result, quantum mechanical (QM) methods play an integral role in calculating radical enthalpies and proposing reaction mechanisms. However, even relatively efficient QM methods such as DFT scale exponentially with basis set size, often taking hours or days to obtain a single BDE value. This conventional workflow requires the geometry of a reactant and its radical products to be optimized and the Hessian of each species evaluated. For flexible compounds this process must be repeated for several alternative conformations. The integration of BDE calculations in molecular design efforts, including quantitative structure-property relationship (QSPR) models, has thus been limited by these computational demands, and the use of BDE calculations for the screening of thousands or millions of candidate structures remains impractical. In this manuscript we describe a new computational workflow that overcomes these limitations.
The rise of machine learning (ML) in quantum chemistry has led to the development of highly-accurate empirical modelsthat have accelerated traditionally difficult QM calculations for predicting enthalpy, optoelectronic properties, and forces. In particular, the rise of graph neural networks (GNNs)in modeling chemical properties has enabled 'end-to-end' learning on molecular structure: a ML strategy where traditional feature engineering is replaced by feature learning from a graph-based molecular representation. These approaches have led to best-inclass prediction accuracies on a range of applications, especially as the amount of available training data grows. An open question in molecular machine learning is whether optimized 3D coordinates are required as inputs to the ML algorithm to reach optimal accuracies. For enthalpy prediction on the QM9 dataset, consisting of all small molecules satisfying known valence rules, 3D coordinates appear to lead to superior prediction performance. However, a recent study has shown that for some molecules and properties, 3D coordinates did not necessarily lead to improved results over more simple representations of 2D connectivity and atom types (i.e., SMILES 26 notation). In addition, while precise, absolute QM-derived atomization energies are often inaccurate by up to a full Hartree for common molecules (627 kcal mol −1 ). Direct prediction of reaction energies may therefore be more reliable when compared with experimental values.
For the prediction of BDEs, a previous study leveraged >12,000 DFT calculations and an associative neural network to achieve a mean absolute error (MAE) of 3.4 kcal mol −1 for unseen bonds relative to DFT results 28 . This model is based on fixed molecular descriptors calculated for each target bond, and thus does not allow the model to learn more detailed descriptions of each bond as more molecular structures and data is added. B3LYP values were used to train this model, however, this functional poorly captures the enthalpies of radical reactions. In our own benchmarking studies this level of theory has an average error 2 kcal mol −1 larger than other DFT methods against experimental BDE values (see below, . Other existing work has used neural networks to predict the contribution of each bond to the overall atomization energy of closed-shell molecules without explicitly calculating radical enthalpies. While this technique reproduces general trends in overall bond strength, quantitative comparison with experimental BDEs results in MAEs of~10 kcal mol −1 . More generally, the use of atomization energies as a benchmark for ML algorithms does not guarantee accuracy in predicting more chemically-relevant reaction energies. The development of an accurate ML pipeline to quickly estimate BDEs, with acceptable accuracy compared with experimental values, thus remains a challenge.
In this study, we develop A machine-Learning derived, Fast, Accurate Bond dissociation Enthalpy Tool (ALFABET) to predict homolytic BDEs at close to chemical accuracy with sub-second computational cost. To accomplish this, we first benchmark several quantum chemistry methods on a database of experimentally measured BDEs 33 , finding that the M06-2X/def2-TZVP level of theory has the optimal trade-off between empirical accuracy and computational efficiency. A database of 42,577 closed-shell compounds with nine or fewer heavy atoms and consisting only of C, H, O, and N atoms is then curated from PubChem 34 . Each single bond in the database that was not present in a ring is cleaved to yield two open-shell radicals. DFT enthalpy calculations are then performed on all open and closedshell molecules to yield 290,664 unique BDEs, representing over 80 days of total CPU time. We then train a graph neural network on a subset of these results, achieving a MAE of 0.58 kcal mol −1 when predicting BDEs for unseen closed-shell molecules (compared with DFT results). When compared against experimental values for large molecules not included in the training set, the ML method adds only 1 kcal mol −1 to the MAE of the DFT approach, while completing in less than a second (compared with over a day per molecule for DFT). The utility of the developed prediction tool is subsequently demonstrated on two separate applications where fast, accurate prediction of the weakest bond in a molecule is required. First, the model is used to rapidly and accurately predict the site of C-H oxidative degradation in large, drug-like molecules. The model replicates the results of much more expensive DFT calculations with an MAE of 1.14 kcal mol −1 , and 95% of metabolic sites occur at bonds within 2 kcal mol −1 of the weakest bond in the molecule. Second, the model is used to predict the dominant radicals formed during combustion of fuel molecules, and the identities of these radicals are used as features for a QSPR model of soot formation pathways. These applications demonstrate the broad applicability of the developed tool and demonstrate that bond strength prediction for organic molecules can be reliably performed using fast ML techniques.
# Results
Evaluation of QM methods for calculating homolytic BDEs. In order to ensure that the resulting ML method closely reproduced experimentally determined BDEs, we performed a benchmark study of common DFT and ab initio methods. Computed gasphase BDE values include unscaled vibrational zero-point energies and thermal corrections to the enthalpy at 298 K and 1 atm, using optimized geometries obtained following a conformational search (see below). For a set of 368 experimentally measured BDEs from the iBond database 33 , combinations of three different DFT functionals (B3LYP-D3, ωB97XD 37 , and M06-2X 38 ) and two basis sets (6-31G(d) and def2-TZVP) were compared with DLPNO-CCSD(T)/cc-pVTZ calculations . As expected, the CCSD(T) calculations took the longest to perform and were the most accurate. Of the DFT methods, the choice of basis set appeared to have the greatest impact on accuracy, with the M06-2X/def2-TZVP combination coming very close to CCSD (T) accuracy. MAEs of the three density functionals followed the order of B3LYP-D3 > ωB97XD > M06-2X for both basis sets. This is consistent with previous benchmarks against the stabilization energy of 43 radical species calculated using CCSD(T)/ CBS. The observed MAE of top performing methods approaches the underlying uncertainty in the experimental measurements.
Conformer sampling was performed using the RDKit library 41 , using the MMFF94s force field 42 . Between 100 and 1000 conformers were generated for each molecule, depending on the number of rotatable bonds. The lowest-energy conformer identified by force-field calculations was then used as an initial guess for subsequent geometry optimization at the higher level of theory. For radicals, initial structures were generated by temporarily replacing the radical with a bonded H atom during force field optimizations. The enthalpy of formation of this first conformer was denoted ΔH f ;0 . As a reordering of conformational energies often occurs upon reoptimizing MM geometries with a higher level of theory, we analyzed the typical error introduced by only optimizing the MM global minimum energy conformer at the higher level of theory. By optimizing additional higher-energy (i.e., local minima) MM conformers we can calculate the difference between our initial enthalpy estimation, ΔH f ;0 , and the Boltzmann-weighted enthalpy (at 298 K) of the entire conformer ensemble, hΔH f i. The difference between these quantities is plotted in , indicating that the median error introduced by only optimizing a single conformer (versus an ensemble of over 100) is only~0.5 kcal mol −1 , while requiring 1/ 100th the computational resources. We therefore proceeded with database construction at the M06-2X/def2-TZVP level of theory and the computational pipeline described above (and in more detail in the methods section), optimizing only the most stable MM conformer.
Construction of a machine-learning compatible BDE database. We next developed a large database of BDE values, BDE-db, on which to train ALFABET. To maximize the variety of bond strengths for a minimum computational effort, we limited the initial database construction to molecules with 10 or fewer heavy atoms. In addition, smaller molecules reduce the risk of the geometry optimization finding a local energy minimum substantially higher than the true global minimum.
Construction of BDE-db began with 42,557 parent C x H y O z N m molecules taken from the PubChem Compound database. Only neutral molecules with assigned CAS numbers were used during database construction. Each single, non-cyclic bond in these molecules was then cleaved to generate two child radicals which were also added to the database. Canonicalized SMILES strings with specified configuration at stereogenic centers were used to represent these molecules and remove duplicates. Child radicals were frequently the product of multiple BDE reactions, reducing the number of DFT calculations required. However, this use of the SMILES language presents some complications for database construction. Specifically, bond cleavage occurring within an enantiotopic or diastereotopic group (that are not differentiated by SMILES) forms radicals with a new and unspecified stereocenter in relation to the parent molecule. The creation of new diastereomeric relationships in the products gives rise to non-equivalent BDE values dependent upon the choice of relative configuration. Dissociations resulting in a new stereocenter were omitted from the database.
DFT calculations were then performed for the parent molecules and unique child radicals. A variety of convergence checks were performed to ensure the DFT optimization converged to a stable structure, including checks for imaginary frequencies and ensuring that the molecule did not further decompose into disconnected molecules (e.g., radical fragmentation of an alkoxyacyl radical into an alkyl radical by loss of CO 2 ) or suffer an intramolecular rearrangement (e.g., by a [1,n]-H Benchmark study of DFT methods. a Trade-off between accuracy (left axis, blue) and computational cost (right axis, orange) for a selection of common QM methods. M06-2X/def2-TZVP was selected for subsequent calculations. MAE and CPU time were averaged over 368 different bonds. b Effect of conformer sampling. Molecules were optimized with MMFF94s, and the lowest-energy conformers were used to initialize DFT calculations. The plot shows the difference between the Boltzmann average enthalpy for the entire ensemble and the DFT-calculated enthalpy of the first conformer as a function of the number of optimizations performed. Exhaustive conformer sampling only changes the median resulting enthalpies by <0.5 kcal mol −1 , with a relatively narrow inner-quartile range (IQR). NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-16201-z ARTICLE shift). Approximately 10% of attempted DFT calculations were discarded, primarily due to imaginary frequencies. A total of 249,374 successful calculations were used to build the BDE-db. These calculations resulted in 484,907 total calculated BDEs, of which 290,664 were unique (methane has only one unique BDE value). These numbers highlight the efficiency gains achieved through calculating a large database in parallel and reusing calculation results for child radicals, as typically three QM calculations are required per one BDE.
[formula] B3LYP-D3 6-31G(d) ω B97XD 6-31G(d) M062X 6-31G(d) B3LYP-D3 def2-TZVP ω B97XD def2tzvp [/formula]
Development of a graph neural network for predicting BDE. A graph neural network (GNN) was developed to predict BDE directly from molecular structure. GNNs in the past have been used to predict the enthalpy of molecules from their optimized 3D structure, with MAEs close to 0.3 kcal mol −122 . The application of this technique for the proposed target would require optimized 3D structures of both the parent molecule and child radicals, and prediction errors would likely compound when summing together three separate predictions. We instead sought to develop a model that only required the 2D structure (i.e., SMILES string) of the parent molecule as input. SMILES strings were converted to a graph representation using RDKit (with atoms as nodes and bonds as edges). Each bond in the molecule was represented by two directional edges, pointing in reverse directions between the two bonded atoms.
GNNs operate by mixing information between neighboring nodes and edges. By iteratively updating node and edge internal states depending on the internal states of their neighbors, embedding vectors are generated that serve as a finitedimensional description of each atom or bond's local environment. For BDE prediction, bond embedding vectors at the final layer are reduced through a linear layer to predict the BDE (predictions from both the forward and backward bond edge are averaged together). The overall network structure was inspired by a model from Jørgensen et al., but with a simplified interaction structure. As only 2D inputs are used, atom and bond vectors are initialized with embedding layers based on a number of properties inferred via RDKit. In each message passing layer, bond states are first updated with information from neighboring atoms, and atom states are then updated with information from neighboring bonds. Residual connections were used for each message passing layer in order to aid convergence of deeper models. Six message passing layers were used in the final model, as no improvement in accuracy was seen for additional layers. The final model structure contains 1.06 M parameters. Bond states from the final message passing layer are reduced to a single BDE prediction by passing them through a linear layer. Following SchNet 22 , these predictions were added to a single mean BDE value for each bond class to generate the final prediction. BDE predictions are therefore generated simultaneously for each bond in the input molecule.
Validation (dev) and test sets were each constructed from all BDEs associated with 1000 parent molecules. The training set thus consisted of 40,577 unique parent molecules and 276,717 unique BDEs. A learning curve for the model, comparing performance against the 1000 molecule dev set while varying the number of molecules in the training set, shows a linear log-log relationship. This trend suggests that model accuracies could be further improved through the collection of additional BDE data. Performance of the final model was tested against the held-out test set, consisting of 6948 unique BDEs. The MAE on these bonds was 0.58 kcal mol −1 (vs DFT), with 95% of predictions falling within 2.25 kcal mol −1 of their DFT-calculated values. A breakdown of the model's performance on each individual bond type is shown in. Since the goal of the method is ultimately to reproduce experimental BDE measurements, the speed and accuracy of the GNN in predicting experimental BDEs from the iBond database was compared with similar predictions generated via the DFT method, Supplementary Data 1). For molecules that were a part of the training set, the ML method achieves prediction accuracies versus experimental measurements that rival those of the DFT approach (2.4 kcal mol −1 for ML, 2.1 kcal mol −1 for DFT). These results compare favorably with previous ML predictions of BDE . However, a more difficult test of the ML approach is for molecules larger than 10 heavy atoms that were not a part of the training database. For these larger molecules, typical DFT calculations required more than a day per molecule. However, the accuracy of the ML method remained acceptable, adding <1 kcal mol −1 to the MAE of the DFT method (3.4 kcal mol −1 for ML, 2.5 kcal mol −1 for DFT) when compared against experimentally measured BDEs. For these molecules, ALFABET was able to predict BDEs for all the bonds in the molecule in under 1 ms per molecule.
Analysis of ALFABET prediction outliers. During construction of BDE-db and ALFABET, we conducted error analyses of preliminary data and models to refine the GNN structure and correct common DFT errors. In this section, we present a more extensive analysis of the remaining large prediction errors (>10 kcal mol −1 ) for bonds in the training, validation, and test sets , Supplementary, Supplementary Data 2). In evaluating errors in DFT and ML calculations, additional BDE calculations were performed at the composite G4 level of theory to serve as a ground-truth reference. G4 radical formation enthalpies lie close to experimental values (4.5-6.2 kJ mol −1 ), albeit at an increased computational cost relative to DFT.
ML predictions using deep neural networks have been criticized as being black-box in nature. However, in this study we use the bond embedding vectors from the final message passing layer to interpret the ALFABET predictions, generating a quantitative similarity score to bonds contained in the training database (see methods). These embeddings are subsequently reduced to a single BDE prediction, and thus neighboring bond The loss of stabilizing non-covalent interactions such as intramolecular hydrogen bonds by bond dissociation result in prediction errors . Relative to the internally H-bonded conformer 1a, the G4 BDE value is 90.8 kcal mol −1 . Our DFT reference value was correctly generated using this more stable conformation. However, ALFABET underpredicts this C-H bond strength by 15 kcal mol −1 , and is much closer to the hypothetical BDE value of 79.0 kcal mol −1 for the less stable conformer (1b) lacking an H-bond. We can attribute this prediction error to a failure to account for this strong H-bond in the parent compound. Inspection of nearest neighbor structures in the training database (including a similar bond for a 7-membered cycloheptanone) confirm this to be the case, since optimized structures for these molecules lacked internal H-bonds and have DFT values in the~80 kcal mol −1 range . For molecules where an intermolecular H-bond is lost or disrupted upon bond cleavage, predictions will tend to underestimate the true BDE value.
Conversely, the development of new stabilizing interactions in radical products result in anomalously low BDE values that are overestimated by ALFABET predictions . For example, the carboxyl radical formed from cis-3 undergoes ring-closure to form a stabilized radical that results in an anomalously small BDE value of 51.4 kcal mol −1 . While the DFT value lies close to this, the prediction is an overestimate by more than . b Bond from.
this value and the failure for cis-3 can be attributed to the occurrence of radical cyclization.
In constructing the BDE-db database, we omitted reactions where a bond dissociation resulted in an unstable radical that further decomposed into smaller species. While G4 calculations (which use uB3LYP/6-31G(2df,p) geometries) suggest that O-H dissociation of a carbamic acid group , results in the spontaneous loss of CO 2 , M06-2X calculations result in a weaklybound adduct with a N-C bond length of 1.63 Å. Relative to the G4 value, both DFT and ML predictions in this case are inaccurate.
Another scenario resulting in BDE prediction outliers arises from difficult-to-converge electronic structure calculations for strongly delocalized systems . The O-H BDE values for phenols 7 is predicted by ALFABET as 89.2 kcal mol −1 , whereas the reference DFT value is much higher at 108.3 kcal mol −1 . The G4 value is much closer to the predicted BDE and suggest that in this case, it is the DFT value that is erroneous. Indeed, phenolic O-H bonds of neighboring molecules in the database have similar BDEs to the predicted value and further indicate that the DFT result is the outlier . The overestimate by DFT results from the convergence of open-shell structures to an incorrect electronic state. We found this was sensitive to the input structure used for geometry optimization and difficult to filter automatically (calculations are fully converged with a stable wavefunction) without prior knowledge of an expected BDE value.
In general, the most egregious ML-DFT prediction errors arise for conformations or electronic structures atypical with respect to the rest of the training database. Inspection of neighboring BDE values is therefore a qualitative method of determining whether a given BDE prediction is trustworthy: BDEs with several, similar neighbors with consistent BDEs lends additional confidence that a prediction is valid. The ALFABET webtool therefore includes the option to search for neighboring bonds from the training dataset. Using 3D features as inputs to the ML model might alleviate some of these prediction errors, although this would increase the computational cost of the ML predictions (as 3D coordinates would be required to generate predictions) and the possibility would remain of passing sub-optimal 3D inputs to the ML model and generating correspondingly poor DFT predictions. Additional filtering of DFT results might allow more accurate ALFABET predictions. However, ML prediction methods will likely never be able to appropriately predict the results of medium-to long-range intramolecular interactions without sufficient training examples.
Application to bond dissociation in large molecules. We used ALFABET to predict the C-C, C-O, and C-H bonds in methyl linolenate, an unsaturated fatty acid methyl ester found in biodiesel. BDE values of biodiesel molecules are difficult to obtain experimentally and computational estimates are important for characterizing combustion chemistry, particularly the initial stages of pyrolysis. DFT BDE values have been obtained previously for methyl linolenate, in addition to multireference averaged coupled-pair functional (MRACPF2) values, which due to the large molecular size, were estimated using small surrogate models. The BDE values of weaker C-C and C-H bonds α-to the carbonyl and in allylic (and doubly-allylic) positions, along with those of stronger C(sp 2 )-C and C(sp 2 )-H bonds are all correctly described. This prediction, taking less than a second to complete, demonstrates the utility and accuracy of ALFABET for BDE prediction of larger, flexible hydrocarbons that are challenging to study by DFT and impossible for ab initio methods.
Application to prediction of major sites of drug oxidation. The main advantage of the proposed method is that, due to its computational speed, it can be used in forward screening applications where DFT calculations would be infeasible. We therefore demonstrate the method's applicability to two design challenges where BDEs play an important role in determining a molecule's suitability. The first application is the pharmaceutical development of drug molecules, where predicting how a compound is likely to be metabolized can reduce failure rates in clinical trials. Many xenobiotics are degraded by the cytochrome P450 enzyme, where the site of metabolism has been shown to correlate with the weakest C-H bond in the molecule.
Calculation of C-H BDEs in drug screening, however, is a computationally expensive task, and we thus determined whether ALFABET demonstrates similar accuracy to a DFT-based calculation approach. We constructed a database of 28 drugs and their sites of oxidative degradation. Drugs considered ranged in size from 6 to 32 heavy atoms. DFT calculations were then performed to determine the BDEs of all C-H bonds, and BDEs were also predicted using the developed GNN.
[formula] R 2 0.957 [/formula]
We then developed a site of metabolism classifier using the calculated BDEs. The weakest bonds in the molecule, within a certain energy tolerance, were predicted as possible targets for oxidation. The accuracy of the classifier, for BDEs derived both from DFT and from ALFABET, were quantified using a receiver operating characteristic (ROC) curve,. This curve plots the true positive rate versus the false positive rate as the classifier tolerance is adjusted. The area under the curve (AUC) of the ROC curve thus represents a quantitative measure of the classifier's performance, ranging from 0.5 (random guessing) to 1.0 (perfect predictions). The AUC for the DFT and ML-based classifiers was 0.86 and 0.87, respectively, indicating that the developed GNN is as accurate as DFT-based methods for predicting the site of metabolism, while requiring a fraction of the computational cost. In addition to an ROC curve, we also calculate precision and recall statistics for classifiers based on both DFT and ML bond strengths. Higher precision values indicate that the site of metabolism is present among only a few flagged candidate locations, while high recall values indicate the metabolic sites for most drugs are included among the predicted candidates. DFTderived bond strengths appear yield a slightly higher maximum precision for tolerances <1 kcal mol −1 , which likely represents the additional uncertainty imposed by the ML prediction. However, beyond this threshold precision and recall curves for both DFT and ML-derived bond strengths are similar, despite the substantially lower computational cost of ML. We note that our suggestions for the site of drug oxidation are only based on weakest bonds that do not explicitly account for accessibility of sites to the enzyme. These predictions could be further enhanced by incorporating accessibilities scores.
To verify that ALFABET predictions are accurate for BDEs of drug molecules much larger than those used to construct the training set, DFT calculations then performed for 82 top-selling drug molecules. These molecules ranged in size between 8 and 34 heavy atoms. Only H-atom BDEs were considered, resulting in 748 unique bonds broken. Despite only being trained on smaller molecules, the GNN successfully predicts the BDEs for much larger species, resulting in a MAE of 1.14 kcal mol −1.
Predicting combustion mechanisms from weakest bonds. In addition to metabolite decomposition, BDEs are essential in determining predominant combustion kinetic mechanisms. We next applied ALFABET to construct a mechanistically-inspired model of soot formation during combustion of new fuel chemistries. The yield sooting index (YSI) is an experimental measurement of the amount of soot a substance forms during combustion in a test flame, and is an important parameter to consider during selection of potential fuel blendstocks. While methods to predict YSI quickly from molecular structure exist, these models do not leverage recent mechanistic understandings of how soot formation proceeds. Specifically, formation and growth of polyaromatic hydrocarbons (PAHs), the main component of particulate matter, is governed by the recombination of radicals formed in the combustion process.
In this study, we use our newly developed ML approach to predict the weakest bond in each of a set of 217 different fuel molecules with measured YSI values. The identities of the two radicals that form are then used to construct a QSPR model to predict soot formation. Instead of a series of descriptors or functional groups, each molecule was represented by only two parameters: one for each of the two radicals formed during cleaving of the weakest bond. These parameters are shared between molecules that decompose to form identical radicals . Molecules were chosen such that each radical was the result of at least two molecule decompositions.
We performed a leave-one-out cross-validation to determine the ability of the model to predict YSI for unseen molecules. In each cross-validation fold, a single compound was removed from the dataset and a weighted least-squares regression (with data weighted by their experimental uncertainty) was performed on the remainder of the data. Fitted radical weights are then used to predict the YSI of the held-out molecule. The cross-validated predictive accuracy of the new model, based on ALFABET predictions, achieves a weighted least-squares loss less than half that of a recently developed group-contribution model on the same dataset . These results demonstrate that AFLABET predictions can improve forward screening approaches in which bond energy is an important parameter.
We further verified that ALFABET is accurate for larger molecules outside the training set considered in this application. For the 91 molecules with YSI measurements and between 11 and 20 heavy atoms, DFT calculations were performed to confirm the predicted BDEs. The resulting prediction error was even lower than for the withheld test set predictions , demonstrating the ability of the model to scale to larger molecules.
# Discussion
In this study, we have developed a ML prediction tool to quickly calculate homolytic BDEs for organic molecules containing C, H, O, and N atoms, at an accuracy comparable with state-of-the-art DFT approaches. An interface for the developed prediction tool is available online at https://ml.nrel.gov/bde. Because BDEs are intrinsic properties of covalently bonded molecules, their relative strengths are important parameters in a wide range of chemical studies. We therefore expect our tool to enable high-throughput and accurate development of novel compounds for applications where elemental compositions are restricted to C, H, N, and O atoms and critical properties are determined by the strengths of single, non-ring bonds. Beyond the application areas to drug design and combustion pathways considered in this paper, we expect our tool to be useful in understanding polymer thermal stability, lignin depolymerization pathways, explosives, and high-performance energy-related materials. Future work will expand the training database to include other elements, bond types, and bonds in rings. As has been shown in a recent study, transfer learning may also permit improved accuracies through the incorporation of BDEs from well-curated experimental results. While we have shown that high-accuracy CCSD(T) do not substantially improve accuracy over the chosen M06-2X method, databases of experimental bond dissociation energies do exist. However, careful selection and fitting of experimental data will be required, as experimental BDEs measurements are biased toward the weakest bonds a molecule and sometimes have high uncertainty. More broadly, this study demonstrates the potential for deep learning techniques to accelerate quantum mechanical investigations where high-throughput computations are possible but time-consuming. Future work will look to expand these approaches to transition state structures.
# Methods
Computational details for calculating homolytic BDEs. To sample radical conformations, H atoms were added to radical centers prior to MMFF structure optimization and removed afterward. MMFF94s performs well in conformational and non-covalent benchmarks involving neutral, closed-shell molecules 60 , however, it was not parametrized for radicals. Unrestricted Kohn-Sham DFT calculations of radicals were carried out with careful consideration of electronic structures because M06-2X showed less accurate results in some aromatic radicals. Specifically, spatial and spin symmetry of orbitals were broken by using the initial guess of mixed HOMO-LUMO with assuming no point-group symmetry of the structure. The stability of wavefunctions was also analyzed to confirm that the most stable electronic state had been found. Convergence to the wrong electronic state occurred most frequently for aromatic radicals. Gaussian 16was used for all DFT calculations with a default ultra-fine grid for all numerical integration and for the G4 calculations to analyze outliers. DLPNO-CCSD(T) calculations were carried out with ORCA 4.0 as a single-point energy correction to the B3LYP-D3/6-31G(d) enthalpy using optimized geometries from B3LYP-D3/6-31G(d).
All optimizations were checked for convergence to an energy minimum, which included checking for proper termination flags from Gaussian and ensuring the resulting structure had no imaginary vibrational frequencies. In addition, we verified that the molecule did not decompose into separate molecules during the Gaussian optimization by ensuring that all bond lengths (expected from the Lewis structure) were <0.4 Å plus the sum of the covalent radii of the participating atoms. Finally, statistical tests on the completed database were used to screen for molecules with abnormally large enthalpies. For a given chemical formula (i.e., elemental composition), a linear model was used to predict overall molecule enthalpy. If residuals from this linear fit were >3 inner-quartile ranges from the predicted enthalpy, the molecule was discarded. This step removed a handful of high-energy, hypothetical molecules or ones that converged to unreasonable geometries. The BDE-db dataset has been published in an open-source database available on Figshare 65 .
Graph neural network development. Determining the optimal inputs and structure to the GNN developed in this study was an iterative process in order to find one that yielded the lowest validation error. Nodes and edges were assigned to independent classes depending on a number of features. For nodes, unique classes were assigned based on an atom's symbol, chirality tag, aromaticity, presence in ring or ≥6), number of neighbors, and number of neighbor H's. Edge classes were assigned based on the start atom symbol, end atom symbol, and presence of the bond in ring . The edge interaction network and atom state updating layers from Jorgensen et al.were simplified by removing layers until losses began to increase, and residual connections were added to the end of each message passing layers while batch normalization layerswere added to the beginning of each message passing layer. The number of message passing layers was varied between 2 and 12, with validation losses not decreasing after six layers. Since the number of atoms for molecules in the training set was capped at nine, this allows messages to traverse the entire molecule except in a few select cases.
The loss function optimized the mean absolute error of all BDEs in the molecule, masking bonds for which DFT values were not available. Since edges in the model are directional, each bond has two corresponding edge states. During training, the BDE prediction of each directional edge is separately scored, while at test time the BDE prediction from both edges is averaged. The model was trained for 500 epochs using a batch size of 128 molecules with the ADAM optimizer using a learning rate of 1E−3 and a decay rate of 1E−5.
GNN implementation. GNN models were implemented using the Python nfp library (https://github.com/nrel/nfp), which provides extensions to the Keras deep learning framework for modeling graph-valued systems. Models were trained using a single Nvidia Tesla V100 GPU for~10-12 h.
Calculating neighboring bonds. Intermediate layers in the GNN could be used to search for similar bonds in the DFT database for a given query bond. Embedding vectors for all bonds with calculated BDE values were generated from the output of the final message passing layer, a 128-dimensional vector. For computational efficiency, these vectors were reduced to a 10-dimensional vector through a principal component analysis (PCA). A nearest-neighbors search was then used to find the 10 closest bonds in the BDE-db database. The scikit-learn library 67 was used to perform the PCA and nearest-neighbors searches.
## Data availability
The datasets generated and/or analyzed during the current study are available on figshare with the identifier https://doi.org/10.6084/m9.figshare.10248932.
## Code availability
Weights for the final trained model and python scripts to generate predictions for new molecules has been made available through a Github repository (https://github.com/ NREL/alfabet). Python scripts to train the model and Jupyter notebooks to create the figures in the paper are available at https://github.com/pstjohn/bde_model_methods.
Received: 24 November 2019; Accepted: 15 April 2020; |
Evaluating changes in the prevalence of female genital mutilation/cutting among 0-14 years old girls in Nigeria using data from multiple surveys: A novel Bayesian hierarchical spatio-temporal model
Female genital mutilation/cutting (FGM/C) is considered a public health and human rights concern, mainly concentrated in Africa, and has been targeted for elimination under the sustainable development goals. Interventions aimed at ending the practice often rely on data from household surveys which employ complex designs leading to outcomes that are not totally independent, thus requiring advanced statistical techniques. Combining data from multiple surveys within robust statistical framework holds promise to provide more precise estimates due to increased sample size, and accurately identify 'hotspots' and allow for assessment of changes over time. In this study, rich datasets from six (6) successive waves of the Nigeria Demographic and Health Surveys and Multiple Indicator Cluster Surveys undertaken between 2003 and 2016/17, were combined and analyzed in order to better assess changes in the likelihood and prevalence of FGM/C among 0-14-year old girls in Nigeria. We used Bayesian hierarchical regression models which explicitly accounted for the inherent spatial and temporal autocorrelations within the data while simultaneously adjusting for variations due to different survey methods and the effects of linear and nonlinear covariates. Parameters were estimated using Markov chain Mote Carlo techniques and model fit assessments were based on Deviance Information Criterion. Results show that prevalence of FGM/C among 0-14 years old girls in Nigeria varied over time and across geographical locations and peaked in 2008 with a shift from South to North. A girl was more likely to be cut if her mother was cut, supported FGM/C continuation, or had no higher education. The effects of mother's age, wealth and type of residence (urban-rural) were no longer significant in 2016. These results reflect the gains of interventions over the years, but PLOS ONE
# Introduction
Female genital mutilation/cutting (FGM/C) is recognised globally as a violation of the fundamental human rights of girls and women, which has no medical basis and could lead to severe health problems including increased risk of new born deaths. It is estimated that about 200 million girls and women alive today from over 30 countries mainly in Africa, the Middle East and Asia have been subjected to FGM/C and that approximately 3 million young girls are at risk of being cut each year. Consequences of the practice, which is mostly carried out on young girls under the age of 15 years, are well documented in the literature and range from shock to haemorrhage, and from difficulty in passing urine to inhibited orgasm [bib_ref] The ongoing violence against women: Female Genital Mutilation/Cutting, Muteshi [/bib_ref] [bib_ref] Female genital mutilation, Rymer [/bib_ref] [bib_ref] Female circumcision as a public health issue, Toubia [/bib_ref] [bib_ref] The risk of medical complications after female circumcision, Dirie [/bib_ref] [bib_ref] Sexual side effects of female genital mutilation/ cutting May Be type dependent:..., El-Naser [/bib_ref] [bib_ref] Sexual function in women with female genital mutilation, Alsibiani [/bib_ref] [bib_ref] Female genital mutilation and obstetric outcome: WHO collaborative prospective study in six..., Banks [/bib_ref].
In relation to Nigeria, a report by 28 Too Many in 2018 stated that about 20 million Nigerian women and girls have experienced FGM/C. Also, a recent study by Kandala et al estimated that the national prevalence of FGM/C among women (ages 15-49) stood at 18.4% in 2016/17 representing a decline of 11.2% from 29.6% in 2008, while the national prevalence of FGM/C among girls aged 0-14 years in Nigeria stood at 25.3% in 2016/17, representing a decline of 4.7% from 30.0% in 2008 [bib_ref] Female genital mutilation/cutting in Nigeria: Is the practice declining? A descriptive analysis..., Kandala [/bib_ref]. This implies that while there has been a sharp decline in prevalence among women aged 15-49 years, the practice seems to still hold sway over 0-14 years old girls in Nigeria despite several concerted efforts aimed at the total eradication of the practice.
In general, efforts to accelerate the abandonment of FGM/C in Nigeria have been a mix of legal /policy and advocacy interventions. In response to the international calls and in line with the Sustainable Development Goals (SGDs) Target 5.3 aimed at the eradication of all forms of harmful practices against women and children including FGM/C by the year 2030, the Nigerian government has passed a federal legislation, the Violence against Persons (Prohibition) Act 2015 (VAPP Act), which strongly prohibits FGM/C and other forms of genderbased violence in Nigeria with provisions including the prosecution of the perpetrators and reintegration of victims into society [bib_ref] Male and female viewpoints on female circumcision in Ekpeye, Rivers State, Briggs [/bib_ref]. This, in addition to other national policies such as the 2013/2017 National Policy and Plan for Action for Elimination of FGM/C in Nigeriaare initiatives aimed at bringing the practice of FGM/C in Nigeria to an end. In the same spirit, the civil society organizations (CSOs) have also continued to mobilize people against FGM/C through public awareness by forming a partnership with media and the civil society including traditional and religious leaders to disseminate anti-FGM/C messages at federal and state levels with the aim of turning cutters into anti-FGM/C campaigners [bib_ref] Female genital mutilation cutting in Nigeria: A scoping Review". Evidence, Mberu [/bib_ref].
Nevertheless, it should be noted that the VAPP Act has hardly been enforced in Nigeria [bib_ref] How to Transform a Social Norm, Unfpa-Unicef [/bib_ref] and its implementation varies across the 36 states of the country with some states (especially the high prevalent ones) yet to follow suit. In addition, where it exists, it is even harder to implement the anti-FGM/C law where there is limited presence of law enforcement agents, for example, in the rural areas. Also, the rate of reporting has been low because the perpetrators are almost always family members and it is even possible that the law enforcement agents may sometimes discharge such reports as a family or community matter aimed at preserving one's socio-cultural norms and decide not to meddle.
For the reasons of the variations in the level of commitments in stemming the practice of FGM/C at state level in Nigeria, and the compelling belief that FGM/C is deeply-rooted in cultural and social norms [bib_ref] Ending foot binding and infibulation: A convention account, Mackie [/bib_ref] [bib_ref] Female genital cutting: The beginning of the end, Mackie [/bib_ref] [bib_ref] Dynamics of change in the practice of female genital cutting in Senegambia:..., Shell-Duncan [/bib_ref] , it then becomes apparent that the need to further investigate trends in the practice as well as the roles of the underlying state structures and the effects of socio-cultural norms in the persistence of the practice in Nigeria especially among 0-14 years old girls using quality data can never be overemphasized. However, such data are mainly provided by household surveys, namely, the Demographic and Health Surveys (DHS) and the Multiple Indicator Cluster Surveys (MICS), which employ complex designs that are not completely independent thus necessitating the use of advanced statistical techniques.
Bayesian hierarchical regression models which explicitly account for the inherent spatial and temporal autocorrelation within data, while simultaneously controlling for other linear and non-linear effects in a statistically robust framework are popular in the areas of disease mapping and ecological studies, for identifying high risk geographical locations [bib_ref] Bayesian latent process spatiotemporal regression model for areal count data, Utazi [/bib_ref]. To date, only a few studies have explored this powerful statistical tool in the context of FGM/C [bib_ref] Spatial distribution of female genital mutilation in Nigeria, Kandala [/bib_ref] [bib_ref] Spatially modelling and mapping of female genital mutilation in Kenya, Achia [/bib_ref] [bib_ref] A Spatial Analysis of the Prevalence of Female Genital Mutilation/Cutting among 0-14-Year-Old..., Kandala [/bib_ref]. However, these studies have largely utilised data from only one single survey thereby providing only a 'snapshot' assessment. Combining data from multiple surveys within robust statistical framework holds promise to provide more precise estimates, narrower width of the credible interval of estimates and smaller standard errors due to increased sample size. We could then exploit the merits of each survey and more accurately identify 'hotspots' where the practice is still rife and assess changes over time.
The overarching aim of this study, therefore, is to assess how the prevalence and likelihood of female genital mutilation/cutting has changed over time with respect to the roles of sociocultural norms (operationalized in terms of a woman's FGM/C status and her support for the continuation of the practice), a girl's geographical location and other key determinants in the persistence of FGM/C among 0-14-year-old girls in Nigeria, and identify spatial patterns and 'hotspots'. We combined rich datasets from six [bib_ref] The risk of medical complications after female circumcision, Dirie [/bib_ref] successive waves of the Nigeria Demographic and Health Surveys (DHS) and Multiple Indicator Cluster Surveys (MICS) undertaken between 2003 and 2016 using Bayesian hierarchical regression models which explicitly accounted for the inherent spatial and temporal autocorrelations within the data while simultaneously adjusting for variations due to different survey methods and the effects of linear and non-linear covariates. The different years datasets were combined to simultaneously investigate spatial and temporal trends in the practice of female genital mutilation/cutting (FGM/C) across the Nigeria 36 states and the federal capital territory (FCT). In addition, by combining the various datasets, we gained more statistical power to better estimate parameters with higher precision. It is hoped that the statistical evidence generated from the study would serve to facilitate the development and implementation of tailored programmatic interventions that would ensure the total eradication of FGM/C in Nigeria.
Specifically, we seek to answer the following questions:
1. Which individual-and community-level factors are key drivers of FGM/C among 0-14 years old girls in Nigeria? . Then, all eligible women of reproductive age living in the selected households were interviewed. A brief description of the similarities in the survey methods adopted by the DHS and MICS is given in. Also, further details on the study design, sampling, data collection procedures and availability are found in https://dhsprogram.com/ and http://mics.unicef.org/ for DHS and MICS, respectively. From both the DHS and the MICS datasets, we extracted information on the respondents' daughters aged 0-14 years from the respondents' files and daughters' birth files. In all, the samples consisted of a combined total of 88,319 daughters of 51,141 women who completed the with consistently higher sample sizes. In addition, the data shows consistently higher proportion of cut women than girls across the survey years with a mutual peak in 2008 .
# Materials and methods
## Data sources
## Response and exposure variables
For the questions we address in this study, the primary outcome is the FGM/C status of the respondent's daughter aged 0-14 years, that is, whether the respondent's daughter has been cut or not. The outcome is, therefore, a binary variable taking values from the discrete set {0,1} such that a value of 1 indicates that the respondent's daughter has been cut while 0 indicates that the respondent's daughter was not cut. It should be noted, however, that these values only reflected the current status of the respondent's daughter as at the time of the surveys and do not represent their final FGM/C status. This is because a girl who was uncut as at the time data are collected may still be cut in the future.
Furthermore, to investigate the effects of both individual-and community-level covariates required to address the questions, we included measures of socio-demographic factors such as a woman's region and state of residence, type of place of residence (urban vs rural), age, ethnicity, wealth index, marital status and highest level of education attained [fig_ref] Table 1: Variable names, levels and descriptions [/fig_ref]. Factors of socio-cultural norms are operationalized by a woman's FGM/C status, her support for the continuation of the practice and her beliefs about FGM/C. For the Bayesian regression models, we used state identification numbers ID to capture the unobserved state-level effects of geographical locations, while the type of survey and year are both included to account for the effects of different types of surveys and survey years, respectively [fig_ref] Table 1: Variable names, levels and descriptions [/fig_ref].
Other variables examined are a girl's current age, girl's age at cutting, type of FGM/C, person who performed FGM/C; a woman's religion, belief on the reasons behind the practice of FGM/C; husband/Partner's educational level, household decision making power, and frequency of reading newspapers, listening to the radio or watching television.
Finally, for the purposes of examining spatio-temporal variations, two datasets were extracted, namely, the 2016 MICS and a pooled dataset comprising the 2003 to 2016 surveys data combined across common variables.
# Statistical analysis
Bivariate data analysis. To understand the pairwise association between a girl's FGM/C status and a set of baseline characteristics, we carried out weighted bivariate analyses of both the DHS and MICS datasets. The survey weights allowed us to ensure that our estimates came from samples which were nationally representative. These analyses were conducted in Stata version 14 using the svy commandand weighted outputs were cross tabulated and reported as FGM/C prevalence. Relevant covariates among those that are significantly associated with the response are then included in the Bayesian regression models.
Test for clustering. We examined the local and global clustering in the data using Moran's I test to evaluate the clustering pattern of FGM/C among 0-14 years old girls in Nigeria [bib_ref] Spatial statistical analysis and geographic information systems, Anselin [/bib_ref] [bib_ref] Beyond Moran's I: Testing for Spatial Dependence Based on the Spatial Autoregressive..., Hongfei [/bib_ref]. The key is to determine the geographical locations (states) with significant clustered, dispersed or random structure of the number of girls aged 0-14 years old who have undergone FGM/C. A significantly positive (negative) value of the Moran's I statistics indicates clustered (dispersed) outcome. These analyses were carried out in R statistical programming software version 3.6.1using moran.test() and moran.mc() functions available in the spdep package in R.
Bayesian hierarchical spatial and spatio-temporal modelling. The DHS and MICS data are inherently hierarchical and spatially and temporally autocorrelated largely due to the cluster sampling design adopted by the DHS and MICS. This, in addition to the fact that the values based on a single woman who had at least two children are not independent, justified the need for sophisticated statistical analytical approaches that explicitly allow for autocorrelated response in space and time while simultaneously accounting for linear and non-linear covariates and other potential sources of random errors in the data.
The response variable y i takes value from the set {0,1} such that y i = 1 if girl i aged 0-14 years has undergone FGM/C and 0, otherwise for i = 1,. . ., n, where n = 17529 for the 2016 data and n = 88319 for the pooled dataset. This implies that the y's are realizations from Bernoulli trials or y * Bernoulli(p), with the probability mass function (pmf) f(y; p) = p y (1 − p) 1-y with E[Yjp] = p and Var(Y|p) = p(1 − p), where p is the probability of success. In other words, p is the probability that a randomly selected Nigerian girl aged 0-14 years has been cut.
Our models follow a class of structured additive regression (STAR; [bib_ref] Generalized Structured Additive Regression Based on Bayesian P-Splines, Brezger [/bib_ref] [bib_ref] Structured Additive Regression for Multicategorical Space-Time Data: A Mixed Model Approach, Kneib [/bib_ref] [bib_ref] Geoadditive Models, Kamman [/bib_ref] models such that the response y depends on a set of covariates through a linear predictor η i linked to a function of its mean with a link function h(μ i ) given in Eq (1).
[formula] hðm i Þ ¼ Z i ¼ f 1 ðx i1 Þ þ � � � þ f c ðx ic Þ þ f spat ðs i Þ þ z 0 i oð1Þ [/formula]
where h(μ i ) is a logit link function; f 1 ,. . .,f c are the non-linear (not necessarily smooth) functions of continuous covariates x i1 ,. . .,x ic (e.g., mother's age, survey year); f spat (s i ) is the (nonparameteric) function of the the spatial covariate s i 2{1,. . .,S} corresponding to the consecutively numbered geographical locations for the 36 Nigerian states (state) and the FCT, that is, S = 37, and which accounts for the unobserved effects of geographical locations; and z i 's are categorical variables (e.g., Gender, Educational level, Wealth index, etc) with the coefficients vector ω.
According to the first law of Geography which states that "Everything else is related to each other but near objects are more similar than those further apart" [bib_ref] A computer movie simulating urban growth in the Detroit region, Tobler [/bib_ref] , it makes sense to assume that the geographical locations (states) that are near to each other (neighbouring states) are more similar thus with potentially autocorrelated response and it is no longer appropriate to use statistical models that are only valid when observations are independent. Consequently, we assume that observations that are further apart are independent and do not share common boundaries and characteristics thus spatially heterogenous and uncorrelated. As a result, to simultaneously account for the inherent spatial autocorrelation between states that are neighbours and the spatial independence between states that are further apart, we decompose, the total spatial effect f spat in (1) into a spatially correlated (structured) f str (.) and an uncorrelated (unstructured) funstr (.) effects as in Eq (2) below.
[formula] f spat ðs i Þ ¼ f str ðs i Þ þ f unstr ðs i Þð2Þ [/formula]
This decomposition allows us to explicitly account for and quantify the effects of spatial autocorrelation among neighbouring states and spatial heterogeneity among states that are not neighbours. For these models, all functions are centred on zero for identifiability.
The logit function assumed in (1) allows us to express our full model in terms of log-odds, log(p i /1 − p i ), as
[formula] logit p i ð Þ ¼ log p i 1 À p i � � ¼ Z i ¼ f str State i ð Þ þ f unstr State i ð Þ þ f Year ð Þ þ f Age ð Þ þ Residence þ Education þ � � � þ fgm woman þ Survey þ xð3Þ [/formula]
where the variables Residence, Education, and fgm woman are fixed effects for residence (ruralurban), mother's education and a mother's FGM/C status. The term Survey is a fixed effect accounting for variations due to differences in the different survey methods with DHS used as the reference survey. The term ξ (= State ID × Survey year) represents space-time interaction effects to account for any other source of variation that varies simultaneously in space and time.
Variants of the model described in [bib_ref] The ongoing violence against women: Female Genital Mutilation/Cutting, Muteshi [/bib_ref] are fitted to the 2016 MICS and the pooled 2003 to 2016 datasets within Bayesian statistical framework. These were implemented in R statistical programming software using the R2BayesX package [bib_ref] Structured Additive Regression Models: An R Interface to BayesX, Umlauf [/bib_ref] , the R interface of BayesX a popular statistical software for fitting various classes of generalized additive mixed models [bib_ref] Structured Additive Regression Models: An R Interface to BayesX, Umlauf [/bib_ref] [bib_ref] Flexible smoothing using B-splines and penalized likelihood (with comments and rejoinder), Eilers [/bib_ref]. In BayesX, the non-linear functions f j are modelled as a linear combination of basis functions as
[formula] f ðxÞ ¼ X M m¼1 b m B m ðxÞð4Þ [/formula]
where the basis function B m is known and the unknown vector of regression coefficients β = (β 1 ,. . .,β M )' are to be estimated such that in matrix form, we have
[formula] f j = X j β j , where X j is an n × M design matrix with elements X[i, m] = B m (x i ). [/formula]
The fixed effects parameters ω are given diffuse priors such that, π(ω j ) / constant. Also, a multiplicative normal prior is assumed for the unknown regression parameters such that
[formula] p b j jt j � � / 1 t 2 j ! rankðK j Þ 2 exp À 1 2t 2 j b 0 j K j b j !ð5Þ [/formula]
where K j correspond to the frequentist penalty matrix and τ j is a smoothing parameter. We assign Markov random fields (MRF;priors to the correlated spatial effect f str (s), s = 1,. . ., S, and exploit the neighbourhood structure of the MRF prior to 'borrow strength' from neighbours with more observations to estimate effects in neighbouring states where observations are sparse. Note that the MRF prior is the spatial extension of random walk models and is given by
[formula] f str s ð Þ j f str r ð Þ ; r 6 ¼ s � N X r�s f str ðrÞ N s ; t 2 str N s !ð6Þ [/formula]
where N s is the number of states that are contiguous (share boarder) to state s, and r * s denotes that state r is a neighbour of state s. Thus, the (conditional) mean of f str (s) is the average of functions f str (s) of the neighbouring states, where τ str is a variance parameter. In contrast, we assign a zero-mean independent and identically distributed Gaussian priors to the uncorrelated (unstructured) spatial effect f unstr (s) as
[formula] f unstr ðsÞ j t unstr � Nð0; t 2 unstr Þ;ð7Þ [/formula]
where τ unstr is a smoothing parameter. We modelled the interaction term ξ as a smooth function with a random walk of the second order (RW2) prior assuming autocorrelation in time and space. However, we also modelled the interaction term assuming the interaction in space and time are not autocorrelated such that x � Nð0; t 2 x Þ. Inverse gamma distributed hyperpriors are then assigned to both the variance and the smooth parameters such that pðt 2 j Þ � IGða j ; b j Þ, where j is a generic subscript representing for example, str, unstr, ξ and where a and b are hyperparameters.
To estimate the smooth functions, f 1 ,. . .,f p , we used cubic splines which are twice continuously differentiable piecewise cubic polynomials. The spline can be written as a linear combination of B-spline basis functions B m (x), the Bayesian version of the Penalized-Splines (P-Splines) proposed by Eilers & Marx [bib_ref] Flexible smoothing using B-splines and penalized likelihood (with comments and rejoinder), Eilers [/bib_ref] , such that f ðxÞ ¼ X l m¼1 b m B m ðxÞ. In our approach, this corresponds to 2nd order random walks given by
[formula] b m ¼ 2b mÀ 1 À b mÀ 2 þ m mð8Þ [/formula]
with Gaussian increments μ m~N (0, τ 2 ) which is estimated from data and where the smoothness parameter τ is also estimated from the data. To address our questions, we tested the following six (6) nested models specified in . For these models, models m 1 & m 2 tested the unadjusted effects of a girl's geographical location to her likelihood of being cut while adjusting for the variations due to different survey methods and the year of survey. On the other hand, with models m 3 to m 6 we simultaneously tested for the effects of geographical locations, social norms, mother's age and other key covariates on a girl's likelihood of being cut while adjusting for variations due to survey methods differences (m 4 ) including potential interactions in space and time (m 5 & m 6 ). Note that for the single 2016 MICS data, these models did not include adjustments for survey differences, survey year . Specifications of the model fitted to the datasets.
## Model specification
Remarks
[formula] m 1 fgm i * f str (state i ) + f unstr (state i ) + f(year) m 2 fgm i * f str (state i ) + f unstr (state i ) + f(year) + survey m 3 fgm i * f str (state i ) + f unstr (state i ) + Ethnicity i + Gender i + � � � + f(Age i ) + f(year) m 4 fgm i * f str (state i ) + f unstr (state i ) + Ethnicity i + Gender i + � � � + f(Age i ) + f(year) + survey m 5 fgm i * f str (state i ) + f unstr (state i ) + Ethnicity i + Gender + � � � + f(Age i ) + f(year) + survey + f(ξ) [/formula]
ξ modelled as a smooth function.
[formula] m 6 fgm i * f str (state i ) + f unstr (state i ) + Ethnicity i + Gender i + � � � + f(Age i ) + f(year) + survey + ξ, with x � Nð0; s 2 x Þ [/formula]
ξ modelled as random effect.
https://doi.org/10.1371/journal.pone.0246661.t002 and space-time interactions so that only two models representing unadjusted and adjusted (full) models are fitted on the 2016 MICS data. Posterior samples are drawn from the parameters space θ = ({f}, f unstr , f str , η) and hyperparameter space ϑ = (τ str , τ unstr , τ ξ ) via Markov chain Monte Carlo (MCMC; [bib_ref] Adaptive Rejection Sampling for Gibbs Sampling, Gilks [/bib_ref] simulation. Specifically, we used Metropolis-Hastings (M-H) updating steps with iteratively weighted least square (IWLS) proposal [bib_ref] Efficient Sampling from the posterior distribution in generalized linear models, Gamerman [/bib_ref].
For our study, 20,000 samples were simulated from the posterior distributions with the hyperparameters set to a = 1, b = 0.0005, that is, diffuse priors. The choice of hyperprior values for the variance parameters were informed by a rigorous sensitivity analysis. Bayesian inference was then based on the last 16000 samples obtained after discarding the first 4000 as burnin. The last 16000 samples were further thinned so that only every 20 th value was included among the values from which the posterior estimates were evaluated. Both the burn-in and thinning are used to minimize autocorrelations within the posterior samples.
In addition, we investigated the appropriateness of the Markov Random Field (MRF) priors through sensitivity analysis by fitting the spatial model using Gaussian Random Field (GRF; [bib_ref] Geoadditive Models, Kamman [/bib_ref] priors. However, we found no evidence of a better fit with the GRF. Besides, the sparseness introduced by the neighbourhood structure of the MRF offered a computational advantage and greatly reduced computational costs thus the models fitted with MRF priors were retained and further assessed.
Finally, we assessed how well the models fit the data and the best fit models were selected based on the Deviance Information Criterion (DIC; [bib_ref] Bayesian measures of model complexity and fit, Spiegelhalter [/bib_ref]. Results are presented as posterior odds ratios (POR), graphs and maps. Maps were produced in R statistical software version 4.0.0, while all boundary files (shapefiles) were freely available and downloaded from DIVA-GIS (www.diva-gis.org/gdata) with permission to publish obtained from Global Administrative Areas (GADM). With respect to handling missing values, the bivariate descriptive analysis using Stata and the Bayesian geo-additive multilevel regression models using R employed listwise deletion algorithm for observations with missing values. Note that women included in the analysis were only those with living daughters aged 0-14 years old and those who were able to provide information on their daughters' FGM/C statuses.
# Results
## Descriptive statistics
Results from the bivariate analysis of the baseline characteristics of the women and their daughters are presented in [fig_ref] Table 3: Characteristics surrounding female genital mutilation/cutting in Nigerian girls aged 0-14 from 2003-2017 [/fig_ref]. These results suggest that FGM/C prevalence among 0-14-year-old girls in Nigeria rose to its peak in 2008 with respect to most of the baseline characteristics (p < 0.0001) but showed a general decline in 2016. For example, in 2008, FGM/C prevalence was highest among daughters of older women (aged 30-49). However, in 2016 prevalence declined by 28.1% among daughters of women aged 45-49 years. In contrast, between 2008 and 2016, FGM/C prevalence rose by 14.9% among daughters of younger women aged 15-19 years. Other individual-level characteristics which showed significant association with a girl's FGM/C status are her ages at the time of the survey and at cutting (p < 0.0001). Most of the girls who were found to have undergone FGM/C were aged between 5-14 years as at the time of the surveys, while the vast majority of the girls (up to 97.7% in 2013) were cut before attaining their 5 th birthday with up to 81% of the cutting carried out by Traditional circumcisers in 2016. FGM/C prevalence was highest among daughters of women who alongside their husbands/partners had no formal education. Also, across household wealth indices we found overall decline in FGM/C prevalence between 2008 and 2016, except for the lowest and second indices, which increased by 16.2% and 0.4%, respectively. High FGM/C prevalence was associated with daughters of women who had undergone FGM/C, supported the continuation of the practice, and women who believed that FGM/C was required by religion and prevents premarital sex. Girls who lived in polygamous households and girls whose mothers are the key decision makers in their households in terms of the woman's health and spending, had highest prevalence of FGM/C. Before 2008, FGM/C was highest in South western and South eastern states in Nigeria, but there has been a shift to the North with the North-western zone accounting for more than half of the cut girls in 2016. Similarly, prior to 2008, prevalence of FGM/C was highest among girls who lived in the urban areas and among Yoruba and Igbo girls, but this has changed with the vast majority of the cut girls being rural dwellers and from Hausa ethnic group since 2008. Finally, the results show that the prevalence of FGM/C was highest among daughters of women who never read newspapers nor watched television, however, the nature of the association between how often a woman listens to radio and her daughter's FGM/C status was not clear.
## Spatial distribution of fgm/c across nigerian states and regions
The spatial distribution of FGM/C prevalence (crude prevalence) among 0-14 years old girls in Nigerian is presented in western state of Kano with prevalence in South-western states still high at around 60%. Another rather surprising outcome is the unexpected high FGM/C prevalence found in Katsina state in 2011 and which radically declined to almost 0% afterwards. It is not clear, why such unexpected variations in prevalence of FGM/C existed in Kano and Katsina states. However, it could be that prior to 2008, there were no meaningful survey coverage in the North and data collection in these states has improved ever since. How likely is this? Another potential reason could be due to the different survey methods used by the DHS and MICS and the different definitions of the fgm variables used by the surveys before being standardised in 2010. From 2008 to 2016, FGM/C prevalence remained high among the North-western states with Jigawa state having the highest prevalence in 2016, while a significant decline was found in the Southern states in the same period.and 6236.40, respectively. This indicates that the full model which has the smaller DIC value provided a better fit to the data. However, graphical results from both models are presented for comparison purposes where necessary. Therefore, in [fig_ref] Table 4: Posterior odds ratio estimates from the Bayesian geo-additive regression model [/fig_ref] , only the results based on m � 3 are presented. Daughters of women who did not have up to higher education had significantly higher likelihood of experiencing FGM/C. A girl who lived in the North west is about 6 times more likely to be cut than a girl who lived in the North central states of Nigeria. Also, likelihood of cutting was high among girls whose mothers were formerly married, whose mothers have undergone FGM/C and almost 14 times higher among girls whose mothers supported the continuation of FGM/C in Nigeria. Lower likelihood of FGM/C is associated with girls from other ethnic minorities, and there is no significant difference in likelihood of FGM/C found among girls who lived in households classified under different wealth quintiles suggesting that FGM/C is practised in Nigeria regardless of whether household is rich or poor. In addition, although the likelihood of FGM/C is lower among urban girls, we note that this is not significant suggesting that as of 2016, type of place of residence is not key to determining the likelihood of cutting a Nigerian girl aged 0-14 years.
## Bayesian hierarchical spatial models
Effects of geographical location. [fig_ref] Fig 4: Posterior estimates of the effects of a girl's likelihood of undergoing FGM/C... [/fig_ref] we show the posterior means (top panel) of the effects of geographical location on a girl's likelihood of being cut, and the corresponding maps testing the statistical significance of the estimates based on 95% credible interval (bottom panel) for the unadjusted model (m � 1 ;left) and the adjusted (m � 3 ;right) models fitted to the 2016 MICS dataset. For these maps, dark blue to dark red correspond to lowest risk to highest risk states. While black, white and grey correspond to significantly high-risk states, significantly low risk states, and the states in which the effects of geographical location were not statistically significant. The unadjusted maps show that significantly higher likelihood of cutting existed among girls who lived in the North western states of Zamfara, Kaduna, Kano and Jigawa; South western states of Ekiti, Oyo, Osun, and Ondo; North central state of Kwara and Plateau; South southern state of Edo and South eastern state of Imo. However, after taking account of the effects of other key covariates considered in the model [fig_ref] Table 4: Posterior odds ratio estimates from the Bayesian geo-additive regression model [/fig_ref] , only the effects of geographical location in Jigawa and Plateau states remained significantly high. This suggests that the key factors driving FGM/C among 0-14 years old in Nigeria are mainly state-specific or due to sharing boundary with high prevalent states (e.g., Jigawa and Kano).
Pooled 2003 to 2016 dataset. In this Section, we present the results from the Bayesian hierarchical models fitted to the pooled dataset. Unlike the 2016 MICS data, the pooled data allows us to assess changes over time and space and better estimate prevalence.
DIC. [fig_ref] Fig 5: Boxplot of the Deviance Information Criterion [/fig_ref] estimates of the Deviance Information Criterion (DIC) of the six (6) nested models specified in they adjusted for variations due to survey differences (m 2 ) or not (m 1 ) suggesting that the methods used by the DHS and MICS were not significantly different in a statistical perspective. Similarly, [fig_ref] Fig 5: Boxplot of the Deviance Information Criterion [/fig_ref] that the full models m 3 and m 4 which did not account for potential interactions in space and time, showed no difference in fit regardless of whether variations due to different survey methods were accounted for (m 4 ) or not (m 3 ). However, the overall best fit was offered by the full model (m 6 ) which modelled space-time interactions as a random effect, that is, x � Nð0; t 2 x Þ, suggesting that the changes in the spatial distribution of FGM/C prevalence among 0-14 years old girls in Nigeria happened independently in time or that the changes in prevalence over time happened regardless of the spatial location. Thus, for the pooled datasets only the results based on m 6 are presented and discussed further. However, for the purpose of comparisons, graphs and maps based on the unadjusted model which did not account for survey differences m 1 (simpler model) will be used where relevant. Posterior Odds Ratio (POR). In [fig_ref] Table 5: Posterior odds ratio estimates from the Bayesian geo-additive regression model [/fig_ref] , we present the results based on m 6 as posterior odds ratio (POR) and 95% credible intervals (95% CI). These results show the overall effects of the covariates across the years (2003 to 2016). We found that daughters of women who had no higher education were significantly more likely to be cut than other girls whose mothers had higher education. A girl whose mother had undergone FGM/C was more than 18 times more likely to be cut than a girl whose mother was not cut. Significantly lower risk of FGM/C was found among girls who lived in urban areas than in the rural areas. In support to earlier findings, we found that although the DHS data shows evidence of higher coverage, there were no For these maps, dark blue to dark red correspond to lowest risk to highest risk states. While black, white and grey correspond to significantly high-risk states, significantly low risk states, and the states in which the effects of geographical location were not statistically significant. Based on m 1 higher likelihood of cutting existed among girls who lived in the North western states of Zamfara, Kaduna, Kano and Jigawa; South western states of Ekiti, Oyo, Osun, and Ondo; North central state of Kwara; South southern state of Edo and South eastern states of Ebonyi, Enugu and Imo. However, after taking account of the effects of other key covariates considered in the model [fig_ref] Table 5: Posterior odds ratio estimates from the Bayesian geo-additive regression model [/fig_ref] , the effects of geographical location in Osun state, Zamfara state and all the South eastern states became non-significant suggesting that over the years FGM/C have been sustained in the North western and South western states largely due to state-specific/local factors.
Posterior estimates of FGM/C prevalence from 2003 to 2016. we present the predicted prevalence across the 36 states in Nigeria and the FCT based on the pooled data , calculated as the back transformed posterior estimate of the linear predictor defined in (3), such that p k = exp(η k ) /(1 + exp(η k )) averaged over all girls in location (state) k, for. k = 1,. . .,37. We quantified the uncertainties in the estimation of the posterior prevalence using the deviance maps calculated as
[formula] D ¼ 2 X y k log y k m � � þ n k À y k ð Þlog n k À y k n k Àm � � � �ð9Þ [/formula]
where Non-linear effects of mother's age. [fig_ref] Fig 9: Non-linear effects [/fig_ref] we compare the effects of mother's age based on the pooled dataset from 2003-2016 [fig_ref] Fig 9: Non-linear effects [/fig_ref] to that based on the 2016 MICS dataset [fig_ref] Fig 9: Non-linear effects [/fig_ref]. The combined effects of mother's age from 2003-2016 show that FGM/C prevalence was significantly high among girls whose mothers were more than 38 years old (9A). However, when the effects of mother's age were assessed using the recent 2016 MICS dataset, the effects of mother's age were no longer significant suggesting possible gains of interventions targeting older women.
# Discussion
This study evaluates changes in the prevalence of female genital mutilation/cutting (FGM/C) among 0-14 years old girls in Nigeria drawing upon data from multiple surveys. Specifically, we combined and analysed rich datasets from six (6) successive waves of nationally representative surveys comprising the Nigeria Demographic and Health Surveys (DHS) and Multiple Indicators Cluster Surveys (MICS) covering years 2003 to 2016/17. The objectives of the study were to provide more precise estimates of prevalence and better quantify the effects of the key determinants of FGM/C, and identify 'hotspots' where the practice is still rife to facilitate the development of more effective interventions. Four key questions were addressed: We examined the roles of key socio-economic and socio-demographic factors; the roles of social norms factors operationalised by a woman's FGM/C status, her beliefs, and her support for the continuation of the practice; the roles of a girl's geographical location in the persistence of the practice in Nigeria; and how changes in the practice varied over geographical locations and time. These questions were addressed using a combination of bivariate data analysis and a novel multivariate analysis which simultaneously adjusted for the inherent spatial and temporal autocorrelations within the data and the effects of other linear and non-linear covariates while accounting for potential variations due to the different survey methods, in a coherent Bayesian hierarchical spatio-temporal regression modelling framework.
Findings showed that FGM/C prevalence among 0-14-year-old girls in Nigeria varied geographically and temporally and has generally declined over the years in the South, but there has been a shift to the North since 2008. A girl was more likely to be cut if her mother was cut, supported FGM/C continuation, or had no higher education. Results from the bivariate analysis showed that most of the girls were cut before their 5 th birthday with more than 80% of the cutting carried out by traditional circumcisers in 2016. Results based on a pooled 2003-2016/ 17 dataset showed that daughters of older women (38+ years) and girls who lived in rural areas were significantly more likely to be cut than the daughters of younger women and girls who lived in urban areas. However, evidence based on the 2016 MICS data alone shows that the effects of mother's age and place of residence (rural-urban) are no longer significant suggesting possible gains of interventions that targeted older women and women who lived in the rural areas. Higher likelihood of undergoing FGM/C were found among girls whose mothers believed that FGM/C was a religious obligation and prevents premarital sex. Furthermore, results suggest that the geographical location in which a girl lives matters. We found that over the years, the effects of geographical location were significantly high in most Southern and Northern states with girls who lived in the Northern states of Jigawa and Plateau having significantly higher likelihood of being cut in 2016. This implies that the practice of FGM/C in these states is being sustained largely due to some state-specific factors such as shared socio-cultural norms or due to their proximity to high prevalent neighbouring states (e.g., Kano and Jigawa states). These results agreed with findings from previous studies [bib_ref] Female genital mutilation/cutting in Nigeria: Is the practice declining? A descriptive analysis..., Kandala [/bib_ref] [bib_ref] Female genital mutilation cutting in Nigeria: A scoping Review". Evidence, Mberu [/bib_ref] [bib_ref] Spatial distribution of female genital mutilation in Nigeria, Kandala [/bib_ref] [bib_ref] Spatially modelling and mapping of female genital mutilation in Kenya, Achia [/bib_ref] [bib_ref] A Spatial Analysis of the Prevalence of Female Genital Mutilation/Cutting among 0-14-Year-Old..., Kandala [/bib_ref] [bib_ref] Understanding the association between parental attitudes and the practice of female genital..., Cappa [/bib_ref] [bib_ref] Mother to daughter transmission of Female Genital Cutting in Egypt, Farina [/bib_ref] [bib_ref] The School Nurse's Role in Addressing Female Genital Mutilation, Nowak [/bib_ref]. For example, it was found that education plays an important role in a mother's decision on whether her daughter will be cut or not [bib_ref] Spatial distribution of female genital mutilation in Nigeria, Kandala [/bib_ref] , with mother's who had higher education less likely to cut their daughters in Nigeria [bib_ref] Female genital mutilation/cutting in Nigeria: Is the practice declining? A descriptive analysis..., Kandala [/bib_ref]. This inverse relationship between a Nigerian woman's level of education and the likelihood of FGM/C was also reported in a scoping review of the determinants of FGM/C in Nigeria by Mberu in 2017 [bib_ref] Female genital mutilation cutting in Nigeria: A scoping Review". Evidence, Mberu [/bib_ref]. Also, a recent study by Cappa et al on the roles of parental attitudes on the likelihood of their daughter's FGM/C suggested that the opinions of both parents matter [bib_ref] Understanding the association between parental attitudes and the practice of female genital..., Cappa [/bib_ref]. They found higher likelihood of FGM/C among African girls (including Nigeria) aged 0-14 whose mothers wanted FGM/C to continue supporting the assertion that FGM/C is a social norm. Similar results were found in the context of other countries [bib_ref] Spatially modelling and mapping of female genital mutilation in Kenya, Achia [/bib_ref] [bib_ref] Mother to daughter transmission of Female Genital Cutting in Egypt, Farina [/bib_ref]. A study conducted in Senegal, Burkina Faso and Egypt by Farina and Ortensi found higher FGM/C prevalence among girls whose mothers had undergone FGM/C [bib_ref] Mother to daughter transmission of Female Genital Cutting in Egypt, Farina [/bib_ref]. In the same spirit, studies conducted in the context of Kenya found that a woman's FGM/C status was key to her support for the continuation of the practice [bib_ref] Spatially modelling and mapping of female genital mutilation in Kenya, Achia [/bib_ref] , and that daughters of women who were cut, women who supported FGM/C continuation and women who believed that FGM/C was a religious obligation, were more likely to be cut than their peers [bib_ref] A Spatial Analysis of the Prevalence of Female Genital Mutilation/Cutting among 0-14-Year-Old..., Kandala [/bib_ref].
It should be noted, however, that majority of the studies mentioned above focused largely on the roles of individual-and community-level factors and the effects of the geographical locations in which a girl lives in Nigeria has only been considered in Kandala et al [bib_ref] Spatial distribution of female genital mutilation in Nigeria, Kandala [/bib_ref]. In addition, studies have mostly relied on single survey and single year data thus providing only a 'snapshot' evidence, while the need to understand the trends in the practice on a larger spatial and temporal scale has largely been overlooked. To the best of our knowledge, this is the first attempt to evaluate changes in FGM/C using data from multiple surveys over multiple number of years while accounting for uncertainties due to the different survey methods in a robust Bayesian hierarchical modelling framework. With this novel approach, we were able to exploit the strengths of each survey to examine the temporal trends in the practice, better quantify the effects of individual-and community-level factors, more accurately estimate prevalence and identify spatial patterns and 'hotspots' where the practice is still rife. The identified 'hotspots' included the North western states of Kaduna, Kano and Jigawa and the South western state of Ekiti thereby raising questions on the effectiveness of the intervention measures within these states.
The regional shifts in FGM/C prevalence are noteworthy as most of the intervention programmes have focused on high prevalence states in the South, with more emphasis laid on the education of the general public [bib_ref] Female genital mutilation cutting in Nigeria: A scoping Review". Evidence, Mberu [/bib_ref]. The North may therefore not have experienced the same FGM/C declines as the South because interventions have largely been focused on hotspots, which were thought to be only in the South based on previously available data examined crosssectionally.
Nevertheless, the regional variances in FGM/C prevalence may stem from differences in the adoption of the anti-FGM/C laws. Although Nigeria enacted the 2015 Violence Against Persons (Prohibition) Act (2015 VAPP Act), which criminalises the practice of FGM/C in Nigeria, some Nigerian communities have continued to practice FGM/C unabatedly. As a federal law, the Act only applies in the Federal Capital Territory Abuja while states have to either adopt the law or create their own anti-FGM/C laws. Some states had their own anti-FGM/C laws prior to the passing of the VAPP Act. However, not all states have these laws in place and even where the law exists, the implementation may suffer setbacks due to lack of reporting probably due to family members involvement in the perpetration of the practiceand the limited presence of law enforcement agents, for example, in the rural areas.
Furthermore, the observed higher prevalence of FGM/C in the North since 2008 does not necessarily mean that the practice has recently been adopted. It is possible that data on FGM/C in the North have been better collected in subsequent surveys. Thus, it might be that we are only just recently getting a clearer but not necessarily a new picture of FGM/C practice in the North. In addition, it is possible that these differences were a result of the different strategies used by the DHS and MICS in data collection before being standardised in 2010. However, we accounted for this potential source of error by adding an extra term 'rvey' in the pooled data model to control for potential variations due to different survey methods, thus providing a clearer picture of the patterns of changes over the years.
In general, FGM/C is often viewed as a requirement that prepares a girl for marriage, inhibits urge for premarital sex and promotes marital fidelity. Several theories have been advanced to explain why the practice of FGM/C persists despite sustained efforts geared towards its elimination. Prominent among these theories is the social norms theory which postulates that several decisions made by individuals are influenced by social norms within their community, and that FGM/C persists because those who practise it share mutual expectations (empirical and normative) with those in their reference group [bib_ref] Ending foot binding and infibulation: A convention account, Mackie [/bib_ref] [bib_ref] Female genital cutting: The beginning of the end, Mackie [/bib_ref] [bib_ref] Dynamics of change in the practice of female genital cutting in Senegambia:..., Shell-Duncan [/bib_ref]. Thus, an individual's preferences and attitudes towards FGM/C ought to be understood within the context of the expectations and pressures that others in their reference group put on them [bib_ref] Dynamics of change in the practice of female genital cutting in Senegambia:..., Shell-Duncan [/bib_ref]. As a result, an individual family is left to choose between the costs and benefits of adherence or nonadherence to social norms. Benefits may include enhanced marriageability and acceptance into peer social networks [bib_ref] Female genital cutting: The beginning of the end, Mackie [/bib_ref] while costs may include social exclusion, persecution or ostracization. FGM/C is therefore considered a social norm because it is hard for an individual family to forgo it in isolation as long as most community members do not assent to abandoning it. Consequently, it has been suggested that programmes designed to promote FGM/C abandonment should approach it in terms of social norms theory [bib_ref] Ending foot binding and infibulation: A convention account, Mackie [/bib_ref] [bib_ref] Female genital cutting: The beginning of the end, Mackie [/bib_ref].
Two major limitations of the study were identified. The first one is related to the use of selfreported data on FGM/C without clinical confirmation, which may lead to biased estimates. Although, a previous study on FGM/C in Tanzania suggests that some respondents make false self-reports that they have been cut due to social pressure to be circumcised [bib_ref] Dynamics of change in the practice of female genital cutting in Senegambia:..., Shell-Duncan [/bib_ref] , it is unlikely that this is always true. Nevertheless, for our purpose, it is expected that error due to recall bias will be minimal in that most women are able to confirm if their daughter was cut or not. A second limitation of this study is that we have considered only girls aged 0-14 years thus limiting the generalisation of the findings from the study. It is important to note that a girl who was not cut at age 14 may still be cut in future, hence this must be considered in interpreting our results noting that the FGM/C status of the girl at the time of the survey was not her final FGM/C status.
# Conclusion
Findings from this study have increased our understanding of the extent of success recorded in the fight against the scourge of FGM/C in Nigeria in the recent years. We have also gained further insights on where more resources should be mobilised in order to achieve a permanent abandonment in a social norm driven FGM/C high prevalence community. Such abandonments must be collectively adopted by the community or groups. Other members of the community would then be expected to conform with the collectively adopted change not to have their daughters cut with conformers socially rewarded while defaulters would possibly face social sanctions [bib_ref] Dynamics of change in the practice of female genital cutting in Senegambia:..., Shell-Duncan [/bib_ref]. A successful intervention programme must, therefore, be all-inclusive and involve partnership with both parents, local governments, policymakers, and community and religious leaders. Notable examples of such programmes are the 'FGM-Free Village Model' in Egypt and the Tostan Community Empowerment programme in Senegal. The advanced statistical approach utilised in this study is an important contribution to literature and could be extended to other contexts where there is need to integrate data from multiple sources over large spatial and temporal scales to gain more statistical power.
[fig] Fig 3: Unadjusted estimates of FGM/C prevalence varied across geographical locations (states and the FCT) in Nigeria with the highest prevalence found among the South western states of Ekiti, Ondo, Osun and Oyo, and North-central state of Kwara in 2003 up to 2007. Surprisingly, in 2008, highest prevalence was found in the North [/fig]
[fig] Fig 4: Posterior estimates of the effects of a girl's likelihood of undergoing FGM/C in Nigeria based on the 2016 MICS dataset. Mean (top panel). Significance maps of the posterior estimates based on 95% credible interval (bottom panel). Dark blue to dark red represent lowest to highest FGM/C risk states. Black, white and grey represent significantly high, low, and nonsignificant FGM/C risk states, respectively. Shapefile republished from DIVA-GIS database (https://www.diva-gis.org/) under a CC BY license, with permission from Global Administrative Areas (GADM), original copyright 2018. https://doi.org/10.1371/journal.pone.0246661.g004 Temporal trend. A girl's likelihood of FGM/C varied over time and significantly peaked in 2008 and declined afterwards with significantly lowest likelihood found in 2011 (Fig 6). In addition, Fig 6 shows no significant progress in the reduction of the likelihood of FGM/C among 0-14 years old girls in Nigeria since 2011 suggesting that the practice is still being sustained despite the concerted abandonment efforts. [/fig]
[fig] Fig 5: Boxplot of the Deviance Information Criterion (DIC) for the six (6) models fitted to the pooled 2003 to 2016 dataset. (A) Unadjusted model. (B) Adjusted model. https://doi.org/10.1371/journal.pone.0246661.g005Fig 6. Temporal trends in FGM/C among 0-14-year old girls in Nigeria (m 3 ) based on the pooled 2003 to 2016 dataset. The blue lines are the posterior means; the yellow lines are the 95% credible intervals of the posterior estimates; and the blue band is the width of the 95% credible interval. https://doi.org/10.1371/journal.pone.0246661.g006 significant differences between the effects due to differences in survey methods suggesting that statistical evidence based on the DHS and MICS datasets are not statistically different. Effects of geographical location. To address the question on the roles of geographical location on a girl's likelihood of undergoing FGM/C in Nigeria, we examine the posterior maps of the unadjusted and adjusted spatial effects. Fig 7 presents the posterior means (top panel) and the corresponding maps testing the statistical significance of the estimates based on 95% credible interval (bottom panel) for the unadjusted model (m 1 ;left) and the adjusted model (m 6 ;right). [/fig]
[fig] Fig 7 g007 Fig 8: Posterior estimates of the effects of a girl's likelihood of undergoing FGM/C in Nigeria based on the pooled 2003 to 2016 dataset. Mean (top panel). Significance maps of the posterior estimates based on 95% credible interval (bottom panel). Dark blue to dark red represent lowest to highest FGM/C risk states. Black, white and grey represent significantly high, low, and nonsignificant FGM/C risk states, respectively. Shapefile republished from DIVA-GIS database (https://www. diva-gis.org/) under a CC BY license, with permission from Global Administrative Areas (GADM), original copyright 2018. https://doi.org/10.1371/journal.pone.0246661.Posterior estimates. (A) Predicted prevalence of FGM/C among 0-14-year old girls in Nigeria from 2003 to 2016 based on m 6 . (B) Corresponding deviance maps for the estimates. Dark blue to dark red implies lowest to highest prevalence. Shapefile republished from DIVA-GIS database (https://www.diva-gis.org/) under a CC BY license, with permission from Global Administrative Areas (GADM), original copyright 2018. https://doi.org/10.1371/journal.pone.0246661.g008 western states of Jigawa, Kaduna and Kano, and the South western state of Ekiti are identified as the hotspots where the practice is still rife. [/fig]
[fig] Fig 9: Non-linear effects (log-odds) of mother's age on her daughter's likelihood of being cut based on (A) m 6 of the pooled 2003 to 2016 data and (B) Model B of the 2016 MICS data. The blue lines are the posterior means; the yellow lines are the 95% credible intervals of the posterior estimates; and the blue band is the width of the 95% credible interval. https://doi.org/10.1371/journal.pone.0246661.g009 [/fig]
[table] Table 1: Variable names, levels and descriptions. [/table]
[table] Table 3: Characteristics surrounding female genital mutilation/cutting in Nigerian girls aged 0-14 from 2003-2017. Mother's age difference with husband/ partner (currently married women only) [/table]
[table] Table 4: Posterior odds ratio estimates from the Bayesian geo-additive regression model (adjusted) fitted to the 2016 MICS data. [/table]
[table] Table 5: Posterior odds ratio estimates from the Bayesian geo-additive regression model (adjusted) fitted to the pooled 2003 to 2016 dataset. [/table]
|
Exopolysaccharides produced by marine bacteria and their applications as glycosaminoglycan-like molecules
Although polysaccharides are ubiquitous and the most abundant renewable bio-components, their studies, covered by the glycochemistry and glycobiology fields, remain a challenge due to their high molecular diversity and complexity. Polysaccharides are industrially used in food products; human therapeutics fall into a more recent research field and pharmaceutical industry is looking for more and more molecules with enhanced activities. Glycosaminoglycans (GAGs) found in animal tissues play a critical role in cellular physiological and pathological processes as they bind many cellular components. Therefore, they present a great potential for the design and preparation of therapeutic drugs. On the other hand, microorganisms producing exopolysaccharides (EPS) are renewable resources meeting well the actual industrial demand. In particular, the diversity of marine microorganisms is still largely unexplored offering great opportunities to discover high value products such as new molecules and biocatalysts. EPS-producing bacteria from the marine environment will be reviewed with a focus on marine-derived EPS from bacteria isolated from deep-sea hydrothermal vents. Information on chemical and structural features, putative pathways of biosynthesis, novel strategies for chemical and enzymatic modifications and potentialities in the biomedical field will be provided. An integrated approach should be used to increase the basic knowledge on these compounds and their applications; new clean environmentally friendly processes for the production of carbohydrate bioactive compounds should also be proposed for a sustainable industry.
## Glycopolymers and their importance in glycobiology
Polysaccharides are natural macromolecules composed of osidic monomers and present in all organisms: microorganisms, plants and animals. Polysaccharides, such as starch, glycogen, and cellulose, are common. However, some other specific classes of polysaccharides exist based on their origin, chemical structure and function. Although polysaccharides are ubiquitous, their chemical structure varies greatly from one to the other. They have various roles within the cells, from structural involvement to numerous biological activities including interactions. Therefore, studying the polysaccharide chemical structure and structure-related activities remains a challenge for glycochemists and glycobiologists. Polysaccharides can be distinguished by their osidic composition: homopolysaccharides, which contain a single type of monosaccharide, and heteropolysaccharides composed of different osidic residues and usually displaying a regular backbone structure with a repeating unit. This repeating unit may be linear or branched and may contain up to 10 monomers as well as organic or inorganic substituents such as phosphate, sulfate, and lactic, succinic, acetic and pyruvic acids. The chemical structure including monosaccharide composition and repeating unit sequence as well as non-carbohydrate substituents is species-specific [bib_ref] Microbial exopolymer secretions in ocean enviroments: their role(s) in food webs and..., Decho [/bib_ref] and may vary, most of the time, with production, culture conditions and the physiological state of the organism. The linkages most commonly found between monomers are β-1,4 or β-1,3 giving a more rigid backbone vs. α-1,2 and α-1,6 for more flexible zones. The overall physical properties of polysaccharides are also influenced by the monosaccharide composition, the osidic sequence and the network formed by the single polymer chains [bib_ref] Bacterial exopolysaccharides from extreme marine habitats: production, characterization and biological activities, Poli [/bib_ref]. These polymers are high-molecular weight macromolecules usually above 10 6 g mol −1 .
Most of the polysaccharides used in the industry are extracted from plants, algae, and animals. A few of them come from bacteria and have industrial applications in the paper, food, biotechnological, environmental or health industry [bib_ref] Fermentation technologies for the optimization of marine microbial exopolysaccharide production, Finore [/bib_ref]. The most famous examples of microbial used macromolecules have been listed in [bib_ref] Bacterial polysaccharides -a perception, Kumar [/bib_ref] [bib_ref] Advances in bacterial exopolysaccharides: from production to biotechnological applications, Freitas [/bib_ref]. Owing mainly to their rheological properties, fungal exopolysaccharides (EPS) have also found several applications in the food, cosmetic and pharmaceutical industries as well as in oil recovery [bib_ref] Fungal exopolysaccharide: production, composition and applications, Mahapatra [/bib_ref]. Scleroglucan produced by Sclerotium rolfsii, schizophyllan, a structurally similar molecule produced by Schizophyllum commune and pullulan from Aureobasidium pullulans are the most
## Levans
Various food industry common fungal polysaccharides of high-added value [bib_ref] Scleroglucan: fermentative production, downstream processing and applications, Survase [/bib_ref] [bib_ref] Fungal exopolysaccharide: production, composition and applications, Mahapatra [/bib_ref]. Some of these polysaccharides are homolog counterparts of plant or animal macromolecules. Compared to these sources, microorganisms allow a better controlled production in bioreactors, devoid of no variation due to physiological state or season encountered for the highest organisms [bib_ref] Sargassum filipendula alginate from Brazil: seasonal influence and characteristics, Bertagnolli [/bib_ref] and an easier extraction without any drastic or environmentally toxic compounds. However, downstream processing of bacterial polysaccharides still represents an important cost intensive step [bib_ref] Recent advances in microbial biopolymer production and purification, Kreyenschulte [/bib_ref]. Moreover, microorganisms cultivation in fermenters allows the optimization of the growth and the production yield either by the study of physiology or by genetic engineering. For the high-added value pharmaceutical industry, bacterial polysaccharides can be produced at a viable economic cost. The production in controlled conditions is in agreement with the Good Manufacturing Practices (GMP) such as well-defined medium, controlled environment without the risk for viral or pathogen agents. The advantages of a bacterial source over plant, algal or animal source have made it attractive to obtain macromolecules for various industrial purposes and strengthened their study. Complex carbohydrates and glycoconjugates have a basic importance in biological systems and cellular processes, either physiological or pathological. They play a major role as structural agents in connective tissues, and they are ubiquitously present on cell surfaces, mediating the interaction of the cells with other cells, with the extracellular matrix, with biotic or abiotic surfaces and with other molecules. But their study that falls into the field of the glycobiology was greatly hindered by technical issues; no sequencing tool such as that existing in proteomics or genomics is available to date while glycopolymers potential chemical diversity is far greater than proteins and nucleic acids [bib_ref] Emerging glycomics technologies, Turnbull [/bib_ref] [bib_ref] Glycansby-design: engineering bacteria for the biosynthesis of complex glycans and glycoconjugates, Merritt [/bib_ref]. While genes and proteins syntheses are based on a template, polysaccharide biosynthesis is regulated by a number of physiological and metabolic parameters including the availability of sugar precursors and the expression level of enzymes. The number of osidic residues combined with the anomeric configuration of the linkages and the possibility of branching result in unique complexity and diversity of carbohydrates molecules. Nonetheless, some advances in analytical techniques have revealed the vast and diverse chemical structures of carbohydrates existing in nature; insights into the structure-function relationship allow determining their role in diverse cellular processes. Glycobiology becomes therefore a major research field allowing the understanding of human diseases and the discovery of novel therapeutic compounds.
Glycosaminoglycans (GAGs) are glycopolymers found in animal tissues and composed of uronic acid and neutral or hexosamine residues. They are covalently bound to a core protein and are the major constituent of proteoglycans. Their carbohydrate backbone is unique for each cell type. GAGs are essential for life of animals since they are involved in their development and organogenesis [bib_ref] Chemoenzymatic synthesis of glycosaminoglycans: re-creating, re-modeling and re-designing nature's longest or most..., De Angelis [/bib_ref]. Two GAG macromolecules are frequently used in various industry fields, namely hyaluronic acid and heparin.
Hyaluronic acid is a polysaccharide ranging from 500 to 1000 10 3 g mol −1 , with a disaccharidic linear and non-sulfated repeating unit [fig_ref] Table 2 |: Repeating unit chemical structures of main glycosaminoglycans and of some marine exopolysaccharides [/fig_ref]. It is widely used in osteoarthritis treatment as synovial fluid because of its effect on the cartilage, in ophthalmology treatments and surgery because of its visco-elastic properties, and in wound healing as well as in the cosmetic industry (moisturizing agent, wrinkle filler).
Heparin is a polyanionic molecule of about 80 10 3 g mol −1 with 20 to 40% of sulfates exhibiting several biological activities such as anticoagulant, anti-thrombotic, anti-inflammatory, antiviral and anticancer [bib_ref] Guide to anticoagulant therapy: heparin: a statement for healthcare professionals from the, Hirsh [/bib_ref]. It has been discovered in the early twentieth century and used medically for over 70 years. The commercially available drug is extracted from a mucous lining of pig intestines. In parallel, low-molecular weight (LMW) heparin (6-10 10 3 g mol −1 ) has been developed for use as an anticoagulant to increase benefit-risk ratios [bib_ref] Guide to anticoagulant therapy: heparin: a statement for healthcare professionals from the, Hirsh [/bib_ref]. In 1983, the pentasaccharide unit responsible for the anti-coagulant and anti-thrombin activity within the heparin molecule was identified, chemically synthesized [bib_ref] Structure-activity relationship in heparin: a synthetic pentasaccharide with high affinity for antithrombin..., Choay [/bib_ref] and it is now commercialized [bib_ref] The synthetic pentasaccharide Fondaparinux: first in the class of antithrombotic agents that..., Petitou [/bib_ref] [fig_ref] Table 2 |: Repeating unit chemical structures of main glycosaminoglycans and of some marine exopolysaccharides [/fig_ref]. Although this molecule presents, as the unfractionated and LMW heparin, some undesirable bleeding side effects Heparin antithrombin-binding site pentasaccharide [bib_ref] Structure-activity relationship in heparin: a synthetic pentasaccharide with high affinity for antithrombin..., Choay [/bib_ref]
[formula] →4)-α-D-GlcNp2S6S-(1→4)-β-D-GlcAp-(1→4)-α-D-GlcNp2S3S6S-(1→4)-α-L-IdoAp2S-(1→4)-α-D-GlcNp2S6S-(1→ Hyaluronic acid →3)-β-D-GlcpNAc-(1→4)-β-D-GlcpA-(1→ EPS ST716 [/formula]
from Alteromonas macleodii subsp. fijiensis [bib_ref] Description of a new polymer-secreting bacterium from a deepsea hydrothermal vent, Alteromonas..., Raguenes [/bib_ref] [bib_ref] Structural studies of an exopolysaccharide produced by Alteromonas macleodii subsp. fijiensis originating..., Rougeaux [/bib_ref]
[formula] 4)-β-D-Glcp-(1 4)-α-D-GalpA-(1 4)-α-D-Galp-(1 3 1 α-D-GlcpA 3 1 β-D-GlcpA 4 1 4,6-Pyr-β-D-Manp [/formula]
## Eps hyd721
from Pseudoalteromonas sp. [bib_ref] Structural determination of the exopolysaccharide of Pseudoalteromonas strain HYD 721 isolated from..., Rougeaux [/bib_ref]
[formula] 4)-β-D-Manp-(1 4)-β-D-Glcp-(1 4)-α-D-Galp-(1 4)-β-D-Glcp-(1 2 3 1 1 α-L-Rhap β-D-Galp 3 1 β-D-GlcpA 4 1 [SO 3 H] 3-β-D-Manp [/formula]
## Eps he800
from Vibrio diabolicus [bib_ref] Vibrio diabolicus sp. nov., a new polysaccharide-secreting organism isolated from a deep-sea..., Raguenes [/bib_ref] [bib_ref] Structure of the exopolysaccharide of Vibrio diabolicus isolated from a deep-sea hydrothermal..., Rougeaux [/bib_ref] →3)-β-D-GlcpNAc-(1→4)-β-D-GlcpA-(1 →4)-β-D-GlcpA -(1 →4)-α-D-GalpNAc-(1 →
## Eps gy785
from Alteromonas infernus [bib_ref] Alteromonas infernus sp. nov., a new polysaccharide-producing bacterium isolated from a deep-sea..., Raguenes [/bib_ref] [bib_ref] Structural studies of the main exopolysaccharide produced by the deep-sea bacterium Alteromonas..., Roger [/bib_ref] [ [bib_ref] Guide to anticoagulant therapy: heparin: a statement for healthcare professionals from the, Hirsh [/bib_ref] [bib_ref] Bleeding risk and the management of bleeding complications in patients undergoing anticoagulant..., Crowther [/bib_ref] , it remains a lead molecule to study the structure-activity relationship. A heparin contamination crisis arose in the early 2008; some Chinese heparin preparations caused severe side effects resulting in the death of more than 100 patients [bib_ref] Lessons learned from the contamination of heparin, Liu [/bib_ref]. The impurity was identified as a variant of chondroitin sulfate, in which sulfate groups exhibited an unusual pattern [bib_ref] Oversulfated chondroitin sulfate is a contaminant in heparin associated with adverse clinical..., Guerrini [/bib_ref]. Today, despite the risk of contamination during the extraction process or by animal pathogenic agents, heparin from natural sources continues to be widely used clinically. Still new alternative macromolecules are looked for. In this context, marine microbial polysaccharides stand for a promising source of macromolecules with reduced risk of contamination by mammalian pathogenic agents. New emerging products of high added-value from the marine environment, showing structural homology to heparin such as the presence of sulfate groups, offer the promising opportunity of novel biologically active compounds [bib_ref] Anticoagulant motifs of marine sulfated glycans, Pomin [/bib_ref]. Exopolysaccharides (EPS) are glycopolymers that microorganisms secrete in their surrounding environment [bib_ref] Bacterial Exopolysaccharides, Sutherland [/bib_ref]. They can be capsular polymers, which are attached to the cell membrane through the lipopolysaccharides (LPS) anchored in the membrane or other specific proteins [bib_ref] Microbial exopolymer secretions in ocean enviroments: their role(s) in food webs and..., Decho [/bib_ref]. EPS can also be produced as a slime loosely bound around the cell or dispersed in the environment [bib_ref] Biosynthesis of microbial polysaccharides, Sutherland [/bib_ref]. Bacterial polysaccharides are also present within the membrane cell as a major constituent of the LPS and may be involved in pathogenicity. These glycopolymers have mostly a protective role and permit resistance under extreme environmental conditions by participating in the cell membrane integrity, trapping nutrients, allowing adhesion to surfaces, protecting from toxic compounds and adverse conditions such as freezing [bib_ref] Deep-sea microbiology, Jannasch [/bib_ref] [bib_ref] Microbial exopolymer secretions in ocean enviroments: their role(s) in food webs and..., Decho [/bib_ref] [bib_ref] Fermentation technologies for the optimization of marine microbial exopolysaccharide production, Finore [/bib_ref]. EPS production requires energy from cells representing a carbon investment for microorganisms but the benefits to growth and survival are higher than the production cost [bib_ref] Synthesis, production, and biotechnological applications of exopolysaccharides and polyhydroxyalkanoates by Archaea, Poli [/bib_ref].
[formula] SO 3 Na] ↓ 2 [→4)-β-Glcp-(1→4)-α-GalpA-(1→4)-α-Galp-(1→] 3 ↑ 1 β-Glcp-(1→6)-α-Galp-(1→4)-β-GlcpA-(1→4)-β-GlcpA 2 3 ↑ ↑ 1 1 α-Glcp α-Glcp [/formula]
## The marine biosphere as a source of new glycopolymers
Representing more than 70% of our planet, ocean is an under-explored and under-exploited vast reservoir for biological organisms and chemical compounds. France owns an expanded maritime domain as well as oceanographic ships, submarine and robots allowing to explore the marine biosphere, and especially deep seas. The marine biosphere is heterogeneous because a large range of ecosystems exist such as microbial mats, Antarctic sea ice, hypersaline marine environments, shallow and deepsea hydrothermal vents. They are characterized by physical and chemical parameters such as pressure, temperature, pH, chemical compounds, usually toxic. Within the deep-sea hydrothermal vents, large physico-chemical gradients exist; for example, the temperature varies from 2 - C, the temperature of the surrounding sea water to the hot temperature of the hydrothermal plume, which can reach 350 - C [bib_ref] Submarine hydrothermal vents and associated gradient environments as sites for the origin..., Baross [/bib_ref]. Due to their microbial diversity, these ecosystems might offer promising new biomolecules [bib_ref] Deep ocean environmental biotechnology, Deming [/bib_ref].
## Marine ecosystems with a focus on deep-sea hydrothermal vents
Deep-sea hydrothermal vents result from oceanic plate tectonic and submarine volcanic activities. They appear at the sea ridges or on subduction back-arc areas, at a depth of 500 to 4000 m . Seawater at high temperature (up to 350 - C) and charged with metals and other compounds such as hydrogen sulfide, hydrogen, ammonia, carbon dioxide flows out of structures built from precipitates called chimneys. Depending on the composition of the fluid, the plume appears with different intensity of white or black color (white or black smokers). Due to volcanic activities of the crust, these ecosystems are ephemeral [bib_ref] Evolution and biogeography of deep-sea vent and seep invertebrates, Van Dover [/bib_ref]. Some other active areas with a diffuse emission of warm or cold water also exist. Deepsea ecosystems also include cold seeps and sediments or microbial mats [bib_ref] Deep-sea microbiology, Jannasch [/bib_ref]. The first hydrothermal vent was discovered in 1977 in the Galapagos area at a depth of 2600 m [bib_ref] Deep-sea microbiology, Jannasch [/bib_ref] , although deep-sea biology started with the Challenger expedition (1873-1876). The presence of bacteria in the open ocean seabed is known since the 1880s from the Travaillier and Talisman expeditions (1882-1883) [bib_ref] Deep-sea bacteria, Zobell [/bib_ref] , when the first microorganisms were found at 5000 m deep [bib_ref] Deep-sea microbiology, Jannasch [/bib_ref]. In the 1950s, Galathea expedition collected bacteria-containing sediment samples from 10,000 m deep [bib_ref] Deep-sea bacteria, Zobell [/bib_ref].
Deep-sea hydrothermal ecosystems, the most productive ecosystems from the deep ocean, are based on a primary chimiosynthetic production of organic matter sustaining an abundant fauna which is mainly composed of endemic species [bib_ref] Deep-sea microbiology, Jannasch [/bib_ref]. These oases of life contrast to the vast surrounding desertic seafloors. A large variety of microbial lifestyles has been described in these ecosystems based on the respiratory type (aerobic-anaerobic), energy and carbon sources. It is adapted to extreme conditions (barophiles needing high pressure, acidophiles-alcalophiles, psychrophiles, mesophiles and extremophiles -growing at temperatures from around 3 - C, the temperature of seawater, to 105 - C-) [bib_ref] Deep-sea microbiology, Jannasch [/bib_ref]. New microorganisms with very diverse metabolisms have then been isolated from marine environment especially in deepsea habitats [bib_ref] Recent developments in the thermophilic microbiology of deep-sea hydrothermal vents, Miroshnichenko [/bib_ref]. Halophiles are also encountered in some marine ecosystems. The high biodiversity potentially offers a large chimiodiversity. Besides their fundamental interest in basic studies for life origin understanding, chimiosynthetic metabolisms and mechanisms of resistance to high temperatures, these microorganisms arise a great biotechnological interest for the isolation of new enzymes and biomolecules with new or enhanced activities.
## From marine microbial biodiversity to new bioactive molecules
Marine biodiversity is not known currently because only a very small fraction of the microorganisms can be cultivated [bib_ref] Community genomics among stratified microbial assemblages in the ocean's interior, Delong [/bib_ref]. There are reports on particular marine environments with high microbial diversity; this is the case for deep-sea hydrothermal vents in North Atlantic [bib_ref] Microbial diversity in the deep sea and the underexplored "rare biosphere, Sogin [/bib_ref]. The next generation high throughput DNA sequencing (NGS) methodologies can provide a way to estimate the biodiversity and can be useful to drive some cultivation experiments [bib_ref] Emerging strategies and integrated systems microbiology technologies for biodiscovery of marine bioactive..., Rocha-Martin [/bib_ref] : when the analysis of metagenomic data sets reveals incomplete pathway for a given compound essential for growth, this compound has to be added as an exogenous source. For example, because SAR11 marine alpha-proteobacteria clade has been shown to be deficient in assimilatory sulfate reduction genes, addition of reduced sulfur compounds in the cultivation medium was tested and was revealed necessary for growth [bib_ref] SAR11 marine bacteria require exogenous reduced sulphur for growth, Tripp [/bib_ref].
When the biodiversity increases, the chimiodiversity also increases. Therefore, the chemical study of the marine organisms will most probably result in the discovery of new therapeutic compounds. In particular, it is well recognized that the chemical diversity of marine and microbial compounds is the broadest one [bib_ref] Scaffold diversity of natural products: inspiration for combinatorial library design, Grabowski [/bib_ref]. Although microorganisms from the sea have been studied for several decades, their development for biotechnology started only recently (see [bib_ref] Bio-mining the microbial treasures of the ocean: new natural products, Imhoff [/bib_ref]. Today, only a small fraction of the known species has been studied for their biochemical metabolites. Marine microorganisms together with sponges, molluscs, algae, echinoderms are rich source of polysaccharides. Chitin, the second most abundant polysaccharide after cellulose, is found particularly in the shell of crustaceans. Alginates, another marine-derived polysaccharide, are extracted from the cell wall of brown algae (Phaecophyta). Marine-derived polysaccharides naturally bearing sulfate groups are also industrially relevant: carrageenan (red algae), fucoidans (brown algae) and fucans (marine invertebrates) [bib_ref] Sulfated fucans, fresh perspectives: structures, functions, and biological properties of sulfated fucans..., Berteau [/bib_ref] [bib_ref] Structure, biological activity, and enzymatic transformation of fucoidans from the brown seaweeds, Kusaykin [/bib_ref] , ulvan from green seaweeds (Chlorophyta) [bib_ref] Structure and functional properties of ulvan, a polysaccharide from green seaweeds, Lahaye [/bib_ref]. Microorganisms producing polysaccharides have also been isolated from marine ecosystems especially from deep-sea hydrothermal vents.
Marine chimiodiversity finds applications in human health as therapeutic agents for innovative biological activities. New compounds with new chemical structure can also enable a better understanding of the cellular processes (pathological or not): inhibitors of enzymes involved in inflammatory diseases (manoalide) and phosphorylation processes (okadaic acid), both extracted from sponge, are available for biochemical research [bib_ref] New pharmaceuticals from marine organisms, Fenical [/bib_ref]. Some examples of marine molecules are already commercialized as analgesic (the peptidic Prialt®), anti-viral (mainly from corals and sponges), anti-cancer (the alkaloid Yondelis® by PharmaMar), or as biomaterials (the calcium carbonate chemical compound Biocoral® by Biocoral from coral). Only a few of them are carbohydrates: alginate of plant origin and chitosan obtained from shrimp extracted chitin have been successfully integrated within wound dressings (Stop Hemo® by Brothier Laboratory in France, Nu-Derm® by Johnson & Johnson in USA) and are also used as encapsulation matrices (geniaBeads® by geniaLab in Germany). Most of the marine molecules are still in research or clinical development, among which four are from marine microorganisms [bib_ref] Bio-mining the microbial treasures of the ocean: new natural products, Imhoff [/bib_ref] ; this leaves the marine biotechnology still at a very promising phase; great efforts are still necessary to get advances in clinical applications.
Bioactive compound prospecting starts with the sampling of organisms. The choices of methods to prepare extracts and to screen for the production of bioactive metabolites are critical steps since they can result in different molecules or even in difficulties to find the activity. The diversification of sources by the exploration of new ecological niches can maintain an effective process leading to innovative results. Compounds should then be isolated, structurally characterized especially including the stereochemistry which can be the main issue for its activity. Biological properties have to be studied in detail, in vitro and subsequently in vivo before clinical development stages. Alternatively, synthetic analogs can be studied instead of natural molecules. Chemical synthesis would offer access to structurally defined oligosaccharides in sufficient purity and quantity to carry out biological studies. In addition, the synthesis can bring a library of analog compounds slightly differing in the chemical structure as well as in the activity. This library would be a useful tool to establish the structure-activity relationship. Synthetic analogs of the active pentasaccharide in heparin have been well studied due to the high value of this drug and its side effects [bib_ref] Synthetic oligosaccharides as heparin-mimetics displaying anticoagulant properties, Avci [/bib_ref]. However, synthetic products are sometimes hard to make because of the complexity of the molecule. This is particularly the case for polysaccharides for which only small oligosaccharides have been obtained to date [bib_ref] Opportunities and challenges in synthetic oligosaccharide and glycoconjugate research, Boltje [/bib_ref]. This issue still limits the progress of glycobiology research and much effort are made to develop oligosaccharide synthesis methodologies. Recently, shown that ionic liquids could be very useful as solvents and promoters of synthetic reactions [bib_ref] Recent developments of ionic liquids in oligosaccharide synthesis: the sweet side of..., Galan [/bib_ref].
## Marine exopolysaccharides microbial polysaccharide diversity
Several EPS-producing marine strains have been studied, which led to the discovery and isolation of novel macromolecules [bib_ref] Fermentation technologies for the optimization of marine microbial exopolysaccharide production, Finore [/bib_ref]. Marine bacterial polysaccharides have been the subject of several reviews [bib_ref] Deep-sea hydrothermal vents: a new source of innovative bacterial exopolysaccharides of biotechnological..., Guezennec [/bib_ref] [bib_ref] Structures of polysaccharides and oligosaccharides of some Gram-negative marine Proteobacteria, Nazarenko [/bib_ref] [bib_ref] Bacterial exopolysaccharides from extreme marine environments with special consideration of the Southern..., Mancuso Nichols [/bib_ref] [bib_ref] Marine polysaccharides in pharmaceutical applications: an overview, Laurienzo [/bib_ref] [bib_ref] Biosurfactants, bioemulsifiers and exopolysaccharides from marine microorganisms, Satpute [/bib_ref] [bib_ref] Advances in bacterial exopolysaccharides: from production to biotechnological applications, Freitas [/bib_ref] [bib_ref] Marine polysaccharides: a source of bioactive molecules for cell therapy and tissue..., Senni [/bib_ref] [bib_ref] Marine-derived bioactive polysaccharides from microorganisms, Colliec-Jouault [/bib_ref] [bib_ref] Marine medicinal glycomics, Pomin [/bib_ref]. Most of the marine derived EPS are bacterial (classical bacteria i.e., mesophilic and heterotrophic; extremophilic microorganisms i.e., psychrophilic, thermophilic and halophilic); archaea have also been shown to produce EPS [bib_ref] Effect of carbon and nitrogen sources on growth dynamics and exopolysaccharide production..., Rinker [/bib_ref] [bib_ref] Exopolysaccharides from extremophiles: from fundamentals to biotechnology, Nicolaus [/bib_ref] [bib_ref] Synthesis, production, and biotechnological applications of exopolysaccharides and polyhydroxyalkanoates by Archaea, Poli [/bib_ref] [bib_ref] Fermentation technologies for the optimization of marine microbial exopolysaccharide production, Finore [/bib_ref]. In the present manuscript, the main focus was put on the extremophilic prokaryotic marine microorganisms that were shown to produce EPS, with an emphasis on bacteria isolated from deep-sea hydrothermal vents.
Mancuso Nichols et al. (2004) described the production of EPS by the marine strains Pseudoalteromonas CAM025 and CAM036 isolated in Antarctica sea water and sea ice. Some other strains from Arctic sea ice have been shown to produce EPS with cryoprotective effect .
The halophilic strains Halomonas maura [bib_ref] Mauran, an exopolysaccharide produced by the halophilic bacterium Halomonas maura, with a..., Arias [/bib_ref] , Halomonas ventosae [bib_ref] Halomonas ventosae sp. nov., a moderately halophilic, denitrifying, exopolysaccharideproducing bacterium, Martinez-Canovas [/bib_ref] , Halomonas alkaliantarctica [bib_ref] Halomonas alkaliantarctica sp. nov., isolated from saline lake Cape Russell in Antarctica,..., Poli [/bib_ref] , Hahella chejuensis [bib_ref] Bacterial exopolysaccharides from extreme marine habitats: production, characterization and biological activities, Poli [/bib_ref] and the archaeal halophilic Haloferax mediterranei [bib_ref] Production of an extracellular polysaccharide by Haloferax mediterranei, Anton [/bib_ref] [bib_ref] The structure of the exopolysaccharide produced by the halophilic archaeon Haloferax mediterranei..., Parolis [/bib_ref] isolated from hypersaline environments were also shown to produce EPS; some of them are sulfated [bib_ref] Bacterial exopolysaccharides from extreme marine habitats: production, characterization and biological activities, Poli [/bib_ref].
Marine thermophilic anaerobes Sulfolobus, Thermococcus and Thermotoga were described to produce EPS [bib_ref] Polysaccharide degradation and synthesis by extremely thermophilic anaerobes, Vanfossen [/bib_ref]. Thermococcus litoralis produces an EPS which contains sulfate and phosphorus substituents [bib_ref] Effect of carbon and nitrogen sources on growth dynamics and exopolysaccharide production..., Rinker [/bib_ref] ; the archaea Sulfolobus solfataricus has also been shown to produce a sulfated polysaccharide [bib_ref] Production of an exopolysaccharide from two thermophilic archaea belonging to the genus..., Nicolaus [/bib_ref]. Geobacillus sp., Bacillus thermodenitrificans and B. licheniformis thermophilic strains have been isolated from shallow marine hydrothermal vents of Vulcano Island (Italy) and the polysaccharide they produce has been analyzed [bib_ref] Bacterial exopolysaccharides from extreme marine habitats: production, characterization and biological activities, Poli [/bib_ref].
Several bacteria belonging to gamma-proteobacteria Alteromonadales or Vibrionales orders and isolated from the deep-sea polychaete annelids Alvinella pompejana and A. caudata tissues have been shown to produce different EPS: HYD1545 [bib_ref] Production and characterization of an exopolysaccharide excreted by a deepsea hydrothermal vent..., Vincent [/bib_ref] , HYD1644 [bib_ref] Structure determination of a novel uronic acid residue isolated from the exopolysaccharide..., Dubreucq [/bib_ref] , HYD721 [bib_ref] Structural determination of the exopolysaccharide of Pseudoalteromonas strain HYD 721 isolated from..., Rougeaux [/bib_ref] , HYD657 produced by A. macleodii subsp. fijiensis biovar deepsane [bib_ref] A novel polymer produced by a bacterium isolated from a deep-sea hydrothermal..., Cambon-Bonavita [/bib_ref] and HE800 produced by Vibrio diabolicus [bib_ref] Vibrio diabolicus sp. nov., a new polysaccharide-secreting organism isolated from a deep-sea..., Raguenes [/bib_ref] [bib_ref] Structure of the exopolysaccharide of Vibrio diabolicus isolated from a deep-sea hydrothermal..., Rougeaux [/bib_ref]. This latter polysaccharide possesses a particular hyaluronic acid-like chemical structure and contains both hexosamines and uronic acids alternating in the repeating unit sequence [fig_ref] Table 2 |: Repeating unit chemical structures of main glycosaminoglycans and of some marine exopolysaccharides [/fig_ref].
Another gamma-proteobacterium isolated from a deep-sea hydrothermal vent shrimp (A. macleodii subsp. fijiensis biovar medioatlantica) was shown to produce the EPS MS907 [bib_ref] A novel, highly viscous polysaccharide excreted by an Alteromonas isolated from a..., Raguenes [/bib_ref] whereas Alteromonas macleodii subsp. fijiensis producing the pyruvated EPS ST716 [bib_ref] Description of a new polymer-secreting bacterium from a deepsea hydrothermal vent, Alteromonas..., Raguenes [/bib_ref] [bib_ref] Structural studies of an exopolysaccharide produced by Alteromonas macleodii subsp. fijiensis originating..., Rougeaux [/bib_ref] and Alteromonas infernus producing the sulfated EPS GY785 were isolated from hydrothermal fluids [bib_ref] Alteromonas infernus sp. nov., a new polysaccharide-producing bacterium isolated from a deep-sea..., Raguenes [/bib_ref] [bib_ref] Structural studies of the main exopolysaccharide produced by the deep-sea bacterium Alteromonas..., Roger [/bib_ref] [fig_ref] Table 2 |: Repeating unit chemical structures of main glycosaminoglycans and of some marine exopolysaccharides [/fig_ref].
Other Pseudoalteromonas strains produce diverse EPS: SM9913 isolated from deep-sea sediments produces an acetylated EPS [bib_ref] Structural characterization and ecological roles of a novel exopolysaccharide from the deep-sea..., Qin [/bib_ref] , SM20310 from the arctic sea ice produces a complex α-mannan .
Pseudomonas, Alteromonas, Paracoccus and Vibrio sp. bacteria producing EPS under laboratory conditions have also been isolated from microbial mats in French Polynesia, another marine ecosystem considered as extreme. Among them, Paracoccus zeaxanthinificiens subsp. Payriae and Vibrio sp. RA 29 are described to produce sulfated polysaccharides and Vibrio sp. MO 245 would produce a polymer very similar to the Vibrio diabolicus one [bib_ref] Microbial mats in French Polynesia and their biotechnological applications, Guézennec [/bib_ref].
When known, the chemical and structural diversity of these few molecules confirms the high value of the marine environment as a source of exciting chimiodiversity.
## Alginate and chitosan, two major marine polysaccharides
Alginate traditionally extracted from brown algae and the most important polysaccharide from this type of seaweeds is also produced by soil bacteria Azotobacter and Pseudomonas species as an exopolysaccharide [bib_ref] Marine derived polysaccharides for biomedical applications: chemical modification approaches, Gomez D'ayala [/bib_ref]. Although they are not marine bacteria, the polysaccharide is usually refered as marine derived. A gene cluster for alginate biosynthesis have been identified in Vibrio sp. QY102 suggesting this polymer would be involved in the formation of biofilm by Vibrio sp. QY102 [bib_ref] Detection of the key enzyme of alginate biosynthesis in Vibrio sp, Shi [/bib_ref].
The bacterial alginates are all composed of mannuronic (M) and guluronic (G) acids arranged in homo-blocks (M or G) and hetero-blocks (MG) as algal alginates; but that from Pseudomonas
## Frontiers in chemistry | chemical biology
October 2014 | Volume 2 | Article 85 | 6
does not have block G; in addition many of bacterial alginates are acetylated. G-rich alginates and M-rich alginates have different physico-chemical and biological properties [bib_ref] Marine-derived bioactive polysaccharides from microorganisms, Colliec-Jouault [/bib_ref]. Although their osidic composition is similar, bacterial alginates vary considerably in their structure depending on the producing strain and in the case of alginate produced by seaweeds, on seasonal and environment variations [bib_ref] Marine derived polysaccharides for biomedical applications: chemical modification approaches, Gomez D'ayala [/bib_ref] [bib_ref] Sargassum filipendula alginate from Brazil: seasonal influence and characteristics, Bertagnolli [/bib_ref]. Industrial production of alginate reaches 30,000 tons annually and it is used as a viscosifier, stabilizer and gelling agent [bib_ref] Microbial alginate production, modification and its applications, Hay [/bib_ref]. Chitin and chitosan, its partially N-deacetylated derivative, are the most abundant marine polysaccharides; they have great interest for biotechnological applications. They are mainly extracted from shellfish waste. To date no bacterium has been described to produce either chitin or chitosan. Nonetheless, some bacteria have been shown to produce PNAG, poly-N-acetyl-glucosamine, a β-1,6-linked N-acetyl-D-glucosamine homopolymer, structurally close to chitin, β-1,4-linked N-acetyl-D-glucosamine homopolymer [bib_ref] Synthase-dependent exopolysaccharide secretion in Gram-negative bacteria, Whitney [/bib_ref] [bib_ref] Effect of sypQ gene on poly-Nacetylglucosamine biosynthesis in Vibrio parahaemolyticus and its..., Ye [/bib_ref].
## Production and physiological conditions influence
Production of polysaccharides by bacteria is a physiological process that takes place under a diversity of environmental conditions [bib_ref] Bacterial polysaccharides -a perception, Kumar [/bib_ref] , usually in response to a stress including the adaptation to extreme environment [bib_ref] Stress-induced production of biofilm in the hyperthermophile Archaeoglobus fulgidus, Lapaglia [/bib_ref]. In the marine environment, bacteria are usually attached to biotic or abiotic surfaces and are embedded in a slime, an exopolymeric matrix composed of proteins, polysaccharides, DNA and small organic molecules; they form biofilms as a protective response to a stress [bib_ref] Stress-induced production of biofilm in the hyperthermophile Archaeoglobus fulgidus, Lapaglia [/bib_ref] [bib_ref] Deep-sea hydrothermal vents: a new source of innovative bacterial exopolysaccharides of biotechnological..., Guezennec [/bib_ref] [bib_ref] Structure and ecological roles of a novel exopolysaccharide from the Arctic Sea..., Liu [/bib_ref]. Therefore, it is important to identify any relationship between the biosynthesis of EPS and biofilm formation regulation in order to better control the production process [bib_ref] Bacterial polymers: biosynthesis, modifications and applications, Rehm [/bib_ref]. Limitations in the availability of nutrients such as nitrogen, phosphorous, sulfur, potassium [bib_ref] Bacterial polysaccharides -a perception, Kumar [/bib_ref] [bib_ref] Exopolysaccharide yield as a kinetic parameter for the statistical optimization of EPS..., Sabra [/bib_ref] , as well as temperature, osmotic or pH shifts [bib_ref] Study of the effects of temperature, pH and yeast extract on growth..., Gorret [/bib_ref] [bib_ref] Bacterial polysaccharides -a perception, Kumar [/bib_ref] are stressing factors enhancing EPS production. Moreover, a suitable carbon source to be added to the production or growth medium is also a main factor for the yield and features of the produced molecule [bib_ref] Microbial exopolysaccharides: main examples of synthesis, excretion, genetics and extraction, Donot [/bib_ref]. Its excess, in fact, is concomitant with another component limitation such as nitrogen [bib_ref] Advances in bacterial exopolysaccharides: from production to biotechnological applications, Freitas [/bib_ref]. In some cases, the osidic composition of the molecule can change with the carbon substrate [bib_ref] Influence of fructose and glucose on the production of exopolysaccharides and the..., Grobben [/bib_ref] [bib_ref] Halomonas alkaliantarctica sp. nov., isolated from saline lake Cape Russell in Antarctica,..., Poli [/bib_ref] [bib_ref] Advances in bacterial exopolysaccharides: from production to biotechnological applications, Freitas [/bib_ref].
Production yield of marine EPS is usually around 1 g L −1 [bib_ref] Microbial exopolymer secretions in ocean enviroments: their role(s) in food webs and..., Decho [/bib_ref]. Industrial development is conceivable for a production yield around 10 g L −1 . Production costs are driven by the yield of polysaccharide, the amount and cost of carbon source, and by the downstream processes needed for molecule separation. Therefore, studies on production optimization are needed even if some bacterial strains are naturally able to produce EPS at a high yield (e.g., 50 g L −1 or over) such as Agrobacterium sp. (curdlan), Xanthomonas campestris (xanthan), Zymomonas mobilis (levan), Alcaligenes faecalis (curdlan), Bacillus sp. (levan) [bib_ref] Microbial exopolysaccharides: main examples of synthesis, excretion, genetics and extraction, Donot [/bib_ref]. Factors such as medium composition and fermentation conditions i.e., temperature, pH and aeration stand for the main studied variables to optimize the fermentation process [bib_ref] Effects of incubation temperature on growth and production of exopolysaccharides by an..., Mancuso Nichols [/bib_ref] [bib_ref] Exopolysaccharide yield as a kinetic parameter for the statistical optimization of EPS..., Sabra [/bib_ref] [bib_ref] Fermentation technologies for the optimization of marine microbial exopolysaccharide production, Finore [/bib_ref]. The nutritional conditions can also affect the molecular weight as well as the osidic composition of the EPS in some cases [bib_ref] Effects of incubation temperature on growth and production of exopolysaccharides by an..., Mancuso Nichols [/bib_ref] [bib_ref] High level synthesis of levan by a novel Halomonas species growing on..., Poli [/bib_ref] [bib_ref] Microbial exopolysaccharides: main examples of synthesis, excretion, genetics and extraction, Donot [/bib_ref] [bib_ref] Fermentation technologies for the optimization of marine microbial exopolysaccharide production, Finore [/bib_ref]. A novel Alteromonas macleodii strain has been shown to produce 23.4 g L −1 exopolysaccharide when grown on 15% lactose which is the highest yield obtained for marine EPS.
Fermentation is usually performed in batch or fed-batch. Heat transfer and oxygen supply limitations are however encountered when EPS are highly viscous [bib_ref] Recent advances in microbial biopolymer production and purification, Kreyenschulte [/bib_ref]. Depending on the producing bacterium, the EPS may be produced during the growth (cellulose, gellan or alginate by Azotobacter vinelandii), when the growth has ended (curdlan by Alcaligenes faecalis), during the stationary phase [bib_ref] Description of a new polymer-secreting bacterium from a deepsea hydrothermal vent, Alteromonas..., Raguenes [/bib_ref] or both (xanthan by X. campestris) [bib_ref] Microbial exopolymer secretions in ocean enviroments: their role(s) in food webs and..., Decho [/bib_ref] [bib_ref] Advances in bacterial exopolysaccharides: from production to biotechnological applications, Freitas [/bib_ref]. The marine EPS HYD1644 is produced only after the initial exponential growth phase [bib_ref] Simultaneous production of two different gel-forming exopolysaccharides by an Alteromonas strain originating..., Samain [/bib_ref]. In Pseudomonas sp., an EPS is produced during the exponential phase and another one with a different chemical structure is produced during the stationary phase [bib_ref] Partial chemical and physical characterization of two extracellular polysaccharides produced by marine,..., Christensen [/bib_ref]. Some reports emphasize the benefit of a growth-production uncoupling by using a bacteriostatic to stop the growth together with the addition of the carbonated substrate and/or shift in temperature or pH [bib_ref] Uncoupling of growth and exopolysaccharide production by Lactococcus lactis subsp. cremoris NIZO..., Looijesteijn [/bib_ref].
## Downstream process for eps recovery
After fermentation, the downtream process for the recovery of EPS is an important step for the production costs and commercial value of the molecule. Process steps will depend on the molecule, on other undesired molecules produced, but also on the level of purity required.
Downstream process, consisting of the removal of insoluble particles, isolation of the product and further purification, is based on classical steps of extraction and purification. Each method has its advantages and disadvantages. Various molecules are usually associated with polysaccharides that act as sponges for some proteins, amino acids, DNA, RNA, salts, metals, fatty acids either dissolved in seawater based medium or released after cell lysis [bib_ref] Microbial exopolymer secretions in ocean enviroments: their role(s) in food webs and..., Decho [/bib_ref] [bib_ref] Advances in bacterial exopolysaccharides: from production to biotechnological applications, Freitas [/bib_ref] [bib_ref] Microbial exopolysaccharides: main examples of synthesis, excretion, genetics and extraction, Donot [/bib_ref]. The presence of degrading enzymes or of a second biopolymer must also be taken into account for a good purification process [bib_ref] Recent advances in microbial biopolymer production and purification, Kreyenschulte [/bib_ref]. Therefore some isolation procedures may be better adapted than others as reviewed by [bib_ref] Microbial exopolysaccharides: main examples of synthesis, excretion, genetics and extraction, Donot [/bib_ref].
Centrifugation or filtration step is usually used to remove cells from the culture broth without lysis and EPS isolation [bib_ref] Recent advances in microbial biopolymer production and purification, Kreyenschulte [/bib_ref]. Culture broth is sometimes subjected to heating at the end of the fermentation to kill the bacteria and denature potential polymer-degrading enzymes [bib_ref] Advances in bacterial exopolysaccharides: from production to biotechnological applications, Freitas [/bib_ref] but this may result in some cell lysis and release of compounds in the soluble medium. The deactivation of the cells is alternatively carried out by chemical, enzymatic or mechanical treatment [bib_ref] Recent advances in microbial biopolymer production and purification, Kreyenschulte [/bib_ref].
Separation and purification of the biopolymer can be obtained by precipitation with the addition of alcohol or by complexing www.frontiersin.org
October 2014 | Volume 2 | Article 85 | 7 metal ions [bib_ref] Recent advances in microbial biopolymer production and purification, Kreyenschulte [/bib_ref]. However, ethanol precipitation may trap and co-precipitate proteins and ions [bib_ref] Bacterial polysaccharides -a perception, Kumar [/bib_ref] and needs a large amount of alcohol. Filtration or ultrafiltration is recognized as a good method to separate high molecular weight from other small adsorbed compounds and has been used since a long time [bib_ref] The fractionation of polysaccharides by the method of ultrafiltration, Wilkie [/bib_ref] [bib_ref] Recent advances in microbial biopolymer production and purification, Kreyenschulte [/bib_ref]. Some additional extractions to remove contaminating compounds may be appropriate such as new precipitation, chemical extraction or enzymatic treatment but they may decrease the recovery yield [bib_ref] Microbial exopolymer secretions in ocean enviroments: their role(s) in food webs and..., Decho [/bib_ref] [bib_ref] Advances in bacterial exopolysaccharides: from production to biotechnological applications, Freitas [/bib_ref]. These separation steps can in turn be hindered when polysaccharide is highly viscous; higher temperature or dilution in water may facilitate the process however increasing the costs [bib_ref] Advances in bacterial exopolysaccharides: from production to biotechnological applications, Freitas [/bib_ref]. The choice of the whole procedure has to be adapted to polymer characteristics as well as to the desired recovery yield, purity and integrity degrees. After isolation, the polysaccharide is freeze dried for a better conservation. Polysaccharides are highly hydrophilic due to hydroxyl groups, especially when they are polyanionic (carboxyl groups), a widespread feature in marine environment, or when they bear sulfate groups; therefore they always conserve a content of water [bib_ref] Chemoenzymatic synthesis of glycosaminoglycans: re-creating, re-modeling and re-designing nature's longest or most..., De Angelis [/bib_ref].
## Putative pathways of biosynthesis
The genetics of the EPS biosynthesis begins to be better understood; however, information appears disparate because it depends on the concerned microorganism and the polysaccharide that it produces. However, depending on the type of polysaccharide, some general mechanisms can be described [bib_ref] Bacterial polymers: biosynthesis, modifications and applications, Rehm [/bib_ref]. An extracellular glycosyltransferase (GT) is responsible for the biosynthesis outside the cell of homopolysaccharides such as dextran, levan, mutan. This particular GT (usually sucrase) cleaves a disaccharide substrate (usually sucrose) and transfers one of the two obtained monomers to the polymer chain [bib_ref] Bacterial polymers: biosynthesis, modifications and applications, Rehm [/bib_ref]. The occurrence of these enzymes is however limited in marine microorganisms [bib_ref] Microbial exopolymer secretions in ocean enviroments: their role(s) in food webs and..., Decho [/bib_ref].
Except for mechanisms involving sucrase, the biosynthesis starts with the production of nucleotide sugars which will be linked in the repeating unit of the molecule. They are biosynthesized within the central cellular metabolism with usually known enzymes [bib_ref] Bacterial polymers: biosynthesis, modifications and applications, Rehm [/bib_ref]. The repeating unit is then synthesized by appropriate GTs [bib_ref] Biosynthesis and assembly of capsular polysaccharides in Escherichia coli, Whitfield [/bib_ref] [bib_ref] Bacterial polymers: biosynthesis, modifications and applications, Rehm [/bib_ref]. After completion, the repeating unit is exported outside the cell and polymerized on the growing EPS chain. Based on biosynthesis and export mechanisms, three pathways have been described. Some simple heteropolysaccharides (two different residue types maximum such as hyaluronic acid) as well as some homopolysaccharides (cellulose, chitin) are synthesized by a synthase enzyme which polymerizes nucleotide sugars while exporting the growing polymer chain outside the cell [bib_ref] Hyaluronan synthases: a decadeplus of novel glycosyltransferases, Weigel [/bib_ref] [bib_ref] Bacterial polymers: biosynthesis, modifications and applications, Rehm [/bib_ref]. The second pathway relies on the ABC transporter for exportation of the entire polysaccharide synthesized on a lipid carrier [bib_ref] Synthase-dependent exopolysaccharide secretion in Gram-negative bacteria, Whitney [/bib_ref]. The third mechanism depends on Wzx-Wzy proteins.
The Wzx-Wzy dependent mechanism has been widely studied in Gram negative bacteria especially for heteropolysaccharide production. Biosynthesis is catalyzed by a membrane-spanning multiprotein complex [bib_ref] Bacterial polymers: biosynthesis, modifications and applications, Rehm [/bib_ref]. A particular GT involved in the initiation step, the phosphoglycosyltransferase (pGT), anchors the first osidic residue to a membrane lipid carrier (undecaprenyl phosphate) through a phosphoryl bond [bib_ref] Recent developments in the biosynthesis and applications of heteropolysaccharides from lactic acid..., Vuyst [/bib_ref] [bib_ref] Biosynthesis and assembly of capsular polysaccharides in Escherichia coli, Whitfield [/bib_ref]. After completion of the repeating unit by successive GTs, it is exported outside the cell across the inner membrane by Wzx and subsequently polymerized by the Wzy protein [bib_ref] Biosynthesis and assembly of capsular polysaccharides in Escherichia coli, Whitfield [/bib_ref]. The lipid carrier anchor is also recognized for the translocation of the repeating unit across the inner membrane [bib_ref] Bacterial polymers: biosynthesis, modifications and applications, Rehm [/bib_ref]. The final translocation across the outer membrane involves a member of the outer membrane polysaccharide export protein family such as Wza [bib_ref] Functional analysis of conserved gene products involved in assembly of Escherichia coli..., Reid [/bib_ref]. Heteropolysaccharides from the Gram-positive lactic acid bacteria are synthesized by a very similar mechanism [bib_ref] Recent developments in the biosynthesis and applications of heteropolysaccharides from lactic acid..., Vuyst [/bib_ref].
Besides these four types of biosynthesis mechanisms (extracellular, synthase-, Wzx-Wzy-and ABC-transporter), some other gene clusters showing a peculiar mechanism have been described. In general, the overall regulation of the biosynthesis is not completely understood as well.
Sphingans are heteropolysaccharides produced by Sphingomonas bacteria, some of whom are of marine origin [bib_ref] Sphingomonads from marine environments, Cavicchioli [/bib_ref]. These polymers are characterized by a tetrasaccharide backbone structure containing rhamnose or mannose (1), glucose (2) and glucuronic acid (1). Gellan, diutan, welan, rhamsan, and sphingan S-88 are examples of them differing by their side chain and substituents [bib_ref] Organization of genes required for gellan polysaccharide biosynthesis in Sphingomonas elodea ATCC..., Harding [/bib_ref] [bib_ref] Advances in bacterial exopolysaccharides: from production to biotechnological applications, Freitas [/bib_ref]. The biosynthesis of sphingans has been recently reviewed by [bib_ref] A comparison of genes involved in sphingan biosynthesis brought up to date, Schmid [/bib_ref]. The organization of genes required for diutan, welan, gellan, S-88 and S-7 biosynthesis shows similarities; genes for polysaccharide and protein secretion, as well as an operon for the synthesis of dTDP-rhamnose are conserved suggesting a well-conserved mechanism for polysaccharide biosynthesis and secretion [bib_ref] Linkage of genes essential for synthesis of a polysaccharide capsule in Sphingomonas..., Yamazaki [/bib_ref] [bib_ref] Organization of genes required for gellan polysaccharide biosynthesis in Sphingomonas elodea ATCC..., Harding [/bib_ref] [bib_ref] Identification and organization of genes for diutan polysaccharide synthesis from Sphingomonas sp...., Coleman [/bib_ref]. The mechanism starts by the transfer of glucose-1-phosphate on the isoprenylphosphate lipid. Successive GTs transfer the other sugar nucleotide to the repeating unit similarly to the first steps of Wzx-Wzy dependent mechanisms [bib_ref] Identification and organization of genes for diutan polysaccharide synthesis from Sphingomonas sp...., Coleman [/bib_ref] [bib_ref] A comparison of genes involved in sphingan biosynthesis brought up to date, Schmid [/bib_ref]. A putative gene for lyase is only present in diutan gene cluster. Comparison of the three gene clusters allowed also the identification of a candidate gene encoding the protein responsible for the addition of rhamnosyl side chain.
Alginate biosynthesis genes in Azotobacter and Pseudomonas bacteria are similar [bib_ref] Bacterial alginates: biosynthesis and applications, Rehm [/bib_ref] , the 12-core genes are clustered in a single operon. Only slight differences exist especially in the regulation [bib_ref] Microbial exopolysaccharides: main examples of synthesis, excretion, genetics and extraction, Donot [/bib_ref] [bib_ref] Microbial alginate production, modification and its applications, Hay [/bib_ref]. Alginate is synthesized as polymannuronnic acid in an undecaprenolindependent manner [bib_ref] Bacterial alginates: from biosynthesis to applications, Remminghorst [/bib_ref] and is O-acetylated in the periplasm. The gene cluster includes genes for the precursor synthesis (GDP-mannuronic acid), the polymerization, the translocation across the inner membrane and the periplasm where alginate encounters some modifications (O-acetylation and C5-epimerization of mannuronic acid to guluronic acid), and the alginate secretion [bib_ref] Bacterial alginates: from biosynthesis to applications, Remminghorst [/bib_ref] [bib_ref] Microbial alginate production, modification and its applications, Hay [/bib_ref]. In the case of alginate biosynthesis by Azotobacter vinelandii and Pseudomonas fluorescens, the
## Frontiers in chemistry | chemical biology
October 2014 | Volume 2 | Article 85 | 8 molecular mechanisms of polymerization and export are not fully understood [bib_ref] Bacterial polymers: biosynthesis, modifications and applications, Rehm [/bib_ref] , but are based on a synthase-dependent mechanism [bib_ref] Synthase-dependent exopolysaccharide secretion in Gram-negative bacteria, Whitney [/bib_ref]. The groups of genes necessary for the biosynthesis of an heteropolysaccharide are typically clustered at one genetic locus of 12-25 kb including genes for the synthesis of the repeating unit (GTs), genes for export, polymerization and regulation [bib_ref] Biosynthesis, characterisation, and design of bacterial exopolysaccharides from lactic acid bacteria, Laws [/bib_ref]. These clusters are also well known in lactic acid bacteria (LAB) such as Streptococcus spp. [bib_ref] Genomic insights into high exopolysaccharide-producing dairy starter bacterium Streptococcus thermophilus ASCC 1275, Wu [/bib_ref] and Bifidobacterium spp. [bib_ref] Genomic overview and biological functions of exopolysaccharide biosynthesis in Bifidobacterium spp, Hidalgo-Cantabrana [/bib_ref] ; in LAB, EPS biosynthesis gene clusters are often located on plasmids [bib_ref] Bacterial polysaccharides -a perception, Kumar [/bib_ref] [bib_ref] Microbial exopolysaccharides: main examples of synthesis, excretion, genetics and extraction, Donot [/bib_ref]. Furthermore, Escherichia coli polysaccharide biosynthesis pathways have become a reference model in these studies [bib_ref] Biosynthesis of lipopolysaccharide O-antigens, Whitfield [/bib_ref] [bib_ref] Biosynthesis and assembly of capsular polysaccharides in Escherichia coli, Whitfield [/bib_ref] [bib_ref] Structure, biosynthesis, and function of bacterial capsular polysaccharides synthesized by ABC transporterdependent..., Willis [/bib_ref].
The activated precursors are also needed for the synthesis of some other cellular components such as peptidoglycan for membranes [bib_ref] Glycansby-design: engineering bacteria for the biosynthesis of complex glycans and glycoconjugates, Merritt [/bib_ref]. Therefore, it is of high importance to understand the fluxes of carbon, nitrogen and energy leading both to cells and bioactive molecules or, at least, to modulate them by physiological optimizations or genetic engineering [bib_ref] Bacterial polymers: biosynthesis, modifications and applications, Rehm [/bib_ref]. It is highly probable, due to their respective role, that cell wall biosynthesis has priority, LPS and finally EPS synthesis [bib_ref] Microbial exopolymer secretions in ocean enviroments: their role(s) in food webs and..., Decho [/bib_ref].
With the post-genomic era, more clusters involved in biosynthesis of polysaccharides are described together with some regulations issues. An understanding of how the high-molecular weight polymers are biosynthesized may lead to better efficiencies in EPS production at an industrial level and a better comprehension of how varies the composition upon production conditions. New strategies like genetic engineering of producing microorganisms are being developed to enhance polysaccharide yield and allow an economically effective production [bib_ref] Fermentation technologies for the optimization of marine microbial exopolysaccharide production, Finore [/bib_ref]. On the other hand, the prediction of metabolic network can allow the identification of key factors for an enhanced production: from the genome sequence, mannitol was identified as a stimulator for levan biosynthesis by Chromohalobacter salexigens, experimental evidence was achieved later in another halophilic bacterium Halomonas smyrnensis [bib_ref] The stimulatory effect of mannitol on levan biosynthesis: lessons from metabolic systems..., Ates [/bib_ref]. Metabolic engineering approaches can complement ongoing efforts on fermentation process engineering with the aim to optimize EPS production [bib_ref] Glycansby-design: engineering bacteria for the biosynthesis of complex glycans and glycoconjugates, Merritt [/bib_ref]. In this strategy, research on K5 EPS biosynthesis is a good example as reviewed by [bib_ref] Escherichia coli K5 heparosan fermentation and improvement by genetic engineering, Wang [/bib_ref]. Fermentation process has been optimized for a better yield and productivity of heparosan. Conversion of UDP-glucose to UDP-glucuronic acid by the UDP-glucose dehydrogenase together with the UDP-N-Acetyl-glucosamine pathway have been identified as limiting steps to keep balanced supply of nucleotide sugars both for heparosan biosynthesis and cell wall synthesis. Genetic engineering targeting these metabolic reactions has revealed the necessity of a balanced over-expression of KfiA and KfiC GTs. Since a part of the K5 polysaccharide remains linked to the cell membrane, the gene of the K5 lyase capable of breaking this linkage has been genetically modified to increase the amount of K5 released in the supernatant [bib_ref] Escherichia coli K5 heparosan fermentation and improvement by genetic engineering, Wang [/bib_ref]. K5 lyase gene can also be genetically engineered to control the chain length [bib_ref] Escherichia coli K5 heparosan fermentation and improvement by genetic engineering, Wang [/bib_ref].
## Novel strategies for chemical and enzymatic modifications
Overall, physical and bioactive features of polysaccharides are based on molecular chemical structure: the osidic sequence and linkages in the repeating unit, but also the substituents, influence the conformation and geometry of polysaccharide chains as well as polysaccharide-polysaccharide and polysaccharideprotein interactions [bib_ref] Interactions of heparin/heparan sulfate with proteins: appraisal of structural factors and experimental..., Powell [/bib_ref] [bib_ref] Review: an overview about the structure-function relationship of marine sulfated homopolysaccharides with..., Pomin [/bib_ref]. The chemical structure determines the physical properties such as solubility in water or interactions with ions [bib_ref] Uptake of metals by bacterial polysaccharides, Geddie [/bib_ref] [bib_ref] Bacterial polysaccharides -a perception, Kumar [/bib_ref]. For examples, the acetyl content in chitosan sulfate influences the inhibition of propyl endopeptidase which is involved in progressive memory deficits and cognitive dysfunctions [bib_ref] Prolyl endopeptidase inhibitory activity of chitosan sulfates with different degree of deacetylation, Je [/bib_ref]. The sulfate groups in heparin participate in the molecular conformation and influence the binding with calcium cations [bib_ref] The heparin-Ca 2+ interaction: the influence of the O-sulfation pattern on binding, Chevalier [/bib_ref]. They also have a great effect on the anticoagulant activity [bib_ref] Structure-activity relationship of antithrombotic polysaccharide derivatives, Franz [/bib_ref] [bib_ref] Heparin oligosaccharide sequence and size essential for inhibition of pulmonary artery smooth..., Garg [/bib_ref] [bib_ref] Influence of functional groups on the in vitro anticoagulant activity of chitosan..., Huang [/bib_ref] [bib_ref] Anticoagulant heparan sulfate: structural specificity and biosynthesis, Liu [/bib_ref]. The antiproliferative activity of the heparin molecule depends on the molecular size but not on 3-O-sulfo group [bib_ref] Heparin oligosaccharide sequence and size essential for inhibition of pulmonary artery smooth..., Garg [/bib_ref] and requires both N-acetylation and N-sulfation [bib_ref] Heparin antiproliferative activity on bovine pulmonary artery smooth muscle cells requires both..., Longas [/bib_ref]. As found for heparin structure-function relationships, the amount of sulfate groups, their distribution pattern and the molecular weight are of great importance for GAG-like activities. Therefore, marine polysaccharides may be structurally modified e.g., depolymerized and (over-)sulfated to render them active or to enhance already existing activities [bib_ref] Design polysaccharides of marine origin: chemical modifications to reach advanced versatile compounds, Chopin [/bib_ref].
Chemical modifications are widely used for this purpose [bib_ref] Marine derived polysaccharides for biomedical applications: chemical modification approaches, Gomez D'ayala [/bib_ref] [bib_ref] Marine polysaccharides in pharmaceutical applications: an overview, Laurienzo [/bib_ref] [bib_ref] Marine polysaccharides: a source of bioactive molecules for cell therapy and tissue..., Senni [/bib_ref] : acid hydrolysis [bib_ref] A low molecular weight fucoidan fraction from the brown seaweed Pelvetia canaliculata, Colliec [/bib_ref] , radical depolymerization [bib_ref] Anticoagualant low molecular weight fucans produced by radical process and ion exchange..., Nardella [/bib_ref] , N-deacetylation [bib_ref] Oligosaccharide fragments of the type III group B streptococcal polysaccharide derived from..., Zou [/bib_ref] , sulfation [bib_ref] Anticoagulant and antithrombin activities of oversulfated fucans, Nishino [/bib_ref] [bib_ref] Sulfation and depolymerization of a bacterial exopolysaccharide of hydrothermal origin, Guezennec [/bib_ref]. However, several drawbacks such as lack of control and regioselectivity, use of organic solvents, non-homogeneous conditions are identified [bib_ref] Chemical sulfation of small molecules -advances and challenges, Al-Horani [/bib_ref]. More recently, ionic liquids have been used for cellulose sulfation in homogeneous media [bib_ref] Applications of ionic liquids in carbohydrate chemistry: a window of opportunities, El Seoud [/bib_ref] [bib_ref] Tailored media for homogeneous cellulose chemistry: ionic liquid/co-solvent mixtures, Gericke [/bib_ref]. Only few reports deal with chemical modifications on marine EPS [bib_ref] Marine derived polysaccharides for biomedical applications: chemical modification approaches, Gomez D'ayala [/bib_ref]. Low-molecular weight oversulfated derivatives of the EPS GY785 from the deep-sea bacterium A. infernus and of the EPS HE800 from V. diabolicus have been obtained by depolymerization by acid hydrolysis or free-radical reaction followed by sulfation with sulfur trioxide pyridine complex [bib_ref] Characterization, chemical modifications and in vitro anticoagulant properties of an exopolysaccharide produced..., Colliec-Jouault [/bib_ref] [bib_ref] Marine polysaccharides: a source of bioactive molecules for cell therapy and tissue..., Senni [/bib_ref]. The low-molecular weight and oversulfated derivatives thus obtained exhibit biological activities similar to heparin and other GAGs (Ruiz [bib_ref] Effects of a sulfated exopolysaccharide produced by Altermonas infernus on bone biology, Velasco [/bib_ref] [bib_ref] Marine polysaccharides: a source of bioactive molecules for cell therapy and tissue..., Senni [/bib_ref] [bib_ref] Unusual glycosaminoglycans from a deep sea hydrothermal bacterium improve fibrillar collagen structuring..., Senni [/bib_ref].
Target compound yield by chemical process is low after purification steps and sulfation lacks specificity giving undesirable by-products or uncontrolled final product chemical structure resulting in non-homogeneous products [bib_ref] Design polysaccharides of marine origin: chemical modifications to reach advanced versatile compounds, Chopin [/bib_ref]. Enzymes, because of their specificity, may allow a better control of the reactions catalyzed. Moreover, enzymatic reactions are more friendly to the environment without the need of solvent or toxic chemicals.
To date, several strategies combining chemical and enzymatic methods for the synthesis of GAGs have been developed [bib_ref] Chemoenzymatic synthesis of glycosaminoglycans: re-creating, re-modeling and re-designing nature's longest or most..., De Angelis [/bib_ref]. After sugar backbone isolation or synthesis, modifications, usually using enzymes cloned from vertebrates up to now, are performed in vitro.
Escherichia coli K5 polysaccharide has the same chemical structure as the biosynthetic precursor of heparin (heparosan); it has been modified by combinations of chemical and enzymatic methods and converted to "biotechnological heparin" [bib_ref] Biosynthesis of heparin. Use of Escherichia coli K5 capsular polysaccharide as a..., Kusche [/bib_ref] [bib_ref] Toward a biotechnological heparin through combined chemical and enzymatic modification of the..., Naggi [/bib_ref] [bib_ref] Chemically sulfated Escherichia coli K5 polysaccharide derivatives as extracellular HIV-1 Tat protein..., Urbinati [/bib_ref] [bib_ref] Generation of "neoheparin" from E. coli K5 capsular polysaccharide, Lindahl [/bib_ref]. N-deacetylation and N-sulfation were achieved by chemical reactions, and C5-epimerization of glucuronic acid (GlcA) to iduronic acid (IdoA) was carried out using the enzyme epimerase. The sulfation reaction was not regioselective enough and was improved in 2005 to yield the desired 2-O-sulfated IdoA [bib_ref] Generation of "neoheparin" from E. coli K5 capsular polysaccharide, Lindahl [/bib_ref] ; at that time, even if the chemical structure of this neoheparin was still not identical to the heparin AT-binding sequence, biological properties were similar to those of heparin. Other similar chemoenzymatic processes have been patented.
The sulfation step is crucial in these biotechnological processes. Although chemical sulfation is not specific enough, only a few biotechnological processes involve sulfotransferases (STs). In particular, a selectivity toward the sulfation state of the substrate exists [bib_ref] Chemoenzymatic synthesis of glycosaminoglycans: re-creating, re-modeling and re-designing nature's longest or most..., De Angelis [/bib_ref]. Enzymatic sulfations are however limited by the need of the expensive sulfate donor 3phosphoadenosine 5 -phosphosulfate (PAPS). Chondroitin sulfate E has been synthesized from chondroitin sulfate A using a sulfotransferase extracted from squid cartilage [bib_ref] Enzymatic synthesis of chondroitin sulfate E by N-acetylgalactosamine 4-sulfate 6-O-sulfotransferase purified from..., Habuchi [/bib_ref]. In this study, dermatan sulfate has also been oversulfated with the same enzyme. Enzymatic sulfation has been used to prepare heparan sulfate from chemically desulfated N-sulfated heparin [bib_ref] Enzymatic redesigning of biologically active heparan sulfate, Chen [/bib_ref]. Using heparan sulfotransferases (2-OST, 6-OST, 3-OST) expressed in E. coli and immobilized, as well as PAPS regeneration system, the polysaccharide substrate was subjected to different enzymatic modifications resulting in heparan sulfates with distinct biological activities.
However, although promising, chemoenzymatic methods still need more research efforts to allow the synthesis of GAG-like macromolecules. These multiple step processes are expensive due to the PAPS cost and the production capacity remains limited for a demand in heparin of about 100 tons per year (De Angelis, 2012).
Chitin and chitosan have been extensively studied for their biotechnological applications and have been subjected to a large range of modifications to modulate their biological properties, hence determining their applications [bib_ref] Marine derived polysaccharides for biomedical applications: chemical modification approaches, Gomez D'ayala [/bib_ref]. Removal of acetyl groups (deacetylation) from chitin is the first of the studied modifications among others including carboxymethylation, sulfation, acylation [bib_ref] Marine derived polysaccharides for biomedical applications: chemical modification approaches, Gomez D'ayala [/bib_ref].
A new alternative to chemical and enzymatic modifications to obtain targeted polysaccharide chemical structure could also be the genetic engineering. This will be feasible if more EPS structures are known together with biosynthetic genetic clusters as well as relationship between structure and bioactivity. EPS biosynthesis pathway reconstruction is feasible with genome sequence and functional annotation and it can be engineered to obtain tailor-made polymers [bib_ref] A comparison of genes involved in sphingan biosynthesis brought up to date, Schmid [/bib_ref]. If biosynthesis mechanisms are known, the over-expression or inhibition of targeted genes may lead to enhanced EPS production. It may also allow a better insight into protein function as well as the biosynthesis of tailored molecules with desired features in terms of molecular weight, osidic structure and functional substituents. The activity of GT GumK involved into the biosynthesis of xanthan has been modified by protein engineering; this resulted in a variation in the xanthan production yield highlighting the possibility to obtain tailor-made xanthan molecules by GT engineering [bib_ref] Structure and mechanism of GumK, a membrane-associated glucuronosyltransferase, Barreras [/bib_ref]. Attempts to determine genes involved in the molecular weight control of diutan in Sphingomonas sp. were not successful [bib_ref] Identification and organization of genes for diutan polysaccharide synthesis from Sphingomonas sp...., Coleman [/bib_ref] , suggesting that the chain length regulation is very complex whatever the biosynthesis mechanism considered [bib_ref] Polymerases: glycan chain-length control, Whitfield [/bib_ref]. In Pseudomonas aeruginosa, the GDP-mannose dehydrogenase has been identified by over-expression studies as a key regulator protein in alginate biosynthesis [bib_ref] GDP-mannose dehydrogenase is the key regulatory enzyme in alginate biosynthesis in Pseudomonas..., Tatnell [/bib_ref].
On the other hand, genomic data would also allow the identification of new enzymatic tools to modify glycopolymers. Enzymes would become promising biotechnological tools for in vitro synthesis or in vivo biosynthesis engineering. The marine biodiversity has already shown a great potential in providing new biocatalysts [bib_ref] Potential biocatalysts originating from sea environments, Trincone [/bib_ref]. Depolymerising enzymes can be used in in vitro depolymerization process but also as tools to study the chemical structure. Carbohydrate sulfotransferases and other enzymes grafting substituents such as acetate, as well as enzymes catalyzing their removal, may be used in vitro for binding or for elimination of substituents which are of great importance for the final bioactivity of the molecule. GTs specificity are difficult to characterize especially when they belong to the polyspecific family 4 or 2 (http://www.cazy.org/). Identifying both the polysaccharide chemical structure and the genetic cluster for the biosynthesis would allow the knowledge of the GT enzymatic function in the biosynthetic pathway by genetic knockout. It would also provide useful directions to determine enzymatic specificity and characterize the enzymes.
Recombinant production can be envisioned when polysaccharide biosynthesis cluster is known: it consists in the cloning of the entire biosynthesis cluster in an appropriate heterologous host. This is a new strategy developed together with the expansion of synthetic biology [bib_ref] Synthetic biological approaches to natural product biosynthesis, Winter [/bib_ref] [bib_ref] A brief history of synthetic biology, Cameron [/bib_ref] [bib_ref] Realizing the potential of synthetic biology, Church [/bib_ref]. Up to now, only less complex polymers have been studied in recombinant production such as hyaluronic acid [bib_ref] Hyaluronic production by recombinant Lactococcus lactis, Chien [/bib_ref] [bib_ref] Production of glucuronic acidbased polysaccharides by microbial fermentation for biomedical applications, Cimini [/bib_ref] , chondroitin [bib_ref] Molecular cloning and characterization of chondroitin polymerase from Escherichia coli strain K4, Ninomiya [/bib_ref] and heparosan [bib_ref] Overexpression of UDP-glucose dehydrogenase in Escherichia coli results in decreased biosynthesis of..., Roman [/bib_ref].
## Concluding remarks
Natural bioactive molecules attract many interests in the search for new therapeutic drugs. Marine environment shields a high diversity of natural products and can be indeed a treasure chest for industrial and pharmaceutical purposes. Due to their functions in survival and competitiveness of marine bacteria in low nutrients and adverse environments, the EPS are ubiquitous in the marine environment.
In addition to various marine origins (animals, seaweeds, invertebrates), microorganisms provide glycopolymers with original chemical structure and promising biological activities. The
## Frontiers in chemistry | chemical biology
October 2014 | Volume 2 | Article 85 | 10 marine biotechnology has not yet reached an economically significant field but is a promising field for sustaining convenient macromolecules.
It is also assumed that new isolated bioactive molecules are also the matter for bioinspired new molecules obtained by synthesis. For this purpose, both the discovery of new natural bioactive molecules and their structural characterization are still of considerable importance to obtain well characterized molecules with truly identified mechanism of action. Basic and applied research efforts in this field require a close collaboration between biologists and chemists and expertise in marine microbiology, biochemistry, biology, chemistry and computational sciences to fulfill screenings, structural characterization, bioactivity studies and biosynthesis understanding and metabolic engineering issues. With the synthetic biology approach, a new era is available for the polysaccharide production field by simple ways.
[table] Table 2 |: Repeating unit chemical structures of main glycosaminoglycans and of some marine exopolysaccharides. [/table]
|
A case report of nonepisodic angioedema with eosinophilia in a Korean patient and a review of the Korean literature
Episodic angioedema with eosinophilia (EAE) is characterized by recurrent angioedema, peripheral eosinophilia, elevated serum IgM, fever, weight gain, and a benign course lacking any internal organ involvement. Dozens of cases of the nonepisodic variant (NEAE), which is limited to a single attack, have been reported in Japan. These NEAE cases normally have been less severe than the episodic type. In this paper, we describe the case of a Korean patient whose clinical and laboratory findings were consistent with NEAE, and review five other cases of EAE and NEAE reported in the Korean literature. The Korean NEAE cases outlined in this paper demonstrate that, as in Japan, NEAE is not uncommon in Korea, and also suggest that this disease exhibits a cultural predilection for Asian populations.
# Introduction
Episodic angioedema with eosinophilia (EAE), originally described by , is a syndrome characterized by recurrent episodes of angioedema, fever, leukocytosis, eosinophilia, elevated serum IgM, increased body weight, and a benign clinical course lacking any internal organ involvement. Many other cases have been detected since this initial report, and this syndrome has been recognized as a distinct clinical entity [bib_ref] Episodic angioedema with eosinophilia, Wolf [/bib_ref] [bib_ref] Increased interleukin-6 production during the acute phase of the syndrome of episodic..., Gosset [/bib_ref]. More than 40 cases with symptomatic presentations and laboratory findings similar to EAE have been reported, but these cases have not involved recurrent attacks, and thus have been subsequently designated as nonepisodic angioedema with eosinophilia (NEAE) [bib_ref] Nonepisodic angioedema associated with eosinophilia: report of 4 cases and review of..., Chikama [/bib_ref] [bib_ref] The cytokine profile in a transient variant of angioedema with eosinophilia, Mizukawa [/bib_ref] [bib_ref] Nonepisodic angioedema with eosinophilia: a report of two cases and a review..., Matsuda [/bib_ref] [bib_ref] A case of angioedema associated with eosinophilia, Ko [/bib_ref] [bib_ref] Two different cases of episodic angioedema with eosinophilia, Park [/bib_ref]. The majority of these NEAE cases have been reported in Japanese patients, with the exception of three cases reported in the Korean literature [bib_ref] A case of angioedema associated with eosinophilia, Ko [/bib_ref] [bib_ref] Two different cases of episodic angioedema with eosinophilia, Park [/bib_ref] [bib_ref] A case of nonepisodic angioedema associated with eosinophila, Jung [/bib_ref]. No cases have yet been reported in Western populations. This ethnic or regional distribution may be an important characteristic of NEAE. In this paper, we describe a patient whose clinical and laboratory features were similar to those observed in previous cases of NEAE, and also present our review of cases thus far reported in Korea.
## Case report
A 26-year-old Korean woman presented with edema of her hands and lower legs, which had developed 2 weeks prior to her hospital visit. Her body weight had increased by 2 kg over those 2 weeks, and this was her first episode of the reported symptoms. The edema was distributed symmetrically on both of the patient's upper and lower extremities, and was pitted. She denied any febrile sensation and had a pruritic rash on her lower legs, and had also experienced mild joint pain in both knees 1 week after the onset of the edema. Her family and past medical histories, including any allergic diseases, were unremarkable. She also denied taking any drugs, including Oriental medications, with the exception of one acetaminophen tablet (300 mg) 1 week prior to the onset of edema. Her symptoms did not appear to be related to her menstrual cycle.
Laboratory tests revealed leukocytosis (12,250/mm 3 ; normal range: 4,000-10,000/mm 3 ) with 51.5% eosinophils (normal range: 0～7%) in the peripheral blood. The eosinophil count was 6,300/mm [bib_ref] Increased interleukin-6 production during the acute phase of the syndrome of episodic..., Gosset [/bib_ref]. Serum IgE had increased to 1,423 IU/mL (normal range: 0-170 IU/mL), but IgA, IgG, and IgM levels were all within normal limits. Antinuclear antibody was negative, as were stool ova and serum antibodies to parasites. No clinical findings were suggestive of autoimmune disease or parasite infestation. No abnormal findings were observed in the serum chemistry that would be reflective of renal or hepatic dysfunction. No proteinuria was detected. The rash on the patient's lower legs was examined histologically, and identified as perivascular eosinophil infiltration. Lymphocyte phenotyping was conducted, but did not indicate the presence of any abnormal lymphocyte clones.
We diagnosed the patient with NEAE, and prescribed an oral antihistamine to control the pruritus. One week after the patient began taking antihistamines, the edema and rash began to evidence improvements. The knee arthralgia also clearly improved. The eosinophil count decreased to 1,530/mm 3 after 2 weeks and to 736/mm 3 after 8 weeks. The patient was free of symptoms at 4 months, and had an eosinophil count of 550/mm [bib_ref] Increased interleukin-6 production during the acute phase of the syndrome of episodic..., Gosset [/bib_ref]. We followed the patient for 18 months, and observed no recurrence of either the symptoms or the eosinophilia (the eosinophil count was 180/mm [bib_ref] Increased interleukin-6 production during the acute phase of the syndrome of episodic..., Gosset [/bib_ref].
# Discussion
Our patient presented with eosinophilia with angioedema localized to the upper and lower extremities, and evidenced features typical of the NEAE cases thus far reported in Japan. That is, the patient was a young female exhibiting a mild weight gain, and her condition lacked internal organ involvement, improved without corticosteroids, involved no increase in IgM levels, and did not recur over the course of 18 months. The skin biopsy indicated perivascular eosinophil infiltration. EAE is characterized by recurrent edema and requires the administration of corticosteroids, and has also been reported principally in Europe and the United States. However, in Japan and Korea, more than 40 cases have been reported with clinical symptoms similar to those observed in EAE, but without recurrent episodes; these patients recovered without corticosteroid treatment [bib_ref] Nonepisodic angioedema associated with eosinophilia: report of 4 cases and review of..., Chikama [/bib_ref] [bib_ref] The cytokine profile in a transient variant of angioedema with eosinophilia, Mizukawa [/bib_ref] [bib_ref] Nonepisodic angioedema with eosinophilia: a report of two cases and a review..., Matsuda [/bib_ref] [bib_ref] A case of angioedema associated with eosinophilia, Ko [/bib_ref] [bib_ref] Two different cases of episodic angioedema with eosinophilia, Park [/bib_ref]. Chikama et al. first gave this distinct form of EAE the designation NEAE [bib_ref] Nonepisodic angioedema associated with eosinophilia: report of 4 cases and review of..., Chikama [/bib_ref] , which has been loosely identified as a less severe form of EAE [bib_ref] Nonepisodic angioedema with eosinophilia: a report of two cases and a review..., Matsuda [/bib_ref]. The most notable aspect of NEAE is the absence of recurrence, but several other differences also exist with regard to the clinical features of NEAE and EAE. One of these differences is the distribution of the edema. In NEAE, the angioedema is normally localized to the hands and the proximal portion of the lower legs, as in the present case, whereas in EAE, the angioedema frequently extends farther down the extremities, sometimes including the face [bib_ref] A case of episodic angioedema associated with eosinophilia, Shikiji [/bib_ref]. The edema in EAE is attended by a substantial increase in body weight. An increase in serum IgM levels is also normally detected in cases of EAE, but is seldom observed in NEAE [bib_ref] Nonepisodic angioedema with eosinophilia: a report of two cases and a review..., Matsuda [/bib_ref] [bib_ref] A case of episodic angioedema associated with eosinophilia, Shikiji [/bib_ref]. Almost all EAE patients require corticosteroid therapy, whereas spontaneous remission frequently occurs in patients suffering from NEAE. Additionally, all NEAE patients thus far have been young females, as in the case described herein [bib_ref] Nonepisodic angioedema associated with eosinophilia: report of 4 cases and review of..., Chikama [/bib_ref] [bib_ref] The cytokine profile in a transient variant of angioedema with eosinophilia, Mizukawa [/bib_ref] [bib_ref] Nonepisodic angioedema with eosinophilia: a report of two cases and a review..., Matsuda [/bib_ref] [bib_ref] A case of angioedema associated with eosinophilia, Ko [/bib_ref] [bib_ref] Two different cases of episodic angioedema with eosinophilia, Park [/bib_ref].
## B c a a b
The pathogenesis of EAE and NEAE remains to be thoroughly elucidated. The cause of the sudden increase of eosinophils remains unknown, as with idiopathic hypereosinophilic syndrome. The identification of abnormal clonal CD3 -CD4 + cells in the blood of some EAE sufferers [bib_ref] Indolent course of a patient with hypereosinophilic syndrome associated with clonal T-cell..., Zenone [/bib_ref] indicates that EAE and NEAE may be, in some if not all patients, a disease that involves abnormal T cell clones, similarly to the lymphocytic variant of hypereosinophilic syndrome [bib_ref] Recent advances in pathogenesis and management of hypereosinophilic syndromes, Roufosse [/bib_ref]. The predilection to affect the skin with an absence of internal organ involvement in EAE and NEAE is also a clinical feature reminiscent of the lymphocytic variant of hypereosinophilic syndrome [bib_ref] Recent advances in pathogenesis and management of hypereosinophilic syndromes, Roufosse [/bib_ref]. In the case described herein, we conducted lymphocyte phenotyping in order to determine the presence of any aberrant clonal lymphocytes, as has been observed in cases of the lymphocytic variant of hypereosinophilic syndrome [bib_ref] Abnormal clones of T cells producing interleukin-5 in idiopathic eosinophilia, Simon [/bib_ref]. However, our results in this regard were negative. The angioedema of EAE and NEAE is likely associated with a temporal increase in the vascular permeability of local capillaries via some mediators, including antiendothelial cell antibodies and cytokines [bib_ref] Increased interleukin-6 production during the acute phase of the syndrome of episodic..., Gosset [/bib_ref] [bib_ref] Presence of antibodies against endothelial cells in the sera of patients with..., Lassalle [/bib_ref] [bib_ref] Episodic angioedema with eosinophilia: a case associated with T cell activation and..., Putterman [/bib_ref] [bib_ref] Clonal T-helper lymphocytes and elevated IL-5 levels in episodic angioedema and eosinophilia..., Morgan [/bib_ref]. Several reports have shown that histamine levels are within normal limits, and that IL-1, IL-5, and IL-6 generated by monocytes, endothelial cells, T cells, and eosinophils increase during the acute phase of EAE [bib_ref] Increased interleukin-6 production during the acute phase of the syndrome of episodic..., Gosset [/bib_ref] [bib_ref] The cytokine profile in a transient variant of angioedema with eosinophilia, Mizukawa [/bib_ref] [bib_ref] Indolent course of a patient with hypereosinophilic syndrome associated with clonal T-cell..., Zenone [/bib_ref] [bib_ref] Episodic angioedema with eosinophilia: a case associated with T cell activation and..., Putterman [/bib_ref] [bib_ref] Clonal T-helper lymphocytes and elevated IL-5 levels in episodic angioedema and eosinophilia..., Morgan [/bib_ref]. According to a recent report, IL-5 levels were determined to be lower, and TNF-αlevels were higher in the acute phase of NEAE than in the acute phase of EAE. Thus, the cytokine profile may be related with the existence or absence of recurrent attacks [bib_ref] The cytokine profile in a transient variant of angioedema with eosinophilia, Mizukawa [/bib_ref].
Recently, several cases of EAE and NEAE have been reported in Korea [bib_ref] A case of angioedema associated with eosinophilia, Ko [/bib_ref] [bib_ref] Two different cases of episodic angioedema with eosinophilia, Park [/bib_ref] [bib_ref] A case of nonepisodic angioedema associated with eosinophila, Jung [/bib_ref] [bib_ref] A case of episodic angioedema with eosinophilia, Yang [/bib_ref]. shows these six cases of EAE and NEAE, including the case described in this paper. Cases 1, 2, and the present case were readily diagnosed as NEAE, whereas cases 3 and 5 were thought to be EAE. In case 4, certain aspects, such as the elevated serum IgM level and the necessity for corticosteroid therapy to control the symptoms, supported a diagnosis of EAE. However, the presence of other aspects, including the fact that it occurred in a young female, the absence of weight gain, and the absence of a previous episode or recurrence, was suggestive of NEAE. This series of NEAE and EAE case reports from Korea demonstrates that NEAE and EAE are not rare diseases within the Korean population, and also suggests that NEAE may have a racial predisposition toward Asian populations.
In this work, we have described the case of a Korean patient with NEAE and have reviewed five other cases of NEAE and EAE occurring in Korean patients. As in the Japanese population, NEAE is not an uncommon disease within the Korean population. Further studies will be required in order to clarify the pathogenesis of NEAE.
[fig] Figure 1: (A) Edema of both hands. (B) Edema of both lower legs and feet. Pruritic eruptions were also noted. (C) Pruritic eruptions on the lower legs. [/fig]
[fig] Figure 2: (A) Skin biopsy section, taken from a skin lesion on the left lower extremity, shows perivascular eosinophil infiltration. (H&E, x100). (B) (H&E, x400). [/fig]
|
Alemtuzumab in Multiple Sclerosis: Short- and Long-Term Effects of Immunodepletion on the Peripheral Treg Compartment
Treatment with alemtuzumab is followed by an early increase in Treg frequencies.Whether naïve and memory subsets are differentially affected and how depletion influences dysfunctional MS-Treg is unclear. In this study, we analyzed the effect of alemtuzumab on regulatory T-cells (Treg) in patients with multiple sclerosis (MS). For this purpose 182 blood samples from 25 MS patients were taken shortly before treatment and serially for up to 24 months after two alemtuzumab cycles. We studied Treg by flow cytometry (quantitation, phenotypical characterization), real-time polymerase chain reaction (T-cell receptor (TCR) excision circles [TREC] content), CDR3-spectratyping (clonal distribution), and proliferation assays (suppressive function). CD52-mediated cytolysis of Treg and conventional T-cells was determined by a complement-dependent cytolysis assay. Our studies revealed that 1 week post-alemtuzumab, Treg were depicted at constant frequencies among CD4 + T-cells. In contrast, Treg frequencies were massively increased at month 1. Post-depletional Treg exhibited a CD45RO + memory phenotype, a skewed TCR repertoire, and contained minimum TREC numbers. Naïve Treg, thymic markers, and TCR-variability commenced to rise after 6 months but did not attain baseline levels. In vitro, Treg exhibited higher susceptibility to lysis than Tcon. Treg suppressive function constantly increased within 1 year when co-cultured with syngeneic T-cells, but remained stable against allogeneic T-cells from normal donors. Our findings suggest that (1) Treg are not spared from alemtuzumab-mediated depletion and thymopoiesis does not considerably contribute to long-term recovery, (2) either homeostatic proliferation and/or conversion from residual Tcon contributes to Treg expansion during the early post-treatment phase (3) the enhanced inhibitory effect of Treg following alemtuzumab is due to altered composition and reactivity of post-depletional Tcon.
Treatment with alemtuzumab is followed by an early increase in Treg frequencies.
Whether naïve and memory subsets are differentially affected and how depletion influences dysfunctional MS-Treg is unclear. In this study, we analyzed the effect of alemtuzumab on regulatory T-cells (Treg) in patients with multiple sclerosis (MS). For this purpose 182 blood samples from 25 MS patients were taken shortly before treatment and serially for up to 24 months after two alemtuzumab cycles. We studied Treg by flow cytometry (quantitation, phenotypical characterization), real-time polymerase chain reaction (T-cell receptor (TCR) excision circles [TREC] content), CDR3-spectratyping (clonal distribution), and proliferation assays (suppressive function). CD52-mediated cytolysis of Treg and conventional T-cells was determined by a complement-dependent cytolysis assay. Our studies revealed that 1 week post-alemtuzumab, Treg were depicted at constant frequencies among CD4 + T-cells. In contrast, Treg frequencies were massively increased at month 1. Post-depletional Treg exhibited a CD45RO + memory phenotype, a skewed TCR repertoire, and contained minimum TREC numbers. Naïve Treg, thymic markers, and TCR-variability commenced to rise after 6 months but did not attain baseline levels. In vitro, Treg exhibited higher susceptibility to lysis than Tcon. Treg suppressive function constantly increased within 1 year when co-cultured with syngeneic T-cells, but remained stable against allogeneic T-cells from normal donors. Our findings suggest that (1) Treg are not spared from alemtuzumab-mediated depletion and thymopoiesis does not considerably contribute to long-term recovery, (2) either homeostatic proliferation and/or conversion from residual Tcon contributes to Treg expansion during the early post-treatment phase (3) the enhanced inhibitory effect of Treg following alemtuzumab is due to altered composition and reactivity of post-depletional Tcon.
Keywords: regulatory t cells, human, immunodepletion, alemtuzumab, multiple sclerosis INTRODUCTION Alemtuzumab, a humanized anti-CD52 monoclonal antibody (mAb), has consistently been shown to provide higher efficacy than the baseline disease-modifying agent interferon-beta1 and is approved in more than 65 countries for use in patients with active relapsing-remitting multiple sclerosis (RRMS). Treatment with alemtuzumab leads to a depletion of circulating B-and T-lymphocytes, which is followed by reconstitution and rebalancing of the immune system, resulting in prolonged RRMS disease suppression. This process is characterized by differing alterations in numbers and proportions of different lymphocyte subsets. In contrast to the clear-out of conventional CD4 + T-cells (Tcon), an increase in frequencies of regulatory T-cells (Treg) early after initiating alemtuzumab therapy has been reported. This selective enrichment in Treg of the CD4 + CD25 hi CD127 low FOXP3 + phenotype is of particular interest, as these cells are important regulators of the immune system. In vitro, Treg suppress Tcon immune responses in a dose-dependent manner, suggesting a potential impact of treatment-induced changes in the Treg/Tcon ratio. Alemtuzumab predominantly affects CD4 + T-cells exhibiting a CD45RO memory phenotypeand, thus, may also exert different differential effects on naïve and memory Treg subsets. Elucidating this aspect is important because Treg are functionally deficient in patients with MS, as we and others have previously shown. Of note, the loss of Treg suppressive properties is precipitated by contraction of CD45RA + naïve and CD45RA + CD31 + recent thymic emigrant (RTE) subtypes and reciprocal expansion of memory phenotypes (CD45RO + ) within peripheral Treg, a feature which possibly arises from premature immunosenescence. To decipher in detail the effects of alemtuzumab on the Treg compartment, we undertook a longitudinal study of 25 MS patients undergoing two cycles of treatment with alemtuzumab and assessed different parameters of Treg neogenesis and Treg suppressive function.
# Methods
## Human samples
Our study included 25 patients with RRMS established according to the 2011 McDonald criteria, all of whom were scheduled for treatment with alemtuzumab, and 21 healthy donors. Alemtuzumab was administered as primary immunotherapy in 15 and as escalating treatment in 10 individuals, respectively. The mean disease duration was 6.3 years (range 2-13), the median age 32.3 years (17-53), and the median Expanded Disability Status Scale (EDDS) score 3.5 (1.0-6.5). During the observation period a total of nine relapses (year 1 post-treatment: n = 1, year 2 post-treatment: n = 8) occurred in eight patients, and one patient developed secondary autoimmune thyroiditis within 10 months after the second alemtuzumab administration. All patients were recruited in the Department of Neurology, University Hospital Heidelberg. Samples from 14/25 patients were repeatedly assessed over a period of 12 months after the second cycle of treatment. A total of 182 peripheral blood specimens (50-70 ml of EDTA blood and 10 ml of serum) were taken, directly before infusion and repeatedly thereafter (at day 7 and months 1, 3, 6, and 12 after each cycle). Plasma and serum samples were immediately stored at −70 - C. The protocol was approved by the University Hospital Heidelberg ethics committee; all individuals gave written informed consent.
## Cell separation
Peripheral blood mononuclear cells (PBMCs) were isolated by Ficoll-gradient centrifugation (Biochrom, Berlin, Germany). Total Treg and Tcon were highly enriched from freshly isolated PBMCs by means of immunomagnetic separation utilizing Dynabead technology as previously described.
## Flow cytometry
For quantitative and phenotypic characterization of Treg, Treg subsets, and Tcon subsets we used established multicolor flow cytometry protocols and gating strategies. In short, freshly isolated PBMCs were immediately stained with a panel of mAbs specific for human Treg markers To determine surface expression of CD52 on Treg and Tcon, PBMCs obtained from five healthy donors were costained with mAbs specific for Treg and Tcon and naive or memory phenotypes (see above) and a mAb specific for human CD52 (Alexa Fluor R 488 conjugated, BD Biosciences). Mean fluorescence intensities (MFI) for CD52 were then determined in gated Tcon and Treg and in Tcon and Treg subtypes. Detection and quantification of Treg with two different T-cell receptor (TCR) Vα chains were achieved using a previously described protocol. Briefly, fresh PBMCs were stained for Treg markers (see above) and with mAbs specific for human TCR-Vα2 (FITC-conjugated, Pierce) and Vα12 (APC-labeled with Zenon Mouse IgG Labeling Kit, Molecular Probes), identifying Vα2 + , Vα12 + as well as Vα2 + Vα12 + (double-positive) cells in gated Treg to calculate proportions of dual TCR cells as described. To quantify alemtuzumab-induced cytolysis in vitro, Tcon or Treg tested in complement-dependent cytolysis assays were washed and then co-stained with mAbs specific for Treg and Tcon subtypes (see above) and propidium iodide (PI). Cytolysis rates for each subtype were calculated from the proportions of PI-positive cells after exposure to alemtuzumab minus the proportions of PI-positive cells after exposure to a control antibody.
## Cell pro-liferation assay
To measure inhibitory capacities, patient-or donor-derived Treg were tested using in vitro proliferation assays against (a) syngeneic Tcon (obtained from the same patient) and (b) allogeneic Tcon (obtained from a frozen pool of Tcon from eight healthy donors). An amount of 4 × 10 4 Tcon (either syngeneic or allogeneic) was incubated alone or in coculture with 1 × 10 4 Treg (Treg/Teff ratio 1:4) and polyclonally activated by adding soluble anti-CD3 (1 µg/ml) and anti-CD28 mAbs (1 µg/ml). After 4 days, the cells were pulsed for 16 h with 1 µCi of[H]-thymidine per well. After harvesting, T-cell proliferation was measured with a scintillation counter.
## Complement-dependent cytolysis assay
To screen for possible differences in alemtuzumab effects on Tcon and on Treg subsets, 2 × 10 5 freshly isolated total Treg or Tcon obtained from five healthy subjects were incubated for 3 h at 37 - with 10 µg/ml alemtuzumab (Genzyme) or control human IgG (Jackson ImmunoResearch) and 10% normal human complement (Quidel) in X-vivo 15 medium (BioWhittaker) supplemented with 10% FBS (Biochrom). After the incubation period, cells were washed and re-suspended in medium for subsequent flow cytometry analysis.
## Quantification of t-cell receptor excision circles
Total DNA was extracted from freshly isolated Treg (1-3 × 10 5 cells) using a QIAmp Blood Mini Kit (Qiagen) according to the manufacturer's protocol. Numbers of Tcell receptor-excision circles (TRECs) were determined by real time PCR as described elsewhereand expressed as TRECs/10 6 Treg.
## Cdr3 spectratyping
First, total RNA was isolated from freshly isolated Treg (2-5 × 10 5 cells) using a Rneasy Mini Kit (Qiagen) according to the manufacturer's protocol and then converted to cDNA using a SuperScript R III First-Strand Synthesis System (Invitrogen). For real-time PCR detection of all 24 TCR Vβ subfamilies, 100 ng of cDNA was set in eight parallel PCR reactions, each containing three different Vβ specific primer sets and a 6-FAM-labeled C-primer for subsequent CDR3 spectratyping. Size distribution of 6-FAM-labeled PCR products was determined by laser-induced capillary electrophoresis with an automated DNA analyzer A310 and GeneScan software (Applied Biosystems). The complexity score within a Vβ subfamily was determined by counting the number of peaks per subfamily. Normal transcript size distribution consists of eight peaks for each V subfamily. A complexity score (CS) was calculated as the sum of scores of all 24 TCR Vβ subfamilies.
# Statistical analysis
To determine whether differences in cell counts, inhibitory capacities, TREC levels, and TCR diversities were statistically significant, we performed non-parametric one-sample Wilcoxon sign rank tests using a two-tailed distribution with paired samples. No analysis of variance [ANOVA] was carried out, because data sets were not complete for all patients and all time points analyzed. A p value of ≤ 0.05 was considered to show be statistically significant difference in a descriptive sense. Two-sided t-tests and paired t-tests were used to compare normally distributed samples (cytolysis assay, CD52 expression). Again, a p value of ≤ 0.05 was considered to be statistically significant.
# Results
## Treg recovery
Shortly after immune cell depletion by alemtuzumab (week 1) a concomitant reduction in numbers of circulating CD4 + T cells and Treg (2 ± 2/µl [Treg cell count]), was observable resulting in constant relative proportions of Treg (3.5 ± 2.4% [of CD4 + T cells]) among circulating CD4 + T-cells compared with baseline (19 ± 6/µl, 2.7 ± 1.5%; p = 0.008 and 0.285).
In contrast, Treg numbers had recovered to 11 ± 5/µl at 1 months, translating into a massive increase in percentages among CD4 + T cells (18.1 ± 9.3%; p = 0.022 and < 0.0001). Treg expansion then decreased, but proportions remained consistently higher at baseline throughout year 1 and year 2;;. At month 1, almost all Treg exhibited a CD45RO + memory phenotype (Treg memory ). The dominance of Treg memory persisted throughout the study period, and CD45RO + Treg accounted for ≥ 90% of CD4 + T-cells until month 12 of each treatment cycle;. In contrast, frequencies of Treg with a naïve or RTE phenotype (Treg naïve , Treg RTE ) were barely detectable one month posttreatment and slowly recovered thereafter without returning to baseline levels. Whereas, total Treg at baseline comprised on average 23 and 12% Treg naive and Treg RTE , Treg naive accounted for 8% (p < 0.001) and 7% (p < 0.001) of T-cells and Treg RTE for 6% (p < 0.01) and 5% (p < 0.01) of circulating CD4 + Tcells after 12 and 24 months, respectively;.
Hence, these observations indicate that Treg are lysed together with other CD4 + T-cells and-in response to treatment-induced lymphopenia-then either expand to preferentially acquire a memory phenotype and/or are converted from post-depletional Tcon. Together with the decline in frequencies of Treg RTE , these changes-similar to what has been reported for CD4 + T-cellsdo not favor replacement of Treg by thymic neogenesis.
## Treg origin and clonal diversity
The post-depletional Treg population contained invariably decreased proportions of dual-TCR cells and markedly contracted TREC numbers compared with Treg assessed at baseline, and the decline in these thymic-dependent markers persisted throughout months 1-12 after each alemtuzumab cycle;. Furthermore, as determined by CDR3 spectratyping post-treatment, Treg exhibited a reduced mean complexity score, reflecting a more pronouncedly constricted TCR repertoire than at baseline, a feature that lasted until month 12 of each treatment period;.
Altogether, these findings are consistent with the notion that homeostatic proliferation of cells that have escaped depletion by alemtuzumab, rather than meaningful release of newly generated cells from the thymus, dominates the recovery phase in the Treg compartment.
## Treg suppressive function
When tested in cocultures with syngeneic effector T-cells, the suppressive function of post-depletional Treg remained constantly increased vs. baseline over the entire study period. In contrast, when co-cultured with allogeneic T-cells from normal donors, Treg inhibitory capacities remained more or less stable compared with their performance at baseline (BT: 29.9 ± 6.3%; year 1: 1 month: 33.5 ± 5.5%, p = 0.28; 6 months: 33.2 ± 5.5%, p = 0.33; 1 year: 30.8 ± 7.7%, p = 0.80; year 2: 1 month: 33.4 ± 5.6%, p = 0.29; 6 months: 32.2 ± 3.8%, p = 0.45; 1 year: 30.7 ± 6.0%, p = 0.82) and clearly reduced as compared to their counterparts from healthy control donors (50.2 ± 4.6%, p < 0.001;. Thus, repopulated Tcon obviously exhibited a decreased ability to resist regulation, thereby mitigating the MS-associated Treg defect previously described.
## Repopulation of ccr7 + and ccr7 − tcon
It was reported that treatment with alemtuzumab affects the balance between CD4 + T cells either expressing or lacking the C-C chemokine receptor type 7 (CCR7). CCR7 is involved in homing of T cells to various secondary lymphoid organs including lymph nodes. We could previously demonstrate that in fingolimod-treated MS patients CCR7negative Tcon expand and are more prone to Treg mediated suppression when tested in in vitro proliferation assays, thereby indirectly upregulating Treg efficiency. We therefore analyzed frequencies of circulating CCR7 + and CCR7 − CD4 + T-cell subsets before and post treatment with alemtuzumab in a subcohort of eight patients in the 1st year post-alemtuzumab by flow cytometry. At month 1, proportions of CCR7expressing Tcon phenotypes (CCR7 + CD45RO + central memory cells, TCM; CCR7 + CD45RO − naive cells, TCRA) were massively decreased (BT: 67.9 ± 10.6% (mean percent of Tcon); 1 month 24.5 ± 6.1%, p < 0.0001) and were still reduced at 6 and12 months (35.6 ± 5.9%, p < 0.001). In contrast, the frequencies of CCR7 − Tcon subtypes (CCR7 − CD45RO + effector memory cells, TEM; CCR7 − CD45RO − cells, TEMRA) were expanded (BT: 21.8 ± 6.9%; 1 month: 71.9 ± 11.8%, p < 0.0001; 6-12 months: 42.8 ± 8.5%, p < 0.001);.
## Cd52 surface expression and cd52-mediated cytolysis of treg and tcon
Cellular CD52 expression on Treg, Tcon, and their subtypes was determined by flow cytometry of PBMCs obtained from five healthy donors. CD52 expression levels were higher on Tcon than on Treg (Tcon: MFI 1230 ± 203; Treg: 952 ± 188; p < 0.05), and were significantly higher on naïve vs. memory subsets (Treg naive 1,030 ± 192, Treg memory 886 ± 174, p < 0.05; Tcon naive 1,281 ± 201, Tcon memory 903 ± 177, p < 0.01).
To assess the effect of alemtuzumab on Treg and Tcon in vitro, freshly isolated Treg and Tcon obtained from five healthy donors were cultured alone for 3 h in the presence of 10 µg/ml alemtuzumab and 10% normal human complement. Cytolysis rates for each subtype were measured by flow cytometry as outlined in the methods section above, revealing that both Treg and Tcon were abundantly depleted in response to alemtuzumab (Treg: 88.2 ± 2.9% [proportion of PI-positive cells], Tcon: 86.7 ± 3.5%; p = 0.064;. Moreover, cells with a naïve phenotype contained in both CD4 + populations were depleted more efficiently, although the difference did not attain statistical significance (Treg naive 89.6 ± 4.2%, Treg memory 87.8 ± 5.1%, p = 0.101; Tcon naive 86.5 ± 3.6%, Tcon memory 84.2 ± 4.0%, p = 0.071;.
# Discussion
Alemtuzumab-induced lymphocyte depletion is followed by an asymmetric repopulation of circulating T-and B-cells which also impacts Treg. Previous studies have shown that Treg rapidly increase relative to other CD4 + T-cell subsets and remain enriched up to several months after therapy. It is obvious that such numeric changes between Treg and Tcon or alterations affecting the homeostatic composition of subtypes comprised in both compartments are of high clinical relevance, as Treg are important mediators in the maintenance of peripheral immunotolerance by suppressing potentially autoaggressive Tcon clones. In patients with MS, a more detailed understanding of alemtuzumab-triggered depletion and recovery in the Treg population is of particular interest, since MS patient-derived Treg are less potent inhibitors of Tcon responses as a result of altered prevalence of naïve and mature Treg subtypes compared to healthy controls.
Here, we found that Treg are lysed together with other CD4 + T-cells, as reflected by equivalent relative proportions of CD4 + CD25 high CD127 low FOXP3 + cells contained in total CD4 + T-cells at baseline and at 1 week after alemtuzumab administration. Accordingly, when exposed to alemtuzumab and human complement ex vivo, Treg were destroyed at least as efficiently as Tcon despite the somewhat lower membrane expression of CD52 on Treg. Again, the small difference in surface densities of CD52 featured by naïve vs. memory subsets translated into a slightly higher susceptibility of both Treg naive and Tcon naive to undergo complement-mediated cell lysis compared with the respective memory counterparts. Hence, Treg and Treg subsets are not spared from the effects of alemtuzumab.
The peak expansion of Treg at month 1 during posttreatment lymphocyte recovery featured a substantial dominance of Treg memory over Treg naïve and thus mirrored the dynamics of reconstituting memory and naïve phenotypes in CD4 + and CD8 + T-cells described earlier. The imbalance between Treg memory and Treg naïve persisted until month 12 after both alemtuzumab cycles and was accompanied by a continued deficit in relative frequencies of Treg RTE . As numbers of TREC-containing and dual TCR-Treg-both markers intrinsic to thymocytes-remained low throughout the study period and coincided with a compromised TCR repertoire as documented by a reduced CDR3 complexity score, these observations altogether confirm that thymopoiesis is not or barely induced after alemtuzumab therapy, possibly due to the premature decline in thymic function attributed to MS.
Whether the abundance of cells with a Treg phenotype among post-depletional total CD4 + T-cells, detectable from month 1 through month 12 independent of the underlying disease, as shown in our study and former studies, is driven by Treg formed extrathymically in the peripheral immune system by expansion of residual Treg and/or conversion from memory Tcon (induced or iTreg) during treatment-induced homeostatic proliferationremains elusive, since currently, no features or phenotypic markers, including the Ikaros family transcription factor Helios, are able to distinguish iTreg from Treg generated by homeostatic proliferation. However, we observed in occasional patients (n = 2, data not shown) that Treg obtained shortly after treatment are not highly proliferative, as depicted by barely detectable intracellular expression of Ki-67, a nuclear antigen exclusively expressed during active cell cycling. As similar results were reported in a previous study, where only one out of four individuals assessed after treatment with Campath-1 showed a post-depletional slight increase in active cycling of Treg, these findings support the notion that homeostatic expansion does not solely account for the increased Treg/Tcon ratio.
A predominance of Treg memory negatively affects the immunoregulatory properties of total Treg, as we and others have shown previously. We therefore sought to establish how post-alemtuzumab Treg and Tcon interact to promote the restored long-term suppressive function of Treg observed in a former study. Of note, despite the more pronounced imbalance in Treg naive and Treg memory proportions inherent to post-treatment cells, the suppressive performance of total Treg appeared unexpectedly raised throughout months 1-24 when tested against Tcon isolated from the same patient. When employing allogeneic instead of syngeneic Tcon, however, Treg activity per se remained stable vs. baseline and decreased compared with counterparts from healthy donors. This, in turn, clearly indicates that post-alemtuzumab Treg remain dysfunctional, but perform better as an indirect result of a treatment-induced redistribution in Tcon phenotypes. Accordingly, the post-treatment Tcon population turned out to be enriched in CCR7 − TEM and TEMRA cells at the expense of CCR7 + naive TCM and TCRA subsets. The preponderance of such effector memory cells parallels our own observations in fingolimod-treated patients with MS, where Tcon lacking CCR7 become elevated during treatment and are less proliferative, thereby indirectly upregulating Treg efficiency. Importantly, it has been recently suggested, that protracted T-cell recovery and, hence, long-lasting lack of T-cell regulation along with earlier and even hyperproliferative B-cell subset reconstitution might serve as a main driver in the emergence of secondary autoimmunity in response to alemtuzumab therapy. Whether homeostatically disturbed Treg/Tcon subsets favor breakthrough MS activity remains elusive. In this study, the disease unequivocally remained clinically stable in the early post-treatment phase, i.e., when the ratio of absolute numbers of Tcon and Treg was lowest (month 1), whereas the Tcon/Treg ratio of recovering lymphocytes had reached at least around 50% of that measured at baseline in all nine relapses documented. However, the numeric relation between Tcon and Treg did not significantly differ in stable patients when assessed at the same points in time.
# Conclusions
Taken together, Treg are not spared in alemtuzumab therapy and thymopoiesis does not considerably contribute to postdepletional long-term recovery. Reconstitution may be driven by homeostatic proliferation and/or by conversion from residual Tcon beginning after day 7 of treatment. This results in a relative Treg expansion among total CD4 + T-cells, which is, however, accompanied by a marked and long-lasting predominance of Treg memory along with a contraction in Treg naive and Treg RTE as a response to therapeutic lymphopenia. As a result of these changes and the concomitant accumulation of effector memory populations in the Tcon compartment, the suppressive capacity of dysfunctional patient Treg is paradoxically restored. Importantly, intra-individual differences in repopulated Treg and Tcon may impact the therapeutic efficacy of alemtuzumab with respect to stabilization of MS activity. By showing that Treg undergo distinct and long-lasting homeostatic changes after alemtuzumab therapy, we challenge the beneficial influence deduced from the previously described apparent sparing of these cells from treatment-induced immune cell lysis.
# Ethics statement
The protocol was approved by the University Hospital Heidelberg ethics committee; all individuals gave written informed consent.
# Author contributions
JH and BW contributed to the conception and design of the study. JH, TR, AV, SP, SM, and MK-K performed research and contributed clinical samples. SJ organized the database. CW performed the statistical analysis. JH wrote the first draft of the manuscript. BW and SJ wrote sections of the manuscript. All authors contributed to the manuscript revision, and read and approved the submitted version. |
Pancreatic ductal adenocarcinoma (PDAC) is one of the leading causes of cancer mortality. Although advances have been made in understanding the pathogenesis of PDAC, the outcome still remains poor. The aim of this study is to conduct a metaanalysis to evaluate the precise association between SMAD4 loss and clinicopathological significance in PDAC. A literature search was made in PubMed, Web of Science, Google scholar, and EMBASE for related publications. The data were extracted and assessed by two reviewers independently. Analysis of pooled data was performed, Odds Ratio or Hazard Ratio with corresponding confidence intervals was calculated and summarized. 12 relevant articles were included for full review in detail and metaanalysis. The frequency of SMAD4 protein loss was significantly increased in PDAC than in nonmalignant pancreatic tissue, Odd Ratio was 0.05 with 95% confidence interval 0.01-0.23, p<0.0001. SMAD4 loss was significantly associated with poor overall survival in patients with PDAC, Hazard Ratio was 0.61 with 95% confidence interval 0.38-0.99, p=0.05. SMAD4 loss was not correlated with the size, grades, and lymph node metastasis of PDAC. In conclusion, SMAD4 is a biomarker for the diagnosis of PDAC. SMAD4 loss is significantly related to poor prognosis in patients with PDAC.
# Introduction
Pancreatic ductal adenocarcinoma (PDAC) is one of the leading causes of cancer mortality. Although improvement in clinical management has been made last two decades, the prognosis of PDAC remains poor, and a 5-year survival rate is approximately 5% [bib_ref] Cancer statistics, 2011: the impact of eliminating socioeconomic and racial disparities on..., Siegel [/bib_ref] [bib_ref] Loss of the acinar-restricted transcription factor Mist1 accelerates Kras-induced pancreatic intraepithelial neoplasia, Shi [/bib_ref] [bib_ref] Systematic review of peri-operative prognostic biomarkers in pancreatic ductal adenocarcinoma, Petrushnko [/bib_ref]. Because of a lack of specific symptoms and appropriate markers for early stages, most PDAC patients are diagnosed at advanced stages, when radical pancreatic resection is not possible. Therefore, it is critical to identify biomarkers for early diagnosis and development of gene targeted therapy. SMAD4 was also known as the deleted in pancreatic carcinoma 4 (DPC4), is located on chromosome 18q21 [bib_ref] Evaluation of candidate tumour suppressor genes on chromosome 18 in colorectal cancers, Thiagalingam [/bib_ref] [bib_ref] DPC4, a candidate tumor suppressor gene at human chromosome 18q21.1, Hahn [/bib_ref]. SMAD4 is a co-factor that facilitates gene transcription and tumor suppression through the TGF-beta signaling pathway. TGF-beta/SMAD4 signaling pathway regulates tumor development through mediating growth arrest and inducing apoptosis [bib_ref] DPC4, a candidate tumor suppressor gene at human chromosome 18q21.1, Hahn [/bib_ref] [bib_ref] The Smad family and its role in pancreatic cancer, Singh [/bib_ref] [bib_ref] Molecular pathology of pancreatic cancer, Saiki [/bib_ref] [bib_ref] Proliferation, cell cycle and apoptosis in cancer, Gi [/bib_ref] [bib_ref] Cytostatic and apoptotic actions of TGF-beta in homeostasis and cancer, Siegel [/bib_ref] [bib_ref] TGF-beta Tumor Suppression through a Lethal EMT, David [/bib_ref] [bib_ref] TGFbeta signaling in growth control, cancer, and heritable disorders, Massague [/bib_ref]. Previous studies have attempted to correlate the alteration of SMAD4 with clinical features and prognosis in patients with PDAC [bib_ref] Tumoral epithelial and stromal expression of SMAD proteins in pancreatic ductal adenocarcinomas, Handra-Luca [/bib_ref] [bib_ref] Contribution of CXCR4 and SMAD4 in predicting disease progression pattern and benefit..., Bachet [/bib_ref] [bib_ref] Reduced expression of bone morphogenetic protein receptor IA in pancreatic cancer is..., Voorneveld [/bib_ref] [bib_ref] Phase II trial of cetuximab, gemcitabine, and oxaliplatin followed by chemoradiation with..., Crane [/bib_ref]. However, a clear correlation has not been established. We conducted a meta-analysis to investigate the association of SMAD4 status with clinicopathlogical significance and prognosis in patients with PDAC.
# Results
12 studies were included for meta-analysis [fig_ref] Figure 1: Schematic flow diagram for selection of included studies [/fig_ref]. The main characteristics were listed in [fig_ref] Table 1: Main Characteristics of included studies [/fig_ref].
The rate of SMAD4 protein loss was significantly higher in PDCA than in nonmalignant pancreatic tissue (including normal pancreas and hyperplasia), OR was 0.05 with 95% CI 0.01-0.23, z=3.97, p<0.0001, I 2 =0% [fig_ref] Figure 2: Forest plot for SMAD4 protein expression in PDAC and nonmalignant pancreatic tissue [/fig_ref]. The frequency of SMAD4 protein loss was similar between high and low grade of PDCA, OR was 0.92 with 95% CI 0.41-2.05, z=0.20, p=0.84, I 2 =61% [fig_ref] Figure 3: Forest plot for SMAD4 protein expression in different grade of PDAC [/fig_ref]. Loss of SMAD4 protein was not associated with lymph node metastasis status, OR was 0.71 with 95% CI 0.42-1.21, z=1.25, p=0.21, I 2 =57% [fig_ref] Figure 4: Forest plot for SMAD4 protein expression in different lymph node metastasis status [/fig_ref]. There
## Research paper
was no significantly difference of SMAD4 protein loss rate between large and small size of PDCA tumor, OR was 0.94 with 95% CI 0.70-1.25, z=0.42, p=0.68, I 2 =0% [fig_ref] Figure 5: Forest plot for SMAD4 protein expression in different size of PDAC [/fig_ref]. Loss of SMAD4 protein was significantly correlated to overall survival in patients with PDCA, HR was 0.61 with 95% CI 0.38-0.99, z=1.99, p=0.05, I 2 =67% [fig_ref] Figure 6: Forest plot for the association of SMAD4 protein expression with overall survival... [/fig_ref].
The methodological quality of each study was assessed separately and independently by JW and XC using Newcastle Ottawa Quality Assessment Scale (NOQAS). This scale for non-randomized case controlled studies and cohort studies was used to allocate a maximum of nine points for the quality of selection, comparability, exposure, and outcomes for study participants. Among 12 studies, two scored 6, six scored 7, and four scored 8. Therefore, the studies were relatively high quality [fig_ref] Table 2: Quality assessment according to the Newcastle-Ottawa scale of the included studies [/fig_ref]. A sensitivity analysis, in which one study was removed at a time, was conducted to assess the result stability. The pooled ORs and HR were not significantly changed, suggesting the stability of our analyses. The funnel plots showed largely symmetric [fig_ref] Figure 7: Funnel plot for publication bias [/fig_ref] which indicated there was no publication biases in the meta-analysis of SMAD4 protein expression and clinicopathological features.
# Discussion
Although tremendous progress has been made last two decades in understanding the pathogenesis of PDAC, five-year survival rates are still less than 5% [bib_ref] Molecular pathology of pancreatic cancer, Saiki [/bib_ref]. The development and progression of PDAC involve various gene alterations including oncogene activation and loss of tumor suppressor gene function [bib_ref] Overexpression of p53 protein in adenocarcinoma of the pancreas, Digiuseppe [/bib_ref] [bib_ref] Immunohistochemical evaluation of K-ras, p53, and HER-2/neu expression in hyperplastic, dysplastic, and..., Apple [/bib_ref] [bib_ref] Loss of expression of Dpc4 in pancreatic intraepithelial neoplasia: evidence that DPC4..., Wilentz [/bib_ref]. The understanding of molecular biology in PDAC contributes to the development of new approaches to its prevention, diagnosis and treatment. Previous studies investigated the relationship of loss of SMAD4 with the features of PDAC and its prognosis in patients, however, the correlation was not clear due to small power of the samples [bib_ref] Tumoral epithelial and stromal expression of SMAD proteins in pancreatic ductal adenocarcinomas, Handra-Luca [/bib_ref] [bib_ref] Contribution of CXCR4 and SMAD4 in predicting disease progression pattern and benefit..., Bachet [/bib_ref] [bib_ref] Loss of DPC4 expression and its correlation with clinicopathological parameters in pancreatic..., Hua [/bib_ref].
Our data showed the loss of SMAD4 protein expression was significantly increased in PDAC than nonmalignant pancreatic tissues. TGF-beta/SMAD4 signaling regulates tumor development because of its effects on growth arrest and induced apoptosis [bib_ref] DPC4, a candidate tumor suppressor gene at human chromosome 18q21.1, Hahn [/bib_ref] [bib_ref] Proliferation, cell cycle and apoptosis in cancer, Gi [/bib_ref] [bib_ref] Cytostatic and apoptotic actions of TGF-beta in homeostasis and cancer, Siegel [/bib_ref]. Mitogenic growth signals, which are regulated by TGFbeta/SMAD4, are required during the process of cell cycle [bib_ref] Proliferation, cell cycle and apoptosis in cancer, Gi [/bib_ref] [bib_ref] TGFbeta prevents proteasomal degradation of the cyclin-dependent kinase inhibitor p27kip1 for cell..., Lecanda [/bib_ref]. Therefore, SMAD4 functions as a tumorsuppressor through growth arrest during tumorgenesis. Previous study indicated that TGF-beta could affect growth arrest through upregulating p21 in a SMAD4dependent manner in colon cancer [bib_ref] Effects of activin and TGFbeta on p21 in colon cancer, Bauer [/bib_ref]. Additionally, TIEG, a Zinc-finger encoding gene regulated by TGFbeta/SMAD4 signaling was reported to induce apoptosis in pancreatic cells (PCs) [bib_ref] Overexpression of the TGFbeta-regulated zinc finger encoding gene, TIEG, induces apoptosis in..., Tachibana [/bib_ref]. Moreover, SMAD4 downregulation caused TGF-beta-induced cell cycle arrest and apoptosis, and the restoration of SMAD4 by gene therapy reversed the invasive phenotype as well as reduced the proliferation in PC cell lines [bib_ref] Restoration of SMAD4 by gene therapy reverses the invasive phenotype in pancreatic..., Duda [/bib_ref] [bib_ref] Restoration of Smad4 in BxPC3 pancreatic cancer cells attenuates proliferation www.impactjournals.com/oncotarget without..., Yasutome [/bib_ref] [bib_ref] Inhibition of pancreatic carcinoma cell growth in vitro by DPC4 gene transfection, Shen [/bib_ref]. An increase in G1 phase fraction was observed in a PDAC cell lines after inducing SMAD4 expression through a tetracycline system construct. The frequency of SMAD4 protein loss was similar between different sizes as well as different grades of PDAC. More studies are needed to confirm the correlation between SMAD4 status and PDAC features. A number of studies have reported the association between SMAD4 protein loss and lymphatic metastases [bib_ref] Contribution of CXCR4 and SMAD4 in predicting disease progression pattern and benefit..., Bachet [/bib_ref] [bib_ref] Loss of DPC4 expression and its correlation with clinicopathological parameters in pancreatic..., Hua [/bib_ref] [bib_ref] It's a SMAD/SMAD World, Taniguchi [/bib_ref] [bib_ref] Loss of SMAD4 staining in pre-operative cell blocks is associated with distant..., Boone [/bib_ref] , however, the results were inconsistent [bib_ref] Loss of SMAD4 staining in pre-operative cell blocks is associated with distant..., Boone [/bib_ref] [bib_ref] The SMAD4 protein and prognosis of pancreatic ductal adenocarcinoma, Tascilar [/bib_ref] [bib_ref] Immunohistochemically detected expression of 3 major genes (CDKN2A/p16, TP53, and SMAD4/DPC4) strongly..., Oshima [/bib_ref]. Pooled OR suggested that the frequency of SMAD4 protein loss was not associated with lymphatic metastasis. TGFbeta/SMAD4-independent signaling pathway is activated due to SMAD4 loss [bib_ref] Human transforming growth factor-beta complementary DNA sequence and expression in normal and..., Derynck [/bib_ref]. TGF-beta activates the PI3K/Akt/ mTOR pathway and leads to increased migratory capacity and invasiveness during Epithelial-mesenchymal transition (EMT) which facilitates tumor progression and metastasis in PDAC. TGF-beta also suppresses PTEN through NFkappa B and enhanced cell motility and invasiveness [bib_ref] RAS/ERK modulates TGFbeta-regulated PTEN expression in human pancreatic adenocarcinoma cells, Chow [/bib_ref].
Another activated downstream of TGF-beta/SMAD4independent signaling is STAT3 which has been proven to be an important regulator for PADC growth, invasion and angiogenesis [bib_ref] STAT3 knockdown reduces pancreatic cancer cell invasiveness and matrix metalloproteinase-7 expression in..., Li [/bib_ref] [bib_ref] Lentivirus-mediated shRNA interference targeting STAT3 inhibits human pancreatic cancer cell invasion, Yang [/bib_ref] [bib_ref] STAT3-targeting RNA interference inhibits pancreatic cancer angiogenesis in vitro and in vivo, Huang [/bib_ref] [bib_ref] Inhibition of STAT3 Tyr705 phosphorylation by Smad4 suppresses transforming growth factor beta-mediated..., Zhao [/bib_ref]. On the other hand, SMAD4 serves as a tumor-suppressor and mediates cell cycle arrest as well as induced apoptosis in a SMAD4-dependent TGFbeta signaling pathway. The converse role of SMAD4 could be the reason that loss of SMAD4 is not associated with lymph node metastasis in PDAC. Recently Hingorani and colleagues used genetically engineered mouse model and found RUNX3 expression together with the DPC4/ SMAD4 haploinsufficiency could inform metastasis status in PDAC [bib_ref] RUNX3 Controls a Metastatic Switch in Pancreatic Ductal Adenocarcinoma, Whittle [/bib_ref]. More studies need to be finished in future for clinical application.
Although substantial progress has been made in understanding the pathogenesis of PDAC over last two decades, the prognosis is still poor. Multiple studies have been conducted showing inconsistencies of the association between SMAD4 expression status and prognosis in PDAC patients due to small samples [bib_ref] Loss of DPC4 expression and its correlation with clinicopathological parameters in pancreatic..., Hua [/bib_ref] [bib_ref] Immunohistochemically detected expression of 3 major genes (CDKN2A/p16, TP53, and SMAD4/DPC4) strongly..., Oshima [/bib_ref] [bib_ref] Vascular endothelial growth factor and DPC4 predict adjuvant therapy outcomes in resected..., Khorana [/bib_ref] [bib_ref] DPC4/Smad4 expression and outcome in pancreatic ductal adenocarcinoma, Biankin [/bib_ref]. In present study, the pooled data from ten studies indicated SMAD4 loss is significantly associated with prognosis in patients with PDAC. Finally our study selected all the published articles written in English and Chinese, did not include some relevant articles written in other languages or unpublished papers which may result in certain publication bias. Therefore, the result should be interpreted carefully.
In conclusion, SMAD4 is a biomarker for the diagnosis of PDAC. SMAD4 loss is significantly related with poor prognosis of patients with PDAC. SMAD4 loss is not associated with lymph node metastasis of PDAC.
# Materials and methods
The meta-analysis was performed by using PRISMA checklist as a guide (Supplementary Checklist 1).
## Selection criteria and study search
Systematic review of several databases was conducted in July 2016 with no lower limit set for date of publication. Following electronic databases were searched for relevant articles without any language restrictions: PubMed, Web of Science, Google scholar, and EMBASE. The keywords "SMAD4" or "DPC4" and "pancreatic cancer" or "pancreatic ductal adenocarcinoma" or pancreatic adenocarcinoma" were used for relative articles searching. There were 500 articles identified from PubMed, 452 articles from Web Science, 895 articles from EMBASE. 15,800 articles were identified from Google Scholar, first 200 of them were screened since the rest of them were not related to present study. A total of 2047 articles were initially identified by the search strategy, and 43 full-text articles were retrieved after screening. Forward and backward citation chasing of each selected article was performed in case they included another study of interest that had not been indentified. Studies were selected based on the following criteria: 1) The association between SMAD4 protein expression and the clinicopathological features of PDAC; 2) The association of SMAD4 protein expression and prognosis in patients with PDAC. SMAD4 protein expression was examined by immunochemistry. The following exclusion criteria were used: 1) the studies investigated the association between SMAD4 mRNA expression and clinicopathological significance 2) the studies utilized the same population or overlapping database, 3) the studies utilized cell lines or mice.
## Data extraction and study assessment
Two reviewers (JW and XC) extracted data from selected studies independently. Any disagreement was discussed and reached a consensus for all issues. The following items were collected from each study: first author's name, year of publication, geographical location, age of patients, sample size, grades, size of the tumor, lymph node metastasis status, immunohistostaining, SMAD4 antibodies used, and HRs with 95% CIs from multivariate analyses.
# Statistics analysis
Odds ratios (OR) and hazard ratio with their 95% confidence intervals were calculated. Heterogeneity among studies was estimated using the Cochran's Q statistic and I 2 tests. The I 2 statistic was used to examine the difference for between study variability due to heterogeneity rather than chance, with a range from 0 to 100 percent. A fixed effect model was used for I 2 <50%, while a random effect model was used for I 2 >50%. The analysis was performed to compare the frequency of SMAD4 protein expression between PDCA and nonmalignant prostate tissue. In addition, we evaluated the frequency of SMAD4 protein expression in different grades, different size of the tumor, and the correlation between SMAD4 expression and lymph node metastasis status, as well as the relationship between SMAD4 expression and prognosis in patients with PDCA. All p values were two sided. Funnel plots were used for detection of publication bias. All analysis was performed with Review Manager 5.2 (Cochrane Collaboration, Software Update, Oxford, UK).
had full access to all data and the final responsibility for the decision to submit the article for publication. All authors read and approved the final manuscript.
[fig] Figure 1: Schematic flow diagram for selection of included studies. www.impactjournals.com/oncotarget [/fig]
[fig] Figure 2: Forest plot for SMAD4 protein expression in PDAC and nonmalignant pancreatic tissue. [/fig]
[fig] Figure 3: Forest plot for SMAD4 protein expression in different grade of PDAC. [/fig]
[fig] Figure 4: Forest plot for SMAD4 protein expression in different lymph node metastasis status. [/fig]
[fig] Figure 5: Forest plot for SMAD4 protein expression in different size of PDAC. [/fig]
[fig] Figure 6: Forest plot for the association of SMAD4 protein expression with overall survival of patients with PDAC. www.impactjournals.com/oncotarget [/fig]
[fig] Figure 7: Funnel plot for publication bias. Each circle represents the weight of individual study. In X axes, Log(OR)=natural logarithm of OR, Log(HR)=natural logarithm of HR, In Y axes, SE=standard error. a. SMAD4 protein expression in PDAC and nonmalignant pancreatic tissue; b. SMAD4 protein expression in different grade of PDAC; c. SMAD4 protein expression in different lymph node metastasis status; d. SMAD4 protein expression in different size of PDAC; e. the association of SMAD4 protein expression with overall survival of patients with PDAC. [/fig]
[table] Table 1: Main Characteristics of included studies [/table]
[table] Table 2: Quality assessment according to the Newcastle-Ottawa scale of the included studies [/table]
|
Prevalence of HCV NS3 pre-treatment resistance associated amino acid variants within a Scottish cohort
a b s t r a c tBackground: Protease inhibitors (PI) including boceprevir, telaprevir and simeprevir have revolutionised HCV genotype 1 treatment since their introduction. A number of pre-treatment resistance associated amino acid variants (RAVs) and polymorphisms have been associated with reduced response to treatment. Objectives: We measured the prevalence of RAVs/polymorphisms in a PI treatment-naïve HCV genotype 1 Scottish cohort using Sanger sequencing. Study design: Chronically infected, treatment-naïve, HCV genotype 1 patients (n = 146) attending NHS Greater Glasgow and Clyde clinics were investigated for RAVs/polymorphisms to the PIs boceprevir, telaprevir and simeprevir. The NS3/4A region was amplified by nested polymerase chain reaction. The 1.4 kb amplified product was sequenced using an ABI 3710XL DNA sequencer. Sequence analysis was performed using web-based ReCall (beta 2.10). Amino acid positions36,41,43, 54, 55, 80, 109, 122, 155, 156, 168 and 170 were analysed for RAVs/polymorphisms. Results: Overall, 23.29% (34/146) of patients had an RAV or polymorphism detected. Overall, 13.69% (20/146) of patients had HCV virus that contained the Q8 K polymorphism. Other RAVs detected were: V36 M 0.70% (1/146), V36L 0.70% (1/146), T54S 6.85% (10/146), V55A 3.42% (5/146) and V/I170A 0.68% (1/146). Four patients had dual combinations of mutations (T54S + V36L; T54S + V55A and 2 patients with T54S + Q80K). Conclusions: Q80K was the most prevalent baseline polymorphism detected in the Scottish cohort. Simeprevir treatment is not recommended in patients infected with the Q80K genotype 1a variant. This highlights the need for baseline sequencing prior to administration of this drug in this population.Crown
# Background
Traditionally genotype 1HCV infections have been the hardest to treat with sustained virological response (SVR) rates to the standard of care treatment of ribavirin (RBV) co-administered with pegylated-interferon alpha (IFN) in the region of 42-50% [bib_ref] Peginterferon alfa-2b plus ribavirin compared with interferon alfa-2b plus ribavirin for initial..., Manns [/bib_ref] [bib_ref] Peginterferon alfa-2a plus ribavirin for chronic hepatitis C virus infection, Fried [/bib_ref]. The development of non-structural protein 3 (NS3) protease inhibitors (PIs) including telaprevir, boceprevir and simeprevir, has substantially improved outcome in these patients with SVR rates now approaching 80% in both treatment-naive patients and relapsers [bib_ref] Telparevir for previously untreated chronic hepatitis C virus infection, Jacobson [/bib_ref] [bib_ref] Boceprevir for untreated chronic HCV genotype 1 infection, Poordad [/bib_ref] [bib_ref] Simeprevir with pegylated interferon alfa 2a plus ribavirin in treatment-naive patients with..., Jacobson [/bib_ref] [bib_ref] Simeprevir with peginterferon and ribavirin leads to high rates of SVR in..., Forns [/bib_ref]. Newer PI based IFN-free regimens show even greater potential and lower toxicity. For example, the combination of simeprevir and the NS5B polymerase inhibitor sofosbuvir increased the SVR * Corresponding author. Tel.: +44 141 2018722.
E-mail address: [email protected] (S.J. Shepherd).
to over 90% in genotype 1 patients [bib_ref] Simeprevir plus sofosbuvir, with or without ribavirin, to treat chronic infection with..., Lawitz [/bib_ref]. NS5A inhibitors daclatasvir or ledipasvir when used with sofosbuvir have also generated SVR rates >90% [bib_ref] Daclatasvir plus sofosbuvir for previously treated or untreated chronic HCV infection, Sulkowski [/bib_ref] [bib_ref] Ledipasvir and sofosbuvir for untreated HCV genotype 1 infection, Afdhal [/bib_ref]. Further breakthroughs are expected as other combinations of antivirals become available which promise to offer improvements in SVR, shortened duration of treatment and lower pill burden.
A number of pre-treatment resistance associated amino acid variants (RAVs) within NS3 are associated with reduced response to PI-IFN regimens. For example, RAVs at position 156 (A156T/V) and R155K have been shown to reduce the effectiveness of all current PIs [bib_ref] Identification and analysis of fitness of resistance mutations against the HCV protease..., Tong [/bib_ref] [bib_ref] Telaprevir and pegylated interferon-alpha-2a inhibit wild-type and resistant genotype 1 hepatitis C..., Kieffer [/bib_ref] [bib_ref] Dynamic hepatitis virus genotype and phenotypic changes in patients treated with the..., Sarrazin [/bib_ref] [bib_ref] In vitro resistance profile of the hepatitis C virus NS3/4A protease inhibitor..., Lenz [/bib_ref]. Substitutions at the D168 locus (D168T/Y/H/A/V/I) result in high-level resistance to simeprevir (>300 fold) and the other 2nd generation PIs only [bib_ref] In vitro resistance profile of the hepatitis C virus NS3/4A protease inhibitor..., Lenz [/bib_ref] [bib_ref] Hepatitis C virus resistance to protease inhibitors, Halfon [/bib_ref] [bib_ref] Importance of HCV genotype 1 subtypes for drug resistance and response to..., Wyles [/bib_ref] [bib_ref] Telaprevir and pegylated interferon-alpha-2a inhibit wild-type and resistant genotype 1 hepatitis C..., Kieffer [/bib_ref] [bib_ref] In vitro resistance profile of the hepatitis C virus NS3/4A protease inhibitor..., Lenz [/bib_ref] [bib_ref] Hepatitis C virus resistance to protease inhibitors, Halfon [/bib_ref] [bib_ref] Characterisation of resistance mutations against HCV ketoamide protease inhibitors, Tong [/bib_ref] [bib_ref] Characterisation of resistance to protease inhibitor boceprevir in hepatitis C virus-infected patients, Susser [/bib_ref]. However, their significance is currently uncertain with most reports suggesting that they only have a minor effect on overall SVR rates. Only a few studies have examined the prevalence of the aforementioned RAVs at baseline [bib_ref] Susceptibility of treatment-naive hepatitis C virus (HCV) clinical isolates to HCV protease..., Bae [/bib_ref] [bib_ref] Naturally occurring hepatitis C virus (HCV) NS3/4A protease inhibitor resistance-related mutations in..., Vicenti [/bib_ref] [bib_ref] Hepatitis C virus variants with decreased sensitivity to direct-acting antivirals (DAAs) were..., Bartels [/bib_ref] [bib_ref] Implications of baseline polymorphisms for potential resistance to NS3 protease inhibitors in..., Palanisamy [/bib_ref] [bib_ref] Patients eligible for treatment with simeprevir in a French center, Morel [/bib_ref]. Knowing their frequency, can be used to plan treatment policies and will determine the usefulness of baseline testing prior to treatment.
## Objectives
We measured the prevalence of natural resistance polymorphisms in a protease inhibitor treatment-naïve HCV genotype 1 Scottish cohort using Sanger sequencing.
## Study design
## Patients
Stored plasma samples, taken between August 2013 and March 2014 for 146 chronically infected HCV genotype 1 patients attending clinics within NHS Greater Glasgow & Clyde, were used in this study. The patients consisted of 141 treatment naïve patients and 5 treatment relapsers who had previously been treated with pegylated IFN and RBV. The majority of the patients (n = 140) were subtype 1a and six subtype 1b. All patients had a detectable HCV RNA tested by Abbott RealTime HCV (detection limit 12 IU/ml).
## Rna extraction
RNA was extracted using the NucliSens easyMag (BioMerieux). Using the on-board lysis protocol, 1000 l of sample was eluted to 60 l.
## Pcr amplification and sequencing procedure
The NS3/4A region was amplified by nested polymerase chain reaction (PCR) using a method and primers supplied by Dr Richard Harrigan (British Columbia Centre for Excellence in HIV/AIDS). The 1st round primer sequences were: 5 TTCAGCCTGGACC-
# Results
There was no evidence of RAVs at position 155, 156 and 168 in any of the sequences analysed . The polymorphism Q80K was found in 13.69% (20/146) of patients sequenced. Other RAVs were found at the following frequencies: 0.70% (1/146) V36M, 0.70% (1/146) V36L, 6.85% (10/146) T54S 3.42% (5/146) V55A and 0.68% (1/146) V/I170A. Four patients were identified as having dual combinations of mutations (T54S + V36L; T54S + V55A and 2 patients with T54S + Q80K).
# Discussion
This study analysed sequences from the NS3/4A serine protease region of 146 genotype 1 patients (140 were genotype 1a and 6 were genotype 1b). The low level of subtype 1b patients in this study is a reflection of the Scottish population, where subtype 1a predominates. This is also a reflection of the UK population as a whole [bib_ref] Changing epidemiology of hepatitis C virus infection in Europe, Esteban [/bib_ref]. Overall 23.29% of patients tested had NS3 RAVs/polymorphisms without prior exposure to PIs. No high-level resistant RAVs were detected at positions 155, 156 or 168. Other prevalence studies in treatment-naive patients have shown that these three key resistance mutations either occur at a very low level (<0.9%) or not at all . The majority of patients had the naturally occurring polymorphism Q80K (13.69%). The prevalence of Q80K in the Scottish cohort is similar to that found in other European studies; France 10.5%; Italy 10.1%; London 16% and Sweden 5.7% [bib_ref] Naturally occurring hepatitis C virus (HCV) NS3/4A protease inhibitor resistance-related mutations in..., Vicenti [/bib_ref] [bib_ref] Implications of baseline polymorphisms for potential resistance to NS3 protease inhibitors in..., Palanisamy [/bib_ref] [bib_ref] Patients eligible for treatment with simeprevir in a French center, Morel [/bib_ref] [bib_ref] Natural NS3 resistance polymorphism occur frequently prior to treatment in HIV-positive patients..., Leggewie [/bib_ref]. Q80K prevalence in the USA has been reported at higher prevalence levels of 37% and 47% [bib_ref] Susceptibility of treatment-naive hepatitis C virus (HCV) clinical isolates to HCV protease..., Bae [/bib_ref] [bib_ref] Hepatitis C virus variants with decreased sensitivity to direct-acting antivirals (DAAs) were..., Bartels [/bib_ref]. Mutational differences between genotype 1 subtypes and clades within subtype 1 may reflect differences seen between American and European patients [bib_ref] Susceptibility of treatment-naive hepatitis C virus (HCV) clinical isolates to HCV protease..., Bae [/bib_ref] [bib_ref] GT-1a or GT-1b subtype-specific resistance profiles for heaptitis C virus inhibitors telaprevir..., Mcgown [/bib_ref] [bib_ref] Update on hepatitis C virus resistance to direct-acting antiviral agents, Proveda [/bib_ref]. Studies have also highlighted that Q80K is more likely to occur in patients with subtype 1a HCV than subtype 1b [bib_ref] Naturally occurring hepatitis C virus (HCV) NS3/4A protease inhibitor resistance-related mutations in..., Vicenti [/bib_ref] [bib_ref] Hepatitis C virus variants with decreased sensitivity to direct-acting antivirals (DAAs) were..., Bartels [/bib_ref].
The V36L/M, T54S, V55A and V/I170A mutations detected in this study are low level resistance RAVs that have little effect on SVR rates in patients treated with triple therapy [bib_ref] Hepatitis C virus variants with decreased sensitivity to direct-acting antivirals (DAAs) were..., Bartels [/bib_ref] [bib_ref] Analysis of boceprevir resistance associated amino acid variants (RAVs) in two phase..., Barnard [/bib_ref]. These indeterminate or low level RAVs have been reported at a prevalence of between 0.2% and 11% [bib_ref] Naturally occurring hepatitis C virus (HCV) NS3/4A protease inhibitor resistance-related mutations in..., Vicenti [/bib_ref] [bib_ref] Hepatitis C virus variants with decreased sensitivity to direct-acting antivirals (DAAs) were..., Bartels [/bib_ref] [bib_ref] Natural prevalence of hepatitis C virus variants with decreased sensitivity to NS3...., Bartels [/bib_ref] [bib_ref] NS3 protease polymorphism and natural resistance to protease inhibitors in French patients..., Vallet [/bib_ref] [bib_ref] Natural NS3 resistance polymorphism occur frequently prior to treatment in HIV-positive patients..., Leggewie [/bib_ref]. The mutations V36M, V36L and V/I170A do not appear to be detrimental to viral fitness compared with high level resistance mutations and may explain the presence of these mutations within untreated populations [bib_ref] Identification and analysis of fitness of resistance mutations against the HCV protease..., Tong [/bib_ref] [bib_ref] Protease inhibitor-resistant hepatitis C virus mutants with reduced fitness from impaired production..., Shimakami [/bib_ref]. In this study, T54S was found at a prevalence of 7.53% within the Scottish cohort. This mutation confers low level resistance to both boceprevir and telprevir but not simeprevir [bib_ref] In vitro resistance profile of the hepatitis C virus NS3/4A protease inhibitor..., Lenz [/bib_ref] [bib_ref] Hepatitis C viral evolution in genotype 1 treatment-naive and treatment-experienced patients receiving..., Kieffer [/bib_ref] [bib_ref] In vitro phenotypic characterization of hepatitis C virus NS3 protease variants observed..., Jiang [/bib_ref]. T54S has been identified in 7.5% treatment-naive patients in Sweden and 2.8% in Italy [bib_ref] Naturally occurring hepatitis C virus (HCV) NS3/4A protease inhibitor resistance-related mutations in..., Vicenti [/bib_ref] [bib_ref] Implications of baseline polymorphisms for potential resistance to NS3 protease inhibitors in..., Palanisamy [/bib_ref].
In this study four (2.74%) subtype 1a patients were found to have RAV combinations, which all contained T54S with another mutation (T54S + V36L; T54S + V55A and T54S + Q80K). Bae et al. [bib_ref] Susceptibility of treatment-naive hepatitis C virus (HCV) clinical isolates to HCV protease..., Bae [/bib_ref] found that the combination of T54S and Q80K did not increase drug resistance to simeprevir but did reduce resistance to boceprevir and teleprevir when compared to the single mutation T54S (<3 fold-5 fold). The combination of V36L + T54S has been reported previously [bib_ref] Naturally occurring hepatitis C virus (HCV) NS3/4A protease inhibitor resistance-related mutations in..., Vicenti [/bib_ref]. It is unclear if this combination substantially increases resistance to PIs but the mutational combination of V36M + T54S increases viral fitness compared to a virus with T54S only [bib_ref] Characterisation of resistance to protease inhibitor boceprevir in hepatitis C virus-infected patients, Susser [/bib_ref]. Combination RAV at positions 54 and 55 have been shown to reduce response to triple therapy containing boceprevir.
This study examined the frequency of NS3 variants detected by Sanger sequencing. Sanger sequencing will only detect these variants at a frequency of >20%. Next generation sequencing (NGS) can detect lower frequency variants by measuring population variants that occur at <1% [bib_ref] Deep sequencing: becoming a critical tool in clinical virology, Quinones-Mateu [/bib_ref]. As a result studies using NGS will likely detect RAVs at an increased frequency. NGS studies of the NS3 region in treatment-naive patients, have again identified Q80K as the most prevalent baseline mutation with ∼42% harbouring this polymorphism [bib_ref] Extent of HCV NS3 protease variability and resistance-associated mutations assessed by next..., Bartolini [/bib_ref] [bib_ref] Hepatitis C virus NS3 sequence diversity and antiviral resistance-associated variant frequency in..., Jabara [/bib_ref].
Currently, European guidelines have six suggested treatment options for genotype 1 patients with simeprevir recommended within two of these treatment protocols. This study confirms that high-level resistance RAVs 155, 156 and 168 are rare within the treatment-naïve population in the West of Scotland. However, NS3/4A mutations detected in a group of PI treatment naive patients. Mutation list was adapted from Lenz et al. [bib_ref] In vitro resistance profile of the hepatitis C virus NS3/4A protease inhibitor..., Lenz [/bib_ref] ; Leggewie et al. [bib_ref] Natural NS3 resistance polymorphism occur frequently prior to treatment in HIV-positive patients..., Leggewie [/bib_ref] ; Forns et al. [bib_ref] Simeprevir with peginterferon and ribavirin leads to high rates of SVR in..., Forns [/bib_ref] ; Povada et al. [bib_ref] Update on hepatitis C virus resistance to direct-acting antiviral agents, Proveda [/bib_ref] ; Schneider and Sarrazin [bib_ref] Antiviral therapy of hepatitis C in 2014: do we need resistance testing?, Schneider [/bib_ref]. Q80K is common (13.69%) and baseline sequencing prior to therapy should be considered when considering simeprevir/IFN treatment in genotype 1a patients. It is possible that such testing will only be a temporary measure since newer dual therapies may largely overcome the negative effect of the Q80K mutation [bib_ref] Once-daily simeprevir (TMC435) plus sofosbuvir (GS-7977) with or without ribavirin in HCV..., Sulkowski [/bib_ref] [bib_ref] Sofosbuvir and ledipasvir fixed-dose combination with and without ribavirin in treatment-naïve and..., Lawitz [/bib_ref].
# Funding
Janssen Pharmaceutical funded the implementation of Q80K testing in the West of Scotland Specialist Virology Centre. TA is funded by the MRC (G0801822) and ECT by the Welcome Trust (WT102789).
## Competing interests
None declared.
# Ethical approval
Not required. |
An improved method for the isolation and identification of unknown proteins that bind to known DNA sequences by affinity capture and mass spectrometry
Transcription of a gene can be regulated at many different levels. One such fundamental level is interaction between protein and DNA. Protein(s) binds to distinct sites on the DNA, which activate, enhance or repress transcription. Despite being such an important process, very few tools exist to identify the proteins that interact with chromosome, most of which are in vitro in nature. Here, we propose an in vivo based method for identification of DNA binding protein(s) in bacteria where the DNA-protein complex formed in vivo is crosslinked by formaldehyde. This complex is further isolated and the bound proteins are identified. The method was used to isolate promoter DNA binding proteins, which bind and regulate the dsz operon in Gordonia sp. IITR 100 responsible for biodesulfurization of organosulfurs. The promoter binding proteins were identified by MALDI ToF MS/MS and the binding was confirmed by gel shift assay. Unlike other reported in vivo methods, this improved method does not require sequence of the whole genome or a chip and can be scaled up to improve yields.Citation: Murarka P, Srivastava P (2018) An improved method for the isolation and identification of unknown proteins that bind to known DNA sequences by affinity capture and mass spectrometry. PLoS ONE 13(8): e0202602.
# Introduction
DNA binding proteins play a crucial role in transcription, DNA replication, repair, recombination and various other cellular activities [bib_ref] Insights from genomic profiling of transcription factors, Farnham [/bib_ref]. To completely understand these fundamental biological processes, it is important to have knowledge about the proteins whose interplay leads to the complex control [bib_ref] The structure, function and evolution of proteins that bind DNA and RNA, Hudson [/bib_ref]. In the past, methods such as electrophoretic mobility shift assay (EMSA) [bib_ref] Electrophoretic mobility shift assay (EMSA) for detecting protein-nucleic acid interactions, Hellman [/bib_ref] , pull down assay etc [bib_ref] In vitro protein-DNA interactions at the human lamin B2 replication origin, Stefanovic [/bib_ref] have been reported to identify DNA binding proteins. These methods are mainly associated with in vitro DNA-protein interactions wherein a purified protein or a crude extract of protein is incubated with labeled DNA. DNA-binding proteins have specific or general affinity for DNA sequences. Some proteins involved in transcriptional regulation or other varied function may bind weakly to the DNA, hence making their isolation and identification difficult. For example, in gel shift assay, the conditions need to be optimized to get an optimal binding [bib_ref] Measurement of protein-DNA interaction parameters by electrophoresis mobility shift assay, Fried [/bib_ref]. Similarly, in an in vitro pull-down assay the DNA fragment is immobilized and the crude extract is allowed to pass through. Thus, the binding conditions in vitro may differ with the binding conditions in vivo. Hence, isolation of all DNA binding proteins is not possible. Another widely used method for identifying DNA a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 binding protein is chromatin immunoprecipitation (ChIP) assay [bib_ref] In vivo interactions of RNA polymerase II with genes of Drosophila melanogaster, Gilmour [/bib_ref] [bib_ref] Chromatin immunoprecipitation assay as a tool for analyzing transcription factor activity, Gade [/bib_ref]. ChIP assay is used for genome wide profiling of DNA-binding proteins, histone modification or nucleosomes. Besides cost and availability, ChIP has its own technical limitations. This method requires antibodies specific to the protein of interest and therefore, it cannot be used when the proteins are not known [bib_ref] Chromatin immunoprecipitation assay as a tool for analyzing transcription factor activity, Gade [/bib_ref]. Also, both ChIP and EMSA have low throughput, which makes these two techniques not suitable for identification of unknown factors that bind to DNA [bib_ref] Chromatin immunoprecipitation assay as a tool for analyzing transcription factor activity, Gade [/bib_ref].
A SILAC (stable isotope labeling by amino acid in cell culture) based method was described by for proteome analysis which was further improvised by to study proteins that interact with DNA. It is an in vitro method in which cells are allowed to grow in presence of labeled amino acids. The protein extract of these cells is passed through column containing immobilized DNA fragment. The DNA fragment is designed in such a way that it consists of a restriction enzyme site which is used for elution of proteins. The peptides generated after in gel digestion of the eluted protein are identified by mass spectrometry in combination with isotope coded affinity tag (ICAT) technology. A major limitation with SILAC based method is that it requires auxotrophy of amino acids employed for labeling, therefore this method cannot be extensively used in all types of bacterial cultures [bib_ref] Stable isotope labeling by amino acids applied to bacterial cell culture. Stable..., Soufi [/bib_ref]. The elution of proteins is based on restriction digestion, therefore designing of DNA fragment is little tricky. There is a possibility that the restriction site is not accessible to the enzyme for digestion if a large complex is bound to DNA due to steric hindrance.
Butala et al., proposed another in vivo method for identification of proteins [bib_ref] DNA sampling: a method for probing protein binding at specific loci on..., Butala [/bib_ref]. The method involves cloning the DNA of interest along with lac operators flanked by ISceI endonuclease sites. The lacI fused to FLAG tag was used for isolating the DNA protein complex. The method is very useful but it requires cloning of the desired DNA in a special type of vector, which has ISceI sites and also requires expression from ISceI meganuclease.
Dejardin and Kingston [bib_ref] Purification of Proteins Associated with Specific Genomic Loci, Déjardin [/bib_ref] proposed a method called Proteomics of Isolated Chromatin segment (PICh) for identifying the proteins associated with telomeres. The major limitation of PICh is that it cannot be used for DNA sequences that are present in single copy or are low in number in the genome. Another drawback of PICh is that it requires large amount of starting material (several hundred litres of culture).
There are only limited reports of methods that can be used to isolate the proteins that bind in vivo with the knowledge of short sequences of DNA. One such example of a known short DNA fragment is the dsz promoter of desulfurizing organisms. The genes for biodesulfurization are present in the form of an operon under the control of this promoter. The sequence of this promoter has been identified but the proteins that bind to this promoter have not been reported.
In this study, we describe a method for isolation and identification of transcription factors that bind to DNA inside the cell. It involves cross-linking of the proteins to dsz promoter DNA with formaldehyde when the cells are present in the log phase followed by harvesting and sonication of the cells. The promoter DNA is then subjected to digestion with exonuclease to generate overhangs, which further binds to specific biotinylated primer attached to streptavidin beads. Bound proteins are eluted based on pH. The eluates from the column are run on SDS-PAGE gel, and the protein bands, which are observed, are in gel digested and analyzed by MS/MS for identification. By this method we were able to isolate and identify transcriptosome complex consisting of 7 proteins whose interaction was evident by the interactome analysis.
# Materials and methods
## Media and growth conditions
Gordonia sp. IITR100 (MCC2877)was cultured in minimal media. The composition of Minimal salt medium per litre was: Na 2 HPO 4 (2.0 g), KH 2 PO 4 (1.0 g), ammonium oxalate (4.25 g), MgCl 2 (0.4 g) and sucrose (50 mM). Trace elements composition for 1 L was: KI (0.05 g), LiCl (0.05 g), MnCl 2 .4H 2 O (0.8 g), H 3 BO 3 (0.5 g), ZnCl 2 (0.1 g), CoCl 2 .6H 2 O (0.1 g), NiCl 2 .6H 2 O (0.1 g), BaCl 2 (0.05 g), (NH 4 ) 6 Mo 7 O 24 . 2H 2 O (0.05 g), SnCl 2 .2H 2 O (0.5 g), Al (OH) 3 (0.1 g). All the chemicals were dissolved in double distilled water. The medium was supplemented with 3 mM sodium sulfate as sulfur source. Single colony is used to inoculate 50 ml medium and incubated at 30˚C and 180 rpm for about four days after which the OD 600 was 1.
For cloning and expression studies, Escherichia coli (E. coli) DH5α and BL21 (DE3) and BL21 (DE3) pLysS cells were used respectively. E. coli was inoculated in LB media, incubated at 37˚C at 180 rpm and the antibiotic used was kanamycin in the concentration of 50 μg/ml where required.
## Cross linking of dna-protein complexes
For crosslinking of DNA protein complex, formaldehyde was used. Three different concentrations (0.5, 0.75 and 1%) of formaldehyde were used. After incubating at 25˚C for 10 min, formaldehyde was quenched by addition of Tris (final concentration 250 mM, pH 8) and incubated at RT (25˚C) for 10 min. As a control experiment, in one set of culture, the cross-linking was reversed by addition of Rapigest, 0.5 M β-mercaptoethanol and Tris (250 mM, pH 8.8), incubated at 99˚C for 25 min. Four different concentrations of Rapigest 0.5, 0.75, 1 and 2% were used.
Cells from both the cultures (cross-linked and reversal of cross-linked) were harvested at 3,500 rpm for 10 min at 4˚C. Supernatant was discarded and the pellet was resuspended in 1 ml sonication buffer which consisted of 20 mM Tris-Cl, 150 mM NaCl, 0.1mM PMSF, 1mM DTT, 0.1mM EDTA and 10% glycerol, pH 7.5. Sonication was performed on Q125 sonicator (Qsonica sonicators, USA) by giving 20 sec bursts and 30 sec rest for 5 min at an amplitude of 30%. After sonication, the cellular debris is removed by centrifugation at 12,000 rpm for 20 min at 4˚C. The supernatant was collected in a fresh eppendorf tube and used for further experiments. The remaining was stored in -20˚C.
## Isolation of dna-protein complexes
Supernatant (88 μl) prepared above was used for performing digestion with T5 exonuclease. T5 exonuclease, 2 μl (0.2unit/μl) enzyme was added to the above supernatant and incubated at 37˚C for 45 min. The activity of the enzyme was stopped by EDTA (final concentration 10mM). A control experiment was set without treatment of supernatant with T5 exonuclease. This mixture of about 100 μl DNA-protein complex is mixed with streptavidin solution containing biotinylated oligonucleotide and incubated for 30 min in a rotating wheel in cold room. To attach the biotinylated oligonucleotide to the streptavidin beads, 5' biotin labeled PS19 primer (5 μM) (5' CAGTCATATGCGCGTATGTGTCCTCTAACCGTAAATAGCG 3') was attached to 200 μl streptavidin solution in a microfuge tube and incubated in a rotating wheel for 10 min at room temperature. The primer used is 40 bp in length and specific to the promoter of our interest. As a control beads without bound oligonucleotides were used.
The mixture was centrifuged at 12,000 rpm for 2 min. at 4˚C and the supernatant was collected. The supernatant should contain unbound proteins. The DNA protein complex bound to streptavidin beads was washed with 200 μl binding buffer three to four times by repeated incubations in a rotating wheel for 5 min. each in the cold room. The binding buffer consisted of 12% glycerol, 12 mM HEPES (pH 7.9), 4 mM Tris-Cl (pH 7.9), 60 mM potassium chloride, 1 mM EDTA, 1 mM DTT. The bound proteins were eluted by elution buffer and four fractions each of 200 μl were collected. The elution buffer consisted of 12% glycerol, 20 mM Tris-Cl (pH 5.8), 1 M potassium chloride, 5 mM magnesium chloride, 1 mM EDTA, 1 mM DTT, 20 μg/ml BSA.
## Sds-page separation, in gel digestion and identification of proteins
The eluates obtained in pull down assay were analyzed on 12% SDS-PAGE gel with 5% stacking gel [bib_ref] SDS-page Laemmli method, Laemmli [/bib_ref]. The gel was silver stained and the different bands obtained in the elution lane of the gel were cut and in-gel digested with trypsin following the protocol described by Bruker Daltonics, Bremen, Germany adapted from . The digested peptides were dried in speed vac and the samples were analyzed by MS/MS on ABI Sciex 5800 TOF/ TOF system, USA.
The parameters set for the MS/MS analysis were as follows: Fixed modifications: Carbamidomethyl (C), Variable modifications: Deamidated (NQ), Oxidation (M), Peptide Mass Tolerance: ± 150 ppm.
To determine the specificity of the method, a non-specific DNA, biotinylated primer of kanamycin promoter was used. The remaining procedure was identical as was performed with primer of dsz promoter. The elution fractions were analyzed on an SDS-PAGE and identified by MALDI-ToF MS/MS analysis.
# Recombinant dna methods
The methods followed for cloning were according to protocol mentioned in Sambrook and Russell. The gene for XRE (Xenobiotic response element) family transcription regulator was cloned between the restriction sites NdeI and HindIII in the vector pET29a. This recombinant plasmid DNA was transformed to expression strain for studying overexpression of XRE. The restriction enzymes, T4 DNA ligase, Taq DNA polymerases were purchased from New England Biolabs, USA. IPTG and other chemicals used were of molecular grade.
Cloning and expression of XRE family transcription regulator (Xenobiotic response element) family transcription regulator. The gene encoding for XRE like protein was amplified using primer pairs 5' CTAGCATATGGTGAGCGAGCAACGCCGAATCGGGTACCAC 3' (XRE-F) and 5' CTAGAAGCTTCGTGGTCGTGCCATCGTCACCATCGCTGCG 3' (XRE-R). The amplified PCR product was digested with NdeI and HindIII and cloned in pET29a vector between NdeI and HindIII. The vector so constructed was named pPM4. The recombinant plasmid was transformed to expression strains BL21 (DE3) and BL21 (DE3) plysS for overexpression of the desired protein. Single colony of the transformed cell was inoculated in LB media containing kanamycin (50 μg/ml). The cells were induced with IPTG (final concentration 1 mM) at an OD 600~0 .5, and the samples were collected at different time intervals (5 hrs and overnight after induction) and analyzed on a 12% SDS PAGE gel to check the expression.
## Purification of xre family protein
To validate the binding of XRE protein to the dsz promoter, the protein was purified partially and EMSA was performed. For purification, a single colony of the BL21 (DE3) plysS cells harboring the plasmid containing the gene encoding for XRE protein was inoculated in 200 ml of LB media and induced for overexpression with 1mM IPTG. After 5 hrs of induction, the cells were harvested by centrifugation at 3500 rpm for 10 min. The cells were resuspended in 4 ml denaturing buffer A (100 mM Na 2 PO 4 , 10 mM Tris, 8 M urea, pH 8) and sonicated on Q125 sonicator (Qsonica sonicators, USA) by giving 30 sec bursts and 20 sec rest for 5 min at an amplitude of 25%. The cell debris was removed by centrifugation at 10,000 rpm for 15 min and the supernatant was collected in a fresh tube. The sonicated supernatant was added to the Ni-NTA column equilibrated with the buffer A. The suspension was incubated at 4˚C on a rotating wheel for 2 hrs to allow the binding of the protein to Ni-NTA agarose resins. The flow through was collected and the column was washed with buffer B (100 mM Na 2 PO 4 , 10 mM Tris, 8 M urea, pH 6.5). The proteins bound to the column was eluted by elution buffer C and D (100 mM Na 2 PO 4 , 10 mM Tris, 8 M urea, pH 5.5 and 100 mM Na 2 PO 4 , 10 mM Tris, 8 M urea, pH 4.5 respectively). All the different collected fractions were run on 12% SDS PAGE gel to check the fraction which contains the protein of our interest. The elution fraction with buffer C (contained the XRE protein) was renatured by adding the eluate (5 ml) dropwise to the 50 ml renaturation buffer (10% glycerol, 25 mM Tris-Cl and 150 mM sodium chloride, pH 8) in cold room by stirring continuously with magnetic stirrer. The renatured protein solution was further dialyzed overnight against the renaturation buffer with 2-3 changes to remove the traces of urea. The renatured sample was run on 12% SDS PAGE gel to check the purity of the protein.
## Emsa with partially purified xre
EMSA was performed to determine the binding of XRE with the dsz promoter. For this purpose, dsz promoter was Cy5 labeled by PCR amplification using Cy5 PS19 primer (5' CAGTC ATATGCGCGTATGTGTCCTCTAACCGTAAATAGCG 3'). Reaction mixture consisting of Tris binding buffer (20 mM Tris Cl, 5 mM MgCl 2 , 0.1 mM EDTA, 6% sucrose, 100 mM KCl and 1 mM DTT, pH 7.5), poly d(I-C), poly L-Lysine, 12 ng of labeled promoter and increasing concentration of partially purified protein (5-20 μg) was incubated for 4 hrs at 4˚C and was then run on 6% native TBE gel for 2 hrs at 100V. The gel was then scanned in a fluorimager (GE typhoon 9000) to see the shift. Two types of controls were used. First contained all the components of the reaction mixture mentioned above expect the protein and the second reaction mixture consisted of tris binding buffer, poly d(I-C), poly L-Lysine, 12 ng of Cy5 labeled dsz promoter and 20 μg of renatured washing eluate from purification (it contains other non-specific proteins present in the partially purified eluate but the protein of our interest in negligible amounts (beyond detection limit)). The reaction mixture was incubated for 4 hrs at 4˚C and was then run on 6% native TBE gel for 2 hrs at 100V. The gel was then scanned in a fluorimager (GE typhoon 9000) to see the shift.
# Results and discussion
To mimic in vivo conditions outside the cellular environment is complicated thereby making it difficult for all the DNA binding proteins bind to their respective DNA. Here, we propose an in vivo method for the isolation of the DNA binding proteins. Our improved method for isolation of DNA binding proteins is based upon the method originally proposed by Wang, 2009 [bib_ref] Exploring DNA-binding proteins with in vivo chemical cross-linking and mass spectrometry, Qiu [/bib_ref] , modified by Wu et al. [bib_ref] Sequence-specific capture of protein-DNA complexes for mass spectrometric protein identification, Wu [/bib_ref]. The method differs in the following aspects. 1) We propose to use T5 exonuclease instead of exonuclease III. T5 exonuclease chews the DNA from 5'-3' direction whereas exonuclease III chews from 3'-5' direction. In any promoter DNA, regulatory elements bind to either the 3' end or the upstream region or both. The extra sequences at the 5' end can be taken for the isolation of all the promoter DNA binding proteins. In this way, it is likely that the user would not miss out any protein. [bib_ref] Transcription factor binding site positioning in yeast: proximal promoter motifs characterize TATA-less..., Erb [/bib_ref] , 2) biotinylated oligo is used for pairing with the DNA bound protein complex. 3) After elution, the eluted proteins were loaded on an SDS-PAGE followed by in gel digestion and analysis by MALDI-ToF MS/MS [bib_ref] Advances in proteome analysis by mass spectrometry, Griffin [/bib_ref] , [bib_ref] Rapid identification of DNA-binding proteins by mass spectrometry, Nordhoff [/bib_ref]. 4) The method proposed worked well for bacterial cells and the DNA-protein complex is formed in vivo unlike the method described by Wu et al, 2011 where a PCR amplified DNA is used for DNA protein complex formation. 5) Microarray is not required and therefore the proposed method is cost effective and does not require the knowledge of whole genome sequence. Also, sufficient amount of protein can be obtained by packing an avidin agarose column and binding the DNA protein complex and further elution of the bound proteins. The method is schematically described in [fig_ref] Fig 1: Schematic diagram showing the experimental procedure followed [/fig_ref]
## Isolation of dna binding proteins
In the present study, the method is used to isolate the proteins that bind to the dsz promoter in a biodesulfurizing bacterium, Gordonia sp. IITR100. In desulfurizing bacteria, the desulfurizing genes (in the order dszA, -B and-C) are present as an operon under the control of the dsz promoter. The proteins which bind to the dsz promoter have not been reported to date. Here, we have employed in vivo cross-linking of proteins with DNA followed by affinity capture method to isolate the proteins that specifically bind to the DNA of our interest.
## In vivo formaldehyde cross-linking and its reversal
Formaldehyde was used for crosslinking of DNA protein complex [bib_ref] Studies on histone organization in the nucleosome using formaldehyde as a reversible..., Jackson [/bib_ref] [bib_ref] Formaldehyde crosslinking: a tool for the study of chromatin complexes, Hoffman [/bib_ref]. An exponentially growing culture of Gordonia sp. IITR100 with an OD 600~1 was selected as the cells are transcriptionally most active and majority of the cells would have been in actively transcribing state [bib_ref] Analysis of growth-phase regulated genes in Streptococcus agalactiae by global transcript profiling, Sitkiewicz [/bib_ref].After crosslinking, the cells were harvested, sonicated and the supernatant was collected. An aliquot of this supernatant was digested with T5 exonuclease and was passed through the streptavidin column containing biotinylated primer (PS19) specific to the dsz promoter. The proteins bound to the DNA were collected as mentioned above. The formaldehyde crosslinked proteins were eluted because of low pH as it is known that formaldehyde crosslinking is highly dependent on pH [bib_ref] Importance of fixation in immunohistochemistry: use of formaldehyde solutions at variable pH..., Berod [/bib_ref].
To get optimal yields various parameters such as concentration of formaldehyde and Rapigest were optimized. Out of the three different concentrations of formaldehyde used (0.5%, 0.75% and 1%), the highest band intensity of eluted proteins was observed at a concentration of 0.75%, therefore this concentration of formaldehyde was used further. The concentration of eluted proteins as determined by Bradford assay is shown in the [fig_ref] Fig 2: Graph depicting the concentration of protein in eluates when different amount of... [/fig_ref] As a control experiment, the cross linking was reversed by adding Rapigest before harvesting the cells. SDS could not be used because it can interfere with MALDI [bib_ref] Sequence-specific capture of protein-DNA complexes for mass spectrometric protein identification, Wu [/bib_ref]. Four different concentrations of Rapigest 0.5, 0.75, 1 and 2% were used [fig_ref] Fig 2: Graph depicting the concentration of protein in eluates when different amount of... [/fig_ref]. After reversal of cross linking, similar protocol as mentioned above was followed to isolate the proteins bound to the DNA. Low Rapigest gave more protein because reversal is not complete and all the proteins remain bound to the DNA that is eluted based on pH. On the other hand, high concentration of Rapigest results in low concentration of proteins because the reversal is complete and protein-DNA complex is not available, instead free DNA is available for binding. Thus, only few proteins were eluted. For maximum reversal 1% rapigest was used in the experiment.
## Characterization of dna-binding proteins with in vivo cross-linking
The eluates from both the experiments (cross linked and reversal of cross linking) were run on 12% SDS PAGE. The number of bands observed by silver staining was 15 in case of eluates from crosslinked samples whereas after reversal of crosslinking the number of bands reduced to 8. The bands were in the size range of 10 to 270 kDa [fig_ref] Fig 3: A 12% SDS-PAGE gel showing different elution fractions after crosslinking by formaldehyde... [/fig_ref]. In control experiments where supernatant without digestion with T5 exonuclease was used, about 5 bands were observed [fig_ref] Fig 4: A 12% SDS-PAGE gel showing different elution fractions after crosslinking by formaldehyde... [/fig_ref]. In another control experiment where beads without bound oligonucleotides were used, 4 bands were observed [fig_ref] Fig 4: A 12% SDS-PAGE gel showing different elution fractions after crosslinking by formaldehyde... [/fig_ref]. To determine the specificity of the method, a non-specific DNA, biotinylated primer of kanamycin promoter was used. Only few proteins were detected because kanamycin promoter is an E. coli promoter and is absent in Gordonia. It is likely that the absence of complementarity to the primer of kanamycin promoter resulted in detection of fewer proteins [fig_ref] Fig 5: A 12% SDS-PAGE gel showing different elution fractions after crosslinking by formaldehyde... [/fig_ref]. The bands obtained in the elution lane of each of the experiments were cut, in gel digested and processed for MS/MS analysis. Based on the Mascot database search, we could identify a transcriptosome complex. On string analysis, we found that a number of identified proteins form an interactome, out of which 3 proteins were found to directly interact with RNA polymerase subunit omega [fig_ref] Fig 6: Interactome of proteins obtained after affinity chromatography [/fig_ref].
## Xre family transcription regulator binds to dsz promoter
To validate the method, one of the genes encoding for XRE like protein obtained in the above assay was selected. The XRE family of transcription regulators is the second most abundant family of transcription regulators in bacteria. It is known for regulating diverse processes in bacteria such as metabolism of nitrogen and carbon containing compounds, lysogeny etc. The XRE family transcription regulator contains a helix turn helix motif which is common in DNA binding proteins [bib_ref] Biochemical characterization of the transcriptional regulator BzdR from Azoarcus sp, Durante-Rodríguez [/bib_ref] [bib_ref] Structural classification of HTH DNA-binding domains and protein-DNA interaction modes, Wintjens [/bib_ref]. The reason for choosing XRE family transcription regulator is: 1) XRE family transcription regulator was obtained in the elution fractions from both crosslinked and after reversal of crosslinking, although the concentration was reduced after reversal, 2) The gene encoding for XRE family transcription regulator is located at a distance of 8-10 kb from the dsz operon in all desulfurizing strains. Thus, to determine whether it binds to the promoter DNA an expression vector pPM4 was used. The expression of XRE family transcription regulator was observed in E. coli BL21(DE3) pLysS cells [fig_ref] Fig 7: A 12% SDS-PAGE gel showing overexpression of our protein in different expression... [/fig_ref]. Gel shift assay was https://doi.org/10.1371/journal.pone.0202602.g008 performed using Cy5 labeled dsz promoter [bib_ref] A fluorescence based nonradioactive electrophoretic mobility shift assay, Ruscher [/bib_ref] and partially purified XRE (37% pure). Elution fraction which did not contain XRE (or it might be present at a very low concentration which is not visible in the SDS PAGE gel) when used, did not give any shift. At protein concentrations 5 and 10 μg two shifted bands were observed, whereas on increasing the protein concentration further, free DNA disappeared and a major more retarded band was observed [fig_ref] Fig 8: Gel shift assay with partially purified XRE family transcription regulator [/fig_ref].
# Conclusions
The method developed in the present study can be used to isolate unknown proteins that bind to known DNA sequences under in vivo conditions from bacterial cells. It confers advantage over other existing methods as it allows the DNA-protein complex formed in vivo to be isolated by a simple affinity-based method. Unlike other method as reported by Wu et al [bib_ref] Sequence-specific capture of protein-DNA complexes for mass spectrometric protein identification, Wu [/bib_ref] , microarray is not required thereby making this method cost effective. Moreover, knowledge of whole genome sequence of the bacteria or the protein sequence (as required in ChiP) is not required [bib_ref] Chromatin immunoprecipitation assay as a tool for analyzing transcription factor activity, Gade [/bib_ref]. Also, sufficient amount of protein can be obtained by packing an avidin agarose column and binding the DNA protein complex and further elution of the bound proteins. Thus, the method is reliable, simple and cost effective for the isolation and identification of DNA-binding proteins. Using the method, we have isolated a transcription regulator XRE, which binds to the dsz promoter. This is the first report on the isolation of a transcription regulator which binds the dsz operon.
## Supporting information
[fig] Fig 1: Schematic diagram showing the experimental procedure followed. The left panel shows the experiment procedure without reversing the crosslinking between DNA and the protein. In the right panel shows the method followed for control where the cross linking is reversed with Rapigest before harvesting the cells. https://doi.org/10.1371/journal.pone.0202602.g001 [/fig]
[fig] Fig 2: Graph depicting the concentration of protein in eluates when different amount of formaldehyde is used for cross linking (A) and Rapigest for reversal of crosslinking (B). https://doi.org/10.1371/journal.pone.0202602.g002 [/fig]
[fig] Fig 3: A 12% SDS-PAGE gel showing different elution fractions after crosslinking by formaldehyde and affinity assay (A) and affinity assay after reversal of crosslinking by Rapigest (B). Lane1: broad range protein marker, lane 2-3: elution fractions. The proteins identified and sequence coverage corresponding to each band is shown. https://doi.org/10.1371/journal.pone.0202602.g003 [/fig]
[fig] Fig 4: A 12% SDS-PAGE gel showing different elution fractions after crosslinking by formaldehyde and affinity assay. (A) The affinity assay was performed using biotinylated primer specific to the dsz promoter attached to streptavidin coated dynabeads. The crosslinked crude extract was passed through the above column without digestion with T5 exonuclease. (B) The affinity assay was performed without attaching any oligonucleotide to streptavidin coated dynabeads. The crosslinked crude extract was digested with T5 exonuclease, then passed through the above column and the different fractions were collected. Lane1: broad range protein marker, lane 2: crude protein extract, lane 3: flow through, lane 4: elution fraction. The proteins identified and sequence coverage corresponding to each band is shown.https://doi.org/10.1371/journal.pone.0202602.g004 [/fig]
[fig] Fig 5: A 12% SDS-PAGE gel showing different elution fractions after crosslinking by formaldehyde and affinity assay. The affinity assay was performed using biotinylated primer of kanamycin promoter attached to streptavidin coated dynabeads. The crosslinked crude extract was digested with T5 exonuclease, then passed through the above column and the different fractions were collected. Lane1: broad range protein marker, lane 2: flow through, lane 3-4: elution fractions. The proteins identified and sequence coverage corresponding to each band is shown. https://doi.org/10.1371/journal.pone.0202602.g005 [/fig]
[fig] Fig 6: Interactome of proteins obtained after affinity chromatography. XRE family protein, WhiB, MerR family protein, Ubiquitin like protein pup, RNA polymerase -70 sigma factor, rpoZ (transcriptase subunit omega) and rbpA (RNA polymerase binding protein). https://doi.org/10.1371/journal.pone.0202602.g006 [/fig]
[fig] Fig 7: A 12% SDS-PAGE gel showing overexpression of our protein in different expression strains. Lane1: marker, lane2: BL21(DE3) uninduced, lane3: BL21(DE3) 5hr induction, lane4: BL21(DE3) overnight induction, lane5: BL21 (DE3) pLysS uninduced, lane6: BL21(DE3) pLysS 5hr induction, lane7: BL21(DE3) pLysS overnight induction. https://doi.org/10.1371/journal.pone.0202602.g007 [/fig]
[fig] Fig 8: Gel shift assay with partially purified XRE family transcription regulator. Lane 1: free Cy5 labeled dsz, lane 2-5: Cy5 labeled dsz + partially purified XRE (1, 2, 3,4 μg respectively), lane 6: control lane (it contains other proteins of partial purified lane except XRE family protein), lane 7-10: Cy5 labeled dsz + partially purified XRE concentrated (5, 10, 15,20 μg respectively). [/fig]
[fig] S1: Fig. Purification of XRE. A) Lane 1: marker, lane 2: sonicated supernatant overexpressing XRE, lane 3: flow through, lane 4: washing, lane 5-7: elution fraction with pH 6.5, 5.5 and 4.5 respectively. (TIF) [/fig]
|
## Supplementary
## Removal efficiency re(d bc )
[formula] D BC4 D BC5 D BC3 D BC1 D BC2 D BC3 D BC4 D BC5 [/formula]
BC-containing particle BC-free particle
[formula] Amount of non-BC coating R D BC -resolved num. conc. N air (D BC ) D BC -resolved num. conc. N rain (D BC ) [/formula]
Hygroscopicity of non-BC material κ
## Ccn number fraction f ccn (d bc ,ss)
Critical supersaturation SS C (D BC ,R)
[fig] F: bc-imp (D BC ) D BC -resolved number fraction of BC in air that will be scavenged by impaction with rain droplets below cloud CCN (D BC ,SS) D BC -resolved number fraction of BC in air that will be activated as CCN at SS ic-imp (D BC ) D BC -resolved number fraction of BC in air that will be scavenged by impaction with cloud or rain droplets in cloud k Hygroscopicity parameter of non-BC aerosols in air N air (D BC ) D BC -resolved number concentration of BC in air N CCN (D BC ) D BC -resolved number concentration of BC in air that will be activated as CCN at SS N rain (D BC ) D BC -resolved number concentration of BC in rainwater R Shell-to-core diameter ratio (relative coating amount) of each BC particle in air RE(D BC ) D BC -resolved removal efficiency; N rain (D BC ) / N air (D BC ) SS Supersaturation of water vapour SS c (D BC ,R) D BC -resolved critical supersaturation of BC in air SS est An SS value providing the best agreement of D BC -dependence between CCN (D BC ) and RE(D BC ) Supplementary igure S1. A flow chart illustrating the sampling, measurements, 26 and data analysis in our observational method. See the main text for details. [/fig]
[fig] D 0 D: Same as Fig. 3 in the main text but for the other 6 precipitation events. Since both the 38 estimated F ic-imp (D BC ) and F bc-imp (D BC ) are much smaller than F CCN (D BC ,SS) for the size 39 range of D BC = 185-370 nm, they are not shown for illustrative convenience. BC (nm) RE (D BC ) F CCN (D BC , 0.08 %) F CCN (D BC , 0.05-BC (nm) RE (D BC ) F CCN (D BC , 0.12 %) F CCN (D BC , 0.07-0.20 %) Event No.8 failing rain droplet is considered. The number size distribution of cloud droplets and the 48 residence time of an air parcel in cloud and below cloud are assumed (Figure S5. D BC -resolved number fraction scavenged by impaction 55 in cloud (F ic-imp ) and below cloud (F bc-imp ) for a selected precipitation event. The 56 results for the event No. 4 are shown. The solid and dashed lines show the estimated 57 F ic-imp and F bc-imp , respectively. The different markers indicate various assumptions of 58 the number size distributions of water droplets. [/fig]
[table] Table S1: Symbols used in this paper. 21 [/table]
|
Delayed Onset of Pleural Effusion After Thoracic Radiation Therapy for Hodgkin Lymphoma: A Case Report With Over 30-Year Follow-Up
Pleural effusion after thoracic radiation therapy is an important adverse event affecting the patient's quality of life. A 58-year-old woman was referred to our hospital with the complaint of exertional dyspnea for several months. Chest radiography revealed left pleural effusion, most likely related to radiotherapy to mediastinal and left cervical lymph nodes in the treatment of localized Hodgkin lymphoma 25 years ago. She was followed for the next eight years and experienced a slow exacerbation of pleural effusion. Here, we report a case of extremely late-onset pleural effusion after thoracic radiotherapy.
# Introduction
Hodgkin lymphoma is a malignant B-cell lymphoma that is histopathologically characterized by the presence of Reed-Sternberg cells. Localized Hodgkin lymphoma is classified into favorable and unfavorable prognostic groups based on several risk factors [bib_ref] Prognostic factors in Hodgkin lymphoma, Bröckelmann [/bib_ref]. Since a favorable prognosis of early-stage Hodgkin lymphoma implies a long-term survival rate, there has been interest in the clinical management of late events that occur after curative treatment [bib_ref] Late effects following treatment of Hodgkin lymphoma during childhood and adolescence. Results..., Dörffel [/bib_ref]. For example, mantle irradiation alone has been utilized for early-stage Hodgkin lymphoma and has an excellent clinical outcome, with a 10-year overall survival rate of 98.2% [bib_ref] Long-term results of a prospective trial of mantle irradiation alone for earlystage..., Ng [/bib_ref]. However, normal tissue damage caused by radiation therapy results in delayed toxicity, such as in the case of cardiovascular diseases and secondary cancers [bib_ref] Late effects in the era of modern therapy for Hodgkin lymphoma, Hodgson [/bib_ref].
Pleural effusion after chest radiation therapy is also an important side effect affecting a patient's quality of life, causing chest pain, nonproductive cough, orthopnea, and dyspnea [bib_ref] Pleural effusion in adults-etiology, diagnosis, and treatment, Jany [/bib_ref]. In this report, we describe a case of pleural effusion that occurred 25 years after thoracic radiotherapy.
## Case presentation
A 58-year-old woman was referred to a pulmonologist at our institution with complaints of exertional dyspnea. She had a medical history of radiotherapy to the mediastinal and cervical lymph nodes, called mantle field irradiation, for the treatment of stage IIA Hodgkin lymphoma 25 years prior to the time of reference. Her medical history also included hypertension, hypothyroidism, hepatitis C following interferon treatment, and bronchial asthma. Chest radiography revealed left pleural effusion .
## Figure 1: chest radiography images
Chest radiography images taken (A) 25 years, (B) 27 years, and (C) 33 years after thoracic radiation therapy for Hodgkin lymphoma. Computed tomography showed that the bronchi were tractioned toward the mediastinum, suggesting decreased lung volume caused by radiation-related fibrotic changes .
## Figure 2: a computed tomography scan of the bronchi
A computed tomography scan demonstrating that the bronchi in the left lung are tractioned toward the mediastinum, thus indicating decreased lung volume.
Thoracentesis was performed for diagnostic purposes and symptom palliation. Pleural fluid analysis revealed a total protein level of 4.5 g/dL, a glucose level of 112 mg/dL, lactate dehydrogenase level of 108 U/L, albumin level of 3.0 g/dL, and carcinoembryonic antigen of 1.4 ng/ml, whereas blood serum analysis revealed a total protein level of 7.2 g/dL and lactate dehydrogenase level of 223 U/L. According to Light's criteria for pleural fluid [bib_ref] The Light criteria: the beginning and why they are useful 40 years..., Light [/bib_ref] , the patient's pleural fluid was classified as exudate because the pleural fluid/serum protein level ratio was >0.5. Both pleural fluid smear and culture for acid-fast bacilli results were negative. No significant increase in adenosine deaminase or hyaluronic acid levels was observed. The pleural fluid cytology result was class 1 with no malignancy. She showed no liver dysfunction (aspartate aminotransferase of 22 U/L, alanine aminotransferase of 10 U/L, total bilirubin of 0.6 mg/dL, prothrombin time of 89%, albumin of 4.1 g/dL) and her hypothyroidism was well controlled (thyroidstimulating hormone of 4.88 microIU/ml, free triiodothyronine of 1.7 pg/mL, and free thyroxine of 0.98 ng/mL). Echocardiography showed no obvious sign of heart failure with a good left ventricular ejection fraction (EF of 65%) and right ventricular function (right ventricular systolic pressure of 38 mmHg, right atrial pressure of 3 mmHg).
After two years, the pleural effusion on the left side gradually increased , and she underwent thoracentesis once again for cytological examination, which resulted in no malignancy. After more than six years, the pleural effusion had exacerbated, presenting in the right chest . Thoracentesis for cytological examination of the right chest revealed no malignancy (
## 1: pleural fluid and serum analysis of the patient
During the follow-up period, three thoracenteses were performed. No significant changes in their quantitative characteristics were observed.
## Rt -radiotherapy
Treatment included thoracentesis to relieve respiratory distress, followed by the oral administration of tolvaptan, a competitive vasopressin receptor antagonist that functions as a diuretic, at 7.5 mg/day.
# Discussion
Here, we report a case of late-onset pleural effusion after thoracic radiotherapy. The production of pleural effusions after radiotherapy is caused by chronic pleuritis and lymphatic obstruction due to mediastinal fibrosis. Aqeel et al. conducted a retrospective analysis of 96 cancer patients receiving thoracic irradiation, and more than half of the patients reported developing new pleural effusions with a median time to development of six months [bib_ref] Evaluation of radiation-induced pleural effusions after radiotherapy to support development of animal..., Aqeel [/bib_ref]. They also reported that 67% of the patients developed pleural effusion ipsilateral to the irradiated site. Zhao et al. reported that the incidence rate of radiation-inducing pleural effusion in lung cancer patients was 24.9%, and the median onset time from the end of thoracic radiotherapy was 3.7 months [bib_ref] Thoracic radiation-induced pleural effusion and risk factors in patients with lung cancer, Zhao [/bib_ref]. Compared to these values, the timing of the onset of pleural effusions we reported is characteristic. Shen et al. also reported a late onset of pleural effusion that occurred 25 years after radiotherapy for cervical lymphoma [bib_ref] Case report of delayed radiotherapy-related pleural effusion following chest radiotherapy for lymphoma, Shen [/bib_ref].
A consensus on the management of radiation-induced pleural effusion is not yet well-established. Its treatment is usually similar to that of general pleural effusion and includes diuretic medications, thoracentesis, an indwelling pleural catheter, and pleural sclerosis [bib_ref] Update in the management of pleural effusions, Aboudara [/bib_ref]. in the oral administration of 30 mg of prednisolone corticosteroid for pleural effusion due to radiation-induced pleuritis after chemoradiotherapy in a patient with esophageal cancer [bib_ref] Successful corticosteroid treatment for pleural effusion due to radiation-induced pleuritis after chemoradiotherapy..., Kumagai [/bib_ref].
Currently, the standard treatment for early-stage Hodgkin lymphoma appears to be a combination of systemic chemotherapy and radiation therapy, with a trend toward smaller doses and more localized irradiation fields called involved-field irradiation that can avoid severe late adverse events [bib_ref] Reduced treatment intensity in patients with early-stage Hodgkin's lymphoma, Engert [/bib_ref]. Recently, Borchman and colleagues reported excellent clinical outcomes of radiotherapy-omitted treatment strategies using positron emission tomography assessment after receiving two cycles of dose-escalated bleomycin, etoposide, doxorubicin, cyclophosphamide, vincristine, procarbazine, and prednisone (BEACOPP) plus two cycles of doxorubicin, bleomycin, vinblastine, and dacarbazine (ABVD) [bib_ref] PET-guided omission of radiotherapy in early-stage unfavourable Hodgkin lymphoma (GHSG HD17): a..., Borchmann [/bib_ref].
Our report describes an extremely late-onset case of pleural effusion that occurred 25 years after thoracic radiotherapy. Although the role of radiotherapy in early-stage Hodgkin lymphoma has been decreasing in recent decades, we have presented the possibility of late side effects in an era where large-field irradiation is more commonly used. Moreover, there is a lack of knowledge regarding the mechanisms by which radiotherapy affects the development of pleural effusions. Integrated studies involving clinical research, radiobiology, and physiology are important to reduce the risk of post-radiotherapy pleural effusions.
# Additional information disclosures
Human subjects: Consent was obtained or waived by all participants in this study. Conflicts of interest: In compliance with the ICMJE uniform disclosure form, all authors declare the following: Payment/services info: All authors have declared that no financial support was received from any organization for the submitted work. Financial relationships: All authors have declared that they have no financial relationships at present or within the previous three years with any organizations that might have an interest in the submitted work. Other relationships: All authors have declared that there are no other relationships or activities that could appear to have influenced the submitted work. |
Baseline mapping of severe fever with thrombocytopenia syndrome virology, epidemiology and vaccine research and development
Severe fever with thrombocytopenia syndrome virus (SFTSV) is a newly emergent tick-borne bunyavirus first discovered in 2009 in China. SFTSV is a growing public health problem that may become more prominent owing to multiple competent tick-vectors and the expansion of human populations in areas where the vectors are found. Although tick-vectors of SFTSV are found in a wide geographic area, SFTS cases have only been reported from China, South Korea, Vietnam, and Japan. Patients with SFTS often present with high fever, leukopenia, and thrombocytopenia, and in some cases, symptoms can progress to severe outcomes, including hemorrhagic disease. Reported SFTSV case fatality rates range from~5 to >30% depending on the region surveyed, with more severe disease reported in older individuals. Currently, treatment options for this viral infection remain mostly supportive as there are no licensed vaccines available and research is in the discovery stage. Animal models for SFTSV appear to recapitulate many facets of human disease, although none of the models mirror all clinical manifestations. There are insufficient data available on basic immunologic responses, the immune correlate(s) of protection, and the determinants of severe disease by SFTSV and related viruses. Many aspects of SFTSV virology and epidemiology are not fully understood, including a detailed understanding of the annual numbers of cases and the vertebrate host of the virus, so additional research on this disease is essential towards the development of vaccines and therapeutics.npj Vaccines (2020) 5:111 ; https://doi.
# Introduction
Severe fever with thrombocytopenia syndrome virus (SFTSV) was first isolated from a patient in China during 2009 1 . DH82 canine macrophage-like cells were infected with the virus and cytopathic effects (CPE), manifested as changes in cell morphology from round to elongated, were observed by light microscopy. After further characterization by electron microscopy and nucleotide sequencing, the virus was shown to be a member of the Bunyaviridae family and Phlebovirus genus 1 . It is currently classified as a member of the order Bunyavirales, family Phenuiviridae, genus Banyangvirus, and a member of the Bhanja virus serocomplex 2 . Early phylogenetic analysis showed that, although SFTSV had been shown to infect ticks, it was distantly related to viruses from both the Phlebotomus fever group as well as the Uukuneimi (UUKV) group, indicating that it was part of a new group of phleboviruses [bib_ref] Systematic review of severe fever with thrombocytopenia syndrome: virology, epidemiology, and clinical..., Liu [/bib_ref] [bib_ref] The ecology, genetic diversity, and phylogeny of Huaiyangshan virus in China, Zhang [/bib_ref]. As more novel phleboviruses continued to emerge, it was determined that SFTSV is closely related to Malsoor (MV) and Heartland viruses (HRTV), isolated in India and the United States, respectively, and is more distantly related to Palma virus (PALV), originally isolated from ticks in Portugal, and Bhanja virus (BHAV), originally isolated from a goat in India [bib_ref] Malsoor virus, a novel bat phlebovirus, is closely related to severe fever..., Mourya [/bib_ref] [bib_ref] First detection of heartland virus (Bunyaviridae: Phlebovirus) from field collected arthropods, Savage [/bib_ref] [bib_ref] Palma virus, a new bunyaviridae isolated from ticks in Portugal, Filipe [/bib_ref]. Since its discovery, SFTSV has been isolated from several species of ticks, including Haemaphysalis longicornis, Rhipicephalus microplus, and Ixodes nipponensis, with one study confirming the ability of H. longicornis to act as a vector for SFTSV [bib_ref] The ecology, genetic diversity, and phylogeny of Huaiyangshan virus in China, Zhang [/bib_ref] [bib_ref] Severe fever with thrombocytopenia syndrome virus in ticks collected from humans, Yun [/bib_ref] [bib_ref] Haemaphysalis longicornis ticks as reservoir and vector of severe fever with thrombocytopenia..., Luo [/bib_ref]. Based upon recent phylogenetic analyses, it has also been proposed that person-to-person transmission is likely [bib_ref] Family cluster analysis of severe fever with thrombocytopenia syndrome virus infection in..., Yoo [/bib_ref] [bib_ref] Nosocomial transmission of severe fever with thrombocytopenia syndrome in Korea, Kim [/bib_ref].
SFTS was first reported in the Hubei and Henan Provinces of central China in 2009 [bib_ref] Fever with thrombocytopenia associated with a novel Bunyavirus in China, Yu [/bib_ref] [bib_ref] Current status of severe fever with thrombocytopenia syndrome in China, Zhan [/bib_ref]. Between 2010 and 2011, it was confirmed that the new disease was caused by a previously unidentified Bunyavirus, now known as SFTSV [bib_ref] Systematic review of severe fever with thrombocytopenia syndrome: virology, epidemiology, and clinical..., Liu [/bib_ref] [bib_ref] Current status of severe fever with thrombocytopenia syndrome in China, Zhan [/bib_ref] [bib_ref] Clinical progress and risk factors for death in severe fever with thrombocytopenia..., Gai [/bib_ref]. SFTSV infection in humans has an incubation period of~7-14 days followed by a clinical disease that typically presents as severe fever with thrombocytopenia and leukocytopenia with symptoms persisting~7-13 days. Patients often also experience lymphadenopathy, gastrointestinal symptoms (nausea, vomiting, diarrhea), central nervous system symptoms (apathy, lethargy, tremors, convulsions, coma), and/or hemorrhagic symptoms (ecchymosis, disseminated intravascular coagulation, pulmonary, and/or gastrointestinal bleeding) [bib_ref] Current status of severe fever with thrombocytopenia syndrome in China, Zhan [/bib_ref] [bib_ref] Clinical progress and risk factors for death in severe fever with thrombocytopenia..., Gai [/bib_ref] [bib_ref] Clinical and laboratory characteristics of severe fever with thrombocytopenia syndrome in Chinese..., Weng [/bib_ref]. In addition, patients often experience multiorgan dysfunction (MOD) in which two or more organ systems are compromised [bib_ref] Clinical progress and risk factors for death in severe fever with thrombocytopenia..., Gai [/bib_ref] [bib_ref] Clinical and laboratory characteristics of severe fever with thrombocytopenia syndrome in Chinese..., Weng [/bib_ref] [bib_ref] Multiple organ dysfunction syndrome, Ramirez [/bib_ref]. Although damage from MOD may be reversible, it can progress to multiorgan failure (MOF), which is typically associated with fatality in SFTS patients [bib_ref] Clinical progress and risk factors for death in severe fever with thrombocytopenia..., Gai [/bib_ref] [bib_ref] Clinical and laboratory characteristics of severe fever with thrombocytopenia syndrome in Chinese..., Weng [/bib_ref] [bib_ref] Multiple organ dysfunction syndrome, Ramirez [/bib_ref]. Although MOD is a relatively common outcome of SFTSV infection, there are reports of mild infections that do not require extensive intervention [bib_ref] Mild clinical course of severe fever with thrombocytopenia syndrome virus infection in..., Ohagi [/bib_ref]. In addition, there are also reports of unusual disease outcomes, including one study of encephalopathy induced by SFTSV that did not result in a fatality, as well as a report of SFTSVinduced reversible myocardial dysfunction [bib_ref] Unusual presentation of a severely ill patient having severe fever with thrombocytopenia..., Kaneko [/bib_ref] [bib_ref] Severe fever with thrombocytopenia syndrome with myocardial dysfunction and encephalopathy: a case..., Kawaguchi [/bib_ref]. SFTSV infections are most often confirmed using RT-PCR, and although plasma exchange and ribavirin therapy have been used as treatments, their success as therapeutics for SFTS is controversial, as not all treated patients recovered and no results from clinical trials of candidate therapies have been reported to date [bib_ref] Current status of severe fever with thrombocytopenia syndrome in China, Zhan [/bib_ref] [bib_ref] Case-fatality ratio and effectiveness of ribavirin therapy among hospitalized patients in china..., Liu [/bib_ref] [bib_ref] Plasma exchange and ribavirin for rapidly progressive severe fever with thrombocytopenia syndrome, Sup [/bib_ref]. In addition, candidate vaccines are still in the discovery phase.
## Molecular virology and phylogenetics
Like other members of the bunyaviruses, SFTSV has a spherical virion that is covered by an envelope, comprised of a lipid bilayer with glycoprotein spikes, that packages the three RNA segments; large (L) and medium (M) are negative-sense RNA and the small (S) RNA is ambisense [bib_ref] Fever with thrombocytopenia associated with a novel Bunyavirus in China, Yu [/bib_ref] [bib_ref] Critical epitopes in the Nucleocapsid protein of SFTS virus recognized by a..., Yu [/bib_ref] [bib_ref] Expression of structural and non-structural proteins of severe fever with thrombocytopenia syndrome..., Lu [/bib_ref] [bib_ref] Biological and phylogenetic analysis of first isolate of Tahyna virus in China, Lu [/bib_ref]. The L segment, with 6368 nucleotides, encodes the RNA dependent RNA polymerase (RdRp) used by the virus in replication [bib_ref] Fever with thrombocytopenia associated with a novel Bunyavirus in China, Yu [/bib_ref] [bib_ref] Critical epitopes in the Nucleocapsid protein of SFTS virus recognized by a..., Yu [/bib_ref]. The M segment, 3378 nucleotides, has one open reading frame that encodes a precursor protein that is cleaved in the endoplasmic reticulum to yield two glycoproteins, Gn and Gc 1 , which are embedded in the envelope surface and are involved in virus entry [bib_ref] Entry of severe fever with thrombocytopenia syndrome virus, Yuan [/bib_ref] [bib_ref] Evidence that processing of the severe fever with thrombocytopenia syndrome virus Gn/Gc..., Plegge [/bib_ref]. The Gc protein is a class II fusion protein and has been shown to be involved in the fusion of SFTS viral and host membranes [bib_ref] Entry of severe fever with thrombocytopenia syndrome virus, Yuan [/bib_ref] [bib_ref] Proteomics computational analyses suggest that the carboxyl terminal glycoproteins of Bunyaviruses are..., Garry [/bib_ref] [bib_ref] Structure of a phleboviral envelope glycoprotein reveals a consolidated model of membrane..., Halldorsson [/bib_ref]. The Gn protein has been shown to be vital to the incorporation of Gc into the envelope surface and also encodes neutralizing epitopes [bib_ref] Evidence that processing of the severe fever with thrombocytopenia syndrome virus Gn/Gc..., Plegge [/bib_ref] [bib_ref] Structure of a phleboviral envelope glycoprotein reveals a consolidated model of membrane..., Halldorsson [/bib_ref]. The S segment is 1746 nucleotides [bib_ref] Reverse genetics system for severe fever with thrombocytopenia syndrome virus, Brennan [/bib_ref] [bib_ref] Molecular genomic characterization of tick-and human-derived severe fever with thrombocytopenia syndrome virus..., Yun [/bib_ref] and encodes the nucleocapsid protein (N) and the nonstructural protein (NSs) [bib_ref] Fever with thrombocytopenia associated with a novel Bunyavirus in China, Yu [/bib_ref]. Although it has not been confirmed for SFTSV, the glycoproteins of bunyaviruses are suspected of interacting with the N as they have been co-immunoprecipitated with the N in studies using UUKV [bib_ref] The cytoplasmic tails of Uukuniemi Virus (Bunyaviridae) G(N) and G(C) glycoproteins are..., Overby [/bib_ref]. The NSs protein is involved in the formation of inclusion bodies that comprise an immune evasion strategy where intracellular vesicles are formed through the interaction of the NSs protein with perilipin A, adipose differentiation-related protein, and synaptogyrin-2 [bib_ref] Synaptogyrin-2 promotes replication of a novel tick-borne Bunyavirus through Interacting with viral..., Sun [/bib_ref]. These vesicles sequester TBK1, IKKϵ, and IRF3, which in turn dampens the interferon (IFN) response to the virus [bib_ref] Hijacking of RIG-I signaling proteins into virus-induced cytoplasmic structures correlates with the..., Santiago [/bib_ref] [bib_ref] Evasion of antiviral immunity through sequestering of TBK1/IKK /IRF3 into viral inclusion..., Wu [/bib_ref]. These vesicles colocalize with multiple early endosomal markers and autophagy markers, suggesting a role of these pathways in SFTSV pathogenesis [bib_ref] Hijacking of RIG-I signaling proteins into virus-induced cytoplasmic structures correlates with the..., Santiago [/bib_ref]. The NSs protein is also involved in inhibition of the IFN response by interfering with the JAK/STAT signaling pathway [bib_ref] Severe fever with thrombocytopenia syndrome virus inhibits exogenous Type I IFN signaling..., Chen [/bib_ref] [bib_ref] Suppression of type I and type III IFN signalling by NSs protein..., Zhang [/bib_ref]. Deletion of the N-terminal 25 amino acids of NSs negated its ability to suppress the IFN response to the virus 38 .
Cell culture and virus entry SFTSV is capable of infecting a variety of cell lines and the DCspecific intracellular molecule 3-grabbing receptor (DC-SIGN) has been shown to be a receptor for SFTSV through interaction with SFTSV glycoproteins, Gc and Gn 39 . The ability of SFTSV to infect cell lines from multiple species, such as hamster, monkey, mouse, and dog, suggests that receptors for the glycoproteins are conserved [bib_ref] Severe fever with thrombocytopenia virus glycoproteins are targeted by neutralizing antibodies and..., Hofmann [/bib_ref]. DC-SIGNR, liver and lymph node sinusoidal endothelial cell C-type lectin (LSECtin), and glucosylceramide synthase have been shown to also be important for viral entry into cells [bib_ref] Severe fever with thrombocytopenia virus glycoproteins are targeted by neutralizing antibodies and..., Hofmann [/bib_ref] [bib_ref] A role for glycolipid biosynthesis in severe fever with thrombocytopenia syndrome virus..., Drake [/bib_ref] [bib_ref] Characterization of severe fever with thrombocytopenia syndrome virus glycoprotein-mediated entry, Tani [/bib_ref]. Gn was shown to bind non-muscle myosin heavy chain IIA (NMMHC-IIA) and overexpression of NMMHC-IIA was found to increase infection in HeLa cells, indicating that NMMHC-IIA is another potential component in SFTSV entry [bib_ref] Nonmuscle myosin heavy chain IIA is a critical factor contributing to the..., Sun [/bib_ref]. It was also found that SFTSV infectivity is irreversibly reduced when the pH is lower than 6.0, indicating that pH may influence the conformation of SFTSV glycoproteins 41 .
Phylogenetic studies As stated above, SFTSV was found to be most closely related to MV and HRTV, and more distantly related to other phleboviruses, including Rift Valley fever virus (RVFV) 4-6 . The SFTSV RdRp was found to have 46.3% and 73.4% amino-acid similarity with that of RVFV and HRTV, respectively [bib_ref] Malsoor virus, a novel bat phlebovirus, is closely related to severe fever..., Mourya [/bib_ref] [bib_ref] A highly pathogenic new bunyavirus emerged in China, Li [/bib_ref]. The N of SFTSV and RVFV are only partially homologous, with 41% similarity, whereas the N of HRTV is 62.8% homologous to SFTSV. SFTSV and HRTV glycoproteins share 62.6% sequence homology [bib_ref] A highly pathogenic new bunyavirus emerged in China, Li [/bib_ref] [bib_ref] RNA encapsidation and packaging in the Phleboviruses, Hornak [/bib_ref] [bib_ref] Evolutionary and molecular analysis of the emergent severe fever with thrombocytopenia syndrome..., Lam [/bib_ref]. However, there are ten conserved cysteine residues in the Gn proteins of RVFV, SFTSV, and other related phleboviruses that are involved in the dimerization of Gn [bib_ref] Structures of phlebovirus glycoprotein Gn and identification of a neutralizing antibody epitope, Wu [/bib_ref]. The NSs proteins of SFTSV and HRTV are 63.5% similar while the NSs of SFTSV and RVFV are only 11-16% similar [bib_ref] A highly pathogenic new bunyavirus emerged in China, Li [/bib_ref]. Maximum likelihood phylogeny for each viral segment further revealed that SFTSV clusters into six genotypes: A, B, C, D, E, and F, whereas earlier analyses also using maximum likelihood divided the strains into five genotypes [bib_ref] Molecular genomic characterization of tick-and human-derived severe fever with thrombocytopenia syndrome virus..., Yun [/bib_ref] [bib_ref] Phylogeographic analysis of severe fever with thrombocytopenia syndrome virus from Zhoushan Islands,..., Fu [/bib_ref] [bib_ref] Increased prevalence of severe fever with thrombocytopenia syndrome in eastern china clustered..., Li [/bib_ref] [bib_ref] Migration, recombination, and reassortment are involved in the evolution of severe fever..., Shi [/bib_ref]. Strains in the F genotype are mostly from China but one analysis included a strain from South Korea 31 . It is also important to note that strains are reported in different genotypes between analyses 4,31 . All genotypes have been found in China, all genotypes but C have been isolated in South Korea, and genotypes B, C, and E have been isolated in Japan [bib_ref] Molecular genomic characterization of tick-and human-derived severe fever with thrombocytopenia syndrome virus..., Yun [/bib_ref] [bib_ref] Molecular evolution and spatial transmission of severe fever with thrombocytopenia syndrome virus..., Liu [/bib_ref]. It has been shown that the L and M segments of the virus are capable of homologous recombination and/or reassortment in instances of coinfection, thereby causing antigenic shift [bib_ref] Increased prevalence of severe fever with thrombocytopenia syndrome in eastern china clustered..., Li [/bib_ref] [bib_ref] Migration, recombination, and reassortment are involved in the evolution of severe fever..., Shi [/bib_ref] [bib_ref] Discovery of severe fever with thrombocytopenia syndrome bunyavirus strains originating from intragenic..., He [/bib_ref]. The three segments of the virus have different rates of evolution with the S segment having the highest rate of evolution and the L segment having the lowest [bib_ref] Molecular evolution and spatial transmission of severe fever with thrombocytopenia syndrome virus..., Liu [/bib_ref].
Owing to the relatively recent discovery of SFTSV, the evolution of the virus has not been fully elucidated. Determination of the phylogenetic history and evolutionary trends of SFTSV would provide information to aid in the development of diagnostics to allow differentiation among related viruses and specific identification of SFTSV in clinical samples.
## Epidemiology
Since its initial discovery, SFTSV has been reported in 23 provinces across China and confirmed human infections were recorded in 18 of these provinces [bib_ref] Current status of severe fever with thrombocytopenia syndrome in China, Zhan [/bib_ref]. Several Chinese provinces have had a higher disease burden than others, including Henan, Shandong, Anhui, and Hubei, and cases in these provinces frequently originate from rural, hilly regions [bib_ref] Current status of severe fever with thrombocytopenia syndrome in China, Zhan [/bib_ref] [bib_ref] A national assessment of the epidemiology of severe fever with thrombocytopenia syndrome, Liu [/bib_ref]. Outside of China, SFTSV infections have also been confirmed in Japan, Vietnam, and South Korea, plus evidence in Taiwan and Pakistan. The first reported cases of SFTS in Japan and South Korea occurred in febrile patients hospitalized in 2012 who were confirmed as SFTSV-positive in 2013 [bib_ref] Severe fever with thrombocytopenia syndrome virus in ticks collected from humans, Yun [/bib_ref] [bib_ref] Current status of severe fever with thrombocytopenia syndrome in China, Zhan [/bib_ref]. Recently, confirmed cases of SFTSV were reported in Vietnam 54 . Serological surveys in Pakistan suggest that a low percentage of the population have neutralizing antibodies indicative of SFTSV infection and seroconversion [bib_ref] Serologic evidence of severe fever with thrombocytopenia syndrome virus and related viruses..., Zohaib [/bib_ref]. Taiwan has also detected serological evidence of SFTSV as well as viral RNA in some of the surveyed ruminants [bib_ref] The first discovery of severe fever With thrombocytopenia syndrome virus in Taiwan, Tl [/bib_ref]. As SFTSV is tick-borne, the typical season for SFTSV infection is from early spring through late autumn [bib_ref] Current status of severe fever with thrombocytopenia syndrome in China, Zhan [/bib_ref] and there are a number of environmental predictors associated with ticks, including cattle density, forest coverage, rainfall, relative humidity, sunshine hours, and temperature elevation 52 . Cumulative data from 2010 to 2016 in China demonstrate that the highest reporting for SFTS cases was between May and July 14 . Reports in South Korea from 2013 to 2015 were highest between July-October, and reports in Japan from 2013 to 2014 were highest between April and August. Serological studies throughout China demonstrate that infected individuals undergo seroconversion during the acute and convalescent course of the disease but reports of SFTSV seropositivity amongst healthy populations vary and overall seroprevalence data indicate that only a small portion of the general population has seroconverted (on average~4.3%) [bib_ref] Current status of severe fever with thrombocytopenia syndrome in China, Zhan [/bib_ref] [bib_ref] Clinical progress and risk factors for death in severe fever with thrombocytopenia..., Gai [/bib_ref]. Although the age of patients diagnosed has ranged widely, from children to the elderly, adults are more commonly infected and older age is associated with worsening disease outcome [bib_ref] Current status of severe fever with thrombocytopenia syndrome in China, Zhan [/bib_ref]. Specifically, most studies report that the majority of SFTS cases occur in patients older than 50 and case fatality rates (CFRs) are higher in the elderly [bib_ref] Current status of severe fever with thrombocytopenia syndrome in China, Zhan [/bib_ref] [bib_ref] A national assessment of the epidemiology of severe fever with thrombocytopenia syndrome, Liu [/bib_ref].
Reports of CFRs from SFTSV infection range between~5 and >30% depending on the region surveyed and the amount of reporting in that area [bib_ref] Current status of severe fever with thrombocytopenia syndrome in China, Zhan [/bib_ref]. Owing to the variation in the dates and locations included in published reports of SFTSV incidence, it is unclear how many cases of SFTSV have been diagnosed since its discovery. Between 2011 and 2016, the Chinese Centers for Disease Control and Prevention reported between~500 and 2500 SFTS cases annually resulting in~50-100 deaths each year (cumulative CFR from 2010-2016 of 5.3%) [bib_ref] Current status of severe fever with thrombocytopenia syndrome in China, Zhan [/bib_ref]. South Korea and Japan tend to report fewer cases and a higher CFR than what is observed in China, however, this may be a result of differences in reporting and surveillance in different regions. South Korea typically reports fewer than 100 cases per year, however, the CFR in South Korea is high (32.6% CFR during 2013-2015). A retrospective analysis following the first case of SFTS in Japan identified 11 confirmed SFTS patients and six deaths in 2012, and through the end of 2015 a total of 161 cases of SFTS had been reported in Japan. Several groups have noted the significant disparity in CFRs between China and other endemic countries. These differences are, in part, attributed to a lack of clinical reporting [bib_ref] Risk factors for death in severe fever with thrombocytopenia syndrome, Yu [/bib_ref]. Several other factors may also play a role in these disparities.
There have been several isolated, putative infections in Greece, the United Arab Emirates, and the United States between 2009 and 2012, in which SFTSV was considered a possible etiological agent, however, laboratory tests have subsequently indicated that other closely related viruses likely caused these infections [bib_ref] A new phlebovirus associated with severe febrile illness in Missouri, Mcmullan [/bib_ref] [bib_ref] Acute thrombocytopenia, leucopenia, and multiorgan dysfunction: the first case of SFTS bunyavirus..., Denic [/bib_ref] [bib_ref] Novel phlebovirus in febril child, Anagnostou [/bib_ref]. In particular, the tick-borne HRTV was identified in the United States in 2012 [bib_ref] A new phlebovirus associated with severe febrile illness in Missouri, Mcmullan [/bib_ref].
A small study comparing fatal and non-fatal cases of SFTS in one region of China identified several differences between clinical symptoms and laboratory biomarkers that correlated with disease outcome [bib_ref] Clinical progress and risk factors for death in severe fever with thrombocytopenia..., Gai [/bib_ref]. During the first 1-7 days after disease onset, there was no significant difference in viral load, fever, or gastrointestinal symptoms between fatal and non-fatal cases [bib_ref] Clinical progress and risk factors for death in severe fever with thrombocytopenia..., Gai [/bib_ref]. During days 7-13 post onset, surviving patients had a decrease in viral load, whereas fatal cases continued to have sustained or increased viremia [bib_ref] Clinical progress and risk factors for death in severe fever with thrombocytopenia..., Gai [/bib_ref]. Both non-fatal and fatal cases were found to develop central nervous system symptoms and hemorrhagic symptoms, but the rate and severity of these manifestations were higher for fatal cases [bib_ref] Clinical progress and risk factors for death in severe fever with thrombocytopenia..., Gai [/bib_ref]. Both fatal and non-fatal cases experienced apathy, lethargy, and tremors, but convulsions and coma were more strongly correlated with a fatal disease outcome [bib_ref] Clinical progress and risk factors for death in severe fever with thrombocytopenia..., Gai [/bib_ref]. Similarly, although both fatal and non-fatal cases could manifest MOD, very few surviving patients progressed to the more severe MOF and all fatal cases were associated with MOF. Laboratory tests found that patients who developed fatal MOD/MOF maintained low platelet counts and high levels of the enzymes aspartate aminotransaminase (AST), creatine kinase (CK), creatine kinase-MB (CK-MB), and lactate dehydrogenase (LDH) beyond 13 days post onset, whereas recovering patients typically were experiencing a return to normal platelet and enzyme levels by day 13 [bib_ref] Clinical progress and risk factors for death in severe fever with thrombocytopenia..., Gai [/bib_ref].
There is ongoing speculation about possible arthropod vectors and reservoir hosts, as studies continue to investigate the transmission cycle of SFTSV. Based on investigation of SFTSV RNA, multiple tick species are implicated as possible vectors while viral RNA has not been detected in mosquitoes, midges, or sandflies [bib_ref] Current status of severe fever with thrombocytopenia syndrome in China, Zhan [/bib_ref] [bib_ref] Severe fever with thrombocytopenia syndrome virus in humans, domesticated animals, ticks, and..., Tian [/bib_ref]. Haemophysalis longicornis ticks often have a high prevalence compared with other tick populations in China, South Korea, and Japan 14 . In China, in addition to H. longicornis, SFTSV RNA has also been detected in H. concinna and Rhipicephalus microplus ticks 4,65 . In South Korea, SFTSV RNA has been identified in H. longicornis, H. flava, Ixodes nipponensis, and Amblyomma testudinarium ticks [bib_ref] Severe fever with thrombocytopenia syndrome virus in ticks collected from humans, Yun [/bib_ref]. In Japan, multiple tick species have been tested for SFTSV but none have yet been positive for infectious virus or viral RNA [bib_ref] Epidemiological survey of severe fever with thrombocytopenia syndrome virus in ticks in..., Hayasaka [/bib_ref]. Prevalence rates of SFTSV in tick populations range from 0.6% to 23.5% based on tick species, geographic area, and sampling size [bib_ref] The ecology, genetic diversity, and phylogeny of Huaiyangshan virus in China, Zhang [/bib_ref] [bib_ref] Severe fever with thrombocytopenia syndrome virus in ticks collected from humans, Yun [/bib_ref] [bib_ref] Severe fever with thrombocytopenia syndrome virus in humans, domesticated animals, ticks, and..., Tian [/bib_ref]. In addition, the various tick species studied are not geographically restricted to the SFTSV endemic regions only. For example, H. longicornis is often implicated as a likely vector and can be found outside of China, South Korea, and Japan in countries including Australia, New Zealand, Fiji, New Caledonia, and Russia 68 . Recently, H. longicornis was introduced into the United States through agricultural trade and has been found in nine states [bib_ref] Multistate infestation with the exotic disease-vector tick haemaphysalis longicornis -United States, Beard [/bib_ref] [bib_ref] Discovery of Haemaphysalis longicornis (Ixodida: Ixodidae) parasitizing a sheep in New Jersey,..., Rainey [/bib_ref]. Although some patients diagnosed with SFTSV report recent tick exposure, not every patient has a known tick-bite history as the apparent infection route [bib_ref] Systematic review of severe fever with thrombocytopenia syndrome: virology, epidemiology, and clinical..., Liu [/bib_ref] [bib_ref] Current status of severe fever with thrombocytopenia syndrome in China, Zhan [/bib_ref]. Therefore, as regions that are endemic for SFTS do not necessarily have a high prevalence of SFTSV in tick populations, studies are in progress to investigate alternative routes of virus transmission.
As it is likely that ticks acquire SFTSV during feeding on an infected reservoir animal, studies across endemic regions have also tested various animal species for serological markers of infection. Multiple wild and domestic animals surveyed in SFTSV endemic regions of China have been found to carry viral RNA and/ or SFTSV-specific antibodies, including sheep, cattle, dogs, pigs, chickens, goats, hedgehogs, geese, and rodents 3,14,65,71-73 . In South Korea, SFTSV RNA and/or antibodies have been detected in Korean water deer, wild boar, feral cats, lizards, and snakes, and in Japan virus-specific antibodies have been detected in cattle and wild boar [bib_ref] Molecular detection of severe fever with thrombocytopenia syndrome virus (SFTSV) in feral..., Hwang [/bib_ref] [bib_ref] Seroepidemiological evidence of severe fever with thrombocytopenia syndrome virus infections in wild..., Hayasaka [/bib_ref] [bib_ref] Detection of SFTS virus in ixodes nipponensis and amblyomma testudinarium (ixodida: ixodidae)..., Suh [/bib_ref]. Although SFTSV markers have been identified in several different animal species, there have been only a few reports of disease associated with SFTSV infection. In Japan, disease caused by SFTSV infection has been observed in domestic cats and captive cheetahs [bib_ref] Fatal tickborne phlebovirus infection in captive cheetahs, Matsuno [/bib_ref]. SFTSV-associated disease in a domestic canine has also been reported, with the companion animal suffering from fever, leukopenia, and thrombocytopenia prior to recovery [bib_ref] Unusual case of severe fever with thrombocytopenia syndrome showing clinical manifestations in..., Nam [/bib_ref]. One group proposed that ticks and migratory birds are responsible for the movement of SFTSV across geographic regions owing to the overlap of bird migratory patterns with the geographic distribution of H. longicornis ticks, but the potential role of birds has not been confirmed [bib_ref] The emergence and cross species transmission of newly discovered tick-borne Bunyavirus in..., Zhang [/bib_ref]. Although there is a notable degree of uncertainty about the natural transmission cycle of SFTSV, multiple reports have demonstrated that the virus can be transmitted person-to-person via contact with contaminated blood [bib_ref] Nosocomial transmission of severe fever with thrombocytopenia syndrome in Korea, Kim [/bib_ref] [bib_ref] Human-to-human transmission of severe fever with thrombocytopenia syndrome bunyavirus through contact with..., Tang [/bib_ref]. The majority of patients who are diagnosed with SFTS are farmers who work in fields and possibly have frequent tick exposure, indicating that agricultural work is a major risk factor for SFTS in endemic areas [bib_ref] Current status of severe fever with thrombocytopenia syndrome in China, Zhan [/bib_ref].
## Diagnostics
## Case definition
Certain laboratory markers, such as elevated AST levels, coupled with relevant patient history, such as time of year, age, and location, can indicate if SFTSV infection is likely; however, the criteria for definitive SFTS diagnosis in China require that SFTSV or viral RNA be detected in patient serum, or patient be positive for anti-SFTSV IgM during active infection, or seroconversion indicated by a fourfold increase in virus-specific IgG in serum drawn during convalescence compared with IgG levels during active infection [bib_ref] Preliminary fast diagnosis of severe fever with thrombocytopenia syndrome with clinical and..., Hu [/bib_ref] [bib_ref] Hemorrhagic fever caused by a novel bunyavirus in china: pathogenesis and correlates..., Zhang [/bib_ref] [bib_ref] Animal models of emerging tick-borne phleboviruses: determining target cells in a lethal..., Matsuno [/bib_ref]. In Japan, the interim diagnostic criteria for SFTSV infection require a fever >38°C, gastrointestinal symptoms, leukopenia with <4 × 10 9 leukocytes/L, thrombocytopenia with <100 × 10 9 platelets/L, increased lactate, alanine aminotransferase, and AST levels, as well as admission to a hospital owing to symptoms, or death 84 . Viral RNA detection Initially, viral RNA was detected in patient sera during active infection by RT-PCR. In 2012, a quantitative real-time RT-PCR assay was developed with primers targeting the L, M, and S segments that were 99% specific to SFTSV and was capable of diagnosing SFTSV in 98.6% of cases [bib_ref] Early diagnosis of novel SFTS bunyavirus infection by quantitative real-time RT-PCR assay, Sun [/bib_ref]. Another study developed an RT-PCR protocol that was capable of detecting SFTSV, dengue, Hantaan, and Seoul viruses with 99% specificity and 100% diagnosis rate [bib_ref] A two-tube multiplex real-time RT-PCR assay for the detection of four hemorrhagic..., Li [/bib_ref]. This two-tube system reduced the cost of running four individual reactions by 40% and provided a quick diagnosis method for multiple viruses with overlapping geographic distributions and similar early clinical presentation [bib_ref] A two-tube multiplex real-time RT-PCR assay for the detection of four hemorrhagic..., Li [/bib_ref]. Several other studies have also successfully developed RT-PCR assays to detect SFTSV RNA with all studies achieving a 10 viral RNA copies/μl lower limit of detection [bib_ref] Development and application of a one-step real-time RT-PCR using a minor-groove-binding probe..., Li [/bib_ref] [bib_ref] Sensitive and specific PCR systems for detection of both chinese and japanese..., Yoshikawa [/bib_ref] [bib_ref] SFTSV RNA detection in sera of patients suffering from fever with thrombocytopenia..., Chi [/bib_ref]. Reverse transcription loop-mediated isothermal amplification assays (RT-LAMP) and RT cross-priming amplification (RT-CPA) have also been used to successfully detect SFTSV RNA [bib_ref] Development and evaluation of a reverse transcription loopmediated isothermal amplification assay for..., Yang [/bib_ref] [bib_ref] Establishment of a novel one-step reverse transcription loopmediated isothermal amplification assay for..., Xu [/bib_ref] [bib_ref] Detection of severe fever with thrombocytopenia syndrome virus by reverse transcription-cross-priming amplification..., Cui [/bib_ref] [bib_ref] Development of a colloidal gold kit for the diagnosis of severe fever..., Wang [/bib_ref]. Multiple reports on the development of RT-LAMP assays to detect SFTSV have shown that those are effective tools for rapid diagnosis with high specificity and sensitivity for SFTSV and the ability to detect multiple strains [bib_ref] Development and evaluation of a reverse transcription loopmediated isothermal amplification assay for..., Yang [/bib_ref] [bib_ref] Development of a real-time loop-mediated isothermal amplification method for the detection of..., Lee [/bib_ref]. The RT-CPA assay had a specificity of 94.1% and a sensitivity of 100% compared with realtime RT-PCR and culturing virus from viral samples isolated from 89 suspected SFTSV-infected donors [bib_ref] Detection of severe fever with thrombocytopenia syndrome virus by reverse transcription-cross-priming amplification..., Cui [/bib_ref]. Both RT-LAMP and RT-CPA assays require~2 hours to obtain results [bib_ref] Development and evaluation of a reverse transcription loopmediated isothermal amplification assay for..., Yang [/bib_ref] [bib_ref] Establishment of a novel one-step reverse transcription loopmediated isothermal amplification assay for..., Xu [/bib_ref] [bib_ref] Detection of severe fever with thrombocytopenia syndrome virus by reverse transcription-cross-priming amplification..., Cui [/bib_ref].
Serological assays According to the SFTS diagnostic guidelines for China, the disease can be diagnosed by the presence of IgM antibodies in sera during acute illness or IgG antibodies in sera in recovering patients. One study developed an assay for the detection of total antibodies against SFTSV that had no cross-reactivity to hantavirus or dengue samples [bib_ref] Development of a colloidal gold kit for the diagnosis of severe fever..., Wang [/bib_ref]. The study used a recombinant SFTSV N conjugated to horseradish peroxidase in a double-antigen sandwich enzyme-linked immunosorbent assay (ELISA) for detection of total serum antibodies against SFTSV and was also shown to be effective for testing sera from a variety of species including goats, hedgehogs, pigs, cattle, and chickens with a specificity of 100% [bib_ref] Preparation and evaluation of recombinant severe fever with thrombocytopenia syndrome virus nucleocapsid..., Jiao [/bib_ref]. Another study developed an immunochromatographic assay to detect IgG and IgM antibodies by conjugating recombinant SFTSV N (based on strain HB29) and streptavidin with colloidal gold [bib_ref] Development of a colloidal gold kit for the diagnosis of severe fever..., Wang [/bib_ref]. The Colloidal Gold test was found to be SFTSV specific and as accurate for detecting anti-SFTSV antibodies as indirect immunofluorescence antibody tests previously used for laboratory confirmation of SFTSV infection, with sensitivity to IgG of 1:512 and IgM of 1:128 [bib_ref] Development of a colloidal gold kit for the diagnosis of severe fever..., Wang [/bib_ref]. An indirect ELISA has also been developed that has comparable performance for detection of anti-SFTSV IgG to that of the reference sandwich ELISA kit detailed above, with a specificity and sensitivity of 100%. The same assay can also detect serum IgM with 100% specificity but only 90.59% sensitivity compared with the sandwich ELISA method 96 . Neutralization assays are also a highly regarded method for detecting the presence of SFTSV neutralizing antibodies in serum.
Although great progress has been made with the development of serological and RT-PCR-based diagnostics, there is a lack of standardized reference reagents and a standardized diagnostic methodology that would improve consistency and comparability of surveillance methods and facilitate evaluation of candidate vaccines.
## Immune response
Innate immune response In vitro and in vivo analyses have begun to elucidate the innate immune responses resulting from SFTSV infection. Studies in the Golden Syrian hamster model support the importance of IFN signaling in disease prevention as STAT2 knockout hamsters are highly susceptible to SFTSV infection while wild-type hamsters are not [bib_ref] Modeling severe fever with thrombocytopenia syndrome virus infection in golden syrian hamsters:..., Gowen [/bib_ref]. Further highlighting the importance of IFN signaling in protection against SFTSV, in vitro studies have found that SFTSV can antagonize type-I IFN signaling by mechanisms involving virus sequestration of components of the IFN pathway including (but not exclusively) TBK1, IKKε, IRF3, STAT1, and STAT2 into inclusion bodies [bib_ref] Evasion of antiviral immunity through sequestering of TBK1/IKK /IRF3 into viral inclusion..., Wu [/bib_ref].
Multiple studies have investigated the human immune response to SFTSV infection in an attempt to identify differential profiles of protective and non-protective immune signaling.
Specifically, several cohort and case-control studies of SFTS patients in China compared serum levels of cytokines, chemokines, and immune cell subsets with those of healthy individuals, SFTS patients experiencing mild disease, and SFTS patients experiencing severe and/or fatal outcome. Generally, SFTSV infection of humans causes an increase in cytokine signaling that correlates with higher viral load and worsening disease outcome [bib_ref] Downregulation of interferon-β and inhibition of TLR3 expression are associated with fatal..., Song [/bib_ref] [bib_ref] Cytokine and chemokine levels in patients with severe ffever with thrombocytopenia syndrome..., Deng [/bib_ref] [bib_ref] Prognostic value of clinical and immunological markers in acute phase of SFTS..., Ding [/bib_ref] [bib_ref] Host cytokine storm is associated with disease severity of severe fever with..., Sun [/bib_ref]. Although there was marked cytokine/chemokine upregulation in all SFTS patients, significant differences in immune signaling between SFTS patients experiencing mild versus severe disease have been identified [bib_ref] Downregulation of interferon-β and inhibition of TLR3 expression are associated with fatal..., Song [/bib_ref] [bib_ref] Cytokine and chemokine levels in patients with severe ffever with thrombocytopenia syndrome..., Deng [/bib_ref] [bib_ref] Prognostic value of clinical and immunological markers in acute phase of SFTS..., Ding [/bib_ref] [bib_ref] Host cytokine storm is associated with disease severity of severe fever with..., Sun [/bib_ref]. Specifically, TNF-α, IFN-γ, IP-10, IL-10, IL-6, MIP-1α, IL-8, IL-15, granzyme B, HSP70, G-CSF, IL-1-RA, and MCP-1 [bib_ref] Downregulation of interferon-β and inhibition of TLR3 expression are associated with fatal..., Song [/bib_ref] [bib_ref] Cytokine and chemokine levels in patients with severe ffever with thrombocytopenia syndrome..., Deng [/bib_ref] [bib_ref] Prognostic value of clinical and immunological markers in acute phase of SFTS..., Ding [/bib_ref] [bib_ref] Host cytokine storm is associated with disease severity of severe fever with..., Sun [/bib_ref] were present at higher levels in patients with severe compared with mild disease. Not all studies identified statistically significant differences in each of the cytokines/ chemokines listed above, however, there is agreement in the literature that severe disease is associated with a "cytokine storm" and increase in pro-inflammatory molecules such as TNF-α, IP-10, and IL-6 [bib_ref] Downregulation of interferon-β and inhibition of TLR3 expression are associated with fatal..., Song [/bib_ref] [bib_ref] Cytokine and chemokine levels in patients with severe ffever with thrombocytopenia syndrome..., Deng [/bib_ref] [bib_ref] Prognostic value of clinical and immunological markers in acute phase of SFTS..., Ding [/bib_ref] [bib_ref] Host cytokine storm is associated with disease severity of severe fever with..., Sun [/bib_ref]. Several cytokines/chemokines are notably downregulated as a result of SFTSV infection. Specifically, tPAI-1, GRO, PDGF-BB, and RANTES have been detected at decreased expression in SFTS patients when compared with healthy individuals; however, some other studies have not observed any change in RANTES expression based on disease severity [bib_ref] Cytokine and chemokine levels in patients with severe ffever with thrombocytopenia syndrome..., Deng [/bib_ref] [bib_ref] Prognostic value of clinical and immunological markers in acute phase of SFTS..., Ding [/bib_ref] [bib_ref] Host cytokine storm is associated with disease severity of severe fever with..., Sun [/bib_ref]. One study measured a significant decrease of IFN-β that negatively correlated with disease severity but did not identify similar correlations for IFN-α, IFN-γ, or IFN-λ [bib_ref] Downregulation of interferon-β and inhibition of TLR3 expression are associated with fatal..., Song [/bib_ref]. The same study found that patients with worsening disease experienced a decrease in IL-1β over time [bib_ref] Downregulation of interferon-β and inhibition of TLR3 expression are associated with fatal..., Song [/bib_ref]. In addition, this study measured host mRNAs in peripheral monocytes of patients with differing disease severity and found that expression of TLR3, IRF3, and IRF7 was each negatively correlated with disease severity, of which TLR3 exhibited the most downregulation in severe cases [bib_ref] Downregulation of interferon-β and inhibition of TLR3 expression are associated with fatal..., Song [/bib_ref].
The role of natural killer (NK) cells in disease outcome is still to be definitively established, as one study found that increased NK cell populations were associated with severe disease, whereas other studies found that NK cells were depleted throughout the first week of symptoms but then began to rise 2 weeks after disease onset [bib_ref] Characterization of immunological responses in patients with severe fever with thrombocytopenia syndrome:..., Lu [/bib_ref] [bib_ref] Dynamic changes of laboratory parameters and peripheral blood lymphocyte subsets in severe..., Liu [/bib_ref] [bib_ref] Detection and evaluation of immunofunction of patients with severe fever with thrombocytopenia..., Sun [/bib_ref]. One study has investigated the plasmacytoid DC (pDC) and myeloid DC (mDC) populations in patients and determined that levels of circulating pDCs were variable among SFTSV-infected patients but distinct patterns of mDCs could be observed that appeared to be associated with outcome [bib_ref] Downregulation of interferon-β and inhibition of TLR3 expression are associated with fatal..., Song [/bib_ref]. Specifically, mDCs were significantly increased 1 week after disease onset but then declined in weeks 2 and 3, especially in patients with severe disease [bib_ref] Downregulation of interferon-β and inhibition of TLR3 expression are associated with fatal..., Song [/bib_ref]. By 3 weeks post disease onset, the surviving patients had significantly more mDCs than deceased patients [bib_ref] Downregulation of interferon-β and inhibition of TLR3 expression are associated with fatal..., Song [/bib_ref].
A regulatory role of A20-binding inhibitor of NF-κB activation 2 (ABIN2), and tumor progression locus 2 (TPL2), have been identified in the inhibition of IFN signaling by SFTSV NSs 106 . Direct interactions between NSs and ABIN2 were shown to promote TPL2 complex formation, which subsequently resulted in increased IL-10 expression. Pharmacological inhibition of TPL2 signaling during SFTSV infection resulted in decreased IL-10 expression. Tpl2 −/− mice, as well as IL-10 −/− mice, survived challenge unlike their wild-type controls. These data show a key role of IL-10 signaling in SFTSV disease severity and outcome.
Adaptive immune response Several studies have also measured differences in T-and B-cell subsets in SFTSV-infected humans. Generally, there is agreement in the literature that severe disease is associated with depleted Tcell populations [bib_ref] Characterization of immunological responses in patients with severe fever with thrombocytopenia syndrome:..., Lu [/bib_ref] [bib_ref] Dynamic changes of laboratory parameters and peripheral blood lymphocyte subsets in severe..., Liu [/bib_ref] [bib_ref] Detection and evaluation of immunofunction of patients with severe fever with thrombocytopenia..., Sun [/bib_ref]. Specifically, patients with severe disease have lower numbers of CD3 + , CD4 + , and CD8 + T cells throughout the acute phase of disease [bib_ref] Characterization of immunological responses in patients with severe fever with thrombocytopenia syndrome:..., Lu [/bib_ref] [bib_ref] Dynamic changes of laboratory parameters and peripheral blood lymphocyte subsets in severe..., Liu [/bib_ref] [bib_ref] Detection and evaluation of immunofunction of patients with severe fever with thrombocytopenia..., Sun [/bib_ref]. A study of Chinese SFTS patients reported that arginine deficiency might contribute to T-cell N.E. Bopp et al. dysregulation during SFTSV infection [bib_ref] Arginine deficiency is involved in thrombocytopenia and immunosuppression in severe fever with..., Li [/bib_ref]. In terms of B cells, there are reports that increased numbers of B lymphocytes relative to other types of lymphocytes were correlated with severe disease and fatality [bib_ref] Dynamic changes of laboratory parameters and peripheral blood lymphocyte subsets in severe..., Liu [/bib_ref] [bib_ref] Detection and evaluation of immunofunction of patients with severe fever with thrombocytopenia..., Sun [/bib_ref].
Overall, adaptive immune mechanisms capable of protecting against SFTSV have not been fully identified. Studies with a related bunyavirus, RVFV, have suggested protective roles for neutralizing antibodies [bib_ref] A single immunization with MVA expressing GnGc glycoproteins promotes epitope-specific CD8+-T cell..., Lopez-Gil [/bib_ref] [bib_ref] The nucleocapsid protein of Rift Valley fever virus is a potent human..., Xu [/bib_ref]. Both the nucleoprotein and the Gn/Gc glycoproteins of RVFV have been identified as CD8+ T-cell targets [bib_ref] A single immunization with MVA expressing GnGc glycoproteins promotes epitope-specific CD8+-T cell..., Lopez-Gil [/bib_ref] [bib_ref] The nucleocapsid protein of Rift Valley fever virus is a potent human..., Xu [/bib_ref] however, the contribution of T cells specific for these proteins to protection against RVFV is incompletely understood.
For SFTSV, there have been a number of studies investigating the induction of virus-specific antibodies, including IgM, IgG, and neutralizing antibodies, resulting from SFTSV infection of humans. A cohort study of 298 Chinese SFTS patients identified some specific patterns of antibody responses following infection that may have relevance to disease pathogenesis [bib_ref] Characterization of immunological responses in patients with severe fever with thrombocytopenia syndrome:..., Lu [/bib_ref]. IgM seroconversion occurred between 4 and 21 days after disease onset, peaked after 4 weeks, and decreased significantly thereafter [bib_ref] Characterization of immunological responses in patients with severe fever with thrombocytopenia syndrome:..., Lu [/bib_ref]. IgG seroconversion occurred between 2 and 9 weeks post disease onset, peaked at 6 months, and then waned with an estimated half-life of 11 months 103 . During the first 4 weeks of disease onset, low IgM levels were measured in patients who were older, had comorbidities, had a higher initial viral load, and had severe disease. Moreover, another study of Chinese SFTS patients found that Nspecific IgM antibodies were associated with lower viral load and less-severe disease, but N-specific IgG, Gn-specific IgM, and Gcspecific IgM did not have the same correlations [bib_ref] Characterization of immunological responses in patients with severe fever with thrombocytopenia syndrome:..., Lu [/bib_ref]. The magnitude and durability of neutralizing antibody responses to SFTSV have not yet been studied in great detail, but a report from China indicates that SFTS patients develop low levels of neutralizing antibodies and although they wane over time, some patients retain neutralizing antibodies through at least 4 years post infection [bib_ref] Neutralizing antibodies to severe fever with thrombocytopenia syndrome virus 4 years after..., Huang [/bib_ref]. Of 25 patients tested, 50% plaque reduction neutralization test (PRNT 50 ) titers varied between 1:5 and 1:640 during the first year of infection and by year 4 the titers generally decreased only slightly, to a range of 1:20 to 1:160 [bib_ref] Neutralizing antibodies to severe fever with thrombocytopenia syndrome virus 4 years after..., Huang [/bib_ref]. Based on this 4-year follow-up data, it was predicted that SFTSV neutralizing antibodies could be protective for up to 9 years 111 .
A study in rhesus macaques found that SFTSV infection of nonhuman primates (NHPs) resulted in increased production of cytokines including IFN-γ and TNF-α similar to that observed in humans [bib_ref] SFTS virus infection in nonhuman primates, Jin [/bib_ref]. In addition, the study in NHPs found that SFTSVspecific IgM appeared~5 days post infection (dpi) and declined after~15 dpi, whereas SFTSV-specific IgG appeared~7 dpi and plateaued~15 dpi [bib_ref] SFTS virus infection in nonhuman primates, Jin [/bib_ref]. Neutralizing antibodies in NHPs were also present at low levels ranging from 1:32 to 1:128. Overall, the immune response in rhesus macaques appeared similar to that of humans but of lower magnitude consistent with the less-severe disease observed in NHPs.
Characterization of humoral responses against other bunyaviruses has indicated that neutralizing antibodies against the viruses typically recognize both Gn and Gc 113 . For SFTSV, neutralizing antibodies to the Gn protein have been identified while non-neutralizing antibodies recognize the N 23,46,114,115 . Using serum samples from four recovered patients as well as SFTSV antisera from mice and rabbits in hemagglutination inhibition (HI) assays, antibodies against SFTSV were found to be cross-reactive with HRTV, with an HI titer of 1:40 to 1:1280, and, to a lesser extent, BHAV, with an HI titer of 1:20 to 1:80 [bib_ref] Characterization of the Bhanja serogroup viruses (Bunyaviridae): a novel species of the..., Matsuno [/bib_ref]. Although the proteins that were bound by antibodies in this study were not specified, cross-reactivity of antibodies against SFTSV with HRTV was corroborated in another study where antibodies to the N from SFTSV bound to HRTV N 117 .
A neutralizing human monoclonal antibody (mAb) for SFTSV was engineered by generating an antibody library from lymphocytes isolated from five individuals infected with SFTSV and was found to bind to the α6 helix in domain III of the Gn protein [bib_ref] Structures of phlebovirus glycoprotein Gn and identification of a neutralizing antibody epitope, Wu [/bib_ref] [bib_ref] Human antibody neutralizes severe Fever with thrombocytopenia syndrome virus, an emerging hemorrhagic..., Guo [/bib_ref].
The mAb was capable of neutralizing several strains of SFTSV, indicating that the epitope is conserved within SFTSV strains and provides a potential vaccine target against SFTSV [bib_ref] Structures of phlebovirus glycoprotein Gn and identification of a neutralizing antibody epitope, Wu [/bib_ref] [bib_ref] Human antibody neutralizes severe Fever with thrombocytopenia syndrome virus, an emerging hemorrhagic..., Guo [/bib_ref]. This mAb was not found to be cross-reactive with RVFV, Forecariah virus (FORV), or PALV [bib_ref] Structures of phlebovirus glycoprotein Gn and identification of a neutralizing antibody epitope, Wu [/bib_ref] [bib_ref] Severe fever with thrombocytopenia syndrome virus antigen detection using monoclonal antibodies to..., Fukuma [/bib_ref]. In addition, mAbs were derived from a phage display library of lymphocytes from two patients with SFTSV. These monoclonal antibodies bound to the N but had no neutralizing activity [bib_ref] Critical epitopes in the Nucleocapsid protein of SFTS virus recognized by a..., Yu [/bib_ref]. It was determined that the N-terminal region of the N is important for antibody binding as well as three epitopes within the N: Glu10 and Phe11; Lys40, Lys41, and Glu44; and Arg243 [bib_ref] Critical epitopes in the Nucleocapsid protein of SFTS virus recognized by a..., Yu [/bib_ref]. Although antibodies that target the N have not been found to have neutralizing activity, they have been demonstrated to be an effective tool for diagnosing SFTSV infection through the recognition of virus in serum [bib_ref] Preparation and evaluation of recombinant severe fever with thrombocytopenia syndrome virus nucleocapsid..., Jiao [/bib_ref] [bib_ref] Application of recombinant severe fever with thrombocytopenia syndrome virus nucleocapsid protein for..., Yu [/bib_ref] [bib_ref] Severe fever with thrombocytopenia syndrome virus antigen detection using monoclonal antibodies to..., Fukuma [/bib_ref].
Animal models of SFTSV Since its discovery, various laboratory animals have been evaluated in an attempt to recapitulate the observed pathogenesis of human SFTSV infection. provides an overview of these studies. Small animal models, including mice, rats, and hamsters, have been tested with various pathogenic outcomes. Newborn mice show high susceptibility to SFTSV infection through multiple routes, including intracranial (IC) and intraperitoneal inoculation (IP) [bib_ref] Infection and pathogenesis of Huaiyangshan virus (a novel tickborne bunyavirus) in laboratory..., Chen [/bib_ref] [bib_ref] The pathogenesis of severe fever with thrombocytopenia syndrome virus infection in alpha/beta..., Liu [/bib_ref]. Full characterization of newborn infections has yet to be performed though they may provide a lethal model for therapeutics testing. Immunocompetent adult mice generally show few clinical signs of infection [bib_ref] Animal models of emerging tick-borne phleboviruses: determining target cells in a lethal..., Matsuno [/bib_ref] [bib_ref] Modeling severe fever with thrombocytopenia syndrome virus infection in golden syrian hamsters:..., Gowen [/bib_ref] [bib_ref] Infection and pathogenesis of Huaiyangshan virus (a novel tickborne bunyavirus) in laboratory..., Chen [/bib_ref] [bib_ref] The pathogenesis of severe fever with thrombocytopenia syndrome virus infection in alpha/beta..., Liu [/bib_ref] , though an in-depth characterization of SFTSV infection in C57BL/6 mice showed similar pathology to mild human infection 120 . Elevated liver enzymes, including AST and ALT, were observed in C57BL/6 mice that were inoculated with 10 5 TCID 50 via the IP route and viral RNA was detected in their livers, spleens, and kidneys. To generate a lethal infection in C57BL/6 mice, a dose of the immunosuppressant Mitomycin C was administered [bib_ref] Pathogenesis of emerging severe fever with thrombocytopenia syndrome virus in C57/BL6 mouse..., Jin [/bib_ref]. In addition to immune suppression by drugs, immunocompromised knockout mice have also been used to study SFTSV infection. Both C57BL/6 mice deficient in the type-I IFNR (IFNα/βR −/− ) [bib_ref] Animal models of emerging tick-borne phleboviruses: determining target cells in a lethal..., Matsuno [/bib_ref] , and strain 129 mice deficient in type-I IFNR (A129) [bib_ref] The pathogenesis of severe fever with thrombocytopenia syndrome virus infection in alpha/beta..., Liu [/bib_ref] have been used for SFTSV model development owing to their immunocompromised status. In both models, lethality was observed with high levels of viral RNA detected in the brain, liver, kidneys and spleen [bib_ref] Animal models of emerging tick-borne phleboviruses: determining target cells in a lethal..., Matsuno [/bib_ref] [bib_ref] The pathogenesis of severe fever with thrombocytopenia syndrome virus infection in alpha/beta..., Liu [/bib_ref]. However, these two models show different mean times to death of 3-4 dpi for A129 mice 119 and 5-7 dpi for IFNα/ βR −/− mice 83 . STAT1 and STAT2 knockout mice have been utilized with only STAT2 knockouts showing lethal infections [bib_ref] Species-specific pathogenicity of severe fever with thrombocytopenia syndrome virus is determined by..., Yoshikawa [/bib_ref]. More detail about these mouse models can be found in .
Rats have also been used to study SFTSV pathogenesis but appear to have limited value as a disease model. In the early characterization of the virus, newborn and adult Wistar rats were used for infections in an effort to recapitulate human disease, however, only newborn rats infected through the IC route showed universal mortality.
Hamsters have previously been employed in the study of other bunyaviruses, such as RVFV, suggesting that they may be a viable model for similar studies with SFTSV 122 . However, much like mouse models, immunocompetent hamsters do not succumb to lethal infections with SFTSV. STAT2 knockout hamsters were found to be susceptible to low doses of virus, succumbing after infection with 10 PFU introduced subcutaneously (SC) [bib_ref] Modeling severe fever with thrombocytopenia syndrome virus infection in golden syrian hamsters:..., Gowen [/bib_ref].
The use of ferrets as a model has shown significant promise owing to their age-dependent disease manifestations [bib_ref] Ferret animal model of severe fever with thrombocytopenia syndrome phlebovirus for human..., Park [/bib_ref] NL Increase in body temperature at 2 dpi and a return to normal by 12dpi. Slight body weight decrease. Slight thrombocytopenia and leukopenia beginning at 2dpi. Elevated AST and ALT between 2dpi and 8 dpi.
Detectable virus in the spleen, liver, kidneys, lungs, and serum with clearance by 8 dpi. [bib_ref] Severe fever with thrombocytopenia syndrome virus antigen detection using monoclonal antibodies to..., Fukuma [/bib_ref] Old ≤48-50 Months 10 7.6 TCID 50 L Increase in body temperature at 2 dpi with death occurring by 6dpi. Significant body weight decrease. Severe thrombocytopenia and leukopenia beginning at 2 dpi. Highly elevated AST and ALT by 2 dpi. Detectable virus in the spleen, liver, kidneys, lungs, brain, spinal cord, intestines, and serum. [bib_ref] Severe fever with thrombocytopenia syndrome virus antigen detection using monoclonal antibodies to..., Fukuma [/bib_ref] Non-human primates Rhesus macaque 4-5 years observed in the liver, spleen, kidneys, brain, spinal cord, and intestine. This model shows promise as it is one of the few immunocompetent models to show lethality during infection. This lethality coincides with the observation that SFTSV has significantly more severe manifestations in older individuals suggesting that age is a direct correlate of disease severity [bib_ref] Ferret animal model of severe fever with thrombocytopenia syndrome phlebovirus for human..., Park [/bib_ref]. Infections of NHPs have been performed in an attempt to develop a more biologically relevant model. Two studies have reported the infection of NHPs with SFTSV, one using rhesus macaques 112 and the other cynomolgus macaques [bib_ref] Animal models of emerging tick-borne phleboviruses: determining target cells in a lethal..., Matsuno [/bib_ref] , though only the infection of rhesus macaques was reported in detail. In that study, elevated cytokines mirroring human levels, such as IFNγ 100 , were observed. Minor lesions in the liver and kidney were also observed in the later stages of the infection with viremia peaking at day 5~10 5 TCID 50 /mL and clearing by day 7. Although the rhesus macaques used in that study did become infected, their disease progression mirrored that of only acute SFTSV infection, including mild fever [bib_ref] SFTS virus infection in nonhuman primates, Jin [/bib_ref].
To date, no animal model completely recapitulates all the observed clinical manifestations of SFTSV nor has a full characterization of any animal model in particular been performed to date. Consequently, further investigation into suitable animal models needs to be performed in order to best model disease progression and identify models for use in vaccine development.
## Therapeutics and vaccines for sftsv
Case management At present, no approved vaccines or therapeutics are available to prevent or treat infections by SFTSV. Care varies from case to case depending on the patient's manifestations. Treatment with antivirals and antisera [bib_ref] Plasma exchange and ribavirin for rapidly progressive severe fever with thrombocytopenia syndrome, Sup [/bib_ref] have shown mixed results. The addition of pulsed steroidal therapy has been shown to have positive effects [bib_ref] Steroid pulse therapy in patients with encephalopathy associated with severe fever with..., Nakamura [/bib_ref]. To date no standard of care has been established.
## Therapeutics
Early in vitro studies suggested that the common antivirals, ribavirin, and favipiravir, show efficacy in inhibiting viral multiplication. Initially, ribavirin was tested in Vero cell cultures against various strains of SFTSV to determine the 99% effective concentration. Cells were pre-treated with Ribavirin before infection with SFTSV (100 TCID 50 ) and cultured for 3 days before viral titers were determined [bib_ref] Combination effects of ribavirin and interferons on severe fever with thrombocytopenia syndrome..., Shimojima [/bib_ref]. Effective concentrations were found to range from 19 to 64 mg/ml depending on the strain 125 . However, treatment of cells with similar concentrations of Ribavirin after infection showed drastically diminished effects. From a few reported case studies, the employment of Ribavirin following the diagnosis of SFTSV infection has shown inconclusive effects on disease outcome [bib_ref] Case-fatality ratio and effectiveness of ribavirin therapy among hospitalized patients in china..., Liu [/bib_ref] [bib_ref] Two treatment cases of severe fever and thrombocytopenia syndrome with oral ribavirin..., Park [/bib_ref]. CFRs and biomarkers of disease, such as viral load and platelet counts, did not show improvement between patients who were given ribavirin and those who were not [bib_ref] Case-fatality ratio and effectiveness of ribavirin therapy among hospitalized patients in china..., Liu [/bib_ref]. These data may suggest that Ribavirin is not a suitable postexposure treatment, although more studies are required to address its suitability as a prophylactic. However, other case reports in which Ribavirin was administered, indicate positive therapeutic effects and increased patient recovery [bib_ref] Case-fatality ratio and effectiveness of ribavirin therapy among hospitalized patients in china..., Liu [/bib_ref] [bib_ref] Two treatment cases of severe fever and thrombocytopenia syndrome with oral ribavirin..., Park [/bib_ref]. These contradictory reports show the need for continued investigation into the therapeutic effects of Ribavirin in the treatment of SFTSV infection. Currently, a clinical trial to determine the efficacy of Ribavirin is taking place in China. In addition, two clinical trials are listed in the Chinese Clinical Trial Registry. The first is investigating Ribavirin only (http://www.chictr.org.cn/showprojen.aspx? proj=10477) and the other is to determine the efficacy of Ribavirin in combination with IFN alpha 2a (http://www.chictr. org.cn/showprojen.aspx?proj=11210).
Favipiravir, a purine analog, has also shown effective inhibition of SFTSV replication in vitro as well as in a susceptible animal model 127 . Using IFNα/βR−/− mice, it was shown that Favipiravir provided 100% protection to animals pre-treated with the drug. Animals infected SC with 1.0 × 10 6 TCID 50 were dosed via either IP or oral routes with Favipiravir for 5 days. Surviving animals showed milder clinical signs with increased platelet counts and lower viremia, compared with the untreated groups. A clinical trial was performed with Favipiravir to determine its ability to inhibit disease progression in infected patients. Patients were to given a dose of 1800 mg twice on the first day and 800 mg twice a day for the subsequent 9-13 days of treatment. Data from this study are yet to be published but a smaller case study showed promise in the use of Favipiravir as a treatment. Two patients, both male, were given a 5-day course of Favipiravir at different stages of disease [bib_ref] Severe fever with thrombocytopenia syndrome (SFTS) treated with a novel antiviral medication,..., Song [/bib_ref]. A notable decline in viral load was observed after administration, suggesting a potential effect of Favipiravir on disease course. Further data are required to show statistical significance of these findings.
In addition to these two common antivirals, several other treatments have been proposed for SFTSV infection. The use of recombinant IFN (α, β, and γ) has shown effects in a Vero cell model in a dose-dependent manner [bib_ref] Combination effects of ribavirin and interferons on severe fever with thrombocytopenia syndrome..., Shimojima [/bib_ref]. However, IFN administration has yet to be well-characterized in the context of patient infections. A clinical trial has been performed using recombinant human IFN α2, however, those results are not published. Case studies employing L-arginine supplementation have also been performed and suggest that arginine supplementation may be beneficial [bib_ref] Arginine deficiency is involved in thrombocytopenia and immunosuppression in severe fever with..., Li [/bib_ref]. In a retrospective clinical investigation, patients treated with calcium channel blockers (CCBs) showed significantly improved clinical outcomes in juxtaposition to patients not receiving CCBs 131 . Animal model testing confirmed that CCBs exhibit anti-SFTSV properties, suggesting that this class of compounds may be a viable option for treatment of patients. Several other molecules have been shown to inhibit SFTSV infection and are shown in [fig_ref] Table 2: Compounds that have been shown to inhibit SFTSV infection a [/fig_ref].
Passive transfer of antisera from human survivors has been investigated in small animal models as well as in patients. Sera from survivors were transferred to SFTSV-infected IFNα/βR−/− mice and shown to increase survival and decrease viremia when given 1 hour after infection 132 . Specifically, sera containing neutralizing antibodies at a titer of 1:2000 were given once every 24 hours to mice infected with 10 6 pfu of the virus. Several case reports have described the use of serum transfer as a treatment for SFTSV infection. One study found that the administration of serum from survivors to patients with rapidly progressing SFTSV infection had a positive effect on disease outcome. All patients treated with serum transfusion showed lower viremia levels as detected by RT-PCR and 13 of the 14 patients treated with serum transfusion survived 132 . These data suggest that serum transfer may be a viable treatment for patients with rapidly progressing SFTSV infection, but the mechanism of protection requires further study, although it is assumed to be neutralizing antibodies.
Overall, Favipiravir, IFN, and antiserum treatment have the potential to be possible therapies for SFTSV infection. Nonetheless, continued research into viable treatments is required and the development of new therapeutics is needed.
## Vaccines
Much like therapeutics, SFTSV vaccine development has been limited and there are no licensed vaccines currently available. Multiple studies, utilizing various approaches, have shown varying levels of efficacy in preventing SFTSV-associated disease in animal models. Recombinant protein strategies [bib_ref] Immunization with recombinant SFTSV/NSs protein does not promote virus clearance in SFTSV-infected..., Liu [/bib_ref] , DNA vaccine strategies [bib_ref] Nano-patterning of a stainless steel microneedle surface to improve the dip-coating efficiency..., Jung [/bib_ref] [bib_ref] Development of a SFTSV DNA vaccine that confers complete protection against lethal..., Kwak [/bib_ref] [bib_ref] Vaccination with single plasmid DNA encoding IL-12 and antigens of severe fever..., Kang [/bib_ref] , and viral vector strategies [bib_ref] Single dose of a rVSV-based vaccine elicits complete protection against severe fever..., Dong [/bib_ref] have all been utilized and several of these studies are discussed below.
The first study evaluated recombinant SFTSV NSs (100ng) given in Freund's adjuvant (which is not an acceptable adjuvant for N.E. Bopp et al. humans) and was administered twice (14 days apart) to C57BL/6 mice, via the SC route. Mice were challenged with 3 × 10 7 pfu SFTSV via the IP route. Despite high titers of anti-NSs antibodies, no inhibition of viral replication nor accelerated viral clearance were observed in vaccinated mice [bib_ref] Immunization with recombinant SFTSV/NSs protein does not promote virus clearance in SFTSV-infected..., Liu [/bib_ref]. The second study investigated a DNA vaccine candidate. Plasmids encoding the NSs and N genes were transfected into mice and immune responses were measured. Significant differences in the release of TNF-α from CD8 + and CD4+ cells were observed in mice transfected with NSs in comparison to N, and both groups of vaccinated animals had increased immunologic response compared with controls. Protective efficacy of this strategy cannot be assessed as a subsequent challenge with the virus was not reported [bib_ref] Nano-patterning of a stainless steel microneedle surface to improve the dip-coating efficiency..., Jung [/bib_ref].
Two other DNA vaccine studies have shown the promise of this strategy. Kwak et al. [bib_ref] Development of a SFTSV DNA vaccine that confers complete protection against lethal..., Kwak [/bib_ref] developed multiple constructs that varied in their expression of each of the five SFTSV proteins. The glycoproteins of SFTSV were found to confer the highest protection when used in the lethal ferret model, with partial protection induced by the NSs, N, and RdRp vaccines [bib_ref] Development of a SFTSV DNA vaccine that confers complete protection against lethal..., Kwak [/bib_ref]. Researchers found that doses as low as 40 µg of plasmid induced a strong SFTSV-specific response with robust protection against intramuscular challenges as high as 10 7.6 TCID 50 . All of the candidate vaccines induced robust T-cell responses and strong, antigen specific antibodies were developed.
Studies by showed that the inclusion of an IL-12 open reading frame in addition to SFTSV genes helped to induce strong cellular responses [bib_ref] Vaccination with single plasmid DNA encoding IL-12 and antigens of severe fever..., Kang [/bib_ref]. Owing to the stability and ease of production, as well as the efficacy observed in these studies, the use of a DNA-based vaccine may be a promising approach to address the need for a SFTSV vaccine.
Of the candidate vaccines reported to date, the use of the viral vector strategy appears to be the most promising. Two studies have demonstrated the effectiveness of viral vectors; one employing a recombinant vesicular stomatitis virus (rVSV) system and another utilizing a recombinant, rationally attenuated SFTSV. The rVSV system encodes the glycoproteins of SFTSV and was given as a single 1 × 10 4 pfu dose to IFNα/βR −/− mice [bib_ref] Single dose of a rVSV-based vaccine elicits complete protection against severe fever..., Dong [/bib_ref]. Immunization with this recombinant VSV vaccine was performed via multiple routes including IP, IV, SC, and intranasal. All immunized animals were protected from an IP challenge of 2 × 10 4 pfu of the Wuhan strain of SFTSV [bib_ref] Single dose of a rVSV-based vaccine elicits complete protection against severe fever..., Dong [/bib_ref]. In comparison with the unvaccinated control group that all succumbed to SFTSV infection, the immunized animals showed minimal clinical signs of infection.
The study of the rationally attenuated recombinant SFTSV demonstrated that two variants, a NSs knockout and a single amino-acid mutant can convey protection in a lethal ferret model [bib_ref] Cross-genotype protection of live-attenuated vaccine candidate for severe fever with thrombocytopenia syndrome..., Yu [/bib_ref]. Both variants inhibited viral replication after post vaccination challenge and decreased most clinical signs. High genetic stability of the NSs knockout virus was observed even after six passages, demonstrating a low likelihood of reversion [bib_ref] Cross-genotype protection of live-attenuated vaccine candidate for severe fever with thrombocytopenia syndrome..., Yu [/bib_ref].
Work has been undertaken to investigate candidate strains of SFTSV on which to base future vaccines so as to maximize protective coverage. Strain HB29 was found to be cross neutralized by rabbit antibodies generated against eight other SFTSV strains, and with sera from 33 SFTS survivors, suggesting that it is broadly neutralized and may be an effective base for candidate vaccines [bib_ref] Identification of a candidate standard strain of severe fever with thrombocytopenia syndrome..., Jia [/bib_ref]. Other groups have taken these studies further and begun to map the antigenic regions of SFTSV's glycoproteins, including identifying subdomain III of Gn as a target for neutralizing antibodies [bib_ref] Structures of phlebovirus glycoprotein Gn and identification of a neutralizing antibody epitope, Wu [/bib_ref]. Determining antigenic regions that also induce neutralizing antibodies may be key in the development of certain types of vaccines.
Overall, the fields of SFTSV therapeutics and vaccines are in early stages and a significant gap in knowledge remains. Determining the correlate(s) of protection, molecular mechanisms of immune subversion, and the development and characterization of effective therapeutics are crucial to combatting SFTSV infection.
# Conclusions
At present, no licensed vaccines are available for SFTS, and research is in the discovery stage with only three published reports of vaccine candidates. Existing drugs are available that may be efficacious in the treatment of SFTSV infection, though more preclinical data and clinical studies are required. In addition, no specific vector control methodologies have been effective at curbing the expansion of SFTSV competent vectors. Animal models for SFTS appear to recapitulate many facets of mild and/ The mention of specific companies or of certain manufacturers' products does not imply that they are endorsed or recommended by the World Health Organization in preference to others of a similar nature that are not mentioned. Errors and omissions, except the names of proprietary products, are distinguished by initial capital letters.
N.E. Bopp et al.
or severe human disease, however, none of the models mirrors all clinical manifestations. There are major gaps in knowledge with insufficient data available on basic immunologic responses, the correlates of protection, and the determinants of severe disease by SFTSV and related viruses.
Received: 23 October 2019; Accepted: 13 October 2020;
[table] Table 2: Compounds that have been shown to inhibit SFTSV infection a . [/table]
[bib_ref] Persistence and gender differences in protection against severe fever with thrombocytopaenia syndrome..., Qi [/bib_ref] |
IgM Augments Complement Bactericidal Activity with Serum from a Patient with a Novel CD79a Mutation
Antibody replacement therapy for patients with antibody deficiencies contains only IgG. As a result, concurrent IgM and IgA deficiency present in a large proportion of antibody deficient patients persists. Especially patients with IgM deficiency remain at risk for recurrent infections of the gastrointestinal and respiratory tract. The lack of IgM in the current IgG replacement therapy is likely to contribute to the persistence of these mucosal infections because this antibody class is especially important for complement activation on the mucosal surface. We evaluated whether supplementation with IgM increased serum bactericidal capacity in vitro. Serum was collected from a patient with agammaglobulinemia and supplemented with purified serum IgM to normal levels. Antibody and complement deposition on the bacterial surface was determined by multi-color flow cytometry. Bacterial survival in serum was determined by colony-forming unit counts. We present a patient previously diagnosed with agammaglobulinemia due to CD79A (Igα) deficiency revealing a novel pathogenic insertion variant in the CD79a gene (NM_001783.3:c.353_354insT). Despite IgG replacement therapy and antibiotic prophylaxis, this patient developed a Campylobacter jejuni spondylodiscitis of lumbar vertebrae L4-L5. We found that serum IgM significantly contributes to complement activation on the bacterial surface of C. jejuni. Furthermore, supplementation of serum IgM augmented serum bactericidal activity significantly. In conclusion, supplementation of intravenous IgG replacement therapy with IgM may potentially offer greater protection against bacterial infections, also in the context of increasing antibiotic resistance.
# Introduction
Patients with agammaglobulinemia have very low or no serum immunoglobulin levels, making these patients highly susceptible to infections. The largest group of patients has X-linked agammaglobulinemia (XLA), which is a caused by a defect in the Btk gene encoding Bruton Tyrosine Kinsase (Btk), which accounts for 85% of agammaglobulinemia patients. A smaller group has defects in for instance IGLL1, CD79A, CD79B, or BLNK, which have clinical findings that are similar from those seen in patients with mutations in Btk but tend to have a more severe onset of disease [bib_ref] Primary B cell immunodeficiencies: comparisons and contrasts, Conley [/bib_ref].
Treatment relies on the use of IgG replacement therapy (IgGRT), sometimes in conjunction with prophylactic antibiotic use. IgGRT was first introduced in 1952 by Colonel Ogden Bruton and applied to the first patient with agammaglobulinemia. IgGRT is still the main therapy to date and successfully reduces the frequency of invasive bacterial infections in patients with immunoglobulin deficiency [bib_ref] Common variable immunodeficiency disorders: division into distinct clinical phenotypes, Chapel [/bib_ref]. However, despite IgGRT, agammaglobulinemia patients still have increased infection rates and lowered life expectancy [bib_ref] X-linked agammaglobulinaemia: outcomes in the modern era, Shillitoe [/bib_ref]. Currently used immunoglobulin preparations contain only IgG. As a result, concurrent IgM and IgA deficiency present in a large proportion of antibody deficient patients persists. Especially patients with IgM deficiency, such as agammaglobulinemia patients, remain at risk for recurrent infections of the respiratory tract [bib_ref] X-linked agammaglobulinemia: report on a United States registry of 201 patients, Winkelstein [/bib_ref]. Gastrointestinal manifestations also present a significant burden; 35% of XLA patients suffer from complications including recurrent infections and inflammatory bowel disease [bib_ref] Gastrointestinal manifestations in X-linked agammaglobulinemia, Barmettler [/bib_ref]. The lack of especially IgM in the current IgGRT is likely to contribute to the persistence of these mucosal infections because this antibody class is especially important for complement activation on the mucosal surface.
Patients with selective IgM deficiency (sIgMD) represent another model disease where IgM is absent. The immunological and clinical phenotype of sIgMD is very heterogeneous, and patients can even remain asymptomatic [bib_ref] Selective IgM deficiency: clinical and laboratory features of 17 patients and a..., Chovancova [/bib_ref]. In a series of 17 sIgMD patients, recurrent upper respiratory tract infections were observed in five out of six patients with undetectable IgM levels (< 0.05 g/L) [bib_ref] Selective IgM deficiency: clinical and laboratory features of 17 patients and a..., Chovancova [/bib_ref]. Patients with sIgMD can also present with invasive manifestations such as sepsis and meningitis [bib_ref] Primary selective IgM deficiency: an ignored immunodeficiency, Louis [/bib_ref]. Although for most sIgMD patients, IgGRT was not required [bib_ref] Selective IgM deficiency: clinical and laboratory features of 17 patients and a..., Chovancova [/bib_ref] ; it is suggested to be beneficial for patients with recurrent or severe infections [bib_ref] Selective IgM deficiency-an underestimated primary immunodeficiency, Gupta [/bib_ref].
Previous observational work demonstrated an important protective effect of IgM against colonization of the respiratory tract by non-typeable Haemophilus influenzae (NTHi) in patients with hypogammaglobulinemia [bib_ref] Protective effect of IgM against colonization of the respiratory tract by nontypeable..., Micol [/bib_ref]. We showed that especially recognition of NTHi lipooligosaccharide by IgM efficiently increased complement-and neutrophil-mediated killing [bib_ref] Modified lipooligosaccharide structure protects nontypeable Haemophilus influenzae from IgM-mediated complement killing in..., Langereis [/bib_ref] [bib_ref] Shielding of a lipooligosaccharide IgM epitope allows evasion of neutrophil-mediated killing of..., Langereis [/bib_ref] , which might contribute to the protective effect of IgM against colonization [bib_ref] Protective effect of IgM against colonization of the respiratory tract by nontypeable..., Micol [/bib_ref]. We hypothesized that the superior efficacy of complement activation by IgM compared to IgG may explain this observation. This may be especially relevant for protection against mucosal Gram-negative bacteria.
Here, we describe a patient with a CD79A (Igα) deficiency due to compound heterozygous mutations in the CD79a gene. Igα is the signal transduction molecule of the pre-B cell receptor. Absence of Igα results in a block from pro-B cells to pre-B cells in the B cell differentiation pathway, resulting in agammaglobulinemia [bib_ref] Primary B cell immunodeficiencies: comparisons and contrasts, Conley [/bib_ref].
We show that IgM supplementation of serum from a CD79A deficient agammaglobulinemia patient on IgG substitution therapy increases serum bactericidal activity against Campylobacter jejuni in vitro. This finding suggests that IgM containing immunoglobulin preparations may improve protective efficacy compared to the current IgGRT.
# Material and methods
# Ethics statement
The study was approved by the ethics committee of the Radboudumc, Nijmegen, the Netherlands. Written informed consent was obtained from the patient. Collection of blood from healthy volunteers was approved by the ethics committee of the Radboudumc, Nijmegen, the Netherlands. All experiments were carried out in accordance with local guidelines and regulations and comply with the Declaration of Helsinki and the Good Clinical Practice guidelines.
## Serum preparation and storage
Blood was drawn in vacutainer gel serum tubes (BD Biosciences) and kept on ice within 30 min following collection. Serum was prepared by centrifuging blood for 10 min at 4000×g, and serum aliquots were stored at − 80°C.
Serum was diluted in HBSS without phenol red containing Ca 2+ and Mg 2+ + 0.1% gelatin (Hepes3+). For complementinactivation, serum was heat-inactivated (HI) by incubating 20 min at 56°C. Purified IgM from human serum (Sigma-Aldrich, I8260) was washed with PBS on a Amicon Ultra-0.5 Centrifugal Filter Unit column (Millipore) to remove the preservative sodium azide and suspended into PBS at a concentration of 1 mg/mL.
## Bacterial growth conditions and storage
The patient's C. jejuni isolate was used for in vitro functional assays. Bacteria were grown on blood agar plates in a microaerobic environment at 42°C. Bacteria were harvested directly from plate and suspended into PBS containing 16% glycerol for storage.
## Flow cytometry
Bacteria were washed with Hepes3+ and diluted to an OD 620 of 0.1. For IgG, IgM, C3, and C5b9 surface opsonization, 50 μL bacteria were incubated with 50 μL 10% serum in Hepes3+ for 15 min at 37°C. Opsonized bacteria were washed, fixed for 20 min with 2% paraformaldehyde, and incubated with 1:500diluted C3-specific polyclonal antibody (MP Biomedical, Cappel), 1:100-diluted C5b-9-specific monoclonal antibody (Santa Cruz biotechnology), 1:50-diluted Fcγ Fragment Specific PerCP AffiniPure Goat Anti-Human IgG (Jackson Immunoresearch), and 1:500-diluted Fc5μ fragment specific Alexa Fluor® 647 AffiniPure Goat Anti-Human IgM (Jackson Immunoresearch) in PBS with 2% BSA for 15 min at room temperature. Bacteria were washed and incubated with 1:200 PE-labeled goat anti-mouse IgG (ThermoFisher Scientific) in PBS with 2% BSA for 15 min at room temperature. Bacteria were washed and suspended in PBS for flow cytometry using a FACS LSR II instrument (BD Biosciences). Data were analyzed using the software program FlowJo version 10.2. Data is presented as geometric mean fluorescence intensity (MFI) in arbitrary units (AU).
## Bacterial survival assays
Bacteria were washed with Hepes3+ and diluted to a concentration of ∼ 20.000 colony forming units (CFU)/mL. Fifty microliter of bacteria was mixed with 50 μL serum or HIserum diluted in Hepes3+ and incubated 30 min at 37°C. Serial dilutions were plated on blood agar plates and incubated in a candle jar at 42°C. Survival was determined by dividing the CFU counts in serum with the CFU count in HI-serum.
# Results
## Case report campylobacter jejuni spondylodiscitis in a patient with agammaglobulinemia associated with novel cd79a pathogenic insertion variant
Here, we describe a 16-year-old male, who was diagnosed with agammaglobulinemia associated with CD79a deficiency at 2 years of age. Initially, he presented with severe bacterial pneumonia associated with nasopharyngeal carriage of Haemophilus influenzae type B. In addition, he suffered from chronic otitis media with effusion. Initial immunologic assessment demonstrated agammaglobulinemia (IgG < 0.5 g/L, IgA < 0.13 g/L, and IgM < 0.06 g/L). Immunophenotyping of peripheral blood revealed total absence of CD19 + B cells. Bone marrow aspirate showed decreased percentages of CD22+ B cells (20%, normal < 5 years of age 36 ± 7%), with absence of CD79A expression. Sequence analysis confirmed the diagnosis of CD79a deficiency due to compound heterozygous mutations: in one allele, we found the NM_001783.3:c.380-2A > G pathogenic variant as described previously [bib_ref] Mutations in Igalpha (CD79a) result in a complete block in B-cell development, Minegishi [/bib_ref] ; in the second allele, we found a novel insertion in exon 2 of a thymine (c.353insT). This variation creates a frame shift starting at p.(Cys119Leufs*63). The new reading frame ends in a STOP codon 63 positions downstream. This is the sixth patient with CD79a mutations described to date.
After the diagnosis, subcutaneous IgGRT was started and combined with antibiotic prophylaxis. The patient received weekly subcutaneous IgG at a dose of 320 mg/kg/28d (Hizentra; CSL Behring AG, Swiss) (IgG through level 8-10 g/L) and was on prophylactic amoxicillin 500 mg orally twice daily. To facilitate travel during summer holidays, subcutaneous IgGRT was interrupted for maximally 3 weeks at a time and intravenous IgGRT at a dose of 365 mg/kg/28d was administered before traveling. The patient had a mild subsequent clinical course with episodes of gastrointestinal parasitic infections and chronic conjunctivitis.
At age 16, the patient presented with a 3-week history of lower back pain without fever. Laboratory evaluation demonstrated leukocytes of 12.10 9 /L, CRP 37 mg/L, and ESR 15 mm/h; MRI examination showed spondylodiscitis of lumbar vertebrae L4-L5. He was initially started on empiric flucloxacillin and gentamicin. Blood culture was positive for a Campylobacter jejuni strain, which was resistant for tetracycline (8 mg/L), but susceptible to erythromycin/azithromycin (4 mg/L) and ciprofloxacin (0.125 mg/L) according to the EUCAST criteria. The strain was additionally tested for amoxicillin (256 mg/L, resistant) and amoxicillin-clavulanic acid (2 mg/L, susceptible), which indicates the presence of betalactamase activity. The patient was switched to azithromycin 500 mg once daily and treated orally for 9 months in total with good clinical recovery. As preferred by the patient, IgGRT was switched from subcutaneous to intravenous at 500 mg/kg/ 28 days. Additional tests of complement function showed normal classical (CH50 102% (67-149%)) and alternative pathway (AP50 68% (67-133%)) activity. Since patients with agammaglobulinemia receiving IgGRT only have corrected IgG serum levels, we evaluated whether supplementation with serum IgM increased serum bactericidal capacity in vitro.
## Campylobacter jejuni-specific igg at different igg concentrations
Six months previous to presentation with spondylodiscitis, subcutaneous IgGRT was briefly interrupted for summer holidays, while the patient was protected with an equal dose intravenous IgG. Serum samples were taken before and after intravenous IgG administration as part of a research protocol. IgG, IgA, and IgM serum levels were determined, and IgG concentration was 8.08 g/L pre-intravenous IgG, and 14.8 g/L post-intravenous IgG, but both samples were within normal range (5.2-15.6 g/L), whereas IgA and IgM concentrations were below the detection limit in both samples (< 0.07 g/L).
To determine the capacity of the patient's low-and highconcentration IgG serum to opsonize C. jejuni, we incubated the bacteria with 5% low-and high-concentration IgG serum followed by detection of IgG, IgM, complement factor 3 (C3), and membrane attack complex C5b9 opsonization by flow cytometry [fig_ref] Figure 1: Similar recognition of Campylobacter jejuni by IgG low-and high-concentration IgG patient serum [/fig_ref]. Despite the large difference in IgG levels in the low-and high-concentration IgG samples, similar IgG opsonization of the bacterial surface (412 ± 73 and 462 ± 111, respectively) was detected, whereas IgM was absent [fig_ref] Figure 1: Similar recognition of Campylobacter jejuni by IgG low-and high-concentration IgG patient serum [/fig_ref]. As functional read out for complement activation, we determined C3 and C5b-9 deposition on the bacterial surface. In accordance with IgG opsonization, no differences between low-and high-concentration IgG serum for C3 (11.043 ± 2.818 and 10.840 ± 3.183, respectively) or C5b-9 (1.011 ± was determined by flow cytometry. Measurement of geometric mean fluorescence intensity (MFI) was depicted in arbitrary units (AU (n = 6, mean ± standard error of the mean). A two-way paired Student t test was used for statistical analysis. NS not significant 281 and 738 ± 214, respectively) were detected [fig_ref] Figure 1: Similar recognition of Campylobacter jejuni by IgG low-and high-concentration IgG patient serum [/fig_ref]. These data indicated that the patient's C. jejuni strain was recognized by IgG present in low-and high-concentration IgG serum, which was able to activate complement. Based on these data, we concluded that bactericidal activity of the low-and high-concentration IgG serum was similar.
## Augmented complement activation and serum killing of campylobacter jejuni with serum igm supplementation
We determined whether IgM supplementation would improve complement activation and bactericidal activity against C. jejuni. For these experiments, a pool of low-and highconcentration IgG serum was used as patient serum and serum from four healthy individuals was included as normal reference. We measured IgG and IgM opsonization of the bacterial surface with 5% patient serum with or without IgM supplementation (5 mg/dL IgM in 5% serum, equivalent for 100 mg/dL IgM in normal serum). Opsonization with IgG was similar for 5% patient serum with or without IgM (369 ± 56 and 479 ± 92, respectively), which was within the normal range found for four healthy individuals (363 ± 70, range 208-521) . Opsonization with IgM was absent with patient serum and clearly present with patient serum supplemented with IgM (1.952 ± 502), which reached the normal range found for four healthy individuals (1.908 ± 48, range 1.803-1.993) . Although IgG in serum alone was able to activate complement as determined by C3 deposition on the bacterial surface, supplementation of IgM augmented C3 opsonization significantly (7.317 ± 2.805 to 24.295 ± 1.521), which reached levels close the levels found for four healthy individuals (27.877 ± 1.665, range 25.263-32.551) . Similarly, increased C5b9 deposition in the presence of IgM was observed (606 ± 123 to 3.736 ± 421), which was well within to the normal levels found for four healthy individuals (3.509 ± 347, range 2.612-4.133) . This increased complement activation was also represented in increased serum bactericidal activity because serum survival of C. jejuni in 2, 5, or 10% serum was significantly decreased with IgM supplementation .
# Discussion
Immunoglobulin replacement therapy was introduced by Colonel Ogden Bruton and applied to the first patients with agammaglobulinemia in 1952and remains the main therapy to date. Despite immunoglobulin therapy, agammaglobulinemia patients have increased infection rate and lowered life expectancy [bib_ref] X-linked agammaglobulinaemia: outcomes in the modern era, Shillitoe [/bib_ref]. There have been no major advancements aside from subcutaneous immunoglobulin therapy in the early 1980s [bib_ref] Subcutaneous immunoglobulin home treatment in hypogammaglobulinaemia, Leahy [/bib_ref]. Recent advances in therapy for patients with agammaglobulinemia are mostly found in gene therapy [bib_ref] Splice-correcting oligonucleotides restore BTK function in X-linked agammaglobulinemia model, Bestas [/bib_ref] [bib_ref] B cell-specific lentiviral gene therapy leads to sustained B-cell functional recovery in..., Kerns [/bib_ref]. Although these methods show great potential, these are still far from being applied. Therefore, limiting infections and thereby increasing patient's quality of life by supplementing IgM to IVIG would be an attractive alternative.
Campylobacter infections are frequently observed in patients with hypogammaglobulinemia, mainly causing gastrointestinal infections, despite IgGRT [bib_ref] Host-pathogen interactions in campylobacter infections: the host perspective, Janssen [/bib_ref]. Besides frequent mucosal infections, Campylobacter species are also seen to cause bacteremia [bib_ref] Recurrent campylobacter lari bacteremia in X-linked agammaglobulinemia: a case report and review, Jirapongsananuruk [/bib_ref] [bib_ref] Erysipelas-like skin lesions associated with cam, Kerstens [/bib_ref] [bib_ref] Relapsing campylobacter jejuni systemic infections in a child with X-linked agammaglobulinemia, Ariganello [/bib_ref] and atypical infections in patient with agammaglobulinemia, including erysipelas-like skin lesions [bib_ref] Erysipelas-like skin lesions associated with cam, Kerstens [/bib_ref] , arthritis [bib_ref] Campylobacter jejuni arthritis in secondary amyloidosis, Simon [/bib_ref] , pericarditis [bib_ref] An unusual case of campylobacter jejuni pericarditis in a patient with X-linked..., Rafi [/bib_ref] , cellulitis [bib_ref] Campylobacter jejuni cellulitis in a patient with panhypogammaglobulinaemia, Hopkins [/bib_ref] , and spondylodiscitis (current report). Secretory IgA may act as an important local immune defense; however, there is no significant increase of Campylobacter infections in patients with selective IgA deficiency [bib_ref] Host-pathogen interactions in campylobacter infections: the host perspective, Janssen [/bib_ref]. Therefore, concurrent IgM deficiency might decrease risk for bacterial infections because this class of antibody is especially good in complement fixation on bacterial pathogens.
Here, we show augmented serum bactericidal activity mediated by human serum IgM, which might increase protection against bacterial infections. Though C3 opsonization on bacteria may partly be explained by alternative complement pathway activation, our data indicate that IgM-mediated classical complement pathway activation is an important determinant. Our results support the use of IgM-enriched IgGRT in patients with IgG deficiency and concurrent IgM deficiency. For example, Borleffs and colleagues successfully treated two hypogammaglobulinemia patients with persistent gastro enteric C. jejuni infections [bib_ref] Use of an immunoglobulin M containing preparation for treatment of two hypogammaglobulinemic..., Borleffs [/bib_ref]. Six dosages of Pentaglobin, an IgMenriched intravenous IgG preparation containing 12% IgM and 12% IgA, at 3-week intervals increased bactericidal activity against C. jejuni, was well tolerated, and there were no cultureconfirmed relapses of C. jejuni infections [bib_ref] Use of an immunoglobulin M containing preparation for treatment of two hypogammaglobulinemic..., Borleffs [/bib_ref]. Pentaglobin was shown to contain both IgG and IgM antibodies recognizing C. jejuni outer membrane proteins. In addition, two patients with relapsing Campylobacter jejuni infections were given fresh frozen plasma, which contains IgA and IgM, for 2 or 4 weeks (500 mL twice weekly) and resulted in cure in both patients [bib_ref] Erysipelas-like skin lesions associated with cam, Kerstens [/bib_ref]. Plasma infusion resulted in detectable serum IgA levels in two patients and detectable serum IgM levels in one patient [bib_ref] Erysipelas-like skin lesions associated with cam, Kerstens [/bib_ref].
The use of IgM-enriched intravenous IgG preparation Pentaglobin has been tested in different patient groups with variable results. Pentaglobin was shown to significantly reduce mortality in patients with sepsis [bib_ref] Cost-effectiveness of immunoglobulin M-enriched immunoglobulin (Pentaglobin) in the treatment of severe sepsis..., Neilson [/bib_ref] [bib_ref] Relationship between the timing of administration of IgM and IgA enriched immunoglobulins..., Berlot [/bib_ref] , but this was not corroborated by another study [bib_ref] The effects of IgM-enriched immunoglobulin preparations in patients with severe sepsis, Tugrul [/bib_ref] or for critical illness polyneuropathy and/or myopathy patients with multi organ failure and sepsis [bib_ref] Early treatment with IgM-enriched intravenous immunoglobulin does not mitigate critical illness polyneuropathy..., Brunner [/bib_ref]. Prophylactic use of IgM-enriched intravenous IgG preparation Pentaglobin has been tested in adults and pediatric patients after hematopoietic stem cell transplantation. In adults, a reduction in risk of dying of an infection after bone marrow transplant was observed [bib_ref] Use of IgM enriched intravenous immunoglobulin (Pentaglobin) in bone marrow transplantation, Poynton [/bib_ref]. However, in pediatric patients, no significant beneficial effects were observed [bib_ref] Comparison of prophylactic use of intravenous immunoglobulin versus Pentaglobin(R) in pediatric patients..., Azik [/bib_ref]. A 2013 Cochrane database review concluded that the trials were too small and the totality of the evidence is still insufficient to support robust conclusions [bib_ref] Intravenous immunoglobulin for treating sepsis, severe sepsis and septic shock, Alejandria [/bib_ref].
A beneficial effect of Pentoglobin in comparison to conventional intravenous IgGRT is its effective reduction of endotoxin [bib_ref] IgM-enriched immunoglobulin attenuates systemic endotoxin activity in early severe sepsis: a before-after..., Wand [/bib_ref] [bib_ref] Endotoxin concentration in neutropenic patients with suspected gram-negative sepsis: correlation with clinical..., Behre [/bib_ref]. Direct effects on bacterial killing are also observed. Pentaglobin had a greater opsonic activity against Pseudomonas aeruginosa, Klebsiella pneumoniae, and Escherichia coli, in comparison to intravenous IgG Sandaglobulin [bib_ref] Opsonic activity of a new intravenous immunoglobulin preparation: Pentaglobin compared with sandoglobulin, Garbett [/bib_ref]. Pentaglobin showed augmented in vitro killing of Staphylococcus aureus, E. coli, and P. aeruginosa, but slightly reduced killing K. pneumoniae and Enterococcus faecium, in comparison to intravenous IgG Intratect [bib_ref] In vitro and in vivo activity of hyperimmune globulin preparations against multiresistant..., Rossmann [/bib_ref]. In a S. aureus mouse sepsis model, the number of bacteria in liver and kidney was significantly lower for animals receiving IgMenriched IVIG Pentaglobin in comparison to animals receiving IVIG Intratect [bib_ref] In vitro and in vivo activity of hyperimmune globulin preparations against multiresistant..., Rossmann [/bib_ref].
Recent advances in IgM-enriched immunoglobulin therapy, Trimodulin (BT-588, predecessor BT-086) containing more IgM (23%), in comparison to Pentaglobin (12%) are promising [bib_ref] IgM-enriched solution BT086 improves host defense capacity and energy store preservation in..., Shmygalev [/bib_ref] [bib_ref] A double blind randomized experimental study on the use of IgM-enriched polyclonal..., Vaschetto [/bib_ref] , and a clinical trial in patients with severe pneumonia is underway [bib_ref] Concept for a study design in patients with severe community-acquired pneumonia: a..., Welte [/bib_ref]. Since IgM concentration present in Trimodulin is higher, we expect a strong additive effect on Bacteria were incubated with 5% patient serum with or without supplementation of IgM or 5% serum from healthy controls. Opsonization with IgG (a), IgM (b), C3 (c), and C5b9 (d) was determined by flow cytometry. Measurement of geometric mean fluorescence intensity (MFI) was depicted in arbitrary units (AU (n = 6, mean ± standard error of the mean). A one-way ANOVA with Dunnett's Multiple Comparison Test was used for statistical analysis. NS not significant; **P < 0.01; ***P < 0.001. (e) Bacterial survival in 2, 5, or 10% patient serum with or without supplementation of IgM was determined after 30 min (n = 4, mean ± standard error of the mean). A two-way ANOVA with Bonferroni post hoc test was used for statistical analysis. ***P < 0.001 bacterial killing. It remains to be determined whether subcutaneous administration of IgM containing products offers sufficient bioavailability.
In conclusion, we show that supplementation of human serum IgM significantly contributes to complement activation on the bacterial surface and increases serum bactericidal activity. Therefore, supplementation of IgGRT with IgM may potentially offer greater protection against bacterial infections, which is also relevant in the context of rising antibiotic resistance. This hypothesis needs to be further tested in clinical studies.
[fig] Figure 1: Similar recognition of Campylobacter jejuni by IgG low-and high-concentration IgG patient serum. Bacteria were incubated with 5% low-and high-concentration IgG patient serum, stained for IgG, IgM, C3, and C5b9 and subjected to flow cytometry. a Flow cytometry gating strategy. Opsonization with IgG (b), IgM (c), C3 (d), and C5b9 (e) [/fig]
|
Nitric oxide and mitochondria in metabolic syndrome
Metabolic syndrome (MS) is a cluster of metabolic disorders that collectively increase the risk of cardiovascular disease. Nitric oxide (NO) plays a crucial role in the pathogeneses of MS components and is involved in different mitochondrial signaling pathways that control respiration and apoptosis. The present review summarizes the recent information regarding the interrelations of mitochondria and NO in MS. Changes in the activities of different NO synthase isoforms lead to the formation of metabolic disorders and therefore are highlighted here. Reduced endothelial NOS activity and NO bioavailability, as the main factors underlying the endothelial dysfunction that occurs in MS, are discussed in this review in relation to mitochondrial dysfunction. We also focus on potential therapeutic strategies involving NO signaling pathways that can be used to treat patients with metabolic disorders associated with mitochondrial dysfunction. The article may help researchers develop new approaches for the diagnosis, prevention and treatment of MS.Keywords: metabolic syndrome, mitochondrial dysfunction, nitric oxide synthase, nitric oxide, nitric oxide synthase type I, nitric oxide synthase type II, oxide synthase type III, mitochondrial nitric oxide synthase www.frontiersin.org
# Introduction
Metabolic syndrome (MS) is a global epidemic, and different environmental, genetic and pathophysiological factors are involved in its development. Mitochondria play central roles in energy metabolism, signaling and apoptosis and its alterations may contribute to the development of metabolic disorders. It has been recently found that mitochondrial biogenesis and function are enhanced by nitric oxide (NO), which is a key signaling molecule in vascular homeostasis. This finding indicates that changes in the production of NO bioavailability and MS may be associated with the mitochondria. However, the link between NO signaling components and mitochondria in MS in different tissues is still not clear.
Today, mitochondria remains an attractive target for the prevention and therapy of MS and its complications. Understanding the mechanisms that lead to decreased mitochondrial activity in insulin-sensitive tissues is important for developing therapies to reverse insulin resistance. A deeper understanding of the role of NO pathway components in this process opens up new opportunities for applied research in this area.
## Metabolic syndrome and mitochondrial dysfunction
Mitochondrial dysfunction is closely associated with obesity, metabolic syndrome and type 2 diabetes mellitus (T2DM). Changes in the mitochondrial membrane potential, a reduction in the ATP level, the inhibition of mitochondrial oxygen consumption and reduced mitochondrial biogenesisare some of the characteristics of mitochondrial dysfunction. The underlying mechanism of mitochondrial dysfunction is very complex, which includes genetic factors from both nuclear and mitochondrial genome and numerous environmental factors. In addition to these factors, excessive formation of reactive oxygen species (ROS) in obesity and T2DMcontributes to mitochondrial dysfunction, while at physiological levels ROS participate in intracellular signaling and regulation as "redox messengers." Particularly, superoxide anion generated by the mitochondria, particularly by complexes I and III of the electron transport chain (ETC), is the precursor of most ROS and a mediator in oxidative chain reactions. Dismutation of superoxide produces hydrogen peroxide, which in turn may be partially reduced to hydroxyl radical causing more damage to various mitochondrial and cellular components. Free radical damage to mitochondria may lead to decreased affinity of mitochondrial proteins for substrates or coenzymes.
The centerpiece of the pathophysiologic mechanism of MS is insulin resistance. The interplay between mitochondrial dysfunction and insulin resistance was first discovered in 1963. Kitt Falk Petersen et al. demonstrated that insulin resistance, which frequently occurs in elderly individuals, is characterized by a decrease in mitochondrial ATP production, a reduction in mitochondrial RNA, decreased activity of oxidative phosphorylation complexes. It is still unclear whether mitochondrial dysfunction results from or causes insulin resistance. One of the key pathophysiological factors underlying the formation of insulin resistance may be the dysregulation of energy metabolism in insulin-sensitive tissues such as skeletal muscle, liver, and adipose . Particularly, in skeletal muscle, decreased mitochondrial respiration capacity, reduced ATP production, and increased ROS levels lead to reduced fatty acid oxidation and increased cytosolic free fatty acid levels, resulting in insulin resistance and T2DM. Tg (sm/p22phox) mice, which produce increased vascular ROS, display impaired spontaneous activity and increased mitochondrial ROS production and mitochondrial dysfunction in their skeletal muscle. As regards liver, there are findings that strongly support that non-alcoholic fatty liver disease can be considered as the hepatic representation of MS. Insulin resistance is a key pathogenic factor in both pathologic states. It has been shown that patients with non-alcoholic fatty liver disease had decreased activities of ETC complexes in liver. Recent findings suggests that reduced SIRT3 activity in fatty liver may result in hyperacetylation of mitochondrial proteins that contributes to mitochondrial dysfunction. Nitrosative stress in nonalcoholic fatty liver disease also contributes to the development of mitochondrial dysfunction that is mediated by the oxidative modifications of mitochondrial DNA (mtDNA), lipids and proteins. Moreover, high-fat diet-induced mtDNA damage correlates with increased oxidative stress in skeletal muscle and liver, which is associated with the induction of markers of endoplasmic reticulum stress, protein degradation and apoptosis. In adipose tissue, mitochondria provide key intermediates for the synthesis of triglycerides and are critical for lipogenesis. Adipose mitochondria are also important for lipolysis through the oxidation of fatty acids, which constitutes an important source of ATP to supply cells with energy.The study of M. Rossmeisl et al. demonstrated that uncoupling of oxidative phosphorylation depresses fatty acid synthesis in white fat. The authors show that reduction of adiposity via mitochondrial uncoupling in white fat not only reflects increased energy expenditure, but also decreased in situ lipogenesis. Therefore, various studies suggest that modulation of mitochondrial function (including oxidative phosphorylation, ATP synthesis and ROS generation) in abovementioned tissues may in turn may affect the development of insulin resistance and obesity.
Mitochondrial dysfunction and insulin resistance promote chronic inflammation and the formation of atherosclerotic lesions, and the activation of the inflammatory Toll-like receptor 2/nuclear factor-κB signaling pathway in monocytes is a central mechanism leading to a vicious cycle between chronic inflammation and mitochondrial ROS production. Tumor necrosis factor (TNF)-α is a key proinflammatory mediator that plays an important role in the chronic inflammation of adipose tissue in obese subjects. It has been concluded that TNF-α also induces mitochondrial dysfunction by reducing complex III activity in the ETC, increasing ROS production, and causing damage to mtDNA in cardiac myocytes. It has been demonstrated in mice with adipocytes deficient in mitochondrial transcription factor A that isolated mitochondrial dysfunction in adipose tissue can cause lipodystrophy and lead to MS and chronic inflammation. Thus, mitochondrial alterations that occur during insulin resistance trigger a cascade of pathological processes in different tissues and form a vicious cycle between chronic inflammation and mitochondrial oxidative stress that eventually contributes to cardiovascular pathologies and T2DM.
## Mitochondrial dna as a molecular marker of metabolic syndrome
MtDNA is a 16569-bp, circular, double-stranded molecule that harbors 37 genes involved in the energy production that takes place in the ETC. This gene set includes 13 structural genes that code for subunits of oxidative phosphorylation complexes as well as genes encoding 22 tRNAs and 2 rRNAs that are involved in the protein synthesis that occurs directly within mitochondria. According to current estimates, the human genome incorporates approximately 2500 mitochondrial proteinencoding genes; however, only fewer than 1000 of these proteins have been identified in isolated mitochondria.
It is now established that mtDNA polymorphisms may be risk factors for T2DM and other metabolic pathologies. The substitution 16189T>C in the D-loop region of mtDNA has been shown to be associated with insulin resistance, obesity and T2DM in both Europeanand Asianpopulations. Analysis of mtDNA has also shown that haplogroup N9A may be a protective factor against MS in Japanese womenand against T2DM in Japanese and Korean individuals . Furthermore, mutations in mtDNA have been independently associated with multiple phenotypes of MS. The T4291C mutation in mtDNA has been shown to be associated with MS in a large Caucasian family, and a common mtDNA variant, T16189C, has been shown to be associated with a lower body mass index at birth, insulin resistance, dilated cardiomyopathy and an increased susceptibility to MS in a Chinese population. Moreover, the mtDNA mutation A8890G has been found in a patient with juvenile-onset MS.
Changes in the amount of mtDNA may be a consequence of genetic defectsand hormonal imbalances in non-hereditary diseases. Our results showing a reduced number of mtDNA copies in peripheral blood and samples of adipose tissue obtained from different locations from patients with MSare consistent with the results of previous studies. It has been previously shown that decreased mtDNA content in peripheral blood occurs before the onset of T2DM. The administration of pioglitazone, which is a mainstay drug for the treatment of T2DM, leads to an increase in the number of mtDNA copies in subcutaneous adipocytes in vitro. As previously mentioned, mtDNA levels may vary in different organs. However, experiments performed in rats have shown that changes in mtDNA copy number in peripheral blood leukocytes reflect similar processes in muscle tissue and hepatocytes. Thus, we can assume that the level of mtDNA in human peripheral blood cells is an indicator of various metabolic disorders. Moreover, the quantitative estimation of mtDNA in various biological
## Frontiers in physiology | mitochondrial research
February 2015 | Volume 6 | Article 20 | 2 samples might become a tool for the prognosis and estimation of treatment efficiency of MS.
## Interconnections between mitochondria and nitric oxide production in different tissues
NO is a relatively stable gas that belongs to a family of gas transmitters with similar inherent intracellular effects. NO is a regulator of many physiological processes. It is synthesized by vascular endothelial cells, is responsible for vasodilatation and is involved in various processes in the nervous, reproductive and immune systems. NO prevents platelet aggregation and the adhesion of neutrophils to the endothelium, and it has cytostatic and cytotoxic properties. In addition to its participation in the regulation of vascular smooth muscle tone, NO directly affects mitochondrial respiration. Its inhibition of ETC enzymes may also be one of the causes of the decrease in oxygen consumption by cardiomyocytes. NO plays important roles in the development of MS components, such as insulin resistance, endothelial dysfunction, hypertriglyceridemia and chronic adipose tissue inflammation. NO is formed from L-arginine, when in the presence of oxygen, by isoenzymes of nitric oxide synthases (NOS), which depend on the cell type. Endothelial NOS (eNOS) and neuronal NOS (nNOS) are constantly active and produce small amounts of NO in response to stimuli that induce an increase in the intracellular Ca 2+ concentration. Inducible NOS (iNOS), which is a calcium-independent form of NOS, is localized to macrophages, neutrophils, and micro-and astroglial cells and is expressed under the influence of factors such as cytokines and bacterial lipopolysaccharide. The mitochondrial production of NO is catalyzed by mitochondrial NOS (mtNOS), which has been identified as the alpha isoform of nNOS that is acylated at a Thr or Ser residue and phosphorylated at the C-terminus. Impaired NOS activity is closely associated with insulin resistance. The work of Nisoli's group has firmly established the function of NO generated by eNOS and nNOS as an upstream regulator of mitochondrial biogenesis. Experiments with homozygous eNOS knockout mice have definitively proven the relationship between NO and insulin sensitivity because these mice show increased blood pressure and insulin resistance.
Endothelial dysfunction, which is defined by a decrease in flow-mediated vasodilatation and vascular insulin resistance, represents an established vascular abnormality in diabetic patients. The possible underlying mechanisms include reductions in the levels of eNOS mRNA expression and protein production. In particular, Alvarado-Vásquez et al. have shown that human umbilical vein endothelial cells isolated from healthy newborns with a strong family history of T2DM exhibit abnormal intracellular synthesis of NO and impaired expression of the eNOS, GLUT1 and p53 genes, which are all associated with NO synthesis. However, studies in animal models and in humans have assumed that T2DM and atherosclerosis are not necessarily associated with reductions in total eNOS. Several mechanisms leading to endothelial dysfunction, such as a lack of enzymatic cofactors for eNOS, the inactivation of NO by the reaction of superoxide (O2-) resulting in peroxynitrite anion (OONO-) formation, the influence of superoxide NADPH oxidaseand "uncoupled" eNOS, have been demonstrated . Peroxynitrite, in turn, has been shown to uncouple endothelial nitric oxide synthase (eNOS), thereby converting an antiatherosclerotic NO-producing enzyme into an enzyme that may accelerate the atherosclerotic process by producing superoxide. It should be noted that these mechanisms are not mutually exclusive and can occur simultaneously. The studies of have demonstrated that the p47phox and NOXO1-dependent activation of NOX1, but not of NOX2, NOX4 or mitochondria, mediates the diabetic uncoupling of eNOS. Interestingly, NOX1-null mice are protected from diabetic endothelial dysfunction, and mice lacking components of NOX2 are protected against hypertension and show reduced atherosclerotic lesion formation when crossed with the apoE-/-background.
ROS derived from mitochondria induce the infiltration and activation of inflammatory cells and increase the apoptosis of endothelial and vascular smooth muscle cells. Excessive formation of peroxynitrite, which is a reactive nitrogen species, has been demonstrated to accelerate the atherosclerotic process by causing direct structural damage and further ROS production. Peroxynitrite also damages DNA and thus triggers the activation of DNA repair systems. A DNA nick-sensor enzyme, poly(ADP-ribose) polymerase-1 (PARP-1), also becomes activated upon sensing DNA breakages. Activated PARP-1 cleaves NAD(+) into nicotinamide and ADP-ribose and polymerizes the latter onto nuclear acceptor proteins. The peroxynitrite-induced overactivation of PARP leads to the consumption of NAD(+) and consequently, ATP, culminating in cell dysfunction, apoptosis or necrosis. This mechanism of cell death has been implicated in the molecular aspects of T2DM-associated cardiovascular disease.
Recent data have indicated the possibility of a significant role for interrelations between PARP-1 and iNOS in human atherosclerotic lesions. The induction of PARP-1 within the nuclei of lesional smooth muscle cells has been correlated with alterations in mitochondrial morphology. Therefore, it should be concluded that plaque formation is one of the consequences of the PARP-1-mediated mechanism that is involved in changes in mitochondrial morphology.
It has been reported that deletion of eNOS or the pharmacological inhibition of NOS impacts baseline and exercisestimulated vascular mitochondrial biogenesis and dynamics. In particular, in cultured human aortic endothelial cells, a similar loss of mitochondrial networks and increased expression of mitochondrial fission 1 protein (Fis1) and dynamin-related protein-1 (Drp1), which are proteins required for mitochondrial fission, have been demonstrated. The silencing of Fis1 or Drp1 expression by siRNA reduces high glucoseinduced alterations in mitochondrial networks, ROS production, eNOS activation, and cGMP production. The increased expression of Fis1 and Drp1 and decreased expression of fusion proteins (Opa1, Mfn1, and Mfn2) in addition to FIGURE 1 | Interrelation between nitric oxide and mitochondria in the pathophysiology of metabolic syndrome (a, in the vascular wall; b, in an adipocyte; c, in a hepatocyte). a: Insulin resistance contributes to decreased activities of signaling molecules, which leads to decreased eNOS activity. Superoxide generated in mitochondria can modestly, and peroxynitrite can strongly, oxidize BH4-BH2, leading to uncoupling of eNOS, increased production of ROS and endothelial dysfunction. Moreover, mtNOS becomes "uncoupled"; it switches from a NO-generating enzyme to a superoxide-producing enzyme. Increased production of peroxynitrite and superoxide anions generated from uncoupled mtNOS and eNOS damage mitochondrial ETC complexes, leading to mitochondrial dysfunction. eNOS then generates superoxide rather than NO, which contributes to vascular oxidative stress and further reduces NO bioavailability. Excessive mtNOS-derived NO can produce reactive nitrogen species, such as peroxynitrite, which may cause tyrosine nitration of mitochondrial components and may play a key role in apoptosis. Under the conditions of oxidative stress, the enhanced activity of iNOS, with less contribution of nNOS, in vascular smooth muscle cells induces the development of chronic metabolic inflammation. decreases in cAMP response element binding protein and PGC-1α have also been observed in aortic tissues of eNOS-null rodents. These data suggest that NOS is a biologically important modulator of basal vascular mitochondrial function and structure.
Enhanced iNOS gene expression and production in different tissues, especially in insulin-sensitive tissues, leads to peroxynitrite formation, causing nitrosative and oxidative stresses. Recent studies have shown the involvement of TNF-α-induced iNOS in the expression of mitochondrial protein uncoupling protein-2 (UCP-2) in adipocytesas well as in NO bioavailability in adipose tissue. It should be concluded that plaque formation is one of the consequences of the PARP-1-mediated mechanism involved in changes in mitochondrial morphology.
Since it was first reported in 1995, the existence of mtNOS has been controversial. It has also been reported that the addition of calcium to mitochondria isolated from hepatocytes stimulates mtNOS, causing peroxynitrite production and cytochrome c release, which are associated with the peroxidation of mitochondrial lipids . Mitochondrial lipid peroxidation may be an important event in mitochondria-mediated apoptosis, but there is no direct evidence that lipid peroxidation mediates NOinduced apoptosis. The NO-synthesizing capacity of mtNOS is higher than all of the cytoplasmic NOS isoforms combined. The discovery of Ca-dependent mtNOS was an outstanding achievement that elucidated late aging-related metabolic alterations that had taken years to become apparent. The participation of mtNOS appears to be of primary importance but remains to be fully elucidated.
## Potential therapeutic targets affecting mitochondrial dysfunction and the nitric oxide pathway in metabolic syndrome
Different research groups have studied various ways to regulate NO signaling pathways to modulate mitochondrial dysfunction. NO signaling in mitochondria underlies many of the metabolic effects of NO, and at low physiologic levels, links the cellular energy demand with the mitochondrial energy supply, while beneficially affecting mitochondrial oxidative stress and calcium handling. Increasing the mitochondrial concentrations of antioxidant drugs by the selective targeting of mitochondria should represent a practical approach for the treatment of a wide range of human diseases. The stimulation of cultured cardiac myocytes with TNF-α or angiotensin II increases ROS generation and myocyte hypertrophy, and treatment with antioxidants inhibits both of these effects. Mitochondriatargeted antioxidants have been developed as pharmaceuticals and have been shown to be safe and effective in phase II clinical trials. Despite these facts and a number of pre-clinical and clinical lines of evidence, the results of studies assessing the effects of classical antioxidants, such as vitamin C, vitamin E, or folic acid in combination with vitamin E, have been disappointing.
Medications used to treat T2DM focus on reducing blood glucose levels. Oral antihyperglycemic medications are used in T2DM to increase organ sensitivity, increase insulin secretion, or decrease glucose absorption during digestion. Insulin and metformin represent the first-line treatments for T2DM, and their effects on mitochondrial function may contribute to the mechanism underlying dysfunction in T2DM. Different hypoglycemic agents, and exercise may reduce oxidative stress in diabetic patients through reducing NOX activity or improving mitochondrial function, which may prevent or postpone the development of cardiovascular complications. Moreover, substances such as statins, angiotensinconverting enzyme inhibitors, or AT1-receptor blockers, which possess indirect antioxidant properties mediated by the stimulation of NO production and by the simultaneous inhibition of superoxide production, have been shown to improve vascular function in pre-clinical and clinical studies and to reduce the incidence of cardiovascular events in patients with cardiovascular disease. These drugs and the NO-releasing β-blocker nebivolol have been shown to possess combined antioxidant properties. Evidence is emerging in support of improved NO bioavailability as a result of exercise training due to its enhanced synthesis and to a reduction in its oxidative stress-mediated destruction. Molecular targets sensitive to the effects of exercise www.frontiersin.org
February 2015 | Volume 6 | Article 20 | 5 FIGURE 3 | Interrelation between nitric oxide and mitochondria in hepatocytes in metabolic syndrome. The enhanced activities of mtNOS and iNOS in hepatocytes during insulin resistance promote peroxynitrite production and cytochrome c release, which are associated with the peroxidation of mitochondrial lipids and apoptosis. Reduced SIRT3 activity contributes to mitochondrial dysfunction, which in turn promotes oxidative stress and chronic inflammation in MS. Taken together, the reviewed studies suggest that NO and the different NOS isoforms play an important role in critical pathological states such as endothelial dysfunction and peroxynitrite-induced cytotoxicity.
training include eNOS and the antioxidant enzyme superoxide dismutase. However, many fundamental details of the cellular and molecular mechanisms linking exercise to alterations in molecular and functional endothelial phenotypes remain to be discovered. Thus, the ideal approach to antioxidant therapy requires the stimulation of NO production and the simultaneous inhibition of vascular superoxide production. Targeting NOS uncoupling to reduce oxidative stress is also possible in humans through the supplementation of the NOS cofactor tetrahydrobiopterin (BH4) or of its precursor sepiapterin. BH4 has been shown to improve endothelial dysfunction in smokers, diabetic subjectsand in patients with hypertension, hypercholesterolemia or coronary artery disease. Folate, which has mild antioxidant properties, can reverse NOS uncoupling, possibly via BH4-dependent mechanisms. Despite these pre-clinical data, the effects of BH4 and related molecules on cardiovascular morbidity and mortality remain unknown, although the possibility of an improved cardiovascular prognosis due to folate supplementation has been definitively ruled out by several randomized controlled trials.
Instead of preventing NO degradation through reducing oxidative stress, it may be possible to provide exogenous NO via administering NO-donating substances, such as organic nitrates. However, research has shown that these mediators cause endothelial dysfunction and increase oxidative stress through at least three different mechanisms, including uncoupling of the ETC, activation of ROS-producing enzymes (including NOS uncoupling), and the direct reaction of nitrate-derived NO with vascular superoxide to form ONOO 2. In turn, these mechanisms accelerate, rather than inhibit, oxidative stress.
Recent evidence has suggested that coenzyme Q10 supplementation may be useful to treat obesity, oxidative stress and the inflammatory process in MS patients because Co-Q10 is an essential component of the ETC. The anti-inflammatory response and lipid-metabolizing effects of coenzyme Q10 are likely mediated by transcriptional regulation of inflammation and lipid metabolism. Curcumin also represents a potential antioxidant and anti-inflammatory agent. Studies have demonstrated that curcumin treatment increases oxygen consumption and significantly decreases lipid and protein oxidation levels in liver mitochondria isolated from high-fat diet-induced obese mice compared with untreated obese mice. Furthermore, curcumin treatment does not cause body weight gain or affect mitochondrial NO synthesis. These findings suggest that obesity induces oxidative stress and mitochondrial dysfunction, whereas curcumin may confer protection against these effects.
Ecdysterone treatment normalizes the levels of stable NO metabolites in addition to the activities of superoxide, iNOS and nNOS in mitochondria in the heart of rats with streptozotocininduced diabetes. These results suggest that ecdysterone treatment attenuates diabetes-induced mitochondrial alterations and protects against oxidative and nitrosative stresses. A study of insulin-resistant male Zucker obese rats by shown that NADPH oxidase inhibitors (apocynin and VAS2870) and BH4 fail to modify bradykinin-induced vasodilatation. This group has shown that superoxide generation is elevated in vessels of insulin-resistant, morbidly obese subjects and reduced by the mitochondrial-targeted superoxide dismutase mimetic mito-TEMPO. Vasodilatation has been shown to improve in insulin-resistant morbidly obese arteries as a result of the activity of the superoxide scavengers, superoxide dismutase and mito-TEMPO.
The blockade of TNF-α with infliximab, but not the inhibition of iNOS or cyclooxygenase, improves endothelial relaxation and decreases superoxide formation.
Endoplasmic reticulum stress has been implicated in the mitochondrial dysfunction-induced apoptosis of pancreatic beta cells. Furthermore, mitochondrial dysfunction increases the expression of the iNOS gene and the production of NO, but NO production is inhibited by compound C or by the dominant-negative mutant form of AMP-activated protein kinase (AMPK) K45R. Moreover, treatment with 1400W, an inhibitor of iNOS, prevents endoplasmic reticulum stress and apoptosis induced by mitochondrial dysfunction. Interestingly, the NO donor S-nitroso N-acetylpenicillamine, the nonspecific NOS inhibitor NG-nitro-L-arginine methyl ester, and the specific nNOS inhibitor 1-(2-trifluoromethylphenyl) imidazole significantly reduce creatine kinase leakage, cell necrosis, and apoptosis in diabetic myocardium. These results demonstrate that both the provision of exogenous NO and the suppression of endogenous NO production result in the potent protection of diabetic human myocardium, overcoming the unresponsiveness of these tissues to cardioprotective therapies. The treatment of cells with the NG-nitro-L-arginine methyl ester (1 mmol/l) has been shown to significantly diminish the TNF-α-mediated sustained downregulation of UCP-2 expression, whereas treatment with the NO donor S-nitroso-L-glutathione (1 mmol/l) mimics the effect of TNF-α on UCP-2 expression.
The modulation of NO-mediated pathways through dietary supplementation with L-arginine or its precursor L-citrulline may represent an effective and practical strategy to prevent and treat MS, including obesity, T2DM, and dyslipidemia, in mammals such as humans. Clinical research evaluating the effects of arginine and citrulline in mitochondrial diseases has been limited to uncontrolled, open-label studies, which have demonstrated that administering arginine to subjects with mitochondrial encephalomyopathy, lactic acidosis or strokelike episodes (MELAS) syndrome results in improvement of the clinical symptoms associated with the stroke-like episodes and decreased frequency and severity of these episodes.
Pathways that regulate mitochondrial biogenesis have recently emerged as potential therapeutic targets for the amelioration of the endothelial and vascular dysfunction observed in metabolic diseases. have suggested that lipoamide is a potent stimulator of mitochondrial biogenesis in adipocytes and may have potential therapeutic applications in obesity and T2DM . Lipoamide dose-dependently stimulates the expression of eNOS and the formation of cGMP. In addition, knockdown of eNOS using siRNA prevents lipoamide-induced mitochondrial biogenesis, which is also blocked by the soluble guanylate cyclase inhibitor 1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one and by the protein kinase G inhibitor KT5823. Thus, the stimulation of mitochondrial biogenesis by lipoamide involves signaling via the eNOS-cGMP-PKG pathway.
Enhanced NO production is known to activate SIRT1, which is a histone deacetylase that regulates PGC-1α, a regulator of www.frontiersin.org
February 2015 | Volume 6 | Article 20 | 7 mitochondrial biogenesis and a coactivator of transcription factors that impacts energy homeostasis. Tadalafil-treated diabetic mice exhibit significantly improved left ventricular function that is associated with increased cardiac SIRT1 activity. Tadalafil also enhances plasma NO oxidation levels, myocardial SIRT1, PGC-1α expression, and the phosphorylation of eNOS, Akt, and AMPK in diabetic hearts. Oxidative phosphorylation of the complex I substrate glutamate is decreased by 50% in diabetic hearts compared with non-diabetic controls. Tadalafil protects oxidative phosphorylation and improves glutamate state 3 respiration rates. Additionally, the increase in ROS production from complex I is significantly decreased by treatment with tadalafil. Studies of resveratrol treatment in primary human coronary arterial endothelial cells have shown that resveratrol induces mitochondrial biogenesis in these cells. The formation of new mitochondria is associated with activation of SIRT1, upregulation of eNOS, and induction of specific mitochondrial biogenesis factors. It has also been shown that resveratrol treatment induces mitochondrial biogenesis in the aorta in T2DM mice.
The insulin-sensitizing drug pioglitazone is a ligand of PPARγ that ameliorates the basic problem of insulin resistance and is therefore considered to reduce the risk of cardiovascular disease in patients with T2DM (Yki-Järvinen, 2004). Pioglitazone reduces myocardial infarct size via the activation of the PPARγ, phosphoinositide 3-kinase, Akt, and eNOS pathways but not via opening of the mitochondrial K(ATP) channels. The use of this drug may be a novel strategy for the treatment of T2DM with coronary artery disease.
Thus, a deeper understanding of the role of NO and NOS isoforms in mitochondrial pathology is necessary for the study of MS pathogenesis as well as its complications and clinical applications.
# Conclusions
The main components of MS, particularly visceral obesity, insulin resistance, and T2DM, are associated with changes in mitochondrial metabolism in different tissues. Mitochondrial dysfunction may accelerate the progression of insulin resistance and subsequent organ dysfunction and has different interconnections with eNOS, iNOS, nNOS and mtNOS. The main mechanisms involved in MS development include augmented generation of mitochondrial superoxide that is formed with the participation of NO and disruptions in mitochondrial respiration and biogenesis that lead to apoptosis and other consequences. In addition, alterations in mitochondrial activity that occur in MS may include acquired or inherited reductions in the mitochondrial oxidative phosphorylation capacity, submaximal ADP-stimulated oxidative phosphorylation and plasticity of mitochondria and/or reductions in mitochondrial contents in insulin-sensitive tissues.
Currently, mitochondria remain an attractive target for the prevention and treatment of metabolic disorders. Therapy based on the inhibition of iNOS, on the prevention of eNOS uncoupling and on other effects on the components of the NO pathway may help to protect against mitochondrial dysfunction associated with metabolic disorders. Despite great advances in this area, questions remain about the relationships between mitochondrial function and the NOSs in various tissues, such as skeletal muscle, liver, heart and adipose tissues, particularly with regard to the effects of therapeutic strategies that affect mitochondrial function and insulin sensitivity. |
A review of the therapeutic role of the new third-generation TKI olverembatinib in chronic myeloid leukemia
Several tyrosine kinase inhibitors (TKIs) have been developed as targeted therapies to inhibit the oncogenic activity of several tyrosine kinases in chronic myeloid leukemia (CML), acute lymphoid leukemia (ALL), gastrointestinal stromal tumor (GIST), and other diseases. TKIs have significantly improved the overall survival of these patients and changed the treatment strategy in the clinic. However, approximately 50% of patients develop resistance or intolerance to imatinib. For second-generation TKIs, approximately 30%-40% of patients need to change therapy by 5 years when they are used as first-line treatment. Clinical study analysis showed that the T315I mutation is highly associated with TKI resistance. Developing new drugs that target the T315I mutation will address the dilemma of treatment failure. Olverembatinib, as a third-generation TKI designed for the T315I mutation, is being researched in China. Preliminary clinical data show the safety and efficacy in treating CML patients harboring the T315I mutation or who are resistant to firstor second-line TKI treatment. Herein, we review the characteristics and clinical trials of olverembatinib. We also discuss its role in the management of CML patients. KEYWORDS olverembatinib, chronic myeloid leukemia, tyrosine kinase inhibitors, breakpoint cluster region protein and abelson tyrosine-protein kinase 1, T315I mutationFrontiers in OncologyQian H, Gang D, He X and Jiang S (2022) A review of the therapeutic role of the new third-generation TKI olverembatinib in chronic myeloid leukemia.
# Introduction
Chronic myeloid leukemia (CML) is a clonal hematopoietic stem cell disorder that accounts for approximately 15% of adult leukemia [bib_ref] Epidemiological patterns of leukaemia in 184 countries: A population-based study, Miranda-Filho [/bib_ref]. The incidence of CML varies from 0.4/100,000 to 1.75/100,000 persons per year [bib_ref] Global, regional, and national burden of chronic myeloid leukemia, 1990-2017: A systematic..., Lin [/bib_ref]. Diagnosis is usually suspected from the complete blood count (CBC) and blood smear. Fluorescence in situ hybridization (FISH) for abnormal chromosome t(9;22)(q34;q11.2) and reverse transcriptase quantitative PCR for the BCR-ABL1 fusion gene help in the final diagnosis [bib_ref] Diagnosis and treatment of chronic myeloid leukemia in 2015, Thompson [/bib_ref]. CML is clinically staged into the chronic phase (CP), accelerated phase (AP), and blast phase (BP). In untreated patients, progression to BP occurs at a median of 3-5 years after initial diagnosis [bib_ref] An overview and update of chronic myeloid leukemia for primary care physicians, Granatowicz [/bib_ref].
When a patient is diagnosed with CML, the first-line treatment is a TKI. In an open-label, multicenter trial with a crossover design for patients treated with the first-generation TKI imatinib, the estimated overall survival rate at 10 years was 83.3%, and the cumulative rate of major cytogenetic response (MCyR) at the end of the trial was 89.0%. A total of 82.8% of patients had a complete cytogenetic response (CCyR). Despite the excellent results obtained in clinical trials, approximately 40%-45% of patients discontinue imatinib for various reasons, including unsatisfactory therapeutic outcomes in 16% of patients or primary resistance [bib_ref] Clinical trials in chronic myeloid leukemia, Saussele [/bib_ref]. Dose escalation after failure of imatinib treatment is an important option, though it is likely to be effective only in some subsets of patients. However, it is not recommended in the clinic. For such patients, an important salvage treatment strategy is represented by second-generation TKIs including dasatinib, nilotinib, and bosutinib, which allow recovering a CCyR in about 50% of cases (7-11, [fig_ref] TABLE 1: TKI treatment of CP-CML patients [/fig_ref]. When used as a front-line treatment, second-generation TKIs provide a faster achievement of a higher rate of cytogenetic and molecular responses with respect to imatinib (12-18, 22-24, [fig_ref] TABLE 1: TKI treatment of CP-CML patients [/fig_ref]. Subsequently, second-generation TKIs have been authorized for the first-line treatment of newly diagnosed Philadelphia chromosome-positive (Ph+) adult CP-CML [bib_ref] Has the time for first-line treatment with second generation tyrosine kinase inhibitors..., Gurion [/bib_ref]. Some patients experience treatment failure with second-generation TKIs and require a switch to a different TKI. Intolerance and resistance are major causes of treatment failure. Resistance to TKIs can arise from BCR-ABL1-dependent mechanisms, such as mutations in the kinase domain, overexpression, or amplification of BCR-ABL1, or BCR-ABL1-independent mechanisms (26). The wellstudied and most common mechanism is point mutation involving the BCR-ABL1 kinase domain (27). In one study, which included 175 patients with newly diagnosed CML or resistance to TKIs, 28 different mutations were detected in 54 (30.86%). A total of 14 (8.0%) patients carried the T315I mutation, accounting for the largest proportion in the mutation group [bib_ref] Compound mutations in BCR-ABL1 are not major drivers of primary or secondary..., Liu [/bib_ref]. It has been demonstrated that the T315I mutation is insensitive to all firstand second-generation TKIs. Asciminib and ponatinib (19-21, [fig_ref] TABLE 1: TKI treatment of CP-CML patients [/fig_ref] , as new-generation TKIs, have unique activity against the T315I mutation and are approved to treat patients with resistance or intolerance to prior TKI therapy or the presence of the BCR-ABL1 T315I mutation in all CML phases. Research has shown that the ATP-binding site mutation T315M confers resistance to ponatinib (29). In vitro data indicated that some compound mutations involving T315I also potentially impact ponatinib sensitivity [bib_ref] BCR-ABL1 compound mutations combining key kinase domain positions confer clinical resistance to..., Zabriskie [/bib_ref]. New TKIs have higher sensitivity to the T315I mutation. In November 2021, olverembatinib, as a new third-generation TKI that showed clinical efficacy in CML patients with the T315I mutation, received its first approval in China for the treatment of adult patients with TKI-resistant CP-CML or AP-CML harboring the T315I mutation [bib_ref] Olverembatinib: First approval, Dhillon [/bib_ref]. It is the first and only approved TKI designed for the T315I mutation in China.
Here, we review the characteristics and clinical studies of olverembatinib and its application value in the clinic.
## Characteristics and inhibitory activity of olverembatinib
Olverembatinib tightly binds to the ATP-binding sites of native BCR-ABL1 and multiple BCR-ABL1 mutants, including Q252H, E255K, F317L, F317I M351T, H396P, and the most refractory gatekeeper mutant T315I [bib_ref] Identification of GZD824 as an orally bioavailable inhibitor that targets phosphorylated and..., Ren [/bib_ref]. It binds to phosphorylated and non-phosphorylated forms of BCR-ABL1 kinase, as well as to several other kinases, including KIT, FLT3, fibroblast growth factor receptor 1 (FGFR1), and platelet-derived growth factor receptor a (PDGFRa). In the predicted binding pose of olverembatinib (GZD824) with activated BCR-ABL1 T315I , the 1H-pyrazolo pyridine core occupies the adenine pocket of BCR-ABL1 kinase to form a hydrogen bond donor-acceptor network. The alkyne moiety of GZD824 presents favorable van der Waals interactions with BCR-ABL T315I [bib_ref] Identification of GZD824 as an orally bioavailable inhibitor that targets phosphorylated and..., Ren [/bib_ref].
In vitro, olverembatinib strongly inhibits the proliferation of Ba/F3 cells expressing BCR-ABL1 T315I , as well as K562/Ku812 CML cells and SUP-B15 ALL cells expressing BCR-ABL1. In mouse allograft leukemia models, it inhibited tumor growth in the allograft model using Ba/F3 cells expressing BCR-ABL1 WT or BCR-ABL1 T315I in a dose-dependent manner and showed overall survival benefits [bib_ref] FLT3 inhibition by olverembatinib (HQP1351) downregulates MCL-1 and synergizes with BCL-2 inhibitor..., Wang [/bib_ref]. Moreover, in FLT3-ITD mutant AML and xenograft tumor models, cell growth was significantly inhibited with olverembatinib (34, [bib_ref] GZD824 suppresses the growth of human b cell precursor acute lymphoblastic leukemia..., Ye [/bib_ref]. Olverembatinib also exhibits antileukemic activity in imatinib-resistant/sensitive (GIST) cell lines and a GIST mouse model by inhibiting the phosphorylation of KIT and its downstream proteins, including AKT, ERK1/2, and STAT3 [bib_ref] Preclinical development of HQP1351, a multi-kinase inhibitor targeting a broad spectrum of..., Liu [/bib_ref]. The above studies demonstrate that olverembatinib has wide antitumor activities and exerts a strong inhibitory effect.
## Clinical trials of olverembatinib cp-cml
In an open-label, multicenter phase 1/2 trial, 127 Chinese patients with TKI-resistant CP-CML were enrolled [bib_ref] Olverembatinib (HQP1351), a well-tolerated and effective tyrosine kinase inhibitor for patients with..., Jiang [/bib_ref]. The characteristics of the patients are depicted in [fig_ref] TABLE 2: Characteristics of patients in the olverembatinib trial [/fig_ref]. Forty-one (43.6%) patients had a single T315I mutation, and 15 (16.0%) patients carried T315I and additional mutations. A total of 20.2% of patients had no BCR-ABL1 mutation. Compound mutations were detected in 7 (7.4%) patients. The median follow-up was 37 months. In evaluable patients without baseline responses, MCyR and CCyR were achieved in 79.3% and 69.4% of patients, respectively, at a median of 3 months. The cumulative 3-year incidences of MCyR, CCyR, MMR, MR4.0, and MR4.5 were 78.6%, 69.0%, 55.9%, 43.5%, and 38.6%, respectively. The probabilities of sustained MCyR, CCyR, and MMR at 3 years were 77.3%, 72.2%, and 76.0%, respectively. Seven patients died of disease progression, other diseases, or unknown reasons. The probabilities of PFS and OS at 3 years were 92% and 94%, respectively.
## Ap-cml
A phase 1/2 study (37) included 38 patients with TKI-resistant BCR-ABL1 T315I AP-CML, 12 patients with a single T315I mutation, and four with T315I and additional mutations. Among 37 patients without baseline MaHR, 27 patients experienced CHR at a median of 3 months. The three-year cumulative incidences of achieving MCyR, CCyR, MMR, MR4.0, and MR4.5 were 47.4%, 47.4%, 44.7%, 39.3%, and 32.2%, respectively. The probabilities of PFS and OS at three years were 60% and 71%, respectively. A total of 11 patients had CML that progressed to the blast phase, and 2 patients died. We also summarize some characteristics in [fig_ref] TABLE 2: Characteristics of patients in the olverembatinib trial [/fig_ref]. These studies show that olverembatinib is as efficacious and well tolerated as monotherapy in patients with TKI-resistant BCR-ABL1 T315I AP-CML.
## Ongoing clinical trials in cml
In the phase I dose escalation study (SJ-0002), patients with TKI-resistant CML were enrolled in 11 dose escalation cohorts ranging from 1 mg to 60 mg every other day. The maximum tolerated dose (MTD) has been established as 50 mg (38). Olverembatinib is also being assessed in patients with CP, AP, and BP-CML or with Ph+ ALL who have experienced resistance or intolerance to at least two TKIs or in subjects with Ph+ BCP ALL or LBP CML who have experienced resistance or intolerance to at least one second-or later-generation TKI (NCT04260022). In a phase 2 study, the efficacy and safety of Olverembatinib in CP-CML patients who are resistant or intolerant to firstand second-generation TKI are being evaluated (NCT04126681). In this study, the available treatment including hydroxyurea or interferon-based therapy, homoharringtonine, imatinib, dasatinib, or nilotinib will be selected by the investigator for each participant. The efficacy of olverembatinib will be determined by evaluating the patients' event-free survival (EFS). In addition, in one phase III study (NCT05311943), 40 patients are recruited to evaluate the efficacy and safety of olverembatinib in patients with CP-CML who are resistant or intolerant to at least two second-generation TKIs. The molecular response (MR), MMR, OS, PFS, and adverse events will be measured. We summarize these trials in [fig_ref] TABLE 3: Clinical studies of olverembatinib [/fig_ref].
## Adverse effects in patients treated with olverembatinib
In CP-CML with olverembatinib treatment, the treatmentrelated adverse events were mainly grade 1 or 2, and the most f re qu e nt n on h em at o l og ic a d ver se e ve nt w a s sk in hyperpigmentation (85%). Grade 3 or 4 nonhematologic adverse events included hypertriglyceridemia (8.7%), increased creatine phosphokinase (7.9%), and hypertension (5.5%). The most common hematologic treatment-related adverse event was thrombocytopenia (73.2%, including 48.8% of patients with grade 3 or 4). A total of 16.5% of patients had grade 3 or higher leukopenia, while 20.5% of patients had anemia (grade 3 or higher). In AP-CML, the common nonhematologic adverse events included skin pigmentation (81.6%), hypocalcemia (60.5%), proteinuria (50%), hyperuricemia (50%), hypertriglyceridemia (34.2%), and hyperphosphatemia (28.9%), of which most were grade 1 or 2. Thrombocytopenia, anemia, leukopenia, and neutropenia were the common hematological adverse events, with incidences of 86.8%, 63.2%, 50%, and 21.1%, respectively; these adverse events were manageable [bib_ref] Olverembatinib (HQP1351), a well-tolerated and effective tyrosine kinase inhibitor for patients with..., Jiang [/bib_ref].
## Role of olverembatinib in cml
The goals of treatment in CML are the prolongation of survival and prevention of progression toward AP and BP. Over time, the goals of treatment have evolved to achieve deep molecular responses. The focus for patients living with CML has shifted toward improving quality of life and saving costs, and the goal of achieving treatment-free remission is becoming more desirable. Some patients show resistance to first-generation TKIs and must switch to second-generation TKIs. Prospective studies report the superiority of second-generation TKIs, particularly in terms of MR rate, speed, and depth [bib_ref] Firstline imatinib vs second-and third generation TKIs for chronic-phase CML: A systematic..., Vener [/bib_ref] [bib_ref] Nilorinib versus imatinib for the treatment of patients with newly diagnosed chronic..., Kantarjian [/bib_ref] [bib_ref] (Scemblix): A third-line treatment option for chronic myeloid leukemia in chronic phase..., Kantarjian [/bib_ref]. Actually, four TKIs, namely, imatinib, nilotinib, dasatinib, and bosutinib, are recommended for first-line treatment. In patients with TKI treatment-naïve BCR-ABL AP/BP-CML, the standard treatment is the abovementioned four TKIs. In the case of intolerance or failure to one TKI, the choice of the subsequent therapy should also consider whether the patients develop ABL1 point mutations (42). The secondline treatment is a switch to any of the other TKIs. If the patients are resistant or intolerant to one of the second-generation TKIs, in such situations, ponatinib is also recommended. If patients are heavily pretreated and resistant to at least two second-generation TKIs, asciminib as the third-line treatment is an important option (43). If disease progresses to AP or BP during TKI treatment, we recommend choosing third-generation TKIs, bone marrow transplantation or experimental treatment [bib_ref] Treatment recommendations for chronic myeloid leukemia, Baccarani [/bib_ref]. In CML patients with the BCR-ABL1 T315I mutation, the thirdgeneration TKIs ponatinib and asciminib are preferentially recommended [bib_ref] Spotlight on ponatinib in the treatment of chronic myeloid leukemia and Philadelphia..., Anagnostou [/bib_ref]. Nevertheless, there is no consensus about the choice of treatment strategy. According to our previous studies, olverembatinib shows safety and efficacy in CP and BP-CML patients with at least two generations of TKI resistance or T315I mutations. We further recommend olverembatinib as the third-line treatment [bib_ref] Olverembatinib (HQP1351), a well-tolerated and effective tyrosine kinase inhibitor for patients with..., Jiang [/bib_ref]. In one study, the third-generation TKI ponatinib combined with fludarabine, cytarabine, granulocyte colony-stimulating factor, and idarubicin (FLAG-IDA) was used in Ph+ BP-CML patients, with 69% of patients in the second chronic phase after one cycle of treatment [bib_ref] Ponatinib with fludarabine, cytarabine, idarubicin, and granulocyte colonystimulating factor chemotherapy for patients..., Copland [/bib_ref]. Further studies are needed to investigate the efficacy and safety of olverembatinib in combination with other chemotherapies in AP-CML and BCR-ABL1T315I mutation patients.
# Conclusion
Resistance to TKIs is a complex and multifactorial process that presents the selection of leukemia clones with the ability to evade treatment [bib_ref] Three novel patient derived BCR/ABL mutants show different sensitivity to second and..., Soverini [/bib_ref]. Some specific mutations result in resistance to one particular TKI. Nilotinib is resistant to Y253H/F, E255K/V, F359V/C/I, and T315I mutations, whereas bosutinib works for all identified mutations with exceptions of V299L and T315I mutations (48, 49). The most common resistance mechanism is the ABL1 kinase domain point mutation. T315I mutation is the most aggressive point mutation identified in BCR-ABL1. Ponatinib was designed to overcome this mutation, while compound mutations, including T315I, cause resistance to ponatinib. Overall, olverembatinib shows a significant inhibitory effect in the presence of the T315I mutation or compound mutations because it does not form a hydrogen bond with the hydroxyl group at this residue [bib_ref] Preclinical development of HQP1351, a multi-kinase inhibitor targeting a broad spectrum of..., Liu [/bib_ref] ; it has been demonstrated to be safe and efficacious in phase I/II clinical trials in CP-CML patients resistant to at least two TKIs. In China, olverembatinib is the only third-generation TKI approved for the treatment of CML patients with T315I mutation. Therefore, it is preferentially recommended in CML patients harboring the T315I mutation in China. Besides, olverembatinib as a third-line therapy might represent a viable option for patients failing firstor second-generation TKIs.
[table] TABLE 1: TKI treatment of CP-CML patients. [/table]
[table] TABLE 2: Characteristics of patients in the olverembatinib trial(37). [/table]
[table] TABLE 3: Clinical studies of olverembatinib.NCT04260022CP/AP/BP-CML and Ph+ ALL, Resistance or intolerance to at least two TKIs, BCP-ALL and LBP-CML, Resistance or intolerant to at least one second or later generation TKI EFS, event-free survival; PK/PD, pharmacokinetic or pharmacodynamic; BCP-ALL, Ph+ B-cell precursor (BCP) ALL; LBP, lymphoid blast phase CML (CML LBP); AUC, area under the curve; MaHR, major hematologic response. Each cycle is 28 days. [/table]
|
Ataxia-telangiectasia mutated (ATM) gene mutation in a patient with primary cutaneous marginal zone lymphoma
# Introduction
Primary cutaneous marginal B-cell lymphoma (PCMZL), a type of extranodal marginal zone lymphoma (MZL) of mucosa-associated lymphoid tissue, is a common indolent non-Hodgkin lymphoma that typically presents with solitary or multifocal papules/ plaques or tumors. [bib_ref] The 2018 update of the WHO-EORTC classification for primary cutaneous lymphomas, Willemze [/bib_ref] Two types of PCMZL have been noted pathologically. [bib_ref] The 2018 update of the WHO-EORTC classification for primary cutaneous lymphomas, Willemze [/bib_ref] The more common subtype is characterized by class-switched immunoglobulins without C-X-C chemokine receptor 3 expression and a small proportion of neoplastic B cells. [bib_ref] The 2018 update of the WHO-EORTC classification for primary cutaneous lymphomas, Willemze [/bib_ref] A smaller subset shows a diffuse proliferation of neoplastic B cells that express immunoglobulin M (and often C-X-C chemokine receptor 3) and has a higher likelihood of extracutaneous disease manifestations. [bib_ref] The 2018 update of the WHO-EORTC classification for primary cutaneous lymphomas, Willemze [/bib_ref] Although some authors assert that many cases of PCMZL are clonal reactions to intradermal antigens, such as tick bites or tattoos, 1,3,4 the t(14;18)(q32;q21);IGH-MALT1 translocation is reported in up to 25% of PCMZLs, with t(3;14)(p14.1;q32);IGH-FOXP1 rearrangements less commonly reported. [bib_ref] How I diagnose primary cutaneous marginal zone lymphoma, Gibson [/bib_ref] We present a patient with PCMZL with a known monoallelic mutation in ataxia-telangiectasia mutated (ATM) gene. ATM is a serine/threonine kinase that activates p53 in response to double-stranded DNA (dsDNA) damage. [bib_ref] ATM mutations in cancer: therapeutic implications, Choi [/bib_ref] Biallelic mutations in ATM result in the autosomal recessive disorder ataxia telangiectasia syndrome, which is associated with a 20%-30% increased risk of lymphoid and other cancers. [bib_ref] ATM mutations in cancer: therapeutic implications, Choi [/bib_ref] In addition, germ-line mutations of ATM have been associated with an increased risk of many types of cancer, including around a 14% risk of non-Hodgkin lymphoma. [bib_ref] ATM mutations in cancer: therapeutic implications, Choi [/bib_ref] We discuss the relevance of this ATM mutation in terms of prognosis and potential treatment options in patients with PCMZL.
## Case report
A 73-year-old woman presented to the Duke Dermatology Cutaneous Lymphoma Clinic for evaluation of an 8-month history of skin lesions on her back and arms, gradually increasing in number and unresponsive to topical hydrocortisone. On examination, she had pink papules on her left and right shoulder, left upper arm, and left aspect of the back without associated bleeding, pain, or pruritus [fig_ref] Fig 1: Clinical presentation of primary cutaneous marginal zone lymphoma [/fig_ref]. Three previous outside biopsies revealed a perivascular and periadnexal infiltrate of small-sized lymphocytes, CD20 positive B cells (85%), and CD3 positive T cells. Further immunostains showed an aberrant expression of B-cell lymphoma 2, a background of plasmacytoid cells with scattered plasma cells-entirely Kappa restricted-with no expression of Lambda, and B-cell lymphoma 6 and CD10highlighted, focally disrupted germinal centers. Molecular testing showed clonal immunoglobulin heavy chain gene rearrangement on a polyclonal background. These histologic findings were felt diagnostic of MZL, most consistent with the B cellpredominant cutaneous subtype. The outside bone marrow biopsy was negative for lymphoma, and computed tomography with contrast of her chest, abdomen, and pelvis revealed only heterogeneousenhancing lesions in the liver, potentially representing hemangiomas. Based on the clinical presentation and imaging as described above, the overall stage of her PCMZL was T3bN0M0. [bib_ref] The 2018 update of the WHO-EORTC classification for primary cutaneous lymphomas, Willemze [/bib_ref] The patient was prescribed Class I topical steroids with occlusion prior to consideration of radiation therapy.
As part of the patient's history, she noted prior testing for (and discovery of) an ATM mutation due to a family history of malignancy; an ATM mutation was found in her brother, who had prostate cancer, and in her niece, who did not have cancer but was tested secondary to a history of malignancy in the patient's brother and his wife. No other family members have been tested for the mutation, including a first cousin with a history of lymphoma.
# Discussion
Although the testing for an ATM mutation in this patient preceded the diagnosis of PCMZL, it presents a possible genetic susceptibility in a common cancer with many questions about its etiology. There are no reports on the prevalence of ATM mutations in patients with PCMZL, but loss of ATM has been noted in both nodal and splenic MZLs. [bib_ref] Genomic analysis of marginal zone and lymphoplasmacytic lymphomas identified common and disease-specific..., Braggio [/bib_ref] ATM mutations have also been reported in both mantle cell lymphoma 7 and chronic lymphocytic leukemia, where its presence has been noted in 12% of the cases. [bib_ref] Clinical utility of recently identified diagnostic, prognostic, and predictive molecular biomarkers in..., Onaindia [/bib_ref] An ATM mutation may also prove to be a prognostic factor for certain cancers, as they have been associated with shorter progression-free survival in chronic lymphocytic leukemia. [bib_ref] Clinical utility of recently identified diagnostic, prognostic, and predictive molecular biomarkers in..., Onaindia [/bib_ref] Although cutaneous relapses are common, PCMZL has a 5-year survival rate around 99%, with rare extension to extracutaneous sites. [bib_ref] The 2018 update of the WHO-EORTC classification for primary cutaneous lymphomas, Willemze [/bib_ref] While there is no curative treatment for this lymphoma, there may be curative treatment for individual lesions, including surgical excision, local radiotherapy, 9 or a combination of both. 9,10 Other initial treatments for limited stage I lesions include topical or intralesional steroids, 10 intralesional rituximab, [bib_ref] Primary cutaneous marginal zone lymphoma, Dalle [/bib_ref] [bib_ref] Innovative therapeutic approaches in primary cutaneous B cell lymphoma, Lang [/bib_ref] or intralesional interferon-alpha. [bib_ref] Primary cutaneous marginal zone lymphoma, Dalle [/bib_ref] [bib_ref] Innovative therapeutic approaches in primary cutaneous B cell lymphoma, Lang [/bib_ref] Recurrent or extensive skin lesions may be managed with intravenous or intralesional rituximab and, rarely, chemotherapy. [bib_ref] Primary cutaneous marginal zone lymphoma, Dalle [/bib_ref] [bib_ref] Primary cutaneous marginal zone B-cell lymphoma: response to treatment and disease-free survival..., Servitje [/bib_ref] In the cases of B. burgdorferi infection-associated PCMZL, treatment with antibiotics (cephalosporins or tetracyclines) may be effective, although specifics on indication, treatment regimens, and antibiotic benefit are not definitive based on available literature. [bib_ref] Primary cutaneous marginal zone lymphoma, Dalle [/bib_ref] [bib_ref] Innovative therapeutic approaches in primary cutaneous B cell lymphoma, Lang [/bib_ref] Mutations in ATM also have potential implications in treatment regimens, as cells deficient in ATM may have increased sensitivity to ionizing radiation due to inadequate double-stranded DNA repair mechanisms. [bib_ref] ATM mutations in cancer: therapeutic implications, Choi [/bib_ref] In our patient, who presented with PCMZL and an ATM mutation, treatment with radiation may be particularly successful due to this mechanism; however, sensitivity may lead to increased radiation-associated skin toxicity, as reported for patients with breast cancer and ATM mutations. [bib_ref] Severe late toxicity after adjuvant breast radiotherapy in a patient with a..., Dosani [/bib_ref] Despite this finding, the role of ATM mutations and risk of toxicity from radiation therapy remains controversial in the literature. [bib_ref] Clinical radiosensitivity in breast cancer patients carrying pathogenic ATM gene mutations: no..., Bremer [/bib_ref] Other treatment approaches that exploit ATM mutations include synthetic lethality, in which cancer cells are killed by targeting their compensatory mechanisms of DNA repair. Although more germane to nodal cancers or chronic lymphocytic leukemia, early preclinical and animal studies suggest that the synthetic lethality of platinum drugs, nucleoside analogs (such as sapacitabine), ataxia telangiectasia and RAD3-related inhibition (using compounds such as VE821, VE822, and AZD6738), and poly (ADP-ribose) polymerase inhibitors (including olaparib) may have utility in cancers associated with ATM mutation. 5
[fig] Fig 1: Clinical presentation of primary cutaneous marginal zone lymphoma (PCMZL), left shoulder and left arm. Scar from prior biopsy of similar papule. [/fig]
|
The role of long non-coding RNAs in cardiac development and disease
Cells display a set of RNA molecules at one time point, reflecting thus the cellular transcriptional steady state, configuring therefore its transcriptome. It is basically composed of two different classes of RNA molecules; protein-coding RNAs (cRNAs) and protein non-coding RNAs (ncRNAs). Sequencing of the human genome and subsequently the ENCODE project identified that more than 80% of the genome is transcribed in some type of RNA. Importantly, only 3% of these transcripts correspond to protein-coding RNAs, pointing that ncRNAs are as important or even more as cRNAs. ncRNAs have pivotal roles in development, differentiation and disease. Non-coding RNAs can be classified into two distinct classes according to their length; i.e., small (<200 nt) and long (>200 nt) noncoding RNAs. The structure, biogenesis and functional roles of small non-coding RNA have been widely studied, particularly for microRNAs (miRNAs). In contrast to microRNAs, our current understanding of long non-coding RNAs (lncRNAs) is limited. In this manuscript, we provide state-of-the art review of the functional roles of long non-coding RNAs during cardiac development as well as an overview of the emerging role of these ncRNAs in distinct cardiac diseases.
# Introduction
The cell transcriptome can be defined as the set of RNA molecules present on it at one time point, reflecting thus the cellular transcriptional steady state. It is basically composed of two different classes of RNA molecules; protein-coding RNAs (cRNAs) and non-coding RNAs (ncRNAs) [fig_ref] Figure 1: Schematic representation of the distinct classes of RNA molecules [/fig_ref]. For decades, scientists focused their attention on coding RNAs, while the non-coding RNAs were defined as the "dark matter" of the genome and were not considered important until recently. An exception to the rule was represented by tRNAs and rRNAs, which were widely studied given their NcRNAs can be classified into two distinct classes according to their length: (1) Small noncoding RNAs, i.e., smaller than 200 nucleotides; including therein microRNAs, small nucleolar RNAs, piRNAs as well as transfer RNAs, and (2) Long non-coding RNAs, i.e., longer than 200 nucleotides [bib_ref] Non-coding RNAs in human disease, Esteller [/bib_ref] , including a extensively variety of types as detailed below [fig_ref] Figure 1: Schematic representation of the distinct classes of RNA molecules [/fig_ref]. The structure, biogenesis and functional roles of small non-coding RNA have been widely studied, particularly for microRNAs (miRNAs). MiRNAs have an average size of 20-22 nucleotides and act by binding to a target mRNA inducing thereafter degradation and/or inhibition of its translation, and thus negatively regulating gene expression. The functional role of miRNAs have been studied and demonstrated in multiple organisms and within multiple cellular contexts (see for a review; [bib_ref] MicroRNAs in Cardiovascular Disease, Barwari [/bib_ref] [bib_ref] MicroRNA regulatory networks in cardiovascular development, Liu [/bib_ref].
In contrast to microRNAs, our current understanding of long non-coding RNAs (lncRNAs) is limited. The development of new massive sequencing techniques has led to the discovery and annotation of a large number of long non-coding RNAs. GENCODE annotation initially estimated the existence of 9640 lncRNA genes in the human genome [bib_ref] The GENCODE v7 catalog of human long noncoding RNAs: analysis of their..., Derrien [/bib_ref] while recently the NONCODE database has increased this number up to 96,308 lncRNA genes [bib_ref] NONCODEv5: a comprehensive annotation database for long non coding RNAs, Fang [/bib_ref]. Such estimates indicate that the number of lncRNAs is twice that coding genes, supporting an important role of these lcnRNA transcripts in multiple biological contexts. Recently, lncRNAs have emerged as major players in regulating gene expression, both at transcriptional and post-transcriptional level and they have been implicated in development, stem cell differentiation, cellular homeostasis and disease [bib_ref] Mechanisms of long noncoding RNA function in development and disease, Schmitz [/bib_ref] [bib_ref] Long noncoding RNA regulation of pluripotency, Rosa [/bib_ref] [bib_ref] Long noncoding RNAs and human disease, Wapinski [/bib_ref].
## Structure and cellular localization of lncrnas
Long non-coding RNAs display essentially no potential to code for proteins, although structurally are similar to mRNAs. They are transcribed using the same pathways; i.e., RNA polymerase II, have typical histone modifications, 5' terminal cap and 3' terminal poly(A) tails. LncRNAs are constituted by exons and intronsand are often spliced. Curiously, a minority of non-polyadenylated lncRNAs is transcribed from RNA polymerase III [bib_ref] The expanding RNA polymerase III transcriptome, Dieci [/bib_ref]. Unlike mRNAs, lncRNAs have lower number of exons (42% of lncRNA transcripts is composed by two exons compared with 6% protein-coding transcritps) are less conserved between species and on average slightly shorter. Interestingly, lncRNA promoters are more conserved than their exons and in fact almost as conserved as protein-coding gene promoters. Conversely, introns from lncRNAs are longer that those from protein-coding transcripts and are normally flanked by canonical splice sites (GT/AG), showing no differences in splicing signals as compared to protein-coding transcripts [bib_ref] The GENCODE v7 catalog of human long noncoding RNAs: analysis of their..., Derrien [/bib_ref]. Although the vast majority of lncRNAs are located in the nuclear genome, lcnRNAs are also reported within the mitochondrial DNA. Mitochondrial encoded lncRNAs are transcribed and processed by mitochondrial transcriptional machinery but regulated by nuclear-encoded proteins [bib_ref] Long noncoding RNAs are generated from the mitochondrial genome and regulated by..., Rackham [/bib_ref].
Importantly, although lncRNAs they are referred as "non-coding", several lncRNAs contains short ORFs, can be engaged by ribosomes and thus can generate oligopeptides. Until recently, examples of lncRNAs that could generate small peptides were limited to sporadic cases. However recently, it is becoming increasingly acknowledged that a significant fraction of currently annotated lncRNAs is predicted to be capable of generating short peptides [bib_ref] Toddler: an embryonic signal that promotes cell movement via Apelin receptors, Pauli [/bib_ref] [bib_ref] Identifying (non-) coding RNAs and small peptides: Challenges and opportunities, Pauli [/bib_ref]. Among these long non-coding RNAs coding for small peptides, there are several examples, such as Toddler and Dworf, which are involved in mesoderm development [bib_ref] Toddler: an embryonic signal that promotes cell movement via Apelin receptors, Pauli [/bib_ref] [bib_ref] Identifying (non-) coding RNAs and small peptides: Challenges and opportunities, Pauli [/bib_ref] [bib_ref] A peptide encoded by a transcript annotated as long noncoding RNA enhances..., Nelson [/bib_ref].
LncRNAs display low expression levels yet with increased tissue and time specificity as compared to the protein-coding genes [bib_ref] Long non-coding RNAs: spatial amplifiers that control nuclear structure and gene expression, Engreitz [/bib_ref]. Such specificity suggests an important role of these transcripts in tightly defined cellular events as supported by several reports [bib_ref] The specificity of long noncoding RNA expression, Gloss [/bib_ref] [bib_ref] Long Noncoding RNAs in Cardiovascular Pathology, Bär [/bib_ref]. At the cellular level, lncRNAs can be located both in the cytoplasm and nucleus. Cytoplasmic enriched lncRNAs have mainly a role in post-transcriptional regulation whereas nuclearly located lncRNAs predominantly play a role in transcriptional gene regulation. Importantly, lncRNAs are dynamic molecules that can be located in the nucleus but translocate and act in the cytoplasm [bib_ref] Linking Long Noncoding RNA Localization and Function, Chen [/bib_ref]. An example of such a dynamic behavior is represented by antisense Uchl1 lncRNA, that partially overlapping Uchl1 protein-coding gene, moving from the nucleus to the cytoplasm where it bind the 5' end of Uchl1 mRNA promoting its translation under stress [bib_ref] The transcriptional landscape of the mammalian genome, Carninci [/bib_ref].
Other feature defining lncRNAs is their poor RNA sequence conservation across species, as exemplified by Braveheart, a mouse-specific lncRNA involved in early cardiogenesis, as detailed below [bib_ref] Braveheart, a long noncoding RNA required for cardiovascular lineage commitment, Klattenhoff [/bib_ref]. However, despite poor conservation between species, the comparison of splice sites suggests that lncRNAs are evolutionarily conserved, showing that the majority of lncRNAs are, as least, as old as mammalian lineage [bib_ref] The GENCODE v7 catalog of human long noncoding RNAs: analysis of their..., Derrien [/bib_ref].
## Classification of lncrnas
Classification of long non-coding RNA species differs between authors but at least five distinct groups can be distinguished. Broadly lcnRNAs are classified according to both, their position within the genome and relative location to neighboring genes. (a) lncRNAs transcribed from the same promotor as the adjacent protein coding gene. They are transcribed in both sense and antisense orientation and can be located in the same strand or opposite strand of protein coding gene. The expression of both is correlated and usually these lncRNAs modulates the expression of adjacent protein coding gene. (b) Long intergenic long non-coding RNAs (lincRNAs) are located between two protein-coding genes usually at distance of approximately 10 kb or in genomic desert as stand-alone genes. Since lincRNAs can be transcribed by their own promoter, they are classified as promoter-associated lncRNAs [bib_ref] lincRNAs: genomics, evolution, and mechanisms, Ulitsky [/bib_ref]. (c) LncRNAs can arise from intronic regions of protein coding genes, i.e., intronic lncRNA or enhancer regions, i.e., enhancer-associated lncRNA. There is a subclass of intronic lncRNAs, derived from these, known as sno-lncRNAs. Sno-lncRNAs do not have the typical structure observed in the majority of lncRNAs. They are not capped and nor polyadenilated and are flanked by small nucleolar RNAs at both extremes. The enhancer-associated lncRNAs are transcribed from enhancer region and their expression correlates with the expression of active enhancers. Also the expression enhancer lncRNAs correlate with expression of target genes showing a dynamic and specific patterning throughout differentiation and development [bib_ref] The promise of enhancer-associated long noncoding RNAs in cardiac regeneration, Ounzain [/bib_ref] [bib_ref] Discovery and functional characterization of cardiovascular long noncoding RNAs, Ounzain [/bib_ref]. (d) Alternative splicing of protein coding genes can generate a circular lncRNA, named circRNAs. These ncRNAs have a great regulatory potential but additional studies are required to fully understand their regulatory mechanisms [bib_ref] Circular RNAs are a large class of animal RNAs with regulatory potency, Memczak [/bib_ref]. For example Hrcr, a cardiac enriched circular lncRNA is a protective RNA against distinct molecular mechanisms leading to cardiac hypertrophy [bib_ref] A circular RNA protects the heart from pathological hypertrophy and heart failure..., Wang [/bib_ref]. (e) Finally, there is a subclass of lncRNAs harboring microRNAs within their genetic structure, such as H19, which encode miR-675 in its first exon [bib_ref] The H19 long noncoding RNA is a novel negative regulator of cardiomyocyte..., Liu [/bib_ref].
## Function role of long non-coding rnas
Long non-coding RNAs are defined as complex non-coding RNA molecules given their particularities affecting their structure as well as their dynamic expression pattern. Such a complexity is reflected in a wide variety of functions. LcnRNAs can act at both transcriptional and post-transcriptional regulation in multiple cellular processes as detailed below.
At the transcriptional level, lncRNAs can modulate the epigenetic landscape of the cell acting as different class of molecules. Currently, four different types of actions have been demonstrated as detail below. Some lncRNAs acting as guide, binding to transcription factors or protein subunits of chromatin remodeling complexes and direct them, as ribonucleoprotein complex, towards their genomic targets, promoting or suppressing gene activity depending on whether the guided complexes are activate (as MLL complex) or repressive complexes (as PRC2 complex). This class of lncRNAs can act in cis (i.e., Xist) or trans (i.e., HOTAIR) [bib_ref] The matrix protein hnRNP U is required for chromosomal localization of Xist..., Hasegawa [/bib_ref] [bib_ref] Long non-coding RNA HOTAIR reprograms chromatin state to promote cancer metastasis, Gupta [/bib_ref]. For example, Fendrr, a cardiac regulatory long non-coding RNA, acts as a guide for PCR2 and Trx/MLL directing them to Foxf1 and Pitx2 promoters and setting active and repressive marks [bib_ref] The tissue-specific lncRNA Fendrr is an essential regulator of heart and body..., Grote [/bib_ref].
Scaffold lncRNAs can acts as scaffold molecules for different complexes facilitating their assembly and being a functional component of it [bib_ref] Long noncoding RNA as modular scaffold of histone modification complexes, Tsai [/bib_ref]. For example, ANRIL, a long non-coding RNA described as a risk factor for coronary disease acts as platform recruiting and interacting with polycomb complex (PRC1 and PRC2) to the INK4b-ARF-INK4a locus promoting its silencing [bib_ref] Long noncoding RNA, polycomb, and the ghosts haunting INK4b-ARF-INK4a expression, Aguilo [/bib_ref].
Many of lncRNAs are dubbed enhancer lncRNAs since they can act as enhancers of transcription, promoting and maintaining the genomic3D conformation necessary for the transcriptional machinery to get access to promoter regions [bib_ref] eRNAs promote transcription by establishing chromatin accessibility at defined genomic loci, Mousavi [/bib_ref]. Similarly, other long non-coding RNAs can repress the formation of this genomic structures and therefore gene expression. An example of this functional role is Playrr, a long non-coding RNA encoded upstream of Pitx2 gene, that represses the expression of this homeodomain transcription factor in asymmetry pathway by interfering with the Pitx2 promoter [bib_ref] Chromatin architecture of the Pitx2 locus requires CTCF-and Pitx2-dependent asymmetry that mirrors..., Welsh [/bib_ref]. The last class of transcriptional regulatory lncRNAs exert their function as decoy molecules competing with transcription factors or chromatin remodeling complexes for their genetic targets. Such interactions avoid that the latter can exert its function and thus indirectly inhibit transcription. Terra, telomeric repeat-containing RNA, physically binds to telomerase inverse transcriptase blocking the action of this enzyme and inhibiting telomere elongation [bib_ref] The non-coding RNA TERRA is a natural ligand and direct inhibitor of..., Redon [/bib_ref]. In the cardiovascular context, Myheart, a cardiac lncRNA located in the murine myosin heavy chain 7 locus, sequestering BRG1-BAF complex avoiding that this complex can bind to targets [bib_ref] A long noncoding RNA protects the heart from pathological hypertrophy, Han [/bib_ref].
On the other hand, long non-coding RNAs can also regulate gene expression at post-transcriptional level interacting with mRNAs, the translational machinery or other non-coding RNAs such as microRNAs [bib_ref] Posttranscriptional gene regulation by long noncoding RNA, Yoon [/bib_ref]. Some nuclear lncRNAs can participate on pre-mRNA maturation by interacting with pivotal alternative splicing factors. For example, Malat1 interact with SR (serine/argenine) splicing factors in the nuclear speckle domains and modulates the concentration and distribution of those factors on the nucleus providing proper alternative splicing [bib_ref] The nuclear-retained noncoding RNA MALAT1 regulates alternative splicing by modulating SR splicing..., Tripathi [/bib_ref]. As a subclass, these lncRNAs, are dubbed as specific regulators-alternative splicing (sno-lncRNAs). Sno-lncRNAs are located within the nucleolus and in Cajal bodies. This subclass of lncRNAs are associated a several FOX proteins, such as Fox2 and regulate mRNA alternative splicing in stem cells [bib_ref] Long Noncoding RNAs with snoRNA Ends, Yin [/bib_ref].
Also, long non-coding RNAs can affect mRNA stability by base-pairing with them and altering their half-lives. Depending on efficacy of base pairing, the interaction can promote decoy or mRNA stability. Incomplete base-pairing normally promotes mRNA decoy whereas a full base pairing between both usually promotes mRNA stability and thus protein translation [bib_ref] Mammalian Staufen1 recruits Upf1 to specific mRNA 3'UTRs so as to elicit..., Kim [/bib_ref] [bib_ref] Staufen1 regulates diverse classes of mammalian transcripts, Kim [/bib_ref] [bib_ref] Expression of a noncoding RNA is elevated in Alzheimer's disease and drives..., Faghihi [/bib_ref] [bib_ref] Evidence for natural antisense transcript-mediated inhibition of microRNA function, Faghihi [/bib_ref].
Interestingly many lncRNAs are associated with ribosomes and can therefore affect mRNA translation. For example, LincRNA-21, an lncRNA co-distributed with ribosomes, represses the translation of different mRNA targets by base pairing at distinct transcript regions, including coding and non-coding regions of mRNAs. This incorrect base pairing generates a complex between LincRNA-21target mRNA that interacts with different translation repressors [bib_ref] Evidence for natural antisense transcript-mediated inhibition of microRNA function, Faghihi [/bib_ref]. Also, distinct lncRNAs can interact with the translation machinery by modulating its function. LcnRNABC1 interacts, through 3'UTR region, with several translation repressors, inhibiting the assembly of translation initiation complex [bib_ref] LincRNA-p21 suppresses target mRNA translation, Yoon [/bib_ref]. On the contrary, Uchl1, an antisense lncRNA of Uchl1 gene, promotes the active polysome generation at the Uchl1 mRNA enhancing its translation [bib_ref] Dendritic BC1 RNA in translational control mechanisms, Wang [/bib_ref]. In the cardiac context, some cytoplasmic lncRNAs act as decoy of mRNA such as Hcrc or Chrr. Also, lncRNAs can interact with the translation machinery by modulating this. So, BC1 interacts, through 3' untranslated region, with several represses translation inhibiting the assembly of translation initiation complex [bib_ref] LincRNA-p21 suppresses target mRNA translation, Yoon [/bib_ref].
Finally, several lncRNAs can act upon microRNAs interacting with them and modulating post-transcriptional gene expression. Linc-MD1 can sequester miR-133 and miR-135, enhancing the expression of their target mRNAs in skeletal muscle, such as Maml1 and Mef2c, respectively, among others [bib_ref] Long non-coding antisense RNA controls Uchl1 translation through an embedded SINEB2 repeat, Carrieri [/bib_ref]. In addition, several lncRNAs harboring microRNAs in their genome structure have been described. For example, H19 contains miR-675 in his first exon. Interestingly, the expression of this microRNA is regulated by H19 [bib_ref] A long noncoding RNA controls muscle differentiation by functioning as a competing..., Cesana [/bib_ref].
## Figure 2.
Schematic representation of the distinct stages of heart development from the bilateral sets of precardiac mesoderm (cardiac crescents) to the adult stage, illustrating the distinct long non-coding RNAs described to date and their corresponding molecular signaling pathways. Additional, if knock-out mice are available, the corresponding phenotype is briefly summarized.
## The role of long non-coding rna in cardiac development
The development of the heart is a complex morphogenetic process and thus highly regulated. The developing heart arises from sets of precursor cells during gastrulation [bib_ref] The H19 lincRNA is a developmental reservoir of miR-675 which suppresses growth..., Keniry [/bib_ref]. These cells progressively converge in the midline of the embryo, forming the early tubular heart [bib_ref] A common progenitor at the heart of development, Garry [/bib_ref]. The tubular heart grows and suffers a rightward displacement, a phenomenon dubbed cardiac looping [bib_ref] Signal transduction in early heart development (I): cardiogenic induction and heart tube..., Wagner [/bib_ref] , leading to the formation of the different cardiac regions and later a fully functional heart [bib_ref] Heart fields and cardiac morphogenesis, Kelly [/bib_ref]. The molecular mechanisms underlying cardiac morphogenesis include the activation and expression of a several cardiac gene network pathways (CGNPs) evolutionarily conserved. The correct expression of CGNPs is closely regulated in time, specific and spatial patterning by different transcriptional (cardiac-enriched transcription factors, such as Mef2c, Gata4, Nkx2.5 and several members of the T-box family) and post-transcriptional factors (non-coding RNAs such as microRNAs) [bib_ref] Chamber formation and morphogenesis in the developing mammalian heart, Christoffels [/bib_ref] [bib_ref] Long noncoding RNAs in cardiac development and pathophysiology, Schonrock [/bib_ref] [bib_ref] Signaling molecules, transcription growth factors and other regulators revealed from in-vivo and..., Meganathan [/bib_ref] [bib_ref] Small non-coding RNAs in animal development, Stefani [/bib_ref]. Cardiac transcription factors regulate the transcription of different elements that make up the cardiac transcription pathways forming part of them and regulating transcriptionally the cardiac development [bib_ref] Long noncoding RNAs in cardiac development and pathophysiology, Schonrock [/bib_ref]. Post-transcriptionally, the role that non-coding RNAs can exert in the regulation of these pathways as well as in other aspects of cardiac development have been described [bib_ref] Long noncoding RNAs in cardiac development and pathophysiology, Schonrock [/bib_ref] [bib_ref] Signaling molecules, transcription growth factors and other regulators revealed from in-vivo and..., Meganathan [/bib_ref] [bib_ref] Small non-coding RNAs in animal development, Stefani [/bib_ref]. Several microRNAs have been post-transcriptionally regulated in the development of the heart interacting with various elements of cardiac signaling pathways [bib_ref] The role of microRNAs in cardiac development and regenerative capacity, Katz [/bib_ref]. Emerging evidences have pointed out lncRNAs as pivotal players in the heart development regulating transcriptionally and post-transcriptionally distinct cardiac signaling pathways. Transcriptomic analyses have identified a large number of differentially expressed lncRNAs during cardiomyogenic differentiation and proliferation [bib_ref] The Role of a Novel Long Noncoding RNA TUC40-in Cardiomyocyte Induction and..., Li [/bib_ref] [bib_ref] Characterization and analysis of long non-coding rna (lncRNA) in In Vitro-and Ex..., Arnone [/bib_ref]. For example, HBL1, a human cardiac-enriched lncRNA, negatively modulates cardiomyocyte differentiation from pluripotent stem cells by interacting with SOX2 and miR-1 [bib_ref] HBL1 Is a Human Long Noncoding RNA that Modulates Cardiomyocyte Development from..., Liu [/bib_ref]. Morever several studies have been performed exploring the functional role of particular cardiac enriched-lncRNAs during cardiogenesis. In particular, to date, seven cardiac lncRNAs have been analyzed in detailed; Carmn, Braveheart, Fendrr, Alien, Upperhand, LncRNA-uc.167 and Pancr. We provided here state-of-the art review of the functional role of these lncRNAs in heart development.
## (a). carmn
Ounzain et al. [bib_ref] Genome-wide profiling of the cardiac transcriptome after myocardial infarction identifies novel heart-specific..., Ounzain [/bib_ref] profiled the lncRNA transcriptome of human cardiac precursor cells and identified a set of 570 lncRNAs differentially expressed during cardiac differentiation. Many of these lcnRNAs were associated with active cardiac enhancers and super enhancers [bib_ref] Genome-wide profiling of the cardiac transcriptome after myocardial infarction identifies novel heart-specific..., Ounzain [/bib_ref]. Super enhancers are associated with increased production of enhancer-associated ncRNAs and with the enrichment of chromatin remodeling complexes and specific histone modifications [bib_ref] Long non-coding RNAs in heart failure: a promising future with much to..., Ounzain [/bib_ref]. Among these lcnRNAs, Ounzain et al. [bib_ref] CARMEN, a human super enhancer-associated long noncoding RNA controlling cardiac specification, differentiation..., Ounzain [/bib_ref] focused their attention on Carmn, a super enhancer-associated lncRNA. Carmn is located upstream of miR-143 and miR-145, two microRNA involved in cardiovascular development [bib_ref] The miR-143/145 cluster is a novel transcriptional target of Jagged-1/Notch signaling in..., Boucher [/bib_ref] [bib_ref] ) miR-145 and miR-143 regulate smooth muscle cell fate and plasticity, Cordes [/bib_ref]. Although Carmn, miR-143 and miR-145 are located within the same genomic locus, they are expressed as independent transcripts. Carmn is expressed both in fetal and adult hearts and it is well conserved between mammalian species. Carmn directly acts during the earliest steps of cardiac lineage commitment regulating cardiac differentiation from nascent mesoderm by modulating the expression downstream of mesp1-cardiac gene network [bib_ref] CARMEN, a human super enhancer-associated long noncoding RNA controlling cardiac specification, differentiation..., Ounzain [/bib_ref]. Moreover Carmn modulates the expression of key factor of pluripotency suggesting a bivalent role both during early differentiation of nascent mesoderm as well as during pluripotency. Interestingly, the expression of Carmn in mice modulates, but not in human, the expression of other cardiac-associated lncRNAs, such as Braveheart, located downstream of Carmn, within the same genomic locus. Mechanistically, Carmn acts in trans by directly interacting with the polycomb repressive complex 2 (PRC2) through SUZ12, a core protein of this complex, and a EZH2, with mediates methylation of lysine 27 on histone 3, a repressive mark epigenetic that promote the correct cardiac differentiation. The role of this repressive complex in the development of heart has been well reported [bib_ref] Chromatin modifications remodel cardiac gene expression, Mathiyalagan [/bib_ref]. Thus Carmn exerts an epigenetic function altering the transcriptional landscape by interacting with repressive complex chromatin remodeling. Moreover, Carmn is necessary for cardiomyocyte homeostasis, maintaining a differentiated cardiac fate in mature cardiomyocytes [bib_ref] CARMEN, a human super enhancer-associated long noncoding RNA controlling cardiac specification, differentiation..., Ounzain [/bib_ref].
## (b). braveheart
Braveheart is a mouse specific heart-associated lncRNA that plays a pivotal role during cardiac development. Braveheart acts as a key regulator in cardiac lineage commitment and it is required for proper cardiac gene expression in mice. Braveheart acts upstream of Mesp1, and it is nonetheless required for its activation within the same genetic pathway. Depletion of Braveheart results in failure of activation of key cardiac factors necessary for correct heart development and cardiomyocyte differentiation. These evidences suggest that Braveheart is required to active Mesp1-driven gene expression program and to promote cardiac cell fate from nascent mesoderm. Moreover, Braveheart is necessary for fetal and neonatal cardiomyocyte homeostasis and the maintenance of cardiac fate of its [bib_ref] Braveheart, a long noncoding RNA required for cardiovascular lineage commitment, Klattenhoff [/bib_ref]. Similarly evidences have also been demonstrated that Braveheart is necessary for ESC to acquire cardiac lineage commitment and differentiation into cardiomyocytes [bib_ref] Braveheart, a long noncoding RNA required for cardiovascular lineage commitment, Klattenhoff [/bib_ref].
Recently, Xue et al. [bib_ref] A G-rich motif in the lncRNA braveheart interacts with a zinc-finger transcription..., Xue [/bib_ref] have determined the secondary structure of Braveheart and showing that Braveheart adopts a high modular structure with a 5' AGIL motif that is required for correct mode of action of Braveheart. Importantly, this motif is necessary to for cardiovascular lineage commitment and proper ESC differentiation. These authors also demonstrated specific interactions between Braveheart and zinc-finger TF CNBP a negative regulator of the cardiac development program repressing the CM differentiation. Thus, it seems that Braveheart act as antagonist of CNBP to promote cardiovascular lineage commitment [bib_ref] A G-rich motif in the lncRNA braveheart interacts with a zinc-finger transcription..., Xue [/bib_ref].
## (c). fendrr
Fendrr is differentially and transiently expressed at the caudal end of the nascent lateral plate mesoderm, being necessary for the correct development of tissues derived from it, especially the heart and the body walls. Fendrr is located 1250 base pairs upstream of Foxf1 and is co-expressed with this transcription factor. Foxf1 is of vital importance for the proper differentiation of lateral mesoderm in splanchnic mesoderm and somatic mesoderm [bib_ref] The forkhead transcription factor Foxf1 is required for differentiation of extra-embryonic and..., Mahlapuu [/bib_ref]. Gene targeting approach has showed that the lack of Fendrr carries embryonic lethality at E13.75, characterized by an incorrect heart function, blood accumulated in the right heart chambers and a critical decrease in the thickness of the ventral body walls. The incorrect function of the heart is explained by a hypoplasia of the myocardium that leads to the development impair of ventral heart walls and interventricular septum that are too thin to be able to withstand the blood pressure. Mechanistically, Fendrr acts as epigenetic regulatory element by establishing a ratio of repressive and activate marks in the promotors of pivotal transcription factors involved in the mesoderm development, such as Foxf1, Pitx2 and Irx3. Such epigenetic regulation is provided through interaction with chromatin remodeling complexes, the histone-modifying Polycomb repressive complex (PRC2) and Trithorax group/MLL protein complex (TrxG/MLL), respectively. The establishment of this ratio determinates the patterns of expression of Fendrr target genes in the nascent lateral plate mesoderm and in their descendants of by setting long-term epigenetic marks. In silico approaches have showed the existence of 40-nucleotide stretch in the Fendrr structure, which is able to interact with Foxf1 and Pitx2 promoters, thereby Fendrr seem to can be binds directly to those promoters via the formation of a dsDNA:RNA triplex structure. However this interaction needs to be experimentally validated [bib_ref] The tissue-specific lncRNA Fendrr is an essential regulator of heart and body..., Grote [/bib_ref] [bib_ref] The long non-coding RNA Fendrr links epigenetic control mechanisms to gene regulatory..., Grote [/bib_ref].
## (d). alien
Alien is co-expressed in vascular progenitor cells derived from allantoides and lateral plate mesoderm along with genes involved in skeletal muscle development and heart morphogenesis. Gene targeting approaches have showed that the loss of Alien results in impaired development of several mesodermal derivatives. Among them defective vascular patterning and cardiac chamber formation is reported . These observations suggest that Alien specially acts as pivotal regulator in the cardiovascular development by exerting a function in an early stage of cardiovascular differentiation common to both vascular and cardiac progenitors. The molecular aspects of Alien function are unknown to date and thus it will be interesting to study the specific role of this lncRNA in the cardiovascular commitment .
On the other hand, Alien participates in the endoderm differentiation regulating positively the transcription of FOXA2, an important regulator of endoderm development, by facilitating SMAD2/3 recruitment to the FOXA2 promoter [bib_ref] The lncRNA DEANR1 facilitates human endoderm differentiation by activating FOXA2 expression, Jiang [/bib_ref]. Thus Alien acts a versatile RNA molecule during the cardiovascular development.
## (e). upperhand
A recent study has showed that transcription of a promoter-associated lncRNA located near to Hand2 is necessary for the expression of this transcription factor and proper heart development. This lncRNA is known as Upperhand, is located 150 bases pairs upstream of Hand2 and shares a bidirectional promoter with this transcription factor. Interestingly, Upperhand locus contains a Hand2 associated cardiac enhancer within an intron. Upperhand is co-expressed in a temporal and tissue-specific pattern along with Hand2 during embryonic development. Upperhand expression is enriched in the cytoplasm. The function of this lncRNA has been studied by gene targeting approach. Upperhand knockout (KO) mice display similar phenotype asHand2KO mice, characterized by a development impairment of right ventricular chamber and embryonic lethality [bib_ref] The basic helix-loop-helix transcription factor, dHAND, is required for vascular development, Yamagashi [/bib_ref]. Upperhand KO embryo failed to establish H3K27ac marks in the Hand2-Uph locus and binding of GATA4 to the Hand2 cardiac enhancer is also impaired. This molecular interaction is required for the activation of Hand2 cardiac enhancer and thus for Hand2 transcription [bib_ref] A GATA-dependent right ventricular enhancer controls dHAND transcription in the developing heart, Mcfadden [/bib_ref] [bib_ref] Dynamic GATA4 enhancers shape the chromatin landscape central to heart development and..., He [/bib_ref]. Chromatin immunoprecipitation analyses (ChIP) in Upperhand KO embryo have showed that both the loss of the H3K27ac marks and the lack of Gata4 interaction with the Hand2 cardiac enhancer negatively affectsHand2 transcription, preventing the RNA polymerase II elongation within the Hand2 locus. Interestingly, in vitro approaches in HL-1 have showed that the mature Upperhand transcripts are not required for Hand2 expression suggesting that is the Upperhand transcription, the responsible of Hand2 expression. Thus, these findings suggest that Upperhand transcription is required to the Hand2 expression by participating in the establishing of the H3K27 marks and in the binding of Gata4 to the Hand2 associated cardiac enhancer both necessary to the proper hand2 transcription by the RNA polymerase II [bib_ref] Transcription of the non-coding RNA upperhand controls Hand2 expression and heart development, Anderson [/bib_ref].
## (f). lncrna-uc.167
A screening of transcriptome of patients with ventricular septal defect (VSD), has showed the differential expression of a considerable number of lncRNAs [bib_ref] Dynamic GATA4 enhancers shape the chromatin landscape central to heart development and..., He [/bib_ref]. Among them Song et al., [bib_ref] Integrated analysis of dysregulated lncRNA expression in fetal cardiac tissues with ventricular..., Song [/bib_ref] focused on LncRNA-uc.167, given its prominent expression in VSD patients. LncRNA-uc.167 is located in the opposite strand of Mef2c and is well conversed between species. The expression of both follows an inverse pattern throughout cardiac development and also during the process of P19 cell differentiation into cardiomyocytes. Moreover, the overexpression of lncRNA-uc.167 results in inhibition of Mef2c and absence of differentiation of P19 characterized by a higher level of apoptosis and a slower proliferation rate. The effects of lncRNA-uc.167 overexpression are partially reduced by Mef2c overexpression, suggesting a functional relationship between them. Thus, those observations suggest that LncRNA-uc.167 can participate in Mef2c signaling during heart development [bib_ref] LncRNA-uc. 167 influences cell proliferation, apoptosis and differentiation of P19 cells by..., Song [/bib_ref] , yet further experiments are required to fully understand the molecular mechanisms behind lncRNA-uc.167 and Mef2c interaction.
## (g). pancr
Genome-wide association studies (GWAS) have associated several SNPs (risk variants) in distinct genetic loci with atrial fibrillation, including 4q25 genomic locus [bib_ref] Variants conferring risk of atrial fibrillation on chromosome 4q25, Gudbjartsson [/bib_ref] [bib_ref] Meta-analysis identifies six new susceptibility loci for atrial fibrillation, Ellinor [/bib_ref] where the homeobox transcription factor Pitx2 is located. Pitx2 is a pivotal player in embryonic left-right asymmetry pathway and cardiac development . In this context, Gore-Panter et al. [bib_ref] PANCR, the PITX2 Adjacent noncoding RNA, is expressed in human left atria..., Gore-Panter [/bib_ref] exploring the possible relationship between Pitx2 and these AF risk variants, identified a long intergenetic non-coding RNA adjacent to Pitx2, dubbed as Pancr. Pancr is expressed in the adult left atrium and in lower levels in the adult eye, and shows a coordinate expression with Pitx2c, during the differentiation of cardiomyocytes regulating positively the expression of Pitx2c mRNA by a yet unknown mechanism. Interestingly, Pancr have been reported in human tissues but no orthologues are found in other mammalian species such as mice. Thereby, Pancr seems to be a human specific lincRNA [bib_ref] PANCR, the PITX2 Adjacent noncoding RNA, is expressed in human left atria..., Gore-Panter [/bib_ref]. Since, the regulation of Pitx2c by Pancr it will be interesting to explore the role of this lincRNA in the cardiac development to provide more information of left-right asymmetry pathway.
## Long non-coding rnas in cardiac diseases
The role of long non-coding RNAs in disease is recently emerging. Over the last few years, the number of reports that have associated lncRNAs differential expression with some cardiac pathology has considerably increased [bib_ref] Non-coding RNAs in development and disease: background, mechanisms, and therapeutic approaches, Beermann [/bib_ref]. Several transcriptomic studies in different species, using deep sequencing, have identified that multiple lncRNAs are deregulated in distinct cardiac pathologies [fig_ref] Figure 3: Schematic representation of the distinct cardiac pathophysiological conditions and the distinct long... [/fig_ref] , particularly in acute myocardium infarction [bib_ref] Regulatory non-coding RNAs in acute myocardial infarction, Guo [/bib_ref] [bib_ref] Long Noncoding RNAs in Patients With Acute Myocardial InfarctionNovelty and Significance, Vausort [/bib_ref] , heart failure [bib_ref] Circular RNAs in heart failure, Devaux [/bib_ref] [bib_ref] Long noncoding RNA dysregulation in ischemic heart failure, Greco [/bib_ref] [bib_ref] Heart failure: Pilot transcriptomic analysis of cardiac tissue by RNA-sequencing, Schiano [/bib_ref] , cardiac fibrosis [bib_ref] The long noncoding RNA Wisper controls cardiac fibrosis and remodeling, Micheletti [/bib_ref] and atrial fibrillation [bib_ref] Altered long non-coding RNA expression profile in rabbit atria with atrial fibrillation:..., Li [/bib_ref] [bib_ref] Long non-coding RNA expression profile in atrial fibrillation, Ruan [/bib_ref]. These studies have shown that lncRNAs are important players in the maintenance of cellular homeostasis and in disease processes, being their function necessary to correct physiological status.
Several studies have indicated the importance of lncRNAs in cardiac hypertrophy. For example, Myheart, a lncRNA located in the murine myosin heavy chain 7 genomic locus, prevents cardiomyocyte hypertrophy by sequestering BRG1-BAF complex and therefore avoiding that this complex can bind to target genes and induce an hypertrophy response [bib_ref] A long noncoding RNA protects the heart from pathological hypertrophy, Han [/bib_ref]. Hrcr, a cardiac circular lncRNA plays a protective role in the cardiac hypertrophy too, acting a decoy lncRNA by sequestering miR-223, considered as pro-hypertrophy factor [bib_ref] A circular RNA protects the heart from pathological hypertrophy and heart failure..., Wang [/bib_ref]. Also, H19 and his encoded miR-675, play a protective role in hypertrophy cardiac by modulating cardiac CaMKIIδ expression in the hypertrophy response [bib_ref] The H19 long noncoding RNA is a novel negative regulator of cardiomyocyte..., Liu [/bib_ref]. Other lncRNAs promote a pro-hypertrophy response in cardiomyocytes, such as Chast, Chaer, Chrf or Miat. Chast, cardiac hypertrophy-associated transcript, is overexpressed by hypertrophy stimuli and its overexpression, independently of pro-hypertrophic factor, is sufficient to activate the hypertrophy response in the cardiomyocytes both in vivo and in vitro [bib_ref] Long noncoding RNA Chast promotes cardiac remodeling, Viereck [/bib_ref]. Chaer, cardiac-hypertrophy-associated epigenetic regulator, promotes the hypertrophy by interacting directly with PRC2 and inhibiting the formation of transcriptional silent chromatin complex of pro-hypertrophic factors [bib_ref] The long noncoding RNA Chaer defines an epigenetic checkpoint in cardiac hypertrophy, Wang [/bib_ref] Chrf, cardiac-hypertrophy-response factor, acts a sponge of miR-489, which target is mRNA of pro-hypertrophy factor Myd88, thereby promote the hypertrophy response by avoiding that miRNA can degrade Myd88 transcripts [bib_ref] The Long Noncoding RNA CHRF Regulates Cardiac Hypertrophy by Targeting miR-489 Novelty..., Wang [/bib_ref]. Finally, Miat, myocardial infarction-associated transcript, promote partly cardiac hypertrophy by sponging miR-150, an important miRNA with suppressor effect in the cardiac hypertrophy [bib_ref] Lncrna miat enhances cardiac hypertrophy partly through sponging mir-150, Zhu [/bib_ref]. Acute myocardial infarction, ultimately leading to heart failure, is frequently associated with a progressive accumulation of fibrotic depositions, i.e., cardiac fibrosis [bib_ref] The long noncoding RNA Wisper controls cardiac fibrosis and remodeling, Micheletti [/bib_ref]. Differentially expressed lncRNAs have been identified in this fibrotic process. Wisper, a cardiac-fibroblast enriched lncRNA conserved between mouse and human, is necessary for the development of cardiac fibrosis. The expression of Wisper is required for survival and transdifferentiation of cardiac fibroblasts and maintaining the correct pro-fibrotic gene regulatory network. GapmeRs-mediated attenuation in vivo reduces the fibrotic process after myocardium infarct in mice, pointing to Whisper as potential therapeutic target [bib_ref] The long noncoding RNA Wisper controls cardiac fibrosis and remodeling, Micheletti [/bib_ref]. Furthermore, two cardiac-fibroblast enriched lncRNAs modulate the fibrotic process through MMP2 (matrix metalloproteinase-2) modulation, a pro-fibrotic factor expressed in cardiac fibroblasts [bib_ref] Inhibition of the Cardiac Fibroblast-Enriched lncRNA Meg3 Prevents Cardiac Fibrosis and Diastolic..., Piccoli [/bib_ref]. Meg, a lncRNA conserved in human and mice, promote cardiac fibrosis, upregulating MMP2 expression, while Gas5 act as negative effector by sponging miR-21, which, in turn, positively regulates the fibrotic process by PTEN/MMP2 pathway [bib_ref] LncRNA GAS5 controls cardiac fibroblast activation and fibrosis by targeting miR-21 via..., Tao [/bib_ref]. Other examples of cardiac pro-fibrotic enriched sponges lncRNAs is exemplified by Miat, acting upon miR-24 and H19, acting upon miR-455 [bib_ref] MIAT is a pro-fibrotic long non-coding RNA governing cardiac fibrosis in post-infarct..., Qu [/bib_ref] [bib_ref] Long noncoding RNA H19 acts as a competing endogenous RNA to mediate..., Huang [/bib_ref] , respectively.
Interestingly, although the role of lncRNAs in cardiac arrhythmogenesis scarce, two different studies in humans and rabbits, respectively, have identified a subset of differentially expressed lncRNAs in atrial fibrillation, yet their functional roles remains to be established [bib_ref] Altered long non-coding RNA expression profile in rabbit atria with atrial fibrillation:..., Li [/bib_ref] [bib_ref] Long non-coding RNA expression profile in atrial fibrillation, Ruan [/bib_ref].
In addition to those functional roles in adult cardiac pathophysiology, several reports are also emerging in congenital heart diseases. Song et al. [bib_ref] Integrated analysis of dysregulated lncRNA expression in fetal cardiac tissues with ventricular..., Song [/bib_ref] have shown the different expression of up to 1500 lncRNAs between normal hearts and hearts from fetuses with ventricular septal defect, a common congenital heart disease. The deregulation of these long non-coding RNAs supports the possible involvement of long non-coding in the development of CHD, however is necessary a greater knowledge about the biology of lnRNAs to understand the role of these in this kind of disease [bib_ref] Integrated analysis of dysregulated lncRNA expression in fetal cardiac tissues with ventricular..., Song [/bib_ref].
# Conclusions
Emerging evidences suggest an important role of long non-coding RNA in cardiac development by regulating different cardiac gene network pathways. In this line of thinking, knockdown of distinct cardiac-enriched lncRNAs results in embryonic lethality, reflecting the pivotal role of lncRNAs in this process. However, understanding of the functional roles of long non-coding RNAs is still in its infancy. In the next coming years we will witness further insights into the diversity of regulatory roles of long non-coding RNAs and their interactions with epigenetic, transcriptional and post-transcriptional regulatory layers, not only during cardiovascular development, but also in cardiovascular pathology. In this context, several lncRNAs have been point out as important regulators of cardiac pathological processes such as cardiac hypertrophy. However, scarce information is available in electrophysiological disorders such as Brugada syndrome, LQT and SQT syndromes or arrhythmogenic right ventricular dysplasia (ARVD). Importantly, lncRNAs are identified in peripheral blood samples, opening the possibility of serving as diagnostic biomarkers of different cardiac disease [bib_ref] Long Noncoding RNAs in Cardiovascular Pathology, Bär [/bib_ref]. Therefore, the identification of lncRNAs and the study of their functional roles, both in development and disease, highlight the important of non-coding RNAs as key regulatory elements.
[fig] Figure 1: Schematic representation of the distinct classes of RNA molecules. [/fig]
[fig] Figure 3: Schematic representation of the distinct cardiac pathophysiological conditions and the distinct long non-coding RNAs described to date. [/fig]
|
ycf1, the most promising plastid DNA barcode of land plants
A DNA barcode is a DNA fragment used to identify species. For land plants, DNA fragments of plastid genome could be the primary consideration. Unfortunately, most of the plastid candidate barcodes lack species-level resolution. The identification of DNA barcodes of high resolution at species level is critical to the success of DNA barcoding in plants. We searched the available plastid genomes for the most variable regions and tested the best candidates using both a large number of tree species and seven well-sampled plant groups. Two regions of the plastid gene ycf1, ycf1a and ycf1b, were the most variable loci that were better than existing plastid candidate barcodes and can serve as a barcode of land plants. Primers were designed for the amplification of these regions, and the PCR success of these primers ranged from 82.80% to 98.17%. Of 420 tree species, 357 species could be distinguished using ycf1b, which was slightly better than the combination of matK and rbcL. For the well-sampled representative plant groups, ycf1b generally performed better than any of the matK, rbcL and trnH-psbA. We concluded that ycf1a or ycf1b is the most variable plastid genome region and can serve as a core barcode of land plants.
D NA barcoding is a technique used to identify unknown materials of known species based on DNA sequences of standard genome regions (i.e. DNA barcodes) [bib_ref] Biological identifications through DNA barcodes, Hebert [/bib_ref]. Before this technique can be fully utilised, a barcode that is variable enough to discriminate between species of interest and a reliable barcode reference library must be made available. The first of these requirements is more significant as it is relatively easy to build a reference library if DNA materials already exist. Ideally, a barcode should be variable enough to resolve closely related species and short enough for easy experimental manipulation and low cost. The sequences flanking the barcode should be conservative enough to facilitate the design of universal primers for high PCR and sequencing success. Presently, the candidate barcodes are selected from markers used in molecular systematics due to the limited knowledge of genome variations; we know little about mitochondrial genomes, much less nuclear genomes. Fortunately, by the end of 2013, 429 plastid genomes were sequenced, and perhaps they contain a useful plant barcode.
In the past decade, several plastid genome regions such as atpF-H, matK, psbK-I, rbcL, ropC1, rpoB, trnH-psbA, and trnL-F that are frequently used in plant molecular systematics have been extensively evaluated [bib_ref] Molecular phylogeny and intra-and intercontinental biogeography of Calycanthaceae, Zhou [/bib_ref] [bib_ref] Resolving an ancient, rapid radiation in Saxifragales, Jian [/bib_ref] [bib_ref] Phylogenetics of seed plants -an analysis of nucleotidesequences from the plastid gene..., Chase [/bib_ref] , and the rbcL and matK genes were selected as core plant barcodes by the CBOL Plant Working Group 5 . Unfortunately, rbcL seems to be more suitable for barcoding lower plants than for seed plants [bib_ref] Discriminating plants using the DNA barcode rbcLb: an appraisal based on a..., Dong [/bib_ref]. One of the most highly variable regions of the plastid genome, trnH-psbA, undergoes chromosomal rearrangements (e.g., inversions and microsatellite loci) and contains a limited number of informative sites due to its short length [bib_ref] Intraspecific inversions pose a challenge for the trnH-psbA plant DNA barcode, Whitlock [/bib_ref] [bib_ref] Utility of the trnH-psbA intergenic spacer region and its combinations as plant..., Pang [/bib_ref]. Recently, found that two regions of the plastid gene ycf1 were very variable in flowering plants [bib_ref] Highly variable chloroplast markers for evaluating plant phylogeny at low taxonomic levels..., Dong [/bib_ref].
As the second largest gene in the plastid genome, ycf1 encodes a protein of approximately 1,800 amino acids. Recent experiments showed that ycf1 is essential for plant viability and encodes Tic214, a vital component of the Arabidopsis TIC complex [bib_ref] Uncovering the protein translocon at the chloroplast inner envelope membrane, Kikuchi [/bib_ref]. Within the plastid genome, ycf1 spans the small single copy (SSC) and the inverted repeat (IR) regions. The section of ycf1 in the IR region is short (less than one kilobase long) and conserved. In contrast, the section of ycf1 in the SSC region has high sequence variability in seed plants. This region of the ycf1 gene is more variable than matK in most taxa investigated thus far [bib_ref] Chloroplast genome sequence of the moss Tortula ruralis: gene content, polymorphism, and..., Oliver [/bib_ref] [bib_ref] The evolution of chloroplast genes and genomes in ferns, Wolf [/bib_ref] and has been used in molecular systematics at low taxonomic levels [bib_ref] Phylogenetic utility of ycf1 in orchids: a plastid gene more variable than..., Neubig [/bib_ref] [bib_ref] Phylogenetic relationships of Pinus subsection Ponderosae inferred from rapidly evolving cpDNA regions, Gernandt [/bib_ref] [bib_ref] The South American radiation of Lepechinia (Lamiaceae): phylogenetics, divergence times and evolution..., Drew [/bib_ref] [bib_ref] Phylogeny of Opuntia s.s. (Cactaceae): Clade delineation, geographic origins, and reticulate evolution, Majure [/bib_ref] [bib_ref] Testing the monophyly and placement of Lepechinia in the Tribe Mentheae (Lamiaceae), Drew [/bib_ref]. Two regions within ycf1, ycf1a and ycf1b, have been predicted to have the highest nucleotide diversity (p) at the species level within angiosperm plastid genomes [bib_ref] Highly variable chloroplast markers for evaluating plant phylogeny at low taxonomic levels..., Dong [/bib_ref].
Because ycf1 is too long (5709 bp in Nicotiana tabacum) and too variable to permit the design of universal primers [bib_ref] Highly variable chloroplast markers for evaluating plant phylogeny at low taxonomic levels..., Dong [/bib_ref] , it has received little attention for DNA barcoding or molecular systematic purposes at low taxonomic levels; however, the high variability of ycf1 indicates its potential value in DNA barcoding of land plants. This paper reports the results of our evaluations of the ycf1 gene for DNA barcoding purposes. We (1) generated primers for PCR amplification of ycf1a and ycf1b, and (2) tested the performance of ycf1a or ycf1b in discriminating between species compared with the plastid genes rbcL, matK and trnH-psbA.
# Results
The ycf1a and ycf1b regions are the most variable regions of the ycf1 gene. According to , angiosperms have two highly variable regions in the ycf1 gene, ycf1a and ycf1b. These results were verified using 136 genomes belonging to 27 genera (see the electronic supplementary material, . The exact positions of ycf1a vary slightly among plant groups, while the positions of ycf1b are quite consistent, especially in seed plants.
Primer universality. The ycf1 gene is too variable for the design of universal primers for all land plants; however, we were able to design universal primers for Bryophytes, Monilophytes, gymnosperms and angiosperms . The ycf1mF/ycf1mR primer pair amplified the expected fragments in 32 of 34 (94.12%) Bryophyte families (see the electronic supplementary material, .
The ycf1b regions of Monilophytes are too divergent to contain satisfactory primer sites; in contrast, the sequences flanking ycf1a are relatively conserved and we designed three primers (one forward and two reverse primers, . Using these primers, ycf1a fragments were amplified from 82.80% of the samples belonging to 93 genera in 42 families (see the electronic supplementary material, .
The ycf1gF/ycf1gR primer pair was designed for conifers and cycads (see the electronic supplementary material, . The divergence of ycf1 sequences in gymnosperms is remarkable, and it is difficult to identify a single pair of universal primers. Consequently, the species-rich conifers required special attention.
The ycf1bF/ycf1bR primer pair worked the best for angiosperms. The PCR success reached 98.17% samples from 219 genera in 217 families (see the electronic supplementary material, . Moreover, critical mutations at the 39 end that would cause amplification failure were observed in some taxa. To minimise PCR failures, some substitutions to the universal primers listed in were generated for 131 families (see the electronic supplementary material, Table S3) in the event that the universal primers failed.
Performance of ycf1 in identifying BBG woody plants. In total, we obtained 1352 sequences of matK, rbcL and ycf1 from 420 woody plant species representing 179 genera in 76 families. The sequence recoveries for rbcLb, matK, and ycf1b were 99.18%, 91.43%, and 85.31%, respectively (see the electronic supplementary material, . Poor PCR amplification of ycf1b was encountered for Lonicera (13 samples) and Berberis (15 samples). When samples from these two genera were excluded, the ycf1b sequence recovery reached 90.48%. Approximately 79.80% (391) of all samples contained all three sequenced fragments. Therefore, two types of datasets were generated for accurate assessments. One comprised all sequences for each marker, and the other comprised the samples with all three markers. Three two-barcode combinations and one three-barcode combination were also tested for each dataset type.
For the datasets of all sequences, ycf1b showed the highest discriminatory power of the three barcodes, discriminating 73.97% of all the species. The barcodes rbcLb (58.02%) and matK (57.56%) had similar discriminatory power , much lower than ycf1b. The performance of ycf1b was slightly better than the combination of rbcLb and matK (71.31%). Combining ycf1b with either rbcLb or matK increased the discriminatory power to 81.39% and 79.83%, respectively, and the use of all three candidates increased the discriminatory power to 86.33%.
To eliminate the possible effects of sample inequity on the different markers, we used datasets of 391 samples with all three markers to increase the reliability of direct comparisons of species resolution among the markers. Similar patterns were observed for these datasets. ycf1b had the highest species resolution among the three markers at 71.87%, with 54.99% resolution for matK and 55.50% resolution for rbcLb .
Performance of ycf1 in identifying species within seven wellsampled plant groups. Using matK, rbcLb and trnH-psbA as controls, seven relatively well-sampled plant groups were chosen to test the discriminatory powers of ycf1b (see the electronic supplementary material, . The lengths of matK, rbcLb and ycf1b varied among the plant groups. The poly-A/T tracts longer than 9 bp were observed in the trnH-psbA sequences from six of the seven plant groups, and non-homologous inversions of 2 to 21 bp were observed in five groups. The nucleotide diversity (p) of ycf1b was the highest among the four markers in the five plant groups (see the electronic supplementary material, -S12).
Using the distance method, ycf1b exhibited the highest discriminatory power among the four markers in six of the seven plant groups . Combinations of matK, rbcLb and trnH-psbA did not typically increase the discriminatory rates; in contrast, the combination of ycf1b with either matK, rbcLb or trnH-psbA increased the percentage of discrimination success by varying degrees for five of the seven plant groups , and see the electronic supplementary material, , -S12). Without ycf1b, matK and rbcLb did not contribute for either Armeniaca or Paeonia (see the electronic supplementary material, .
# Discussion
Which gene can serve as a barcode of plants? The question of which gene can serve as a barcode for plants remains to be answered even though the combination of matK and rbcL has been suggested as the core barcode of land plants. Marker selection is critically important as time and money can be saved by the quick identification of a suitable barcode. considerations. The definition of DNA barcode 1 implies that the first consideration should be species resolution. The issue of a barcoding gap is not an important issue in DNA barcoding; rather, it is an important question in systematics and taxonomy. Many DNA barcoders are also systematists and the two issues have been investigated simultaneously. If a species is correctly circumscribed, the optimal barcode will show the highest probability to distinguish it from other congeners regardless of the existence of a barcoding gap.
Unfortunately, it is difficult to predict when such an ideal barcode might be found. We can now attempt to identify an improved barcode while also using the old ones for some time and then eventually substitute them.
Imperfections of the existing candidates. Many candidate barcodes that cannot withstand tests of universality are only suitable for specific plant groups and are falling from common use. Only four continue to receive some use: the chloroplast genes rbcL, matK and trnH-psbA, and the nuclear internal transcribed spacer (ITS). The rbcL gene was suggested as a core barcode not as a result of its power in barcoding species, but rather its historical popularity [bib_ref] Selection of candidate coding DNA barcoding regions for use on land plants, Ford [/bib_ref] and possibly experimental ease. The rbcL gene has been subject to considerable criticism as a barcode for seed plants [bib_ref] Testing DNA barcoding in closely related groups of Lysimachia L. (Myrsinaceae), Zhang [/bib_ref] [bib_ref] Phylogenetic relationships and species delimitation in Pinus section Trifoliae inferrred from plastid..., Hernandez-Leon [/bib_ref] [bib_ref] Barcoding success as a function of phylogenetic relatedness in Viburnum, a clade..., Clement [/bib_ref] ; however, it may be useful for lower plants [bib_ref] Discriminating plants using the DNA barcode rbcLb: an appraisal based on a..., Dong [/bib_ref]. Although they are advocates of matK, Ford and his colleagues acknowledged that matK had only modest performance [bib_ref] Selection of candidate coding DNA barcoding regions for use on land plants, Ford [/bib_ref]. The matK gene was not among the top 10 most variable species-level markers [bib_ref] Highly variable chloroplast markers for evaluating plant phylogeny at low taxonomic levels..., Dong [/bib_ref]. Although matK is more useful than rbcL 6 , in most cases it is not the only useful species-level barcode [bib_ref] Barcoding success as a function of phylogenetic relatedness in Viburnum, a clade..., Clement [/bib_ref] [bib_ref] DNA barcode identification of Podocarpaceae-The second largest conifer family, Little [/bib_ref].
The use of trnH-psbA had been well evaluated 8 , and although trnH-psbA is more variable than either matK or rbcL, several problems limit its widespread adoption. The extensive prevalence of inversions and insertions within species, long polystructures that cause sequencing difficulties, and relatively short lengths prevent its use as a core barcode. After a few years of disfavour, the nuclear ribosomal ITS first proposed by Kress et al. [bib_ref] Use of DNA barcodes to identify flowering plants, Kress [/bib_ref] has again become accepted as a core barcode, as exemplified by studies using large data [bib_ref] Validation of the ITS2 Region as a Novel DNA Barcode for Identifying..., Chen [/bib_ref] [bib_ref] Use of ITS2 Region as the Universal DNA Barcode for Plants and..., Yao [/bib_ref] [bib_ref] ITS1: a DNA barcode better than ITS2 in eukaryotes?, Wang [/bib_ref]. With the exception of ycf1, ITS has been shown to have unparalleled species resolution compared with the candidate barcodes proposed thus far; however, it suffers from incomplete concerted evolution in some cases and from experimental complexity for species of hybrid origin.
Is ycf1 good enough to be a barcode of land plants? It is easy to identify the most variable regions in certain taxa at the species level; however, it is difficult, and even unlikely, to identify such regions in all taxa. With a few isolated exceptions, the ycf1a and ycf1b regions are perhaps the most variable regions in most taxa [bib_ref] Highly variable chloroplast markers for evaluating plant phylogeny at low taxonomic levels..., Dong [/bib_ref]. A barcode should be chosen because it shows the highest species resolution in most cases rather than in specific cases. The ycf1 gene meets this criterion and can serve as a barcode of land plants.
Currently, there are two applications for DNA barcoding. One application is for flora, and the other is for specific taxa. The first application is exemplified by barcoding trees in large ecological plots [bib_ref] Plant DNA barcodes and a community phylogeny of a tropical forest dynamics..., Kress [/bib_ref] [bib_ref] Identification of Amazonian trees with DNA barcodes, Gonzalez [/bib_ref] [bib_ref] Exploring tree-habitat associations in a Chinese subtropical forest plot using a molecular..., Pei [/bib_ref] [bib_ref] How effective are DNA barcodes in the identification of African rainforest trees?, Parmentier [/bib_ref] , and a similar example was given in this study. Our test example differs from the barcoding of local flora in that more species are from the same genera, which may show reduced discriminating power. The ycf1b performed satisfactorily when compared to the core barcode combination of matK and rbcL. The second application is becoming increasingly popular for barcoding of medicines, teas, and foods, etc [bib_ref] Commercial teas highlight plant DNA barcode identification successes and obstacles, Stoeckle [/bib_ref] [bib_ref] DNA barcode authentication of saw palmetto herbal dietary supplements, Little [/bib_ref]. As ycf1 was not previously identified as a potential barcode, comparisons are not currently available. In this paper, seven examples were used to show the superior performance of ycf1b compared with other barcodes for distantly related plant groups. Consequently, ycf1b is expected to be suitable for an extensive group of plants.
The ycf1 gene was slowly identified for its potential use as a barcode most likely due to its length and lack of universal primers; however, a few phylogenetic applications had been found for Pinaceae [bib_ref] Phylogenetic relationships of Pinus subsection Ponderosae inferred from rapidly evolving cpDNA regions, Gernandt [/bib_ref] [bib_ref] Increasing phylogenetic resolution at low taxonomic levels using massively parallel sequencing of..., Parks [/bib_ref] , Orchidaceae 13 , Lamiaceae 15,36 and Prunus 37 .
One major concern for the use of ycf1 as a barcode is the absence of ycf1 in some taxa. The ycf1 gene is functional and is not commonly lost [bib_ref] The evolution of the plastid chromosome in land plants: gene content, gene..., Wicke [/bib_ref]. It was erroneously reported to have been lost from Acorales, Poales, and Passiflora; however, it is only absent from Poaceae [bib_ref] Analysis of Acorus calamus chloroplast genome and its phylogenetic implications, Goremykin [/bib_ref] [bib_ref] Implications of the plastid genome sequence of Typha (Typhaceae, Poales) for understanding..., Guisinger [/bib_ref].
Experimental considerations for ycf1 use. Primer universality is an important criterion for an ideal DNA barcode. The primers for rbcLb were recently optimised 6 . Although matK primers had been the subject of several studies [bib_ref] Broad-scale amplification of matK for DNA barcoding plants, a technical note, Dunning [/bib_ref] [bib_ref] High universality of matK primers for barcoding gymnosperms, Li [/bib_ref] [bib_ref] New universal matK primers for DNA barcoding angiosperms, Yu [/bib_ref] , obtaining matK fragments from ferns and mosses continues to be a challenge. At the beginning of our work, hardly any ycf1 sequences had been deposited in GenBank; consequently, we had to generate ycf1 sequences to facilitate primer design. The amplification successes of the ycf1 primers used in this study were quite satisfactory: 98.17% for angiosperms, 90.91% for gymnosperms, 82.80% for Monilophytes, and 94.12% for Bryophytes. Amplification in seed plants would not be very difficult using the universal primers. The primer performances for Monilophytes and gymnosperms were relatively poor due to the significant divergence within these groups. To minimise PCR failure, taxon-specific primers were generated for focal taxa (see the electronic supplementary material, .
# Methods
Reconfirming the variability of the ycf1a and ycf1b regions. Very few ycf1 sequences have been deposited in GenBank. We downloaded 144 whole plastid or chloroplast genomes from GenBank (see the electronic supplementary material, . The ycf1 sequences were extracted from two or more plastid genomes from the same genera of land plants (sorted into four groups: Bryophytes, Monilophytes, gymnosperms and angiosperms), aligned using MAFFT and manually adjusted with Se-Al 2.0 as necessary. The nucleotide diversity (p) was computed using the R package with a 600 bp sliding-window and a 50 bp step size. The averages within each group represent ycf1 variability and the most variable regions were roughly identified.
Primer design. All of the ycf1 sequences deposited in GenBank were downloaded or extracted from the plastid genomes. The sequences were sorted into four groups as described above, aligned in each group using Clustal X ver. 2.0 and then adjusted using Se-Al 2.0. These sequences served as initial templates for the design of several primer pairs spanning the roughly identified regions for each group using Primer Premier 5.0 software (Premier Biosoft International, Palo Alto, CA). The successfully amplified fragments were sequenced, and together with those from GenBank, served as templates for universal primer design and hypervariable region positioning. .
Two types of tests were used to assay the performance of ycf1a, ycf1b and other markers of the same genome. The first test used almost entirely woody plant species cultivated in the Beijing Botanical Garden (BBG) of the Chinese Academy of Sciences (CAS). We used 490 samples belonging to 420 species in 76 families, including 53 samples belonging to 48 species in 5 gymnosperm families and 437 samples belonging to 372 species in 71 angiosperm families (see the electronic supplementary material, . The use of BBG materials represents a common practice of using DNA barcoding for the identification of plant materials from local flora such as from largescale community plots.
The second test used the well-sampled groups of seed plants (see the electronic supplementary material, , namely gymnosperm Pinus (Pinaceae); basal angiosperm Calycanthaceae; monocotyledon Iris (Iridaceae); Saxifragales Paeonia sect. Moutan (Paeoniaceae); rosid Prunus sect. Armeniaca (Rosaceae); Quercus (Fagaceae); and asterid Panax (Araliaceae). We attempted to include a representative from each major angiosperm group. The species of Calycanthaceae, Paeonia sect. Moutan, Prunus sect. Armeniaca, and Panax were completely sampled. Unfortunately, there are too many species in Pinus, Iris, and Quercus for full inclusion in this study.
Experimental details. Genomic DNA was extracted from fresh or silica gel-dried leaves using the mCTAB method [bib_ref] A modified CTAB protocol for plant DNA extraction, Li [/bib_ref]. The 25-mL PCR reactions contained 1 3 PCR buffer (with Mg 21 ), 0.25 mmol/L of each dNTP, 0.25 mmol/L of each primer, 1.25 U Taq polymerase, and 20-30 ng DNA. The PCR program consisted of 4 min at 94uC, 34 cycles of 30 s at 94uC, 40 s at 52uC, and 1 min at 72uC, followed by 10 min at 72uC. The PCR products were examined by electrophoresis on a 1% agarose gel containing ethidium bromide and visualised using an ultraviolet transilluminator. Both strands were sequenced on an ABI 3730xl DNA analyzer (Applied Biosystems, Foster City, U.S.A.) according to the manufacturer's protocols.
The primers used in this study were matK472F and matK1248R [bib_ref] New universal matK primers for DNA barcoding angiosperms, Yu [/bib_ref] for angiosperm matK, Gym_F1A and Gym_R1A 42 for gymnosperm matK, rbcLbF and rbcLbR for rbcLb 6 , and cp001F and cp001R 45 for trnH-psbA.
Data analyses. The sequences were edited and assembled using a Sequencer 4.7 (Gene Codes, Ann Arbor, MI, USA), aligned using Clustal X ver2.0 and manually adjusted with Se-Al 2.0 as necessary.
BLAST, distance and tree-building methods were used to evaluate the performance of ycf1, matK, rbcL and trnH-psbA. BLAST (version 2.2.17) was used for the BBG sequences. The entire data set was used as a reference database and each sequence was used as a query. Only hits having E values , 1 3 10 25 were considered. If the hits with the highest scores included sequences from more than one species, the identification was considered to be a failure, otherwise it was considered successful following China Plant BOL Groups.
In addition to Blast, distance and tree-building methods were used on the seven well-sampled plant groups. The Kimura 2-parameter (K2P) distances were calculated using MEGA 5.0. We considered discrimination to be successful if the minimum uncorrected interspecific K2P distance of focal species was greater than their maximum intraspecific distance. When using the tree-building method, neighbour joining (NJ) and unweighted pair group method with arithmetic mean (UPGMA) dendrograms based on K2P distances were constructed using PAUP 4.0. Species were considered to have been discriminated from one another if all of the individuals of a species formed a single and exclusive clade.
To assess the effects of multiple gene regions on the resolution of species, we compared the resolution of species as a cumulative percentage for each combination of gene regions for both the BBG samples and the seven representative plant groups.
[fig] Figure 1 |Figure 2 |: Performances of matK, rbcLb and ycf1b in resolving BBG tree species using BLAST method. Dataset A includes all 490 samples and dataset Bincludes 391 samples having all three barcode sequences. Performances of matK, rbcLb, trnH-psbA and ycf1b in resolving species in seven well-sampled plant groups representing gymnosperms, basal angiosperms, monocots, Saxifragales, rosids, and asterids. [/fig]
|
Detecting Transforming Growth Factor-β Release from Liver Cells Using an Aptasensor Integrated with Microfluidics
We developed a cell-culture/biosensor platform consisting of aptamer-modified Au electrodes integrated with reconfigurable microfluidics for monitoring of transforming growth factor-beta 1 (TGF-β1), an important inflammatory and pro-fibrotic cytokine. Aptamers were thiolated, labeled with redox reporters, and self-assembled on gold surfaces. The biosensor was determined to be specific for TGF-β1 with an experimental detection limit of 1 ng/mL and linear range extending to 250 ng/mL. Upon determining figures of merit, aptasensor was miniaturized and integrated with human hepatic stellate cells inside microfluidic devices. Reconfigurable microfluidics were developed to ensure that seeding of "sticky" stromal cells did not foul the electrode and compromise sensor performance. This microsystem with integrated aptasensors was used to monitor TGF-β1 release from activated stellate cells over the course of 20 h. The electrochemical response went down upon infusing anti-TGF-β1 antibodies into the microfluidic devices containing activated stellate cells. To further validate aptasensor responses, stellate cells were stained for markers of activation (e.g., alpha smooth muscle actin) and were also tested for presence of TGF-β1 using enzyme linked immunosorbent assay (ELISA). Given the importance of TGF-β1 as a fibrogenic signal, a microsystem with integrated biosensors for local and continuous detection of TGF-β1 may prove to be an important tool to study fibrosis of the liver and other organs.
T he liver is at the center of body's metabolism, and its injury by toxicants or infections is the main cause of several diseases such as cirrhosis, fatty liver, hepatitis, jaundice, and liver cancer.Liver fibrosis is an inflammatory condition that is present during liver injury, cancer, or infection.Transforming growth factor-beta 1 (TGF-β1) is an important factor associated with fibrosis of the liver and other organs.In the liver, TGF-β1 is secreted by the activated hepatic stellate (stromal) cells, causing stellate cells to begin aberrant production of extracellular matrix proteins and leading to loss of differentiated hepatic phenotype. Given that TGF-β1 is a key molecular trigger of fibrosis and liver injury, it is important to know how fast it appears and what its dynamics are over the course of injury or insult. Immunoassays traditionally used for detection of signaling molecules such as TGF-β are limiting when it comes to determining secretion dynamics. We wanted to leverage aptamer-based biosensors for continuous monitoring of TGF-β1 secreted by liver cells. These aptasensors are based on the concept of structure switching pioneered by Plaxco and coworkers.Our lab has been interested in placing aptasensors at the site of cells for local, sensitive, and continuous detection of secreted molecules. Our focus has previously been on detecting inflammatory cytokines secreted from immune cells.In this paper, we wanted to develop an aptasensor for monitoring activation and TGF-β1 release from hepatic stellate cells. The aptamer was based on the DNA sequence described in the literature.Unlike our previous work with anchorage independent immune cells, stellate cells are quite adhesive, capable of attaching to and fouling the electrode surfaces. To remedy this, a reconfigurable microfluidic device was developed to allow for lowering of a microstructured poly(dimethylsiloxane) (PDMS) membrane to protect the electrodes during stellate cells seeding and for raising during cell detection experiments.
The miniaturized aptasensor was constructed by immobilizing methylene blue (MB)-tagged TGF-β1 aptameron top of Au electrodes placed inside of microfluidic devices. Stellate cells were cultivated inside microfluidic devices next to sensing electrodes. The aptamer immobilized electrodes were protected with PDMS microcups in order to avoid nonspecific attachment of cells during seeding. The cells were then activated by infusion of platelet-derived growth factor (PDGF). Onset and progression of TGF-β1 release was monitored using square wave voltammetry (SWV) over the course of 20 h. This TGF-β1 sensor provides highly specific and sensitive detection. The PDGF activation of stellate cells was verified with immunostaining and enzyme linked immunosorbent assay (ELISA).
## ■ materials and methods
Chemicals and Reagents. Chromium (CR-4S) and gold etchants (Au-5) were purchased from Cyantek Corporation (Fremont, CA). Positive photoresist (S1813) and its developer solution (MF-319) were bought from Shipley (Marlborough, MA). Poly(dimethylsiloxane) (PDMS) and silicone elastomer curing agent were purchased from Dow Corning (Midland, MI). Amine functionalized thiolated transforming growth factor (TGF)-β1 DNA aptamer (MW 23 689.9) was purchased from Integrated DNA Technologies, USA. Recombinant human TGF-β1, platelet-derived growth factor (PDGF), 6-mercapto-1hexanol (MCH), triton-X 100, bovine serum albumin (BSA), tris(2-carboxyethyl)phosphine hydrochloride (TCEP), sodium bicarbonate (NaHCO 3 ), collagen (Type I), and 4-(2hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) were purchased from Sigma-Aldrich, USA. Methylene blue (MB), carboxylic acid, and succinimidyl ester (MB-NHS) (MW 598.12) was received from Biosearch Technologies, Inc. (Novato, CA). Dulbecco's modified Eagle's medium (DMEM), sodium pyruvate, fetal bovine serum (FBS), and penicillin/streptomycin (PS) were purchased from Invitrogen (Carlsbad, CA, USA). Paraformaldehyde was purchased from Electron Microscopy Sciences, USA. Anti-α-smooth muscle actin (α-SMA) and goat antirabbit IgG conjugated with Alexa-488 were obtained from Abcam and Invitrogen, respectively. 4,6-Diamidino-2-phenylindole (DAPI) was from Molecular Probes, Invitrogen. All other chemicals were used without further purification. The ELISA kit was obtained from R&D Systems, USA.
In the present study, we utilized a TGF-β1-binding aptamer (IDT Technologies, San Diego, CA) having a loop structure with amino functionality at the 5′ and thiol functionality at the 3′ end.
The backbone of single-stranded DNA aptamers was modified to have phosphorothioates (represented by " * " in the aptamer sequence) on 5′ of both A and C. This modification is believed to provide enhanced nuclease resistance as well as higher affinity than that of a phosphate counterpart.The aptamer stock solution was made in 40 mM HEPES buffer containing 100 mM NaCl, 5 mM KCl, and 5 mM MgCl 2 .
Functionalization and Immobilization of Aptamers. To fabricate an electrochemical biosensor, the TGF-β1 aptamer was tagged with MB, an electroactive redox label. NHS-labeled MB was conjugated to the 5′-end of amino-modified TGF-β1 aptamer through the succinimide ester coupling reported previously.Prior to modification of the Au electrodes, aptamer stock solution was reduced in 10 mM TCEP for 1 h to cleave disulfide bonds. This solution was then diluted in HEPES buffer to achieve the desired aptamer concentration (1 μM, the optimal concentration for TGF-β1 binding). For immobilization of aptamer onto the microfabricated Au electrodes, 1 μM MB-tagged TGF-β1 aptamer solution was infused into the working channels of the microfluidic device followed by incubation for about 18 h at 4°C in the dark. Following incubation, the channels were rinsed with copious amounts of deionized water and HEPES and then incubated with 1 mM MCH for 15 min to displace nonspecifically adsorbed aptamer molecules and to passivate the electrode surface.
Design and Fabrication of Microfluidic Device with Integrated Electrodes. In order to avoid nonspecific attachment of the stellate cell to the sensing electrodes during cell seeding, we designed a reconfigurable microfluidic device of the kind shown in Scheme 1A. This device consisted of two PDMS layers assembled onto a glass slide with micropatterned Au electrodes. The micropatterned Au arrays of eight electrodes of diameter 300 μM were prepared using standard photolithography and a metal-etching process.One PDMS layer contained a linear channel with length × width × height dimensions of 15 mm × 1.7 mm × 0.075 mm. This was the control layer used for actuation of the working microfluidic layer. The latter consisted of eight microchannels, each with length × width × height dimensions of 6 mm × 1.2 mm × 0.075 mm. Each microchannel contained four microcups with inner and outer diameters of 500 and 700 μm, respectively, and the height of 0.060 mm. The device assembly steps included integration of the control PDMS layer (thickness ∼5 mm) on another thin PDMS layer (thickness ∼1 mm) containing eight microchannels with microcups followed by assembling the joint layers of PDMS with microfabricated Au electrodes. The layers of the device were bound together and to the substrate by O 2 plasma treatment. Alignment of each layer during the device assembly was checked optically under the microscope. The device was left for 15 min at room temperature and then filled with deionized water to remove the bubbles. The tubing was attached in the control layer, and glass cylinders (diameter 8 mm) were placed at the inlets of the device using 10:1 ratio of PDMS and curing agent followed by baking for 20 min at 70°C
. This device was utilized to culture stellate cells next to the aptasensing electrodes preventing nonspecific attachment of cells on the sensor by lowering and raising the PDMS microcups (Scheme 1B).
Cultivation of Stellate Cells inside Sensing Microfluidic Devices. A human hepatic stellate cell line (LX2) was maintained in DMEM supplemented with 0.5% FBS, 200 units/mL penicillin, and 200 μg/mL streptomycin at 37°C in a humidified 5% CO 2 atmosphere. Cells were cultured in a tissue culture flask (growth area of 75 cm 2 ) until 90% confluence and then passaged. For cell seeding experiments, aptamer functionalized working Au electrodes were first protected with PDMS microcups by applying positive pressure through the control layer and HEPES buffer was flowed into the microchannels. Subsequently, 0.1 mg/mL solution of collagen (type I) was infused into the channels and kept at room temperature for about 1 h to adsorb collagen on the glass surface around aptasensing electrodes. The channels were again washed with buffer and sterilized under UV. Stellate cells, resuspended at a concentration of 1.6 × 10 6 cells/mL in DMEM described above, were infused into a microfluidic device and incubated at 37°C.
Characterization of TGF-β1 Aptasensor Using Electrochemistry and Surface Plasmon Resonance. In order to obtain the optimal TGF-β1 aptasensor, we characterized aptamer immobilization and target binding via a surface plasmon resonance (SPR) flow through system (Biosensing Instruments, USA). In all the SPR experiments, the system was purged with the running buffer solution (HEPES) to set and stabilize an initial baseline. Subsequently, TGF-β1 aptamer was loaded at the flow rate of 40 μL/min in the SPR flow system and allowed to interact with the Au surface. At the completion of binding, the surface was washed with running buffer and the final baseline was recorded. The SPR angle (m°) difference (represented by the y-axis in, Supporting Information) between the final and initial baselines in the binding curve corresponds to the change in refractive index due to bound TGF-β1aptamer. The aptamer modified SPR surface was challenged with a series of recombinant TGF-β1 concentrations to find the binding affinity of TGF-β1 aptamer with recombinant TGF-β1.
Additional initial optimization studies to find out the response and specificity of TGF-β1 aptamer were carried out by performing square wave voltammetry (SWV) using a potentiostat (CHI instruments, model 842B) at a frequency of 60 Hz over the potential range from 0 to −0.50 V. Au-coated Si substrates were incubated with TGF-β1 aptamer (1 μM) and were then placed into a homemade electrochemical cell creating an electrode area of 1.13 cm 2, Supporting Information). Pt wire counter electrode and Ag/AgCl reference electrode were immersed into the electrolyte solution to complete the three-electrode cell. The aptasensors were also challenged with nonspecific proteins such as IgG, BSA, IL-2, and IFN-γ as well as the analyte of interest TGF-β1. Responses to other isoforms of TGF-β were also studied.
Electrochemical Sensing of TGF-β1 Release from Stellate Cells within the Microfluidic Device. The aptamer functionalized microfluidic device was calibrated by infusing recombinant TGF-β1 ranging in concentration from 0.5 to 300 ng/mL in cell culture media (DMEM supplemented with 0.5% FBS and 1% PS). For cell experiments, miniaturized aptasensing electrodes were protected by lowering down the PDMS microcups (Scheme 1B) via application of positive pressure created in the control layer (filled with water and clamped). The hepatic stellate cells were cultured around aptasensing electrodes inside the microchannels of the device. To stimulate cells for TGF-β1 secretion, mitogenic solution, consisting of 20 ng/mL PDGF spiked in DMEM, was injected into the microchannels. After activation of cells, pressure was released from the control layer raising PDMS cups and exposing the aptasenor to the cytokines released from the activated stellate cells (Scheme 1B). Real-time SWV measurements were performed for up to 20 h to detect the TGF-β1 release from cells by the aptasensor. For the microfluidic electrochemical set up, we used a flow-through Ag/AgCl reference electrode and Pt wire counter electrode connected to the outlet and inlet of the fluid system. A homemade switching system was employed to sample individual electrodes at predefined time intervals. During the electrochemical experiment, the device was kept in a custom-designed environmental box with 5% CO 2 and temperature of 37°C. As shown in Scheme 1, our device contains 8 individually addressable electrodes, one electrode per fluidic channel. However, four microchannels are connected to one inlet and one outlet. We used one set of four channels to measure the TGF-β1 release from PDGF treated stellate cells, whereas the other set of four channels was used to monitor cytokine production in nontreated cells.
Verifying Activation of Stellate Cells by Immunofluorescent Staining and ELISA. Hepatic stellate cells in four of the microfluidic channels were stimulated with 20 ng/mL PDGF for 24 h while the cells in the other four channels were left unstimulated. Cells inside the microfluidic devices were then fixed in 4% paraformaldehyde (Electron Microscopy Sciences) + 0.3% Triton-X100 (Sigma) in PBS for 15 min followed by incubation in blocking solution (1% bovine serum albumin (BSA) in PBS) for 1 h. The channels were washed several times with PBS and exposed to rabbit anti-α-smooth muscle actin (α-SMA, Abcam) antibody for 90 min. The samples were again washed for 5 min and incubated with goat antirabbit IgG conjugated with Alexa-488 (diluted 1:1000) (Invitrogen) for 1 h. After a 5 min washing with PBS, samples were incubated with 4,6-diamidino-2-phenylindole (DAPI) for 15 min. All incubations were performed at room temperature. Fluorescent images were captured by a NIKON Eclipse Ti.
TGF-β1 ELISA was performed using the ELISA kit obtained from R&D Systems. 1.6 × 10 6 cells/mL stellate cells were seeded in each well of the collagen-coated 6-well plate using the same method used for the microfluidic device. Half of the wells were treated with 20 ng/mL PDGF, and the rest were left with nonstimulated DMEM. Samples were collected from each well after 48 h of PDGF treatment. ELISA was conducted in a 96well plate using the protocol provided by R&D Systems.
## ■ results and discussion
The goal of this paper was to develop an electrochemical aptasensor for detection of TGF-β1, an important inflammatory and fibrogenic cytokine. A biosensor specific and sensitive for TGF-β1 was developed, miniaturized, and integrated with microfluidics. This sensing microsystem was then employed to Analytical Chemistry SPR and Electrochemical Analysis of Aptamer-TGF-β1 Interactions. In order to arrive at the ideal TGF-β1 aptasensor, we characterized aptamer immobilization and target binding via SPR studies (Supporting Information). These investigations showed 1 μM to be an optimal aptamer concentration for target binding. To determine the binding affinity of TGF-β1 aptamer with recombinant TGF-β1, the aptamer-modified SPR surface was challenged with a series of TGF-β1 concentrations, Supporting Information). Kinetic analysis of the obtained binding data was done via the "Scrubber" software package provided with the SPR instrument to determine the equilibrium dissociation constant (K d ). The value of K d for aptamer−target binding was found to be 1.07 nM. This data analysis utilizes SPR angle shift, due to the change in refractive index, as a response unit to quantify the binding of macromolecules at the sensor surface.
The surface density (N tot ) of 1 μM TGF-β1 apatmer on the Au surface was determined to be 3.83 × 10 12 molecules/cm 2 via electrochemistry utilizing eq 1. 17
[formula] = + I E nfFN nFE RT nFE RT ( ) 2 sin h( / ) cos h( / ) 1 avg 0 tot(1) [/formula]
where I avg (E 0 ) is the average peak current in a voltammogram, n is the number of electrons transferred per redox event (with MB label n = 2), F is the Faraday current, R is the universal gas constant, T is the temperature, E is the peak amplitude, and f is the frequency. Additional studies to determine specificity of TGF-β1 aptamer were carried out using electrochemistry. Redox-labeled aptamer molecules were assembled on Au electrodes and loaded into a homemade electrochemical cell as described in the previous section of the manuscript. These electrodes were challenged with 50 ng/mL concentrations of IgG, BSA, IL-2, IFN-γ, and TGF-β1.shows that aptasensor response to nonspecific proteins was less than 10% of the signal generated in response to TGF-β1. Comparison of sensor responses to three known isoforms of TGF-β (1, 2, 3) showed ∼15% cross-reactivity with TGF-β3 and 30% cross-reactivity with TGF-β2. This observation is not unexpected in light of homology between TGF-β1 and -β3.Operation of the Microfluidic Device. In the multilayered microfluidic device, the upper most PDMS layer controls the raising and lowering of the microcups fabricated into the other PDMS layer. The up and down action of the PDMS layer was regulated by applying positive or negative pressure via the control layer. Water was injected into the control layer to apply positive pressure and was evacuated to create negative pressure.
In, food dye was used to illustrate the working principle of the microfluidic device. In, the PDMS control channel was filled with black dye while the bottom layer was infused with red dye. This image shows the presence of 8 parallel microchannels, each containing a sensing electrode.,C shows the microchannels that were infused with red dye after the cups where lowered (control layer filled with water). Thus, one can see dye-free white regions around the electrodes, surrounded by red liquid.
The ability to reconfigure the microfluidic channel to separate the environment around the electrodes from the rest of the fluidic channel was used to seed cells.,B shows stellate cells 30 min and 12 h after seeding into the microfluidic channel. One can see an annular region around the electrode that was protected by microcups and remained free of cells after seeding. To highlight the importance of protecting electrodes during the seeding of "sticky" anchorage-dependent stellate cells, aptamer-functionalized electrodes were characterized by SWV after seeding cells with or without microcup protection. As shown in, the redox signal of electrodes directly exposed to cell seeding was about 15-fold lower than the electrode protected with microcups. Thus, the reconfigurable microfluidic device allowed us to ensure that aptasensors were not fouled during collagen coating and cell seeding steps and retained the high sensitivity needed for cell secretion monitoring.
Monitoring of TGF-β1 Release from Stellate Cells. As described in Scheme 1B, it was expected that our biosensor would react to TGF-β1 secreted from stellate cells, providing an electrochemical signal associated with the amount and the rate of cytokine secretion. Prior to carry out cell detection experiments, we wanted to construct calibration curves to determine the linear range of the miniature aptasensors integrated with microfluidics.shows typical response of aptasensor to varying (increasing) concentration of exogenous TGF-β1. The signal decreases with increasing concentration of the analyte and is reported as signal suppression − ((initial current − final current)/initial current).shows a calibration curve of signal suppression vs TGF-β1 concentration. The experimental data shows the linear range for our biosensor from 1 to 250 ng/mL for exogenous TGF-β1.
For TGF-β1 release experiments, stellate cells were seeded into microfluidic channels that were reconfigured so as to protect aptasensing electrodes. Upon attaching and spreading, cells were stimulated by infusion of PDGF, growth factor known to activate stellate cells. The device was reconfigured once again to raise the microcups, allowing for cell secreted factors to diffuse to the nearby sensing electrode. The resultant electrochemical signals were continuously monitored by sequentially addressing aptasensing electrodes. To ensure viability and function, microfluidic devices with stellate cells were kept under physiological temperature and 5% CO 2 during the 20 h experiment.shows TGF-β1 secretion results after PDGF activation. In our experiments, one inlet of the device, connected to four cell seeded microchannels, was infused with stimulant to activate the cells while the other four channels contained quiescent cells. As highlighted by data in, electrodes in proximity to activated stellate cells were recording significantly higher changes in redox current (signal suppression) compared to electrodes near quiescent cells. After 18 h, the electrodes near activated stellate cells reached 80% signal suppression whereas sensors next to quiescent cells showed less than 20% signal suppression. Given that 80% signal suppression corresponds to a saturated aptasensor (see, the electrode sensing activated stellate cells likely became saturated after 18 h of monitoring.
To determine cell secretion rates based on the binding curves shown in, we constructed a numerical model for simulating cytokine production, diffusion, convection, and binding using COMSOL multiphysics (COMSOL Inc., Burlington, MA). The geometry and sensor configuration were the actual sensing devices used for the experiments. The main parameters in the simulations are listed in [fig_ref] Table 1: List of Parameters Obtained Experimentally and Further Used for Determining Secretion Rates... [/fig_ref].
We assumed a constant secretion rate in each 3 h time interval and then determined the secretion rate by the reaction−diffusion−convection model coupled with leastsquares approximation (further details provided in the. From the model, the average cytokine production rate from the stimulated cells within the microfluidic channel was determined to be 0.0140 pg/cell/hwhile for quiescent cells, the rate was 0.0009 pg/cell/h. This rate is comparable to the production of TNF-α by activated monocytes reported by our group previously. Our TGF-β production rate over the course of 24 h (3.36 × 10 −1 pg/cell) is considerably higher than rates reported for stellate cells activated under standard tissue culture conditions (4.2 × 10 −5 to 3.2 × 10 −3 pg/cell).This may be explained by the enhanced sensitivity of our approach where local concentrations are being monitored. It is also possible that confinement of stellate cells inside low volume microfluidic channels enhances the rate of TGF-β secretion.
A number of control experiments were carried out to ensure that the signal observed inwas indeed due to TGF-β1 secretion. Exposure of stellate cells to culture media without PDGF did not cause an appreciable change in electrochemical signal, lower curve). Similarly, there was no effect of PDGF containing media on the aptamer modified Au surface without cells (result not shown).
We conducted another control experiment where activated stellate cells residing inside microfluidic channels were bathed with TGF-β1 antibodies during the sensing experiment. The presence of competing antibody molecules abrogated redox signals at the sensing electrodes, demonstrating once again that electrodes were indeed sensing secreted TGF-β1, Supporting Information).
Characterizing Phenotype and Function of Stellate Cells Using Molecular Biology Approaches. Immunofluorescent staining and ELISA techniques were used to confirm that exposure to PDGF activated stellate cells caused them to release TGF-β1. ,B shows expression of α-SMA, one of the benchmarks of stellate cell activation, in the presence and absence of PDGF.These results clearly demonstrate increased expression of SMA in cells after a 24 h exposure to PDGF, indicating activation of these cells. Another correlate of stellate cell activation is production of TGF-β1. We collected media from activated and quiescent cells cultured in a well plate for 48 h and performed TGF-β1 ELISA. This experiment revealed that stellate cells activated with PDGF produced 6 times more TGF-β1 compared to quiescent cells. While it may not be possible to compare concentrations obtained with ELISA and aptasensors due to differences in analytical methods, the 6-fold enhancement in TGF-β1 activated and resting stellate cells was observed for both approaches (seefor comparison).
# ■ conclusions
The paper describes the use of aptamer-modified electrodes for continuous monitoring of the TGF-β1 release from hepatic stellate cells. While our lab has previously demonstrated detection from immune cells,this study deals with monitoring secretory activity of adhesive stromal cells. Reconfigurable microfluidic devices were used to address electrode fouling problems. Such devices were microfabricated to contain microcups imbedded in the roof of the microfluidic channel. By actuating a control layer, microcups were lowered to protect the electrodes during collagen coating and cell seeding steps. The device was reconfigured once again after seeding and activation of stellate cells, to commence the detection phase of the experiment. These specific and sensitive aptasensors were used to electrochemically monitor TGF-β1 secretion from activated and resting stellate cells and to determine TGF-β1 secretion rates. Furthermore, additional molecular biology assays demonstrated that stellate cells were indeed activated by stimulation with PDGF and that these activated stellate cells produced TGF-β1. In fact, a similar 6fold ratio of TGF-β1 production between activated and resting cells was obtained using aptasensors inside microfluidic channels and ELISA in 6-well plates. These results point to the fact that stellate cells functioned similarly (vis a vis TGF-β1 secretion) inside microfluidic devices and under standard tissue culture conditions. In combination with aptasensor specificity testing, the TGF-β1 antibody competition experiment showed that detection was indeed specific to secreted TGF-β1. Given the importance of TGF-β1 signaling in inflammation, fibrogenesis, T-cell differentiation, and stem cell development, the ability to locally and continuously monitor production of TGF-β1 has high significance. In the future, the duration of the detection experiment will be extended from 20 h to multiple days. Multiplexed aptamer-based detection of several important trophic factors is also envisioned.
[fig] Figure 1: (A) The results of SWV response of aptasensor toward 50 ng/mL of different nonspecific (IL-2, IFN-γ, BSA, IgG) molecules and 50 ng/mL of the specific target molecule (TGF-β1). Plot showing the change in aptasensor current resulting due to each nonspecific and specific molecule. These results show that the biosensor did not respond to nonspecific proteins but did respond to TGF-β1. (B) Response of the aptasensor toward 50 ng/mL of different isoforms of the TGF-β molecule. [/fig]
[fig] Figure 2: (A) Fluidic channels of the reconfigurable microfluidic device are infused with red dye whereas the control layer is filled with black dye. (B, C) Device is reconfigured, lowering microcups around the electrodes. The channels are filled with red dye to highlight that the area around the electrodes remains free of dye. (C) Higher magnification view of the electrodes being sequestered from the rest of the channel. [/fig]
[fig] Figure 3: (A) Stellate cells 30 min after seeding around Au electrodes. (B) The same cells 12 h after seeding spread out around the electrode.(C) SWV curves looking at the MB signal obtained from aptasensor when cells were seeded with and without microcup protection. These results show 15-fold higher peak currents for the electrode protected by microcups during cell seeding. [/fig]
[fig] Figure 4: (A) SWV curves obtained for aptamer-modified electrodes challenged with recombinant TGF-β1 ranging in concentration from 1 to 300 ng/mL. (B) Signal suppression plotted against recombinant TGF-β1 concentration. The aptamer shows good response in the cell culture media with signal saturation after 250 ng/mL TGF-β1. [/fig]
[fig] Figure 5: (A) Continuous monitoring of the TGF-β1 release from stellate cells within the microfluidic device using aptamer modified microfabricated electrodes. The device was kept in the CO 2 and temperature control chamber during the SWV experiments. (B) Change of the TGF-β1 secretion rate over time obtained from the reaction-diffusion model analyzed with COMSOL. [/fig]
[fig] *: and figures giving details of the surface plasmon resonance (SPR)-based characterization of TGF-β1 aptamer, electrochemical characterization of TGF-β1 aptamer, description of the computational model for secretion rate, and inhibition of TGF-β1 detection via anti-TGF-β1 antibodies. This material is available free of charge via the Internet at http://pubs.acs.org. Phone: +1-530-752-2383. Fax: +1-530-754-5739. E-mail: [email protected]. [/fig]
[table] Table 1: List of Parameters Obtained Experimentally and Further Used for Determining Secretion Rates via Simulation [/table]
|
Long-term effects and costs of pelvic floor muscle training for prolapse: trial follow-up record-linkage study
Introduction and hypothesis Pelvic organ prolapse affects around 40% of women aged over 50 years. A multicentre parallel group randomised trial (the Pelvic Organ Prolapse PhysiotherapY (POPPY) trial) demonstrated that pelvic floor muscle training (PFMT) was effective in reducing prolapse symptoms compared with no treatment. However, insight into the longterm impact of PFMT on health outcomes and health-service utilisation is scarce. Methods This study utilised linkage of Scottish administrative health records to follow-up POPPY trial participants resident in Scotland over 11 years. Mixed effects logistic regression determined the likelihood of receiving further prolapse treatment for those in the PFMT and control groups. Analyses were adjusted for age group, prolapse stage, baseline symptom severity and attitude towards surgery. A cost assessment estimated longitudinal costs to the UK National Health Service (in Scotland) of accessing further prolapse treatment for each trial group. Results Two hundred and ninety-three women, aged 25 to 79 years, were followed up. One hundred and forty-one women (48.1%) had received further prolapse treatment: 65 (of 149; 43.6%) in the PFMT group compared with 76 (of 144; 52.8%) in the control group. PFMT was associated with a reduction in the odds of any prolapse treatment during follow-up (AOR 0.61; 95% CI 0.37 to 0.99). Total cost of secondary care was £154,544 (GBP) in the PFMT group and £172,549 (GBP) in the control group. Conclusions Although PFMT did not lead to significant differences in total costs for further prolapse treatment over a postintervention period of more than 10 years, it reduced the overall long-term risk of requiring hospital-based treatment for pelvic floor disorders.Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
# Introduction
Pelvic organ prolapse (POP) is the loss of support for the uterus, bladder or bowel that results in the descent of one or more of these organs into the vagina [bib_ref] An International Urogynecological Association (IUGA)/International Continence Society (ICS) joint report on the..., Haylen [/bib_ref]. Prolapse affects around 40% of women aged over 50 years [bib_ref] Pelvic organ prolapse in the Women's Health Initiative: gravity and gravidity, Hendrix [/bib_ref]. Risk factors include pregnancy, childbirth, and ageing. The symptoms of POP can affect daily activities and quality of life [bib_ref] A systematic review and meta-synthesis of qualitative studies on pelvic organ prolapse..., Rada [/bib_ref] , and women may experience pelvic pain, bladder and bowel dysfunction. The most common treatments for POP in women include surgery, vaginal support pessary and pelvic floor muscle training (PFMT). Women often access PFMT first as a conservative (nonsurgical) measure [bib_ref] Conservative treatment options for women with stress urinary incontinence: clinical update, Imamura [/bib_ref]. PFMT is an effective (majority level 1 evidence [bib_ref] An International Urogynecological Association (IUGA)/International Continence Society (ICS) joint report on the..., Haylen [/bib_ref] non-surgical option that has been shown to improve prolapse symptoms and health-related quality of life for women experiencing POP [bib_ref] Conservative management for female urinary incontinence and pelvic organ prolapse review 2013:..., Dumoulin [/bib_ref] [bib_ref] Individualised pelvic floor muscle training in women with pelvic organ prolapse (POPPY):..., Hagen [/bib_ref] [bib_ref] Can pelvic floor muscle training reverse pelvic organ prolapse and reduce prolapse..., Braekken [/bib_ref]. PFMT is aimed at improving the function of the pelvic floor muscles to increase the structural support for the pelvic organs. This is achieved by targeted exercises to strengthen and improve coordination of the muscles.
Economic analyses are an important basis for determining the cost-effectiveness of different intervention options that might be used. Yet there are limited data on the cost-effectiveness of conservative interventions for POP [bib_ref] Systematic review and economic modelling of the effectiveness and cost-effectiveness of non-surgical..., Imamura [/bib_ref]. Previous modelling analysis of costs and benefits of non-surgical and surgical treatments for POP included only pessary as a non-surgical intervention [bib_ref] Treatment strategies for pelvic organ prolapse: a cost-effectiveness analysis, Hullfish [/bib_ref]. Efforts to establish the economics of PFMT have been conducted within the scope of trials. A systematic review of PFMT for stress urinary incontinence in women identified only one economic evaluation [bib_ref] Pelvic floor muscle training versus no treatment, or inactive control treatments, for..., Dumoulin [/bib_ref]. It identified a cost-utility analysis of PFMT delivered by a mobile phone app, which concluded that the app treatment was cost-effective, estimated at €7,615.50 (Euro) per quality-adjusted life year (QALY) [bib_ref] Mobile app for treatment of stress urinary incontinence: a cost-effectiveness analysis, Sjöström [/bib_ref]. However, the trial that it related to had not collected data on the presence of co-existing POP in the participants [bib_ref] Mobile app for treatment of stress urinary incontinence: a randomized controlled trial, Asklund [/bib_ref]. The cost-effectiveness of PFMT, in comparison with watchful waiting, has been assessed over a 2-year follow-up period in a patient population with symptomatic mild POP [bib_ref] Two-year effects and cost-effectiveness of pelvic floor muscle training in mild pelvic..., Panman [/bib_ref]. Although the study reported a positive effect of PFMT on POP symptoms, and estimated an incremental cost-effectiveness ratio of €31,983 (Euro; 95% CI: 76,652 to 88,078) per QALY, it noted that direct medical costs per person were reasonable and suggested that decisions should be based on clinical need and patient preference. Economic findings of the Pelvic Floor Muscle Training for Secondary Prevention of Pelvic Organ Prolapse (PREVPROL) trial suggested a cost per QALY range from £21,996 to £29,409 (GBP) [bib_ref] Pelvic floor muscle training for secondary prevention of pelvic organ prolapse (PREVPROL):..., Hagen [/bib_ref]. However, the effectiveness and cost-effectiveness evaluations of PFMT have, to date, been limited to relatively short follow-up periods. The economic impact of PFMT is likely to occur over years as opposed to months.
Evidence of the impact of PFMT on women's longer-term treatment outcomes is scarce. The Pelvic Organ Prolapse PhysiotherapY (POPPY) trial was a large, pragmatic trial of PFMT for prolapse (registration number NCT00476892), which concluded that PFMT was effective in reducing reported prolapse symptom severity at 12-month followup in treatment-naive women presenting with stage 1 to 3 prolapse [bib_ref] The POPPY trial participant data linkage study, in Pelvic floor muscle training..., Maxwell [/bib_ref]. The net cost of delivering PFMT during the trial was reported to be £130 (GBP) per woman, but conclusions regarding the potential of PFMT to be a cost-effective treatment were limited by uncertainty about whether PFMT changed the need for subsequent treatment in the longer term or delayed it. Investment in Scotland in health record linkage makes it possible to follow-up original POPPY trial participants using record linkage of hospital admissions and outpatient datasets.
To understand if women who undertook PFMT were less likely to receive further secondary care treatment related to their prolapse during the follow-up period than women who were in the control condition, routinely collected administrative data were used to investigate health service use over a follow-up period of up to 11 years in the 310 original POPPY trial participants resident in Scotland. This accounted for 69.3% of the 447 participants in the original trial cohort. The aim of the study was to investigate longer term associations between PFMT and the need for prolapse intervention; the time to prolapse intervention; and the long-term costs associated with accessing further prolapse treatment.
# Materials and methods
## Record linkage
Estimates of long-term effects and costs associated with accessing further prolapse treatment over time required linkage to routinely collected administrative health care data about individual patients, specifically Scottish Morbidity Recordsfor outpatients (SMR-00) and inpatients and day cases (SMR-01), for participants in the original POPPY trial who were based in Scotland. Approval for linkage was obtained from the National Health Service (NHS) Scotland Public Benefit and Privacy Panel for Health and Social Care. Following approval, NHS National Services Scotland's Electronic Data Research and Innovation Service (eDRIS, now part of Public Health Scotland, www. isdsc otland. org/ produ cts-and-servi ces/ edris/ use-of-the-natio nal-safe-haven) undertook the linkage of the POPPY trial data with the SMR-00 and SMR-01 datasets and information on deaths from National Records Scotland. Anonymised linked data files were made available for analysis via the National Safe Haven.
# Analysis dataset
Longitudinal health care data were available for a postintervention period of 10 years and 9 months (September 2007 to May 2018). The primary outcome measure was the occurrence of any related secondary care treatment for pelvic floor disorders during follow-up, defined as Office of Population Censuses and Surveys Classification of Interventions and Procedures version 4 (OPCS4) procedure code P18, P22-P24, P26, Q07-Q08, Q54, M51-M58, A70 or H57 as the main operation or other operation and/or International Statistical Classification of Diseases and Related Health Problems, 10th revision (ICD10) diagnosis code N81, N99.3, N39.3, N39.4, R32, K62.2, K62.3 or R15 as main or other condition. The primary outcome was expressed as a binary variable.
# Statistical analysis
Analysis was conducted in Stata version 14. Mixed effects logistic regression was used to estimate effect size for the association between PFMT and any related secondary care treatment for pelvic floor disorders (odds ratios [OR] and 95% confidence intervals [CI]). Analyses were adjusted for the variables as used in the original POPPY trial analysis (baseline symptom severity [Pelvic Organ Prolapse Symptom Score, POP-SS], score range 0-28, higher scores indicate worse severity [bib_ref] Psychometric properties of the pelvic organ prolapse symptom score, Hagen [/bib_ref] ; prolapse stage [Pelvic Organ Prolapse Quantification, POP-Q]; stages 1-3 included in the trial, higher stages indicate worse severity [bib_ref] The standardization of terminology of female pelvic organ prolapse and pelvic floor..., Bump [/bib_ref] ; and attitude to prolapse surgery). Attitude towards surgery was assessed with the question "If you were to be offered surgery in the future as a treatment for your prolapse, how would you feel about it?" with response options 'I would like to avoid surgery if at all possible", "I am willing to have surgery but only if it is unavoidable" and "I am keen to have surgery". In addition, treatment centre was included in the model as a random effect. Age group was included as an additional covariate, as over an 11-year follow-up period, the impact of age may be increased. Missing prolapse symptom severity scores were imputed at median because of the relatively small sample size. Results for a complete case analysis are presented as a sensitivity analysis to check the effects of this approach to missing data. As a secondary analysis, Cox regression was used to analyse time-to-event data. That is, time from the intervention to any related secondary care treatment for pelvic floor disorders was compared between treatment and control conditions. The analysis was stratified by attitude to surgery and was adjusted for symptom severity at baseline, prolapse stage, age group and body mass index (BMI).
## Resource utilisation and cost estimation
Resource use data for the predefined OPCS4 main operation codes for pelvic floor disorders (P18, P22-P24, P26, Q07-Q08, Q54, M51-M58, A70, H57) were identified. Records were excluded where a "did not attend" code was present. A simple mapping of predefined OPCS4 main operation codes of interest to Healthcare Resource Groups (HRG) and thereafter to the relevant currency codes was used to extract unit cost data from NHS reference costs. provides a summary of the reference costs for Excess bed days were calculated as the difference between the number of bed days recorded and the average number of bed days for POPPY participants with the same main operation code. The administrative data were assessed for differences in resource use between control and intervention groups, including the influence of women who required more than one episode of care during the time of followup on cost estimates. Information is presented in aggregate to meet the data protection requirements of NHS National Services Scotland's Statistical Disclosure Control Protocol. In line with the original POPPY trial costs base year of 2009, reporting of economic analysis was in 2009 UK pound sterling (GBP £; GDP deflators.
# Results
## Participants
Linked follow-up data were available for 293 of the 447 participants (65.5%) in the original POPPY trial. The characteristics of participants included in the analysis are summarised in, using data measured at baseline in the original POPPY trial. The distribution of characteristics is similar to that of the full POPPY trial sample. During the follow-up period, there were 15 participants (5%) who died. The all-cause mortality rate was 5.4% in the intervention group and 4.9% in the control group. All other outcome measures in this analysis used the hospital episode activity data in the linked dataset, although all episodes with a pelvic floor disorder were identifiable from OPCS4 procedure codes alone (there were no cases identified from ICD10 diagnosis codes alone).
## Women receiving treatment during follow-up
The overall proportion of women receiving any secondary care treatment for pelvic floor disorders was 48.1% (141 out of 293), with the rate being 43.6% (65 out of 149) in the intervention group compared with 52.8% (76 out of 144) in the control group. This corresponds to an absolute risk reduction of 9% and a number needed to treat of 11. That is, 11 women would need to receive the PFMT intervention in order to prevent one further treatment during follow-up. The mixed effects logistic regression on "any treatment during follow-up" (primary outcome measure) with adjustment for baseline prolapse symptom severity, motivation for prolapse surgery, age group and prolapse stage, with random effect of centre, estimates a significant treatment effect (odds ratio 0.61, 95% CI 0.37 to 0.99, p=0.047). The estimated parameters from the model are shown on [fig_ref] Table 4: Mixed effects logistic regression on any treatment during follow-up with random effect... [/fig_ref].
The likelihood ratio test for the mixed effects logistic regression model versus a model with fixed effects only (likelihood ratio test Chi-squared=2.74, p=0.0489) indicated that the random effect of centre is required. The PFMT intervention is associated with a reduction in the odds of any treatment during follow-up. The adjusted odds ratio of 0.61 corresponds to a relative risk reduction of 0.80, indicating that the probability of need for further treatment is 20% lower in the group receiving the PFMT intervention than in the control condition after adjustment for covariates.
## Analysis of time to first treatment
This study examined whether there was any difference between the PFMT and control groups in the time until they received pelvic floor disorder secondary care treatment during follow-up. This was in order to determine whether PFMT was effective in delaying any further pelvic floor disorder treatment. First, the overall follow-up periods of the PFMT and control groups were compared. The time in follow-up of the two groups (i.e. the time at risk and under observation) was found to be comparable: (mean time in follow-up in the intervention group was 3,415.3 days (SD=262.5) and the mean in the control group was 3,403.0 days (SD=267.9)). The survival function for the PFMT and control group is shown in [fig_ref] 1: Unit costs applied to SMR-00 and SMR-0 data, 207-208 prices [/fig_ref]. The survival function shows how the proportion of women in the two groups who have not received any further treatment decays over time.
Cox regression [fig_ref] Table 5: Cox regression on time to any treatment [/fig_ref] was used to test whether women in the control group were likely to receive treatment sooner than women in the PFMT group. The Cox regression included 264 women and 125 events. This model estimated a hazard ratio in favour of the intervention group of 0.65 (95% CI 0.46 to 0.94), p=0.020. This indicates that the intervention did delay the need for further treatment and that during follow-up women in the PFMT group had a 35% reduction in hazard for treatment. . The majority of procedures coded within the episodes of care were for colporrhaphy/vaginal wall repair (P22, P23, P24, Q54). (Note that some patients will have received more than one procedure of interest during a single health care episode.) Owing to issues of confidentiality, SMR-01 data cannot be broken down further. Aggregate costs associated with inpatient treatment received for POP were estimated as £163,267 (PFMT group) and £184,748 (control group) (GBP 2017-2018 prices), inclusive of excess bed days. The difference is potentially an artefact of price deflation adjustment. When costs were estimated with and without women who required more than one episode of care during the follow-up period, 11 women had more than one episode of care during the follow-up period and accounted for 24% of all surgical interventions observed. indicates that average cost per woman with one episode of care was £2,973 (£3,077 in the PFMT group and £2,797 in the control group). Average cost diverged when considering women with more than one episode of care: £6,331 for the PFMT group and £8,252 for the control group. Analysis of POPPY trial participants' outpatient and inpatient health care utilisation costs for procedures of interest (F2 gynaecology) did not find an observed difference between the PFMT group and the control group at 11 years post-PFMT intervention.
# Discussion
The longitudinal data record linkage study examined the long-term impact of PFMT on health outcomes and health service utilisation of POPPY trial participants in Scotland, UK. The record linkage study, which focused on individual women's treatment, provides evidence that PFMT reduces the overall long-term risk of requiring hospital treatment for pelvic floor disorders, over a post-intervention period of more than 10 years. There is also evidence that PFMT extends the time for which hospital treatment is not required. This effect was evident only after adjustment for age suggesting that age is an important predictor of long-term treatment. Older women (over 65 years of age) were more likely to have further treatment and to receive it sooner independent of treatment allocation. Economic interpretation of longitudinal follow-up data for POPPY participants did not indicate a difference in the use of inpatient health care resources between the groups. In an effort to assess potential distortion of resource use cost assessment owing to the small numbers of participants, costs were estimated with and without women who required more than one episode of care during the time of follow-up. Although this indicated that the intervention group had fewer and less costly repeat episodes, small numbers mean that these findings must be treated cautiously. On the basis of the follow-up data obtained from the administrative data sets, it was concluded that there was no clear difference in resource use costs between the intervention group and control group participants at 11 years post-PFMT. Cost estimates for longitudinal follow-up of POPPY participants were based on the main operation code. Some participants received more than one procedure of interest during a single health care episode. This has implications for the confidence that can be placed on resource estimations because the average costs cannot reflect the influences of other procedures and personal characteristics such as other diagnoses and the presence of co-morbidities that affect the resulting costs per patient. It was not feasible to gather information about post-operative care, such as medications, general practitioner or practice nurse support.
The main limitation of our study was the exclusion of records in the Scottish Morbidity Records for outpatient and inpatient datasets owing to the inability to determine whether health care resource use was relevant because of missing main condition or procedure codes recorded. The likely implication of this is that resource use costs are underestimated. This was particularly apparent for SMR-00 outpatient records because procedure codes were recorded in only 16% of records and no main condition was recorded in any of the records. An operation code is non-mandatory for this dataset, preventing the search of only episodes relevant to pelvic floor disorders. It also affects the estimation of costs for POPPY participants related to non-surgical health care in outpatient care settings, such as the use of a vaginal support pessary, which is a main non-surgical option for the management of POP.
As the first study to assess the impact of PFMT on women's longer-term treatment outcomes including the need for pelvic floor disorder intervention, time to intervention, and the long-term costs associated with accessing further treatment, this study provides unique information. For research, the importance of planned long-term follow-up of trial participants is highlighted. For practice, this data linkage study provides further evidence to support recommending PFMT to women as first-line treatment as PFMT may delay, or possibly avoid, the need for more treatment. Over a postintervention period of more than 10 years, PFMT reduced the overall long-term risk of requiring hospital treatment for pelvic floor disorders. Although the total cost was broadly similar for both the intervention and the control groups, assessment of costs when patients required more than one episode of care suggested that targeted planning of care such as PFMT may be advisable for health care resource planning. The study also provides reflections on lessons learned about the benefits and limitations of using routine administrative data, adding to the limited methodological knowledge on the use of routine linked health data for longer-term follow-up.
Award No. 14-04-02: Implementation of an evidence-based pelvic floor muscle training intervention for women with pelvic organ prolapse (PROlapse and PFMT: implementing Evidence Locally-PROPEL).
# Declarations
Ethics approval Approval was obtained from the Scotland A Research Ethics Committee in 2007 (REC reference number: 07/MRE10/4).
Linkage approval process An application was made to the NHS Scotland Public Benefit and Privacy Panel (PBPP) for Health and Social Care in 2017. Following a Tier 1 review (April 2017), the application was initially referred for a Tier 2 (out-of-meeting) review but was subsequently referred to a Tier 2 panel meeting (November 2017), which required the PI to attend and respond to questions. Approval was obtained in January 2018 (PBPP Application 1617-0049/Semple: Using linked data to follow up longer-term outcomes of POPPY trial participants).
## Conflicts of interest none.
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] 1: Unit costs applied to SMR-00 and SMR-0 data, 207-208 prices (GBP) Currency units (average for the year 207) per GBP: = .465 EUR; =.282692 USD; = 43.837608 JPY; = .26325 CHF; = .663342 CAD; = .6733 AUD; = .799358 NZD; = 7.0558 ZAR; = 24.309267 MXN; = 4.73 AED. Source (last accessed online April 2022): https:// www. gov. uk/ gover nment/ publi catio ns/ excha nge-rates-for-custo ms-and-vat-yearly] Source: National Schedule of Reference Costs-Year 207-208-NHS trust and NHS foundation trusts HRG Healthcare Resource Group, SMR Scottish Morbidity Records, OPCS Office of Population Censuses and Surveys Classification of Interventions and Procedures version 4 [/fig]
[fig] 6: Summary of Scottish Morbidity Records (SMR)-01 inpatient episodes of care and related costs (2007-2018), by trial group Information is presented in aggregate to meet the data protection requirements of the NHS National Services Scotland's Statistical Disclosure Control Protocol Version 3.0. (2015). (Accessed 4 March 2019 https:// www. isdsc otland. org/ About-ISD/ Confi denti ality/ discl osure_ proto col_ v3. pdf). Small numbers have been necessarily withheld Intervention (PFMT); control (lifestyle advice leaflet); GBP, 2017-2018 prices deflated to 2009 UK pound sterling Predefined Office of Population Censuses and Surveys Classification of Interventions and Procedures version 4 main operation codes of interest: P18, P22-P24, P2, Q07-Q08, Q54, M51-M58, A70, H57 PFMT pelvic floor muscle training [/fig]
[table] Table 4: Mixed effects logistic regression on any treatment during follow-up with random effect of centre Statistically significant, p≤0.05, in bold CI confidence interval, OR odds ratio, PNFT pelvic floor muscle training, POP-SS Pelvic Organ Prolapse Symptom Score [/table]
[table] Table 5: Cox regression on time to any treatment [/table]
|
A Retrospective Comparison of Ultrasound-Assisted Catheter-Directed Thrombolysis and Catheter-Directed Thrombolysis Alone for Treatment of Proximal Deep Vein Thrombosis
Background Recent studies have suggested that catheterdirected thrombolysis (CDT) reduces development of postthrombotic syndrome (PTS). Ultrasound-assisted CDT (USCDT) might enhance the efficiency of thrombolysis. We aimed to compare USCDT with CDT on efficacy, safety, development of PTS, and quality of life after longterm follow-up. Methods We describe a retrospective case series of 94 consecutive patients admitted with iliofemoral or more proximal deep vein thrombosis (DVT) to the University Hospital from 2002 to 2011, treated either with CDT or USCDT. Scheduled follow-up visits took place between April 2013 and January 2014. Venography measured the degree of residual luminal obstruction of the affected veins. Each patient completed the Short Form 36-item health survey assessment and the Venous Insufficiency Epidemiological and Economic Study-Quality of Life/Symptoms questionnaires. PTS was assessed using the Villalta scale.Results Risk factors of DVT were equally distributed between groups. In the USCDT group, we observed a significant decline in the duration of thrombolytic treatment (\48 h: 27 vs. 10 %), shortened hospital stay (median 6.0 days (IQR 5.0-9.0) vs. 8.0 (IQR 5.8-12.0)), and less implantation of (intravenous) stents (30 vs. 55 %). There was no difference in patency (76 vs. 79 % fully patent), prevalence of PTS (52 vs. 55 %), or quality of life between groups after long-term follow-up (median 65 months, range: 15-141). Conclusions In this observational study, USCDT was associated with shortened treatment duration, shorter hospital stay, and less intravenous stenting, compared to CDT alone without affecting the long-term prevalence of PTS or quality of life.
# Introduction
The post-thrombotic syndrome (PTS) causes considerable morbidity in patients after deep vein thrombosis (DVT) of the lower extremity. After 2 years, 30-50 % of the patients will develop PTS, of which 10 % will have moderate and 3-5 % severe PTS [bib_ref] Determinants and time course of the postthrombotic syndrome after acute deep venous..., Kahn [/bib_ref] [bib_ref] Comparison of the Villalta post-thrombotic syndrome score in the ipsilateral vs. contralateral..., Galanaud [/bib_ref]. Risk factors for (severe) PTS are thrombus proximity and recurrent ipsilateral venous thrombosis [bib_ref] Determinants and time course of the postthrombotic syndrome after acute deep venous..., Kahn [/bib_ref] [bib_ref] The long-term clinical course of acute deep venous thrombosis, Prandoni [/bib_ref] where prevention of the latter is the only effective way to prevent the increased severity and frequency of PTS [bib_ref] Post-thrombotic syndrome: a clinical review, Baldwin [/bib_ref]. The wearing of compression stockings, still recommended by leading guidelines [bib_ref] American college of chest physicians evidence-based clinical practice guidelines, Guyatt [/bib_ref] , has recently been questioned [bib_ref] Compression stockings to prevent post-thrombotic syndrome: a randomised placebo-controlled trial, Kahn [/bib_ref].
Recent studies have shown that catheter-directed thrombolysis (CDT) reduces the development of PTS. The CAVENT study reported 14.4 % [95 % confidence interval (CI) 0. ] and 28 % (95 % CI 14-42) absolute risk reduction of PTS after 2 and 5 years of follow-up, respectively, when CDT was compared with standard treatment (anticoagulation and compression stockings) [bib_ref] Long-term outcome after additional catheter-directed thrombolysis versus standard treatment for acute iliofemoral..., Enden [/bib_ref] [bib_ref] Post-thrombotic syndrome after catheter-directed thrombolysis for deep vein thrombosis (CaVenT): 5-year follow-up..., Haig [/bib_ref]. A recent Cochrane review on CDT reported a relative reduction of 25 % of PTS [bib_ref] Thrombolysis for acute deep vein thrombosis, Watson [/bib_ref]. As PTS is more common in patients with a proximal DVT [bib_ref] Determinants and time course of the postthrombotic syndrome after acute deep venous..., Kahn [/bib_ref] [bib_ref] National Clinical Guideline Centre (UK). Venous thromboembolic diseases: the management of venous..., Tick [/bib_ref] , guidelines recommend the use of CDT in selected patients with iliofemoral venous thrombosis [bib_ref] American college of chest physicians evidence-based clinical practice guidelines, Guyatt [/bib_ref]. However, a recent cross-sectional study of patients with proximal lower-extremity DVT showed that CDT was resource-demanding and increased therapy-related adverse events [bib_ref] Comparative outcomes of catheter-directed thrombolysis plus anticoagulation vs anticoagulation alone to treat..., Bashir [/bib_ref]. Unfortunately, this study did not comprise outcomes such as recurrent venous thrombosis, late mortality, and incidence of PTS in the analyses [bib_ref] Comparative outcomes of catheter-directed thrombolysis plus anticoagulation vs anticoagulation alone to treat..., Bashir [/bib_ref].
Ultrasound-assisted CDT (USCDT), combining CDT with a catheter system that uses high frequency ultrasound, might enhance the efficiency of the thrombolytic process even more. In a case series of 53 patients, treatment with USCDT led to reduced total infusion time of the thrombolytic agent, a greater incidence of complete clot lysis, and a reduction in bleeding rates, compared with historical data [bib_ref] Ultrasound-accelerated thrombolysis for the treatment of deep vein thrombosis: initial clinical experience, Parikh [/bib_ref]. In vitro data indicate that ultrasound facilitates thrombolysis by making more plasminogen receptor sites available for the thrombolytic agent [bib_ref] Binding of tissue-plasminogen activator to fibrin: effect of ultrasound, Siddiqi [/bib_ref]. A single center experience suggests that USCDT may be an equally safe and efficacious treatment for DVT as CDT alone [bib_ref] Ultrasound-accelerated vs standard catheter-directed thrombolysis-a comparative study in patients with iliofemoral deep..., Baker [/bib_ref]. Only one study has previously conducted a direct comparison of USCDT and CDT for the treatment of acute iliofemoral DVT of lower extremity [bib_ref] Ultrasound-assisted versus conventional catheter-directed thrombolysis for acute iliofemoral deep vein thrombosis, Engelberger [/bib_ref]. Forty-eight patients were randomized between CDT and USCDT with a fixed-dose alteplase (20 mg/15 h). The primary outcome of residual thrombus load after 15 h of treatment did not differ between treatment modalities nor did the bleeding rate and quality of life, 3 months after therapy [bib_ref] Ultrasound-assisted versus conventional catheter-directed thrombolysis for acute iliofemoral deep vein thrombosis, Engelberger [/bib_ref].
The incidence and severity of PTS among DVT patients tend to increase during the first 2 years after the DVT [bib_ref] Determinants and time course of the postthrombotic syndrome after acute deep venous..., Kahn [/bib_ref] , and the CAVENT study showed that follow-up had to exceed 6 months in order to detect a benefit of CDT on the incidence of PTS [bib_ref] Long-term outcome after additional catheter-directed thrombolysis versus standard treatment for acute iliofemoral..., Enden [/bib_ref]. Therefore, it is likely assumed that it will take more than 3 months to be able to observe differential impact of USCDT versus CDT on development of PTS. We aimed to compare USCDT with CDT in a case series of patients with DVT on efficacy, safety, degree of PTS, and quality of life after long-term ([12 months) follow-up.
# Methods
Our study is a retrospective case series of consecutive patients admitted to the University Hospital from January 2002 to January 2012 with iliofemoral or vena cava inferior DVT, who were treated with CDT or USCDT. From 2009 on, the patients were preferentially treated with USCDT (85 % of all cases treated after 2009). Due to random lack of US catheters, a few patients were treated with CDT only. Our design thus closely resembles a historical cohort study. Inclusion criteria were an objectively diagnosed DVT using ultrasound examination (64 %) or venography (36 %), extending into vena iliaca externa or vena cava inferior. Exclusion criteria for any catheter-directed thrombolysis were prior cerebral bleeding, a thrombotic or embolic cerebral infarction in the last 3 months, head trauma or major surgery in the last 14 days, bleeding tendency, platelet counts below 100 9 10 9 /L, pregnancy, hypertension (systolic blood pressure above 180 mmHg or diastolic blood pressure above 110 mmHg), renal failure, liver failure, and malignant disease with an expected survival of less than 1 year [bib_ref] Catheter-directed thrombolysis for treatment of deep venous thrombosis in the upper extremities, Vik [/bib_ref]. All patients alive as on April 1, 2013 were eligible for a follow-up visit at the hospital between April 2013 and January 2014, to clinically evaluate the effects of treatment. As it is the obligation of physicians to monitor the efficacy and safety of introduction of new treatment modalities, a review by the Institutional Review Board was not performed. Informed consent was obtained from all the participants included in this study.
All patients were treated with low molecular weight heparin (LMWH) directly after diagnosis of DVT and they underwent an initial venography, followed by thrombolytic therapy. Interventions were patency-driven, and venography was performed at intervals of 24 h. The treatment protocols of CDT and USCDT, including inclusion/exclusion criteria, type and dosage scheme of thrombolytic agent, and evaluations during treatment remained unchanged over the time period from 2002 to 2013, except for the catheters used. The responsible senior consultants at the radiology and hematology departments were experienced physicians and remained unchanged over the time period.
In the USCDT group, the EKOS EkoSonic Ò Endovascular System with MACH 4e (EKOS Corporation, 11911 South Bothell, WA 98011, USA) was used [bib_ref] Catheterdirected ultrasound-accelerated thrombolysis for the treatment of acute pulmonary embolism, Engelhardt [/bib_ref]. Frequency of the ultrasound was 2.05-2.35 MHz with a power of 50.0 W pulse power maximum (30.0 W average) resulting in 0.5 W per ultrasound transducer unit. In the CDT group, a catheter with multiple side holes (UNI-FUSE, AngioDynamics or Cragg-McNamara Ò Valved Infusion Catheter) was used.
Venous access was obtained through popliteal vein in both groups. Balloon dilatation and deflation were performed when deemed necessary while evaluating treatment result. Self-expandable nitinol stents ranging in diameter from 10 to 16 mm were used when deemed necessary by the treating radiologist and hematologist (organized thrombus, clear signs of May-Thurner syndrome or when other external compression was suspected (e-Luminex, Bard; Protégé, EV3 or Sinus-XL stent, Optimed)). These criteria did not change during the study period.
A bolus of 5000 IE unfractionated heparin (UFH) and 5 mg rt-PA (Actilyse; Boehringer Ingelheim, GmbH) was injected through the catheter at the start of both catheterdirected thrombolysis procedures, followed by continuous infusion of rt-PA, 0.01 mg/kg/h, and UFH, 300 IU/kg/ 24 h. The activated partial thromboplastin time (normal range 25-36 s) was measured twice daily and 6 h after adjustment of the UFH dose to keep the APTT between 50 and 70 s during treatment. Plasma fibrinogen was measured twice daily, and if fibrinogen was reduced to or below 1.0 g/L, the infusion of rt-PA was stopped for 2 h and then restarted at 50 % of the original dose. Hemoglobin, platelet count, PT-INR, and d-dimer were measured once daily during treatment. Blood pressure and heart rate were monitored, and inspection of the leg was performed regularly.
After both thrombolytic procedures, anticoagulant therapy proceeded with standard DVT treatment. According to our treatment guidelines, oral anticoagulant treatment with warfarin was recommended for 1 year or indefinitely in patients with stent implantation, with a treatment intensity of 2.0-3.0 PT-INR.
Clot burden was categorized into three groups (length of \10, 10-30, and [30 cm). The degree of thrombolysis after the procedure was visualized by venography and categorized into no effect or progression, less than 50 % lysis (grade I), 50-90 % lysis (grade II), and more than 90 % lysis (grade III). Percentage clot lysis was estimated by the difference in the length of thrombus before versus after treatment. Major bleeding was defined as any clinically overt bleeding that resulted in the cessation of therapy, further hospitalization, death or that required transfusion or surgical intervention. All other bleedings were classified as minor. Lowering of the standard dosing regimen of rt-PA according to the plasma fibrinogen level (see paragraph above) was defined as the outcome 'less thrombolytic dose.'
At follow-up, a new venography was performed to measure the degree of residual luminal obstruction. Patency were defined as open when there was no vessel stenosis and no collateral venous drainage, as stenosis when an open venous segment had significant stenosis (50-90 %) or in case of severe vessel stenosis ([90 %) with collateral venous drainage, and as occluded when total occlusion of a venous segment with collateral venous drainage was observed.
The diagnosis of post-thrombotic syndrome (PTS) was made using the Villalta Scale [bib_ref] Assessment of validity and reproducibility of a clinical scale for the post..., Villalta [/bib_ref]. PTS is categorized by severity: mild (C 5 \ 10 points), moderate (C 10 \ 15 points), and severe (C 15 points). The presence of venous ulcers directly accrues 15 points on the Villalta Scale [bib_ref] Definition of postthrombotic syndrome of the leg for use in clinical investigations:..., Kahn [/bib_ref]. To assess quality of life, we included the Short Form Health Survey-36 (SF-36) [bib_ref] The factor structure of the SF-36 Health Survey in 10 countries: results..., Ware [/bib_ref] and the Venous Insufficiency Epidemiological and Economic Study-Quality of Life/Symptoms (VEINES-QOL/Sym) questionnaires [bib_ref] Evaluation of outcomes in chronic venous disorders of the leg: development of..., Lamping [/bib_ref].
Data were analyzed using IBM SPSS Statistics, version 22 (Armonk, NY, United States of America). Differences in categorical data between treatment modalities were analyzed using Pearson's Chi-squared test, and differences in continuous data were analyzed with Student's t-test or Mann-Whitney U-test when data were not normally distributed. The maximum percentage of missing data for a given variable was 11.8 % within the VEINES or SF-36 questionnaire, and appeared to have a random pattern. Values are expressed as means ± one standard deviation if data were normally distributed and as medians with 25 and 75th percentiles in parenthesis (IQR) if data were not normally distributed.
# Results
We included 94 patients with an iliofemoral or inferior vena cava DVT, accounting for 95 events since one patient experienced a second DVT. Sixty-two patients were treated with CDT (6 of these after 2009) and 33 with USCDT. Characteristics of patients, risk factors for DVT, location of thrombi, and duration of symptoms, stratified by treatment modality are shown in. There were no significant differences between groups.displays the characteristics of venous thrombus and short-term outcomes after treatment. The thrombus burden was equal in both groups before the start of treatment. The proportion of patients that achieved sufficient patency within 48 h was significantly higher after USCDT (27 vs. 10 %, P \ 0.05). The median duration of total hospitalization was shorter for USCDT patients compared to those treated with CDT [6.0 days (IQR 5.0-9.0) vs. 8.0 (IQR 5.8-12.0), P \ 0.05]. Finally, intravenous stenting of residual thrombosis after thrombolysis was less often deemed necessary by the treating team in the USCDT group (30 % vs. 55 %, P \ 0.05). We did not observe any significant differences in the dose of thrombolytic agent used, nor in the degree of thrombolysis achieved immediately after treatment [[90 % of luminal recanalization in 88 % (CDT) vs. 76 % (USCDT)]. Exclusion of patients (n = 8) with symptoms of VTE longer than 14 days, an exclusion criterion in most guidelines [bib_ref] American college of chest physicians evidence-based clinical practice guidelines, Guyatt [/bib_ref] , did not change outcomes (data not shown).
Major non-fatal bleeding occurred in three patients in the USCDT group and in two patients in the CDT group (9 vs. 3 %). There were seven patients (11 %) in the CDT group and one patient (3 %) in the USCDT group who died from any cause during follow-up. In both treatment groups, one patient developed treatment-related recurrent thrombotic event.
Forty-seven of 62 CDT patients (75 %) and 21 of 33 of USCDT patients (64 %) attended the follow-up visit. [fig_ref] Table 3: Long-term outcomes after [/fig_ref] displays long-term follow-up outcomes. Median time to follow-up was 89 months in the CDT group (range 15-141) and 34 in the USCDT group (range 17-51). Twice as much patients were still on anticoagulation in the CDT group than in the USCDT group (70 vs. 33 %, P \ 0.05). Incidence of recurrent events and use of compression stockings (data not shown) was similar between groups.
Fifty-five percent in the USCDT and 52 % in the CDT group developed PTS [fig_ref] Table 3: Long-term outcomes after [/fig_ref]. We observed non-significant lower prevalence of severe PTS in the USCDT group (5 vs. 10 %). There were no significant differences between groups regarding quality of life scores [fig_ref] Table 3: Long-term outcomes after [/fig_ref]. Vascular patency after long-term follow-up did not differ between groups [complete patency 79 (CDT) vs. 76 % (USCDT)].
# Discussion
We found that USCDT was associated with a higher proportion of patients requiring short treatment (\48 h), shorter duration of the hospital stay, and less intravenous stenting of residual thrombosis after thrombolysis. Short-term vessel patency and bleeding complications did not differ between groups. Likewise, long-term vessel patency, prevalence of PTS, and quality of life scores were essentially similar. Thus, we were not able to confirm the previous findings of improved vessel patency [bib_ref] Ultrasound-accelerated thrombolysis for the treatment of deep vein thrombosis: initial clinical experience, Parikh [/bib_ref] [bib_ref] Binding of tissue-plasminogen activator to fibrin: effect of ultrasound, Siddiqi [/bib_ref] [bib_ref] Ultrasound-accelerated vs standard catheter-directed thrombolysis-a comparative study in patients with iliofemoral deep..., Baker [/bib_ref] [bib_ref] Ultrasound reversibly disaggregates fibrin fibers, Braaten [/bib_ref] or the decreased amount of thrombolytic agent used for USCDT. Our findings suggest that USCDT does not have any apparent clinical benefits over CDT alone.
Our findings are in agreement with the randomized study by Engelberger et al. [bib_ref] Ultrasound-assisted versus conventional catheter-directed thrombolysis for acute iliofemoral deep vein thrombosis, Engelberger [/bib_ref] , who neither found any differences in thrombus load when comparing CDT with USCDT nor in vessel patency and incidence of PTS 3 months after treatment. These findings rejected their hypothesis [bib_ref] Postthrombotic morbidity correlates with residual thrombus following catheterdirected thrombolysis for iliofemoral deep..., Comerota [/bib_ref] [bib_ref] Quantity of clot lysed after catheter-directed thrombolysis for iliofemoral deep venous thrombosis..., Grewal [/bib_ref] , assuming that USCDT would improve the reduction of thrombus load directly after diagnosis of an acute DVT, leading to a lower incidence of PTS in the future. Similarly, Baker et al. [bib_ref] Ultrasound-accelerated vs standard catheter-directed thrombolysis-a comparative study in patients with iliofemoral deep..., Baker [/bib_ref] did not find a difference between USCDT (n = 64) and CDT (n = 19) in thrombus resolution directly after the procedure. A possible explanation for the lack of differences between treatment modalities may be too low power of the ultrasound device used [bib_ref] Ultrasound-assisted versus conventional catheter-directed thrombolysis for acute iliofemoral deep vein thrombosis, Engelberger [/bib_ref]. The EKOS MACH4 device that we used has an ultrasound frequency of 2.05-2.35 MHz and 0.5 W power per transducer. Experimental studies have suggested a (7) Acute non-surgical illness 21 (13) 12 (4) Unprovoked 52 (32) 55 (18) With pulmonary embolism 19 (12) 21 (7) Location of DVT CDT catheter-directed thrombolysis, USCDT ultrasound-assisted CDT, IQR interquartile range (25-75th percentile), VTE venous thromboembolic event, DVT deep vein thrombosis a No significant differences between groups thrombolysis optimum of around 2.2 MHz with a higher transmitted power (1, 2, 4, or 8 W per square centimeter) [bib_ref] Binding of tissue-plasminogen activator to fibrin: effect of ultrasound, Siddiqi [/bib_ref] [bib_ref] Characterization of ultrasound-potentiated fibrinolysis in vitro, Blinc [/bib_ref]. Lysis of human clots has been shown to increase significantly when ultrasound was applied at 1.0 or 1.5 W [bib_ref] Ultrasound-assisted thrombolysis, Sehgal [/bib_ref]. The lower power may partly explain the lack of beneficial impact of USCDT treatment.
Our study has some advantages. Most importantly, our study had longer follow-up of patients (minimally 15 months in both groups) than other studies (minimal 3 or 6 months). This allowed us to assess the prevalence of PTS at a time point when most patients are near reaching a stable level of their PTS [bib_ref] Determinants and time course of the postthrombotic syndrome after acute deep venous..., Kahn [/bib_ref] , i.e., looking at real long-term sequels of DVT. Second, we performed a new venography in most attendees after the follow-up, allowing us to associate a subjective outcome as PTS with vessel patency.
Due to the non-randomized design, confounding might have influenced our results. However, a systematic difference is unlikely as the patient and pretreatment characteristics were essentially similar between treatment modalities and between patients admitted for therapy before and after 2009. Second, medical specialists as well as treatment protocols remained the same over the years, thereby reducing the possibility of confounding. However, we cannot exclude the influence of minor unknown confounders that occurred because of time passing. Our study reflects a real life setting and is therefore less prone to selection bias compared to a randomized controlled trial, where participants tend to be younger and healthier compared to the population they are recruited from. An alternative explanation for our finding of a reduced treatment time in the USCDT group might be a learning effect: medical specialists are assumed to be more experienced over time with the techniques and logistics of an intravenous catheter-directed thrombolytic procedure in DVT patients. Shortening of the hospital stay in the USCDT group might reflect a change in general health care recommendations rather than the effect of adding ultrasound to the treatment. This might also be true for our finding that in the USCDT group less intravenous stenting was used, thus reflecting not a true effect of the treatment but more a general growth in insights in interventional radiology.
Another minor drawback of our study is that the adjudicators of the Villalta Scale and quality of life were not blinded for treatment, but this generally increases the chance of a type I error and not that of type II. Finally, the prevalence of PTS (55 % in the CDT group and 52 % in the USCDT group) is somewhat higher than that reported in the literature [bib_ref] Determinants and time course of the postthrombotic syndrome after acute deep venous..., Kahn [/bib_ref]. This might be due to our long followup in both arms, and also by the fact that we included only patients with an iliofemoral or more proximal DVT, who are known to be more prone to develop PTS [bib_ref] Determinants and time course of the postthrombotic syndrome after acute deep venous..., Kahn [/bib_ref]. Selection of participants with PTS at the follow-up clinical examination might also contribute to the high prevalence of PTS.
In conclusion, we found that USCDT leads to a higher proportion of patients that needed short duration of thrombolysis, shorter hospital stay, and less frequent intravenous stenting. In accordance with previous studies, we showed that USCDT was not superior to CDT with regard to short-and long-term vessel patency, the longterm prevalence of PTS, and quality of life. Our findings suggest that USCDT does not have any apparent clinical benefits over CDT alone. However, due to limited available data, a large randomized trial comparing USCDT and CDT with long-term follow-up is warranted. (4) CDT catheter-directed thrombolysis, USCDT ultrasound-assisted CDT, VTE venous thromboembolic event, SD standard deviation * P \ 0.05 Acknowledgments TREC is supported by a grant of the K. G. Jebsen Foundation of Norway.
## Compliance with ethical standards
Conflicts of interest All authors declare that they have no conflicts of interest.
Ethical Standards All procedures performed were in accordance with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. As it is the duty of physicians by law to clinically evaluate the effects of new therapies or procedures, a review by the Institutional Review Board was not performed.
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http:// creativecommons.org/licenses/by/4.0/), which permits unrestricted 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.
[fig] * P \ 0: CDT catheter-directed thrombolysis, USCDT ultrasound-assisted CDT, IQR interquartile range (25-75th percentile), VTE venous thromboembolic event [/fig]
[table] Table 3: Long-term outcomes after (ultrasound-assisted) catheter-directed thrombolysis in 68 cases of iliofemoral or inferior vena cava deep vein thrombosis [/table]
|
Fever as a first presentation of castration-resistant prostate cancer: A case report
Rationale: Cancer is a well-recognized cause of fever, which is related to cytokines produced by malignant cells. Prostate cancer presenting with fever and other inflammatory markers as a paraneoplastic syndrome rarely occurs.Patients concerns and diagnoses: We describe the case of high fever and lower-urinary tract symptoms that progressed 1 month prior to presentation. A 78-year-old man had been diagnosed with prostate cancer 8 months ago. He received androgen deprivation therapy with leuprolide acetate 22.5 mg for every 3 months. Castration-resistant prostate cancer was diagnosed due to elevated prostate specific antigen (1639 ng/mL) and cancer fever.Intervention: The patient received docetaxel-based systemic chemotherapy 50 mg/mm 2 biweekly. Naproxen 500 mg was administered twice a day.Outcomes: After one cycle of systemic chemotherapy, the patient had no major side effects, no more fever was observed, and the systemic condition improved.Conclusion: Differentiating cancer-related fever from infection-related fever is important for appropriate patient management. In this case, fever appeared as the first symptom of castration-resistant prostate cancer and was managed by naproxen and resolved with systemic chemotherapy.Abbreviations: CRPC = castration-resistant prostate cancer, PSA = prostate specific antigen.
# Introduction
Fever is an important host-defense response accompanied by inflammation. It is one of the most common symptoms and signs of outpatients and inpatients. The definition of fever is arbitrary and depends on the purpose for which it is defined. In patients with neutropenia, fever is defined as a single oral temperature of >38.3°C. [bib_ref] Guidelines for evaluation of new fever in critically ill adult patients: 2008..., O'grady [/bib_ref] Cancer can cause fever. The pathophysiology of cancer-induced fever maybe a result of several mechanisms, including release of cytokines from tumor cells, necrosis of tumoral tissue, or obstruction of a hollow duct. Other causes of fever in cancer patients include drug-induced fever (antibiotics or chemotherapeutic agents), thrombotic thrombocytopenic purpura, and deep venous thrombus. [bib_ref] Fever in patients with cancer, Pasikhova [/bib_ref] Cancer-induced fever is associated with lymphoma, leukemia, and solid tumors. Solid tumors that result in cancer-induced fever include renal cell carcinoma, hepatocellular carcinoma, pancreatic carcinoma, and brain tumor. [bib_ref] Malignant causes of fever of unknown origin, Foggo [/bib_ref] Castration-resistant prostate cancer (CRPC) is defined as castrated serum testosterone <50 ng/dL, plus one of following biochemical radiological progressions. [bib_ref] EAU-ESTRO-SIOG guidelines on prostate cancer Part II: treatment of relapsing, metastatic, and..., Cornford [/bib_ref] Primary symptoms of CRPC are bone pain and skeletal events. Fever and inflammatory paraneoplastic syndromes associated with CRPC are extremely rare.
## Case presentation
A 78-year-old Korean man was admitted with high fever (up to 38°C), poor feeding, knee pain, and lower-urinary tract symptoms that progressed over the preceding month. He had been diagnosed with prostate cancer ~8 months ago. At initial admission, the serum prostate specific antigen (PSA) level was over 100 ng/mL and prostate volume was 84 ml. The biopsy result showed Gleason grade was 1 in all cores. A computed tomography showed bladder and seminal vesicle invasion of prostate cancer, multiple enlarged pelvic lymph nodes, and osteoblastic bone metastases in spines, pelvic bones, and both proximal femurs. He was treated with androgen deprivation therapy with leuprolide acetate 22.5 mg every 3 months. He was treated well and had only mild lower urinary tract symptoms. Imaging tests performed 3 months prior to visit showed a stable disease, the serum PSA levels decreased to 13 ng/mL. On physical examination, he had a regular peripheral pulse of 101 bpm and a blood pressure of 140/70 mm Hg. He was febrile with a body temperature of 38.8°C. Laboratory test revealed the following results: hemoglobin 12.8 g/dL; WBC 11,980/μl; platelet 147,000/μL; C-reactive protein 114 mg/dL; serum PSA 1639 ng/mL. Urine analysis revealed no hematuria and pyuria. No bacteria had grown in the urine culture test and blood culture test. COVID-19 test was negative. Serum procalcitonin level had increased to 0.61 ng/mL (<0.1 ng/mL) and so did the serum interleukin 6 level increase to 307 pg/mL (<7.0 pg/mL). According to the opinion of the Infectious Internal Medicine, third-generation intravenous cephalosporin and nonsteroidal anti-inflammatory drugs were administered, but the fever persisted for more than 3 days. Naproxen 500 mg was administered twice a day and the fever improved to a certain extent. No new lesion was observed on the imaging test [fig_ref] Figure 1: Bone scintigraph showing extensive bone metastases [/fig_ref] , but CRPC was diagnosed due to elevated PSA and cancer fever. Hence, docetaxel-based systemic chemotherapy (50 mg/m 2 biweekly) was initiated. After 1 cycle of systemic chemotherapy, the patient had no major side effects, no more fever was observed, and the systemic condition improved. The patient received 22 cycles of systemic docetaxel chemotherapy during the 14-month follow-up and still suffered from a stable disease. The last serum PSA level was 10 ng/mL. This study was approved by Wonkwang University Hospital Institutional Review Board (WKUH 2021-05-018). Patient has provided informed consent for publication for the case.
# Discussion
In cancer patients, fever is a common symptom for various reasons. In certain cancers, cancer itself causes fever, and typical cancers that cause fever are lymphoma, leukemia, and solid tumor. Some specific solid malignancies that result in cancer fever include renal cell carcinoma (by elaboration of interleukin-6), hepatocellular carcinoma, pancreatic carcinoma, bronchogenic carcinoma, and brain tumor. [bib_ref] Malignant causes of fever of unknown origin, Foggo [/bib_ref] Patients with cancer are exposed to a variety of treatment situations, blood products, and medications that may induce fever. Infection is a major cause for it in cancer patients. Risk factors of fever in cancer patients include venous catheters, chemotherapy-induced mucositis, various surgical procedures, and foreign bodies.Deep vein thromboembolism is an important diagnosis that may occur in patients with cancer and should be considered in any patient without clear evidence of infection-associated fever. Diagnostic criteria for cancer fever included a temperature of over 37.8°C at least once each day, lasting for over 2 weeks, a lack of evidence of infection, absence of alleged mechanisms, that is, drug allergy, transfusion reaction, lack of response of fever to an empiric, adequate antibiotics, prompt, complete lysis of fever by the naproxen test.
In any cancer patient presenting with fever, it is very important to distinguish cancer from an infection. Fever can occur when acute prostatitis develops in patients with prostate cancer. Clinically, when acute prostatitis occurs, PSA increases by an average of 20 ng/mL. [bib_ref] Total and free serum prostate specific antigen levels during the first month..., Gamé [/bib_ref] In this case, PSA increased to 1000 ng/ mL or more, and infection could be ruled out because bacteria were not cultured in urine culture tests and blood culture tests. Procalcitonin is an important marker to determine the severity of bacteremia or sepsis. Yaegashi et al suggested that procalcitonin was significantly higher in bacterial infections in advanced urological cancer, which was useful for differential diagnosis. [bib_ref] Differential diagnosis between bacterial infection and neoplastic fever in patients with advanced..., Yaegashi [/bib_ref] In this case, procalcitonin was slightly increased to 0.61 ng/mL, indicating that it was not an infection-induced fever.
Across various solid tumors, fever is caused by several cytokines such as interleukin-2, TNF, and interleukin-6. In particular, interleukin-6 is involved in fever associated with renal cell carcinoma. Inflammatory cytokines are known to play an important role in the progression and prognosis of prostate cancer. Mauri et al report an association between high serum levels of interleukin-6, a pyrogenic molecule, and advanced prostate cancer. [bib_ref] Inflammatory prostate cancer: an underestimated paraneoplastic clinical manifestation, Mauri [/bib_ref] In this case as well, interleukin-6 increased to 307 pg/mL, which can be considered as a factor causing fever. Prostate cancer presenting with fever and other inflammatory markers as a paraneoplastic syndrome is a rare occurrence. All patients with prostate cancer related systemic inflammation had metastatic prostate cancer. This finding suggests that systemic inflammation may be related with high tumor burden and a rapidly progressive course of prostate cancer. Fever and other related symptoms were resolved with androgen deprivation therapy in all cases reported. [bib_ref] Metastatic prostatic cancer as an infrequent cause of fever unknown origin, and..., Yilmaz [/bib_ref] In this case, castration-resistant prostate cancer was diagnosed as a result of a sudden fever in a patient receiving androgen deprivation therapy. In this case, fever was controlled after administration of naproxen and resolved after systemic docetaxel chemotherapy. Although paraneoplastic syndrome is rare in prostate cancer, if fever persists unresponsive to broad spectrum antibiotics, rapid progression of the prostate cancer can be predicted and appropriate treatment decisions should be made.
# Conclusions
Differentiating cancer-related fever from infection-related fever is important for appropriate patient management. In this case, fever appeared as the first symptom of castration-resistant prostate cancer. The fever was controlled by naproxen and resolved with systemic chemotherapy.
[fig] Figure 1: Bone scintigraph showing extensive bone metastases. [/fig]
|
A comprehensive review of the progress of cell migration inducing hyaluronidase 1
The gene cell migration inducing hyaluronidase 1 (CEMIP) is on chromosome 15q25 and codes for a 150-kDa protein with an N-terminal secretion signal, a G8 domain, 2 GG domains, and several repeats. It was first described as a specific protein in the inner ear relating to nonsyndromic hearing loss. Recently, increasing research detected its association in various cancers, determining the progression, metastasis, and prognosis by influencing the proliferation and invasion of the cells. This relation is accomplished through various interacting pathways, such as the Wnt/β-catenin signaling pathway and the epidermal growth factor receptor signaling pathway. Thus, CEMIP could be a novel and potential focus for tumor diagnosis and treatment, but further studies on the regulatory role of CEMIP in vivo and in vitro are still needed. Herein, we summarize the process in recent studies of CEMIP, especially in cancer research.Abbreviations: AMPK = AMP-activated protein kinase, BiP = binding immunoglobulin, CA19-9 = cancer antigen 19-9, CASC19 = cancer susceptibility 19, CCA = cholangiocarcinoma, CDH23 = cadherin-related 23, CEMIP = cell migration-inducing hyaluronidase 1, CRC = colorectal cancer, ECM = extracellular matrix, EGF = epidermal growth factor, EGFR = epidermal growth factor receptor, EMT = epithelial-mesenchymal transition, Fzd = frizzled, GJB2 = gap junction protein beta 2, GSK3β = glycogen synthase kinase 3β, HUGE = human unidentified gene-encoded, MEK1 = mitogen-activated protein kinase 1, NADPH = reduced nicotinamide adenine dinucleotide phosphate, NF-κB = nuclear factor-κB, PKC = protein kinase C, PP2A = protein phosphatase 2A, PTP4A3 = protein tyrosine phosphatase type IVA member 3, SLC26A4 = Solute carrier family 26 member 4, TCF/LEF = T-cell factor/lymphoid enhancer factor, TMEM2L = transmembrane protein 2-like, TUG1 = taurine-upregulated gene 1, WBP11 = WW domain binding protein 11, ZO1 = Zona occludens 1.
# Introduction
Cell migration inducing hyaluronidase 1 (CEMIP), also known as KIAA1199, was recruited in the Human Unidentified Gene-Encoded (HUGE) database as one of 1087 long cDNAs. KIAA family encodes large proteins with an estimated average length of 872 amino acids. [bib_ref] Prediction of the coding sequences of unidentified human genes. XV. The complete..., Nagase [/bib_ref] There are several aliases for KIAA1199, such as colon cancer secreted protein 1 (CCSP1), hyaluronan-binding protein involved in hyaluronan depolymerization (HYBID), and transmembrane protein 2-like (TMEM2L). Currently, KIAA1199 is officially named CEMIP. The gene is situated on chromosome 15q25.1, which also has a brain tumor suppressor gene. [bib_ref] Fine mapping of the keratoconus with cataract locus on chromosome 15q and..., Dash [/bib_ref] [bib_ref] Upregulation of the KIAA1199 gene is associated with cellular mortality, Michishita [/bib_ref] [bib_ref] Analysis of molecular cytogenetic alterations in uterine leiomyosarcoma by array-based comparative genomic..., Raish [/bib_ref] CEMIP was initially reported as a specific protein in the nucleus and cytoplasm of the inner ear. [bib_ref] Repression of KIAA1199 attenuates Wnt-signalling and decreases the proliferation of colon cancer..., Birkenkamp-Demtroder [/bib_ref] It is a 150-kDa protein with a G8 domain in the N-terminal, which contains 5 duplicate β chain pairs and 8 glycine residues . G8 domain is a potential transmembrane structure that is predicted to have signal peptides [bib_ref] The role of CEMIP in tumors: An update based on cellular and..., Chen [/bib_ref]. It had been reported that mutations in the coding region of CEMIP could cause the development of nonsyndromic hearing loss. [bib_ref] Mutations in the gene encoding KIAA1199 protein, an inner-ear protein expressed in..., Abe [/bib_ref] According to recently published studies, in terms of structure, the CEMIP protein has 3 other domains, with a secretion signal at the N terminal. CEMIP has 2 GG structural domains, each consisting of 7 β-strands and 2 α-helices with approximately 100 amino acid residues . According to the phylogenetic tree, these 2 GG domains are derived from separate combinatorial incidents rather than intragene duplication. Additionally, the N-terminal of the 2 GG domains is more homologous to the Dictyostelium protein and the phage gp35 protein. [bib_ref] a domain involved in phage LTF apparatus and implicated in human MEB..., Guo [/bib_ref] such as GG domain, WxxW repeats, and PbH1 repeats . Most G8-containing proteins with transmembrane segments or signal portions are positioned intra-or transmembrane. [bib_ref] a domain involved in phage LTF apparatus and implicated in human MEB..., Guo [/bib_ref] [bib_ref] G8: a novel domain associated with polycystic kidney disease and nonsyndromic hearing..., He [/bib_ref] However, the mature CEMIP is verified to be localized in the cytoplasm instead of cellular membranes. Subcellular localization of CEMIP is limited to the perinuclear space, the ER, and then secreted outside. [bib_ref] The localization of proteins encoded by CRYM, KIAA1199, UBA52, COL9A3, and COL9A1,..., Usami [/bib_ref] CEMIP expression is tightly regulated by genetic and epigenetic regulatory systems. [bib_ref] Early insights into the function of KIAA1199, a markedly overexpressed protein in..., Tiwari [/bib_ref] Two parts around the CEMIP sequence (−1425 to −1135 and −125 to +27) are the enhancers of the gene in disease, specific to bind with NF-κB and AP-1, respectively. In contrast, the primary promoter activity of CEMIP depends on the DNA methylation status, with the active region located in the first intron region, a 1.9-kb long CpG island (−444 to +1509). This region in which methylation level is inversely proportional to the expression of CEMIP can be a potential therapeutic site. The less methylated the base sequence of the CpG island, the more advanced cancer, as evidenced by the demethylation of invasive breast cancer samples. [bib_ref] Early insights into the function of KIAA1199, a markedly overexpressed protein in..., Tiwari [/bib_ref] Therefore, CEMIP regulates its gene expression through genetic and epigenetic mechanisms.
CEMIP can be secreted into the extracellular environment as an exosomal protein. Moreover, CEMIP promotes all 3 branches of the Wnt signaling pathways by modulating Ca 2 + signaling via interacting with the cell-membrane receptor Ephrin A2, WW domain binding protein 11 (WBP11), Protein tyrosine phosphatase type IVA member 3 (PTP4A3) or ER receptor inositol 1,4,5-trisphosphate receptor type 2 (ITPR2). [bib_ref] Early insights into the function of KIAA1199, a markedly overexpressed protein in..., Tiwari [/bib_ref] [bib_ref] Down-regulation of KIAA1199/ CEMIP by miR-216a suppresses tumor invasion and metastasis in..., Zhang [/bib_ref] Furthermore, the survival rate of solid cytoplasmic expression of CEMIP is one-quarter lower than that of nucleus strong expression at 4 years postoperatively. [bib_ref] Repression of KIAA1199 attenuates Wnt-signalling and decreases the proliferation of colon cancer..., Birkenkamp-Demtroder [/bib_ref] The nuclear confinement status of CEMIP illustrates the relative safety of cancer patients. It can also be used for early medical intervention or diagnostic indicators of tumor invasion. [bib_ref] Induction of KIAA1199/CEMIP is associated with colon cancer phenotype and poor patient..., Fink [/bib_ref] [bib_ref] Transcriptional and epigenetic regulation of KIAA1199 gene expression in human breast cancer, Kuscu [/bib_ref] Although CEMIP expression in the nucleus has been found in colorectal adenomas, nuclear localization and overexpression of CEMIP may be associated with the presence of β-catenin regardless of the International Union Against Cancer stage. [bib_ref] Repression of KIAA1199 attenuates Wnt-signalling and decreases the proliferation of colon cancer..., Birkenkamp-Demtroder [/bib_ref] Comparing CEMIP-containing colorectal cancer (CRC) cells and their knockouts suggests that CEMIP is belonged to the Wnt signaling pathway, as changes in 8 Wnt axis genes, including β-catenin expression, share the same regulatory trends.
In this review, we summarize the cellular signaling pathways that CEMIP involves, describe the role of CEMIP in different cancers and hereditary hearing loss, and discuss the prospects of CEMIP research.
## General properties of cemip
It has been reported that CEMIP is involved in several pathways. In the nearly 20 years since the first article on CEMIP and human disease was published, various possible interaction pathways for CEMIP have been reported.
## Cemip/pp2a/stathmin pathway
Protein phosphatase 2A (PP2A) belongs to serine/threonine phosphatases. PP2A is essential in several cellular signaling pathways, primarily regulating cell migration and survival. [bib_ref] Duodenal mucosal risk markers in patients with familial adenomatous polyposis: effects of..., Van Heumen [/bib_ref] Based on co-immunoprecipitation and liquid chromatograph-mass spectrometer peptide sequencing of HCT116 cells, PP2A Cα and B56γ are demonstrated as the putative proteins interacting with CEMIP. Consequent experiments testified that a stable complex could be formed by CEMIP, PP2A Cα, and PP2A B56γ. The formed complex can induce PP2A Cα methylation, which is essential for the enzymatic activation of the PP2A holoenzyme, thus enhancing the phosphatase activity of PP2A. The complex causes the dephosphorylation of the stathmin, a major microtubule-destabilizing phosphoprotein. [bib_ref] KIAA1199 as a potential diagnostic biomarker of rheumatoid arthritis related to angiogenesis, Yang [/bib_ref] Unphosphorylated stathmin binds to and sequesters tubulin dimers to decrease the microtubule polymer, indirectly leading to microtubule instability. [bib_ref] Protein phosphatase 2A: a highly regulated family of serine/threonine phosphatases implicated in..., Janssens [/bib_ref] Microtubules regulate cell migration by driving forces as an integral part of the cytoskeleton. [bib_ref] KIAA1199 promotes metastasis of colorectal cancer cells via microtubule destabilization regulated by..., Zhao [/bib_ref] It has been reported that CEMIP-mediated destabilization of microtubules is essential to enhance the motility, invasion, and metastasis of tumor cells [fig_ref] Figure 2: The proposed model of CEMIP-induced microtubule instability enhanced tumor metastasis [/fig_ref]. [bib_ref] Protein-protein interactions between large proteins: two-hybrid screening using a functionally classified library..., Nakayama [/bib_ref] [bib_ref] Regulation of microtubule dynamic instability in vitro by differentially phosphorylated stathmin, Manna [/bib_ref] 3. AMPK/GSK3β/β-catenin cascade AMP-activated protein kinase (AMPK) has been identified as a cellular energy sensor, which is sensitive to and activated by energy stress. Once activated, AMPK inhibits acetyl-CoA carboxylases, which are responsible for the initial reaction that catalyzes the biosynthesis of fatty acids in all cells. Thus, the consumption of reduced nicotinamide adenine dinucleotide phosphate (NADPH) in fatty acid synthesis is reduced, while the product of NADPH by fueling fatty acid oxidation is elevated. This progress substantially decreases the generation of reactive oxygen species following extracellular matrix (ECM) stripping, averting anoikis and promoting cancer cell survival and migration. [bib_ref] Microtubules in cell migration, Etienne-Manneville [/bib_ref] Elevated reactive oxygen species production leads to glycogen synthase kinase 3β (GSK3β) phosphorylation, thus inhibiting its activation. The GSK3β is the regulator of β-catenin, and phosphorylated (p)-GSK3β can increase β-catenin activity by enhancing its stability. Therefore by contributing to reduced degradation, elevated intracellular concentration, and nuclear cumulation of β-catenin, GSK3β/β-catenin is crucial in the modulation of several genes regulating cellular growth, differentiation, and metastasis. [bib_ref] Regulation of cell migration by dynamic microtubules, Kaverina [/bib_ref] This impact triggers the downstream transcription of specific target genes, enhancing cell survival growth and metastasis.. [bib_ref] AMPK regulates NADPH homeostasis to promote tumour cell survival during energy stress, Jeon [/bib_ref] CEMIP is a putative target gene, and its overexpression appears when the AMPK/ GSK3β/β-catenin cascade is activated [fig_ref] Figure 3: Proposed model of AMPK/GSK3β/β-catenin cascade with CEMIP [/fig_ref]. Such overexpression of CEMIP is detected in colon cancer tissues, indicating that CEMIP may be involved in the progression of local tumors. [bib_ref] The metastasis-associated gene S100A4 is a novel target of beta-catenin/T-cell factor signaling..., Stein [/bib_ref] 4. NF-κb signaling pathway Nuclear factor-κB (NF-κB)-activating cascades play a prominent role in inflammation and immunity response. When aberrantly activated, they increase the expression of proto-oncogenes, anti-apoptotic genes, and cell cycle proteins, promoting cancer metastasis [bib_ref] When prostate cancer meets bone: control by wnts, Emami [/bib_ref] [bib_ref] AMPK/GSK3beta/beta-catenin cascade-triggered overexpression of CEMIP promotes migration and invasion in anoikis-resistant prostate..., Zhang [/bib_ref]. In the NF-κB cascades, NF-κB p50, NF-κB p65, and B-cell lymphoma-3 (BCL-3) are responsible for the carcinogenic potential. [bib_ref] New regulators of NF-kappaB in inflammation, Ghosh [/bib_ref] Based on the study on the transformed keratinocytes with infinite reproduction ability, both BCL-3 and NF-κB p65 promote CEMIP expression levels. [bib_ref] The diverse and complex roles of NF-kappaB subunits in cancer, Perkins [/bib_ref] Figure 1. Schematic representation of the CEMIP secondary structure. SP: Signal peptide. G8 domain: G8 is known for the 8 conserved glycines, and it is predicted to contain an alpha helix and ten beta strands. GG domain: this domain is potentially relevant to CEMIP inactivation, including a beta-beta-alpha fold. It has been identified in members of the eukaryotic FAM3 family members and related proteins with unclear functions. WxxW repeats: this domain with unknown function carries a highly conserved WxxW sequence motif and has no less than 6 conserved cysteine residues. PbH1 repeats: Four parallel beta-helix repeats in the CEMIP. Proteins containing the repeats most often are enzymes with polysaccharide substrates. However, the function of the CEMIP PbH1 domain remains elusive. CEMIP = cell migration inducing hyaluronidase 1.
## Cemip as a linkage between semaphorin 3a/ plexin 2a and egfr signaling
Plexin A2 was identified to bind with the G8 domain of CEMIPthrough yeast 2-hybrid experiments. [bib_ref] Protein phosphatase 2A: a highly regulated family of serine/threonine phosphatases implicated in..., Janssens [/bib_ref] [bib_ref] The diverse and complex roles of NF-kappaB subunits in cancer, Perkins [/bib_ref] Typically, Plexin A2 is responsible for class 3 signaling semaphorin bound to Neurospherins, and sustained activation of class 3 signaling Semaphorin can lead to apoptotic cell death. [bib_ref] Cloning of the chromosome translocation breakpoint junction of the t(14;19) in chronic..., Mckeithan [/bib_ref] [bib_ref] NF-kappaB-induced KIAA1199 promotes survival through EGFR signalling, Shostak [/bib_ref] Mechanistic investigations revealed that CEMIP blocks the semaphorin 3Aand plexin A2-dependent apoptotic pathway in cervical cancer cells as they trigger apoptosis only on CEMIP deficiency. How CEMIP blocks the apoptosis may depend on promoting the stability and signaling of epidermal growth factor receptor (EGFR), and the result supports that cell death rises even more dramatically in response to semaphorin 3A stimulation in EGFR-deficient cells. In addition, it suggests that semaphorin 3A can drive the apoptotic pathway when the activity of EGFR is inhibited. [bib_ref] The diverse and complex roles of NF-kappaB subunits in cancer, Perkins [/bib_ref] Moreover, as an EGFR-binding protein, CEMIP can promote phosphorylations of EGFR, sarcoma gene, and mitogen-activated protein kinase 1 (MEK1), indicating that CEMIP links EGFR to downstream kinases. Of note, as epidermal growth factor (EGF) can trigger epithelial-mesenchymal transition (EMT) in CaSki cells, CEMIP may relate to the EMT [fig_ref] Figure 4: The proposed model of Semaphorin 3A/Plexin 2A and EGFR [/fig_ref]. Results of immunofluorescence analysis demonstrate that deficiency of CEMIP disturbs the re-localization of E-cadherin and zona occludens 1(ZO1) in CaSki cells dependent on EGF-mediated stimulation. [bib_ref] The diverse and complex roles of NF-kappaB subunits in cancer, Perkins [/bib_ref] Therefore, CEMIP promotes EGF-induced EMT and relates to cell invasion in cervical cancer cells.
## Wnt/β-catenin signaling pathway
The Wnt signaling pathway, as a multichannel signaling pathway, is initiated by the binding between ligand Wnt protein and related receptors in the membrane. Wnt proteins primarily bind to Frizzled (Fzd) and low-density lipoprotein receptor-related protein 5/6 (LRP5/6), and both are cell surface receptors, thereby inhibiting GSK-3β and casein kinase 1 (CK1). This repression facilitates the stability of cytoplasmic β-catenin and the ultimate translocation of β-catenin to the nucleus. The interaction with the T-cell factor/lymphoid enhancer factor (TCF/ LEF) happens, and the specific gene is activated. [bib_ref] Semaphorin 3B inhibits the phosphatidylinositol 3-kinase/Akt pathway through neuropilin-1 in lung and..., Castro-Rivera [/bib_ref] The results of chromatin immunoprecipitation analysis showed that 4 binding regions surround the CEMIP locus for TCF4. Since TCF4 protein is a critical transcription factor in the Wnt/β-catenin signaling, CEMIP has been deemed a target gene for this pathway. [bib_ref] The semaphorins: versatile regulators of tumour progression and tumour angiogenesis, Neufeld [/bib_ref] [bib_ref] KIAA1199 and its biological role in human cancer and cancer cells (review), Zhang [/bib_ref] As a core component of this pathway, β-catenin plays a prominent role in regulating multiple genes that participate in cell proliferation, differentiation, and invasion. [bib_ref] KIAA1199 and its biological role in human cancer and cancer cells (review), Zhang [/bib_ref] CEMIP and Wnt are both highly expressed in colorectal adenomas and carcinomas, indicating that CEMIP is a positively regulated protein in this pathway and a presumed marker of transformation in colorectal adenomas. [bib_ref] Identification of CRYM as a candidate responsible for nonsyndromic deafness, through cDNA..., Abe [/bib_ref] Moreover, according to an experiment in NCI-N87 and AGS cells, knockdown of CEMIP decreases β-catenin expression. Thus, in the cytoplasm, the accumulation of β-catenin is disturbed, causing less β-catenin to enter the nucleus. This process suppresses the initiation of the Wnt/β-catenin signaling pathway. Afterward, both decreased expression of c-Myc and cyclin D1, 2 major downstream proteins involved in this pathway, is detected. Based on these results, it is proposed that CEMIP correlates with cancer invasion and metastasis ability through the Wnt/β-catenin signaling pathway [fig_ref] Figure 5: Proposed model of CEMIP in Wnt/β-catenin pathway [/fig_ref]. [bib_ref] Wnt/beta-catenin signaling and small molecule inhibitors, Voronkov [/bib_ref] 6.1. CEMIP and tumorigenesis CEMIP has been associated with several human diseases or conditions characterized by abnormal cell migration and proliferation . Based on the studies, several kinds of tumors are involved. However, although CEMIP upregulation is associated with promoting tumors in most cancers, CEMIP may be antineoplastic in specific cancers.
## Colorectal cancer.
Recent studies have shown that CEMIP upregulation is related to poor survival and neoplastic recurrence in CRC patients. Furthermore, CEMIP is identified as a potential target of miR-140-3p, promoting the proliferation and invasion of CRC cells and reverting miR-140-3p-induced anti-proliferative effects by modulating the expression of several epithelial marker proteins. Additionally, miR-140-3p expression is negatively associated with CEMIP and is an independent prospective factor for tumor regression for CRC patients. [bib_ref] Transcriptome profile of human colorectal adenomas, Sabates-Bellver [/bib_ref] Several studies have demonstrated that down-regulating CEMIP by miR-216a can inhibit the invasion and metastasis of CRC cells, revealing the regulatory role of the miRNA-CEMIP axis in CRC. [bib_ref] KIAA1199 promotes migration and invasion by Wnt/beta-catenin pathway and MMPs mediated EMT..., Jia [/bib_ref] Individual studies have identified the relationship between the expression of CEMIP and EMT. Taurineupregulated gene 1 (TUG1) promotes the expression of CEMIP and accelerates EMT and metastasis of CRC cells by inhibiting the expression of miR-600. [bib_ref] Cell migration-inducing hyaluronan-binding protein is regulated by miR-140-3p and promotes the growth..., Yang [/bib_ref] Furthermore, in vitro experiments have shown that overexpression of cancer susceptibility 19 (CASC19) can enhance the proliferation and migration of CRC cells with the upregulation of CEMIP and EMT markers. [bib_ref] Down-regulation of KIAA1199/ CEMIP by miR-216a suppresses tumor invasion and metastasis in..., Zhang [/bib_ref]
## Breast cancer. cemip expression is upregulated in invasive breast cancer specimens and negatively correlated to patient survival. cemip silencing in mda-mb-435 cancer cells leads to the transformation of mesenchymal cells into epithelial cells, thereby reducing the ability of cells to migrate in vitro.
Gain-of-function experiments provide further confirmation of the role of CEMIP in migration. Enhanced cell migration by CEMIP requires ER localization, which binds to its partner binding immunoglobulin (BiP) to form a stable complex. A novel ER-retention motif is found in CEMIP, which is necessary for ER localization and BiP interaction. It is found that CEMIP mediates ER calcium leakage, resulting in an accumulation of cytoplasmic calcium, thus leading to activation of protein kinase C (PKC) and, ultimately, cell migration. [bib_ref] LncRNA TUG1 promoted KIAA1199 expression via miR-600 to accelerate cell metastasis and..., Sun [/bib_ref] MDA-MB-231 and HS578T cell lines decrease cell motility and proliferation ability after the CEMIP gene is knocked out. In addition, quantitative proteomics analysis shows that CEMIP knockout in MDA-MB-231 affects many cell functions, including apoptosis, metabolism, and cell motility. [bib_ref] Functional role of long non-coding RNA CASC19/miR-140-5p/CEMIP axis in colorectal cancer progression..., Wang [/bib_ref] In general, tumor cells are in a strict microenvironment because of their accelerated and uncontrolled growth rate and demands for high nutritional supplementation. Individual studies have provided insights into the link between CEMIP and BiP expression and their role in promoting survival in hypoxia. BiP signaling downstream of CEMIP regulates cell resistance to hypoxia. Decreasing BiP in cells via regulating CEMIP expression makes the cells sensitive to hypoxia therapy, reduces glucose uptake, and leads to tumor regression in vivo. [bib_ref] Unraveling the role of KIAA1199, a novel endoplasmic reticulum protein, in cancer..., Evensen [/bib_ref] 6.1.3. Pancreatic cancer. In previous studies, CEMIP has been investigated and reported to be correlated to early detection, cancer cell migration, invasion, and poor prognosis. [bib_ref] Functional proteomic Analysis reveals the involvement of KIAA1199 in breast cancer growth,..., Jami [/bib_ref] [bib_ref] CEMIP upregulates BiP to promote breast cancer cell survival in hypoxia, Banach [/bib_ref] [bib_ref] KIAA1199/CEMIP/HYBID overexpression predicts poor prognosis in pancreatic ductal adenocarcinoma, Koga [/bib_ref] Suh et al proposed that CEMIP is probably valuable for the early detection of pancreatic cancer, while Koga et al showed that it is associated with cancer prognosis. [bib_ref] Functional proteomic Analysis reveals the involvement of KIAA1199 in breast cancer growth,..., Jami [/bib_ref] Increased expression of CEMIP may partially promote the aggressive phenotype by increasing the concentration of low molecular weight hyaluronan. A possible link between inflammation and enhanced migration is induced by CEMIP in pancreatic ductal adenocarcinoma. [bib_ref] CEMIP upregulates BiP to promote breast cancer cell survival in hypoxia, Banach [/bib_ref] However, the small sample size makes it difficult to draw a reliable conclusion based on these previous experiments. Recently, a novel point has explored the relationship between CEMIP and pancreatic cancer. CEMIP can be a potential marker for pancreatic cancer. The integrated measurement of serum cancer antigen 19-9 (CA19-9) and CEMIP concentration could be a novel laboratory method for diagnosing pancreatic cancer clinically. [bib_ref] KIAA1199 is induced by inflammation and enhances malignant phenotype in pancreatic cancer, Kohi [/bib_ref] 6.1.4. Gastric cancer. CEMIP is upregulated in gastric cancer tissues and correlates to poorer clinical results in gastric cancer. CEMIP promotes the invasion and migration of gastric cancer cells by enhancing the Wnt/β-catenin pathway and matrix metalloproteinases-mediated EMT progression. [bib_ref] Wnt/beta-catenin signaling and small molecule inhibitors, Voronkov [/bib_ref] 6.1.5. Lung cancer. In nonsmall cell lung cancer cells, reduced CEMIP suppresses proliferation and migration of nonsmall cell lung cancer cells and down-regulated the expression of several transcription factors related to the EMT process and EGFR signaling. [bib_ref] Identification of KIAA1199 as a biomarker for pancreatic intraepithelial neoplasia, Suh [/bib_ref] 6.1.6. Chondrosarcoma. Recently, individual studies have shown that increased expression of CEMIP contributes to antitumor activity. CEMIP expression suppresses cell invasion and migration in rat chondrosarcoma cells rather than affecting cell proliferation and apoptosis. Additionally, due to changes in the tumor microenvironment, such as inhibition of ECM formation, CEMIP expression prominently inhibits the growth of transplanted tumors and suppresses the staining ability of Achillean blue in tumor tissues. [bib_ref] Combined use of CEMIP and CA 19-9 enhances diagnostic accuracy for pancreatic..., Lee [/bib_ref] The antitumor effect of CEMIP provides a new idea for us to carry out new research.
6.1.7. Prostate cancer. CEMIP is reported to involve prostate cancer proliferation, [bib_ref] Long noncoding RNA HCP5 promotes prostate cancer cell proliferation by acting as..., Hu [/bib_ref] migration, and invasion. [bib_ref] AMPK/GSK3β/β-catenin cascade-triggered overexpression of CEMIP promotes migration and invasion in anoikis-resistant prostate..., Zhang [/bib_ref] It was found that CEMIP promotes anoikis resistance by enhancing protective autophagy [bib_ref] ATF4/CEMIP/PKCα promotes anoikis resistance by enhancing protective autophagy in prostate cancer cells, Yu [/bib_ref] [bib_ref] CircCEMIP promotes anoikis-resistance by enhancing protective autophagy in prostate cancer cells, Yu [/bib_ref] and facilitates ferroptosis resistance by promoting cystine uptake [bib_ref] CEMIP promotes extracellular matrix-detached prostate cancer cell survival by inhibiting ferroptosis, Liu [/bib_ref] in prostate cancer cells during ECM, thus promoting metastasis formation. These studies provide new insight into therapeutic strategy development for prostate cancer.
6.1.8. Ovarian cancer. CEMIP was found to play an essential role in ovarian cancer progression. CEMIP expression is significantly upregulated in ovarian cancer tissues. Ovarian cancer cells' migration and invasion capacity are significantly decreased, and the proportion of apoptotic cells increases after silencing CEMIP. [bib_ref] CEMIP promotes ovarian cancer development and progression via the PI3K/ AKT signaling..., Shen [/bib_ref] 6.1.9. Other cancers. CEMIP is also reported to involve in other cancers, such as gallbladder cancer, [bib_ref] Lenvatinib induces anticancer activity in gallbladder cancer by targeting AKT, Ye [/bib_ref] endometrial cancer, [bib_ref] Identification of predictive biomarkers for lymph node involvement in obese women with..., Lopez-Ozuna [/bib_ref] and papillary thyroid cancer. [bib_ref] KIAA1199, a target of microRNA-486-5p, promotes papillary thyroid cancer invasion by influencing..., Jiao [/bib_ref]
## Cemip and hereditary hearing loss
Hereditary hearing loss is a highly genetically heterogeneous sensory disorder. According to the cDNA microarray analysis and semiquantitative RT-PCR experiments, CEMIP has predominantly expressed in Deiters' cells and the spiral ligament of the inner ear, especially. [bib_ref] KIAA1199 and its biological role in human cancer and cancer cells (review), Zhang [/bib_ref] Based on the screening results of the CEMIP gene for mutations in patients with nonsyndromic hearing loss in Japan, 3 possible point mutations related to this disease are selected: an Arg187-to-Cys (R187C) mutation, an Arg187-to-His (R187H) mutation, and a His783-to-Tyr (H783Y) mutation. The H783Y mutation exhibits an abnormal pattern of cytoplasmic distribution that may be the basis of the molecular mechanism of hearing impairment.
Further transiently transfection experiments demonstrated that the R187C and R187H mutations do not affect the subcellular localization of the gene product in vitro, while the H783Y mutation emerged in a worm-eaten pattern, suggesting the potential involvement of hearing impairment. [bib_ref] Mutations in the gene encoding KIAA1199 protein, an inner-ear protein expressed in..., Abe [/bib_ref] However, based on the large-scale screening, in Japanese patients, the primary causes of hearing loss are mutations in gap junction protein beta 2 (GJB2), solute carrier family 26 member 4 (SLC26A4), cadherin-related 23 (CDH23), and mitochondrial DNA 1555A→G mutation. In independent autosomal dominant families, mutations of CEMIP and multiple related genes have been detected. [bib_ref] Downregulation of KIAA1199 by miR-486-5p suppresses tumorigenesis in lung cancer, Wang [/bib_ref] Also, in various supporting cells in the organ of Corti, distribution of CEMIP is detected, but the specific function of CEMIP in these cells is still unclear. [bib_ref] Downregulation of KIAA1199 by miR-486-5p suppresses tumorigenesis in lung cancer, Wang [/bib_ref]
## Prospects of cemip research
Previous results suggested that CEMIP might be a diagnostic marker for several cancers, such as cholangiocarcinoma (CCA), gastric cancer, and CRC. [bib_ref] Serum KIAA1199 is an advanced-stage prognostic biomarker and metastatic oncogene in cholangiocarcinoma, Zhai [/bib_ref] [bib_ref] Down-regulation of KIAA1199/ CEMIP by miR-216a suppresses tumor invasion and metastasis in..., Zhang [/bib_ref] [bib_ref] Clinical significance of KIAA1199 as a novel target for gastric cancer drug..., Oneyama [/bib_ref] As a secreted protein, overexpression of CEMIP is detected in serum levels in CCA patients before curative surgery. Moreover, the attenuated serum CEMIP level may be a marker for estimated poor prognosis in patients with CCA. Based on clinical records, the serum CEMIP as a diagnostic marker is more accurate than the traditional diagnostic marker CA19-9, suggesting that CEMIP can serve as a novel cancer detection marker. [bib_ref] Serum KIAA1199 is an advanced-stage prognostic biomarker and metastatic oncogene in cholangiocarcinoma, Zhai [/bib_ref] However, much previous research relating to CEMIP and cancers only focused on verifying the potential role of CEMIP in diverse cell lines, and the overexpression of secreted CEMIP was always ignored. Additionally, as the potential antitumor ability in chondrosarcoma, the relation between serum level and cancer diagnosis and prognosis becomes slightly unclear. Further research in vitro needs to be performed to verify whether the serum level of CEMIP can serve as a standard marker in clinical medicine, facilitating the early phase diagnosis and prognosis.
In addition, human microRNA genes are detected frequently in cancer-related genomic regions, participating in the diverse biological and physiological processes involving proliferation and migration. [bib_ref] Forced expression of KIAA1199, a novel hyaluronidase, inhibits tumorigenicity of low-grade chondrosarcoma, Koike [/bib_ref] [bib_ref] The responsible genes in Japanese deafness patients and clinical application using Invader..., Usami [/bib_ref] The expression level of CEMIP is negatively associated with miR-216a, and this correlation is verified in CRC tumor tissues. The luciferase reporter assays confirm the association, indicating that CEMIP is the direct target of miR-216a. Also, another microRNA miR-188-5p is reported as the regulator in RA, suggesting that microRNA-KIAA199 interacting plays a prominent role in CEMIP-related diseases. [bib_ref] Repression of KIAA1199 attenuates Wnt-signalling and decreases the proliferation of colon cancer..., Birkenkamp-Demtroder [/bib_ref] However, the detailed mechanism of microRNA still awaits elucidation. Whether the microRNA interact with other binding proteins in the regulation procedure? What domain is the most selected binding site? Can specific microRNAs serve as therapeutic tools for CEMIP-related diseases such as cancer? These open questions need more research to be clarified.
# Conclusion remarks
CEMIP is a prominent element in hearing loss, keratoconus with cataracts, rheumatic heart, osteoarthritis, and cancers as a multidomain protein involved in various interacting pathways. It is usually deemed as a protumor factor in the progression of cancers. Overexpression is related to metastasis and poor prognosis of cancers that several CEMIP and tumors research detected. However, another study indicated that CEMIP has antitumor characteristics, leading to a more complex role in cancer regulation. As a potential marker of cancers, CEMIP can associate with several types of microRNAs to be a novel therapeutic target for cancer treatment. Overall, CEMIP is a novel target for cancers and other human diseases such as CRC, CCA, and prostate cancer, further investigations on the regulation of CEMIP are awaited. The function of CEMIP in different cancers.
## Cancer
## Function
## References in vivo
In vitro
Colorectal cancer Diagnosis, medical intervention, and prognosis Cell proliferation, invasion, and migration [bib_ref] Repression of KIAA1199 attenuates Wnt-signalling and decreases the proliferation of colon cancer..., Birkenkamp-Demtroder [/bib_ref] [bib_ref] Early insights into the function of KIAA1199, a markedly overexpressed protein in..., Tiwari [/bib_ref] [bib_ref] Induction of KIAA1199/CEMIP is associated with colon cancer phenotype and poor patient..., Fink [/bib_ref] [bib_ref] KIAA1199 promotes metastasis of colorectal cancer cells via microtubule destabilization regulated by..., Zhao [/bib_ref] [bib_ref] The metastasis-associated gene S100A4 is a novel target of beta-catenin/T-cell factor signaling..., Stein [/bib_ref] [bib_ref] Semaphorin 3B inhibits the phosphatidylinositol 3-kinase/Akt pathway through neuropilin-1 in lung and..., Castro-Rivera [/bib_ref] [bib_ref] Identification of CRYM as a candidate responsible for nonsyndromic deafness, through cDNA..., Abe [/bib_ref] [bib_ref] Transcriptome profile of human colorectal adenomas, Sabates-Bellver [/bib_ref] [bib_ref] Cell migration-inducing hyaluronan-binding protein is regulated by miR-140-3p and promotes the growth..., Yang [/bib_ref] [bib_ref] Down-regulation of KIAA1199/ CEMIP by miR-216a suppresses tumor invasion and metastasis in..., Zhang [/bib_ref] [bib_ref] LncRNA TUG1 promoted KIAA1199 expression via miR-600 to accelerate cell metastasis and..., Sun [/bib_ref] [bib_ref] Functional role of long non-coding RNA CASC19/miR-140-5p/CEMIP axis in colorectal cancer progression..., Wang [/bib_ref] [bib_ref] Down-regulation of KIAA1199/ CEMIP by miR-216a suppresses tumor invasion and metastasis in..., Zhang [/bib_ref] Breast cancer Diagnosis, medical intervention, and prognosis Cell migration, proliferation, apoptosis, and metabolism [bib_ref] Early insights into the function of KIAA1199, a markedly overexpressed protein in..., Tiwari [/bib_ref] [bib_ref] Transcriptional and epigenetic regulation of KIAA1199 gene expression in human breast cancer, Kuscu [/bib_ref] [bib_ref] Semaphorin 3B inhibits the phosphatidylinositol 3-kinase/Akt pathway through neuropilin-1 in lung and..., Castro-Rivera [/bib_ref] [bib_ref] LncRNA TUG1 promoted KIAA1199 expression via miR-600 to accelerate cell metastasis and..., Sun [/bib_ref] [bib_ref] Functional role of long non-coding RNA CASC19/miR-140-5p/CEMIP axis in colorectal cancer progression..., Wang [/bib_ref] [bib_ref] Unraveling the role of KIAA1199, a novel endoplasmic reticulum protein, in cancer..., Evensen [/bib_ref] [bib_ref] Functional proteomic Analysis reveals the involvement of KIAA1199 in breast cancer growth,..., Jami [/bib_ref] [bib_ref] CEMIP upregulates BiP to promote breast cancer cell survival in hypoxia, Banach [/bib_ref] Pancreatic cancer Diagnosis and prognosis Cell migration and invasion [bib_ref] KIAA1199/CEMIP/HYBID overexpression predicts poor prognosis in pancreatic ductal adenocarcinoma, Koga [/bib_ref] [bib_ref] KIAA1199 is induced by inflammation and enhances malignant phenotype in pancreatic cancer, Kohi [/bib_ref] [bib_ref] Identification of KIAA1199 as a biomarker for pancreatic intraepithelial neoplasia, Suh [/bib_ref] [bib_ref] Combined use of CEMIP and CA 19-9 enhances diagnostic accuracy for pancreatic..., Lee [/bib_ref] Gastric cancer Diagnosis Cell proliferation and migration Lung cancer Cell proliferation and migration [bib_ref] Downregulation of KIAA1199 by miR-486-5p suppresses tumorigenesis in lung cancer, Wang [/bib_ref] Chondrosarcoma Diagnosis and tumorigenesis Cell invasion and migration [bib_ref] Serum KIAA1199 is an advanced-stage prognostic biomarker and metastatic oncogene in cholangiocarcinoma, Zhai [/bib_ref] [bib_ref] Forced expression of KIAA1199, a novel hyaluronidase, inhibits tumorigenicity of low-grade chondrosarcoma, Koike [/bib_ref] Prostate cancer Tumorigenesis Cell proliferation, migration, invasion, and apoptosis [bib_ref] Protein-protein interactions between large proteins: two-hybrid screening using a functionally classified library..., Nakayama [/bib_ref] [bib_ref] Long noncoding RNA HCP5 promotes prostate cancer cell proliferation by acting as..., Hu [/bib_ref] [bib_ref] AMPK/GSK3β/β-catenin cascade-triggered overexpression of CEMIP promotes migration and invasion in anoikis-resistant prostate..., Zhang [/bib_ref] [bib_ref] ATF4/CEMIP/PKCα promotes anoikis resistance by enhancing protective autophagy in prostate cancer cells, Yu [/bib_ref] [bib_ref] CircCEMIP promotes anoikis-resistance by enhancing protective autophagy in prostate cancer cells, Yu [/bib_ref] [bib_ref] CEMIP promotes extracellular matrix-detached prostate cancer cell survival by inhibiting ferroptosis, Liu [/bib_ref] Ovarian cancer Cell migration, invasion, and apoptosis [bib_ref] CEMIP promotes ovarian cancer development and progression via the PI3K/ AKT signaling..., Shen [/bib_ref] CEMIP = cell migration inducing hyaluronidase 1.
[fig] Figure 2: The proposed model of CEMIP-induced microtubule instability enhanced tumor metastasis. CEMIP interacts with PP2A (protein phosphatase 2A) and enhances its phosphatase ability, catalyzing dephosphorylation of stathmin. Unphosphorylated stathmin disturbs the microtubule formation, thus resulting in microtubule instability and, consequently, tumor metastasis. CEMIP = cell migration inducing hyaluronidase 1. [/fig]
[fig] Figure 3: Proposed model of AMPK/GSK3β/β-catenin cascade with CEMIP. AMPK (AMP-activated protein kinase) promotes the phosphorylation of GSK3β (glycogen synthase kinase 3β), thus inhibiting its β-catenin degradation ability. β-Catenin can then enter the nucleus as a transcription factor, enhancing the expression of CEMIP. CEMIP = cell migration inducing hyaluronidase 1. [/fig]
[fig] Figure 4: The proposed model of Semaphorin 3A/Plexin 2A and EGFR (epidermal growth factor receptor) signaling is linked by CEMIP. Semaphorin triggers apoptosis in a plexin A2-dependent manner. At the same time, EGF promotes cell survival and EMT (epithelial-mesenchymal transition) by binding to EGFR. CEMIP serves as a bridge between these 2 pathways by inhibiting plexin A2 and supporting EGFR. CEMIP = cell migration inducing hyaluronidase 1. [/fig]
[fig] Figure 5: Proposed model of CEMIP in Wnt/β-catenin pathway. When the classical Wnt signaling pathway is turned on, the destruction complex formed by Axin, CK1 (casein kinase 1), APC (APC regulator of Wnt signaling pathway), and GSK3β (glycogen synthase kinase 3b) is inactivated to stabilize the β-catenin. Thus, the β-catenin can be transported to the nucleus for modulating transcription of CEMIP. CEMIP = cell migration inducing hyaluronidase 1. www.md-journal.com [/fig]
|
PAI1 is a Marker of Bad Prognosis in Rectal Cancer but Predicts a Better Response to Treatment with PIM Inhibitor AZD1208
Colorectal cancer (CRC) is the third most common cancer worldwide. The standard treatment in locally advanced rectal cancer is preoperative radiation alone or in combination with chemotherapy, followed by adjuvant chemotherapy. Rectal cancer is highly lethal, with only 20% of patients showing a complete remission (by RECIST) after standard treatment, although they commonly show local or systemic relapse likely due to its late detection and high chemotherapy resistance, among other reasons. Here, we explored the role of PAI1 (Serpin E1) in rectal cancer through the analyses of public patient databases, our own cohort of locally advanced rectal cancer patients and a panel of CRC cell lines. We showed that PAI1 expression is upregulated in rectal tumors, which is associated with decreased overall survival and increased metastasis and invasion in advanced rectal tumors. Accordingly, PAI1 expression is correlated with the expression of (Epithelial-to-Mesenchymal Transition) EMT-associated genes and genes encoding drug targets, including the tyrosine kinases PDGFRb, PDGFRa and FYN, the serine/threonine kinase PIM1 and BRAF. In addition, we demonstrate that cells expressing PAI1 protein are more sensitive to the PIM inhibitor AZD1208, suggesting that PAI1 could be used to predict response to treatment with PIM inhibitors and to complement radiotherapy in rectal tumors.
# Introduction
Colorectal cancer (CRC) is the third most common cancer worldwide (10.2%), and the second in Europe (12%). Specifically, cancers of the rectum and rectosigmoid junction represent the 30% of all diagnosed CRC cases. In 2018, there were approximately 200,000 new cases of diagnosed rectal cancer in Europe. Rectal cancers comprise tumors arising within 15 cm of the anal verge. Although these tumors are histologically similar to global CRC tumors, their anatomical situation, invasive growth pattern, surgical approach and treatment outcomes make them a different entity. The standard treatment for patients with locally advanced rectal cancer is preoperative radiotherapy alone or in combination with chemotherapy. The common chemotherapy agents include 5-Fluorouracil
## Immunohistochemistry assays
Tumor samples from HUVR were obtained from surgical resection of rectal cancer performed, stored in TMA blocks. Briefly, four-micrometer-thick tissue sections from paraffin blocks were dewaxed in xylene and rehydrated in a series of graded alcohols. Sections were immersed in 3% H 2 O 2 aqueous solution for 15 min. to exhaust endogenous peroxidase activity and then covered with 1%-blocking reagent (Roche) in PBS to block nonspecific binding sites. Antigen retrieval was performed with a PT Link instrument (Agilent), using citrate/EDTA buffer. Sections were incubated with primary antibody overnight at 4 - C in a humid chamber. Peroxidase-labeled secondary antibodies (visualization reagent-HRP, Dako) and 3,3-diaminobenzidine were applied to develop immunoreactivity, according to the manufacturer's protocol. Slides were then counterstained with hematoxylin and mounted in DPX (BDH Laboratories, Poole, UK). Sections in which primary antibody was omitted were used as negative controls. Immunostaining was evaluated independently by two observers. The percentage of immunostained tumour cells was scored as follows: 0, negative; 1, < 10%; 2, 10-49% and 3, > 50%.
## Cell culture
Cells were obtained from The European Collection of Authenticated Cell Cultures (ECACC) and cultured according to the manufacturer's instructions. Briefly, SW48, SW480, T84, LS180, LOVO, HT29 HCT116 and COLO205 were cultured in the corresponding medium and incubated at 37 - C in 5% CO 2 in a humidified atmosphere. The features of the cell lines used in this study are shown in . All cell lines were authenticated and regularly tested for mycoplasma.
## Cytotoxic mtt assay.
A total of 2 × 10 4 cells were seeded in 96-well plates and treated with Desatinib, Vemurafenib, or AZD1208 after 24 h. 96 h later, cell viability was measured by the MTT assay and confirmed by crystal violet staining using a iMark Microplate reader (BioRad). The IC50 was calculated using GraphPad Prism 7 software (San Diego, CA, USA).
## Western blot analyses
Western blotting was performed as previously described. Briefly, cells were grown in serum-free medium for 48 h and then both total cell extracts and medium were used to determine the intracellular and extracellular protein levels of PAI1, respectively. For intracellular determination, cells were washed twice with PBS and lysed by sonication in lysis buffer (50 mM Tris-HCl, pH 7.5; 1% NP-40; 1 mM Na 3 VO 4 ; 150 mM NaCl; 20 mM Na 4 P 2 O 7 ; 100 mM NaF; 1% Na-deoxycholate; 0.1% SDS; 1 mM EDTA; complete phosphatase inhibitor cocktail (Sigma, San Luis, MO, USA) and complete protease inhibitor cocktail (Sigma)). For extracellular determination, medium was filtered using a 10-kDa Cells 2020, 9, 1071 4 of 17 ultrafiltration tube (Millipore, Burlington, MA, USA) and centrifuged at 3500× g at 4 - C for 30 min. Samples were then run on 6-15% SDS-PAGE gels, transferred to Nitrocellulose membranes (Protran BA83, Whatman, GE Healthcare, Chicago, IL, USA) and immunostained. Primary antibodies used were anti-PAI-1 (1:1000; ab66705) and MAb anti-α-tubulin (1:5000; Sigma 9026); secondary antibody was horseradish peroxidase-labeled rabbit anti-mouse (1:5000; Amersham, GE Healthcare, Chicago, IL, USA). Protein visualization was performed using the ECL detection system (Amersham, GE Healthcare, Chicago, IL, USA). Relative PAI1 protein quantification was performed relative to α-tubulin or total protein (Ponceau staining) for intracellular or extracellular protein, respectively.
## Quantification and statistical analysis.
All statistical analyses were performed using the SPSS statistical software, Chicago, IL, USA (v19), as previously described. Statistically significant differences between study groups was assessed using parametric Student's t-test or non-parametric Mann-Whitney's t-test, as appropriate, for pairwise comparisons, or parametric ANOVA for multiple comparisons, setting a P-value of 0.05 as the cut-off for statistical significance. Experiments were performed a minimum of three times. Survival data from patient databases were analyzed by the Log-rank Mantel-Cox statistical test.
## Analyses of cancer patient databases.
Meta-analyses of the PAI1 expression levels in tumor and no tumor colorectal samples were performed using the PrognoScan public patient datasets, as previously described. Patient survival was analyzed using the R2 Genomics analysis and visualization platform, as previously described. Expression data were downloaded and plotted using GraphPad Prism 6.01. Databases with available survival data were used to generate the Kaplan-Meier plots showing patient survival with the scan method, which searches for the optimum survival cut-off based on statistical analyses (log-rank Mantel-Cox test).
# Results
## Pai1 is upregulated in rectal tumors and associated with reduced overall survival
To study the potential of PAI1 as a marker in rectal cancer, we first analyzed PAI1 expression levels in four public rectal cancer databases: GSE35452, GSE8671 and GSE2109 . Since GSE8671 was the only database containing values of non-tumor tissue and the three databases used the same platform and normalization methods, we used them as a control. We found that the PAI1 mRNA levels were significantly higher in tumor samples than in normal rectal tissue. We extended this analysis to colorectal cancer patients in GSE8671 and four additional databases with paired non-tumor and tumor samples (GSE21510, GSE4183, GSE201916 and GSE33114). For all of them, we observed a statistically significant increase of PAI1 expression in tumor samples; . Moreover, this result was corroborated using 135 samples from our previously published cohort of locally advanced rectal cancer patients (HUVR-IBIS), who had received preoperative chemoradiotherapy and showed a 5-year overall survival (OS) of 75%; , similar to other published datasets. In this cohort, we also analyzed the levels of PAI1 protein by immunohistochemistry, defining a high PAI1 protein level in the tumor samples when the score of the stained tumor was higher than 1 on a scale of 0-3 (see Material and Methods) compared with non-tumor samples, and low PAI1 protein level when the score was lower than or equal to 1. Under these conditions, 35 (29.6%) tumor samples showed low expression, and 100 (70.4%) tumor samples showed high expression of PAI1. DFS, and that PAI1 could determine the response to treatment in rectal cancer, at least for the HUVR cohort. show PAI1 expression levels in rectal tumor tissue (orange) or non-tumor tissue (white) patients. In (A), data were compared using the ANOVA test, except for the GSE35452, whose data were not normally distributed and thus a non-parametric Mann-Whitney's t-test was used. In (B,C), tumoral and non-tumoral data were compared using Student's t-tests. *p < 0.05; **p < 0.01; ***p < 0.001. (D) In (A), data were compared using the ANOVA test, except for the GSE35452, whose data were not normally distributed and thus a non-parametric Mann-Whitney's t-test was used. In (B,C), tumoral and non-tumoral data were compared using Student's t-tests. *p < 0.05; **p < 0.01; ***p < 0.001. (D) Overall survival analysis of the HUVR-IBIS rectal tumor cohort Then, we tested the relevance of PAI1 expression levels in the survival of patients. We found that patients in our cohort with higher PAI1 expression showed a decrease in disease-free survival (DFS) compared to that in patients with lower PAI1 expression, although it did not reach statistical significance; p = 0.076). This was not observed when we analyzed the overall survival data in rectal adenocarcinoma samples from the TCGA database. However, in patients from this database we observed that individuals with higher levels of PAI1 were at a higher risk of colorectal cancer than those with lower levels of PAI1 (p < 0.001), a result that was also observed in the HUVR-IBIS database (p < 0.01). Accordingly, the analysis of the patient response in the HUVR-IBIS cohort showed that non-responder patients had higher levels of PAI1 (p < 0.05), while a non-statistically significant trend was observed in database GSE35452 (p = 0.3). Altogether, these results suggest that PAI1 is upregulated in rectal tumors and may be associated with reduced DFS, and that PAI1 could determine the response to treatment in rectal cancer, at least for the HUVR cohort.
## Pai1 expression correlates with emt-associated genes in rectal tumors
According to previous results in other types of cancer, including breast, head and neck carcinoma and esophageal carcinoma, PAI1 is involved in invasion and metastasis, angiogenesis and proliferation, and it confers a bad prognosis in cancer. To study its role in rectal cancer, we first searched for genes whose expression was correlated with PAI1 expression in rectal tumor samples from the TCGA database. We found that the expression of 2024 genes correlated with that of PAI1 (R < −0.35 or R > 0.35). Gene Ontology term enrichment analyses of these genes showed biological processes associated with cell adhesion, angiogenesis, inflammatory/immune response, cell migration and differentiationand Supplementary Dataset), which was in agreement with the well-known function of PAI1 as a regulator of these processes.
The epithelial-to-mesenchymal transition (EMT) is a process by which epithelial cells lose their polarity and cell-cell adhesion, and increase migratory and invasive properties to become mesenchymal stem cells. Therefore, we wondered whether PAI1 expression could correlate with the expression levels of EMT-associated genes in rectal tumors. To explore this possibility, we visualized in a heatmap the expression levels of EMT-associated genes that significantly correlated with PAI1 expression in two representative patient databases with available microarray expression data: GSE35452 and TCGA(Note that microarray data are not available for the HUVR-IBIS cohort). We observed a clear correlation between PAI1 expression and EMT-associated genes in both rectal tumor databases, suggesting a relation between PAI1 expression and the EMT.
Accordingly, the expression levels of VIM1, TWIST1, FOXC2 and SNAI1, the most representative EMT-associated genes, were significantly correlated with PAI1 expression. Therefore, these data suggest that the upregulation of PAI1 in rectal tumors is correlated with the upregulation of EMT-associated genes, which could explain the higher metastasis and tumorigenesis of tumors expressing high levels of PAI1.
## Upregulation of pai1 is associated with metastasis and invasion in rectal tumors
Next, we wondered whether PAI1 expression levels were associated with metastasis in rectal tumors. For this, we first analyzed PAI1 expression levels in rectal adenocarcinoma patients from our cohort with and without metastasis, finding that it was significantly higher in patients with metastasis. This tendency was not statistically significant for the TCGA database. However, rectal tumor samples from TCGA showing perineural, lymphovascular or vascular invasion showed significantly higher levels of PAI1 expression, suggesting that PAI1 may play a role in metastasis and invasion in rectal cancer.
Invasion and metastasis-dependence on PAI1 may be responsible for the bad prognosis of the rectal tumors. Therefore, it will be interesting to suggest possible therapeutic targets to eliminate these malignant phenotypes.
Cells 2020, 9, x FOR PEER REVIEW 8 of 18
## Upregulation of pai1 is associated with metastasis and invasion in rectal tumors
Next, we wondered whether PAI1 expression levels were associated with metastasis in rectal tumors. For this, we first analyzed PAI1 expression levels in rectal adenocarcinoma patients from our cohort with and without metastasis, finding that it was significantly higher in patients with metastasis. This tendency was not statistically significant for the TCGA database. However, rectal tumor samples from TCGA showing perineural, lymphovascular or vascular invasion showed significantly higher levels of PAI1 expression, suggesting that PAI1 may play a role in metastasis and invasion in rectal cancer.
Invasion and metastasis-dependence on PAI1 may be responsible for the bad prognosis of the rectal tumors. Therefore, it will be interesting to suggest possible therapeutic targets to eliminate these malignant phenotypes.
## Pai1 expression correlates with drug target genes in rectal tumors
We have shown so far that patients of the HUVR-IBIS cohort with tumors showing upregulated PAI1 expression have a worse prognosis and more resistance to treatment in rectal cancer, I-J). An important part of chemotherapy is based on the use of drugs or molecules that inhibit or activate specific targets in tumor cells. Most drug targets are members of phylogenetically conserved protein families and includes G protein-coupled receptors, protein kinases, nuclear hormone receptors, serine proteases and ion channels. Therefore, we analyzed drug target genes whose expression correlated with PAI1 expression in rectal adenocarcinoma patients from the TCGA database, since microarray data were not available for the HUVR-IBIS cohort. Among the found drug target genes, we observed highly significant positive correlations for some of the most representative drug target genes used in the treatment of rectal cancer.
## Pai1 expression correlates with drug target genes in rectal tumors
We have shown so far that patients of the HUVR-IBIS cohort with tumors showing upregulated PAI1 expression have a worse prognosis and more resistance to treatment in rectal cancer. An important part of chemotherapy is based on the use of drugs or molecules that inhibit or activate specific targets in tumor cells. Most drug targets are members of phylogenetically conserved protein families and includes G protein-coupled receptors, protein kinases, nuclear hormone receptors, serine proteases and ion channels. Therefore, we analyzed drug target genes whose expression correlated with PAI1 expression in rectal adenocarcinoma patients from the TCGA database, since microarray data were not available for the HUVR-IBIS cohort. Among the found drug target genes, Cells 2020, 9, 1071 9 of 17 we observed highly significant positive correlations for some of the most representative drug target genes used in the treatment of rectal cancer.
These included platelet-derived growth factor receptors (PDGFRs): PDGFRB (R = 0.71, p < 0.0001;and PDGFRA (R = 0.43, p < 0.0001;. PDGFRs are cell surface type III receptor tyrosine kinases that have been shown to increase proliferation and migration in several tumors. PDGFR-α and PDGFR-β are expressed in CRC tissues and these factors were revealed to stimulate invasion and liver-metastasis formation in mice, having been related with recurrence in this type of cancer. FGFR1 (R = 0.61, p < 0.0001;is a member of the Fibroblast Growth Factor family whose alterations have been recently identified as likely mechanisms of primary and secondary resistance to therapy using anti-EGFR antibody in CRC.
These included platelet-derived growth factor receptors (PDGFRs): PDGFRB (R = 0.71, p < 0.0001;and PDGFRA (R = 0.43, p < 0.0001;. PDGFRs are cell surface type III receptor tyrosine kinases that have been shown to increase proliferation and migration in several tumors. PDGFR-α and PDGFR-β are expressed in CRC tissues and these factors were revealed to stimulate invasion and liver-metastasis formation in mice, having been related with recurrence in this type of cancer. FGFR1 (R = 0.61, p < 0.0001;is a member of the Fibroblast Growth Factor family whose alterations have been recently identified as likely mechanisms of primary and secondary resistance to therapy using anti-EGFR antibody in CRC.
The proto-oncogene serine/threonine-protein kinase PIM1 (R = 0.34, p < 0.0001;is overexpressed in many types of cancers, including CRC, leading to tumor development and progression. The FYN gene (R = 0.49, p < 0.0001;encodes a membrane-associated tyrosine kinase that has been involved in the control of cell growth and in the regulation of EMT and metastasis in CRC. Finally, the BRAF gene (R = 0.38, p < 0.0001;encodes a serine/threonine kinase and BRAF mutations have been associated with poor prognosis and less response to treatment in metastatic CRC. Altogether, these results suggest that the worse prognosis and increased resistance to treatment in rectal cancer patients with upregulated PAI1 may be linked to the overexpression of several drug targetable genes. The proto-oncogene serine/threonine-protein kinase PIM1 (R = 0.34, p < 0.0001;is overexpressed in many types of cancers, including CRC, leading to tumor development and progression. The FYN gene (R = 0.49, p < 0.0001;encodes a membrane-associated tyrosine kinase that has been involved in the control of cell growth and in the regulation of EMT and metastasis in CRC. Finally, the BRAF gene (R = 0.38, p < 0.0001;encodes a serine/threonine kinase and BRAF mutations have been associated with poor prognosis and less response to treatment in metastatic CRC. Altogether, these results suggest that the worse prognosis and increased resistance to treatment in rectal cancer patients with upregulated PAI1 may be linked to the overexpression of several drug targetable genes.
## Resistance to azd1208 is associated with pai1 expression levels in rectal tumors.
To explore this point, we searched for inhibitors specific for the main lines of targetable proteins correlating with PAI1 expression. We selected Dasatinib, vemurafenib and AZD1208, inhibitors of broad-spectrum tyrosine kinases, BRAF and Pim Ser/Thr kinases, respectively. Dasatinib is a multitargeted kinase inhibitor proven to be effective for the treatment of several malignancies, including several types of cancers such as CRC. Dasatinib inhibits a broad spectrum of kinases, including Kit, PDGFR, FGFR1 and many others, although it is relatively specific for the Src and Abl family kinases such as Fyn, Yes, Src, and Lyk. As shown by several preclinical studies, Src inhibitors like dasatinib are able to overcome chemoresistance and resistance to targeted agents, such as the EGFR monoclonal antibody cetuximab. Vemurafenib is the first BRAF serine-threonine kinase protein inhibitor authorized for the treatment of adult patients with unresectable or metastatic melanoma with a BRAF V600E positive mutation. Mutations in the BRAF gene substitute the amino acid valine for glutamic acid at position 600 and lead to the oncogenic activation of BRAF proteins. AZD1208 is a potent and selective inhibitor that affects all three isoforms of serine-threonine Pim kinases.
We wondered whether PAI1 expression levels could determine the resistance to dasatinib, vemurafenib or AZD1208. To test this, we first analyzed the expression levels of PAI1 in a panel of eight CRC cell lines by Western blot, both intra and extracellular, showing a high variability among them. Next, these cells were treated with different doses of dasatinib, vemurafenib and AZD1208 to calculate the IC 50 using the MTT assay as a readout of their sensitivity. We detected similar sensitivity/resistance behavior to the B-Raf inhibitor vemurafenib in all cell lines independently of PAI levels; . For dasatinib, we found more heterogeneous results, although sensitivity was not correlated with PAI1 protein levels either; . Interestingly, the analysis with the Pim inhibitor AZD1208 showed that, although the IC 50 was negatively correlated with PAI1 protein levels, most of the cell lines expressing PAI1 were similarly affected by AZD1208 treatment, in contrast to cell lines expressing undetectable levels; , indicating that PAI1 expression determines sensitivity to AZD1208-mediated PIM inhibition. The mutation status of KRAS, BRAF and PI3KCA did not seem to be related to the degree of sensitivity of these cell lines to AZD1208. Therefore, our data may suggest that PIM inhibition may be a suitable treatment for the patients with a worse prognosis due to high levels of PAI1 in rectal tumors.
# Discussion
In Europe, 200,000 new cases of rectal cancer were diagnosed in 2018. Rectal cancer is a highly lethal cancer, with only 20% of patients showing a complete response after surgery and even responding patients inevitably developing refractory disease. This is due to its late detection and to the development of chemotherapy resistance to common treatments. Thus, it is essential to find new biomarkers to stratify patients with advanced or metastatic rectal cancer and predict their response in order to overcome chemotherapy resistance. Here, we found that PAI1 gene expression is upregulated in rectal cancer and that this upregulation may reduce the DFS and the response to treatment of patients, at least those of our cohort (HUVR-IBIS), and could be used as a bad prognosis marker. Moreover, high expression levels of PAI1 are associated with increased metastasis and invasion of rectal tumors in this cohort and TCGA database, respectively, which is correlated with EMT-associated and drug target genes expression, including PDGFRa, PDGFRb, FYN, PIM1 and BRAF. Strikingly, cells expressing PAI1 protein are more sensitive to PIM inhibitor AZD1208, suggesting that PAI1 could be used to predict increased efficacy of this PIM inhibitor and may complement radiotherapy in rectal tumors.
# Discussion
In Europe, 200,000 new cases of rectal cancer were diagnosed in 2018. Rectal cancer is a highly lethal cancer, with only 20% of patients showing a complete response after surgery and even responding patients inevitably developing refractory disease. This is due to its late detection and to the development of chemotherapy resistance to common treatments. Thus, it is essential to find new biomarkers to stratify patients with advanced or metastatic rectal cancer and predict their response in order to overcome chemotherapy resistance. Here, we found that PAI1 gene expression is upregulated in rectal cancer and that this upregulation may reduce the DFS and the response to treatment of patients, at least those of our cohort (HUVR-IBIS), and could be used as a bad prognosis marker. Moreover, high expression levels of PAI1 are associated with increased metastasis and invasion of rectal tumors in this cohort and TCGA database, respectively, which is correlated with EMT-associated and drug target genes expression, including PDGFRa, PDGFRb, FYN, PIM1 and BRAF. Strikingly, cells expressing PAI1 protein are more sensitive to PIM inhibitor AZD1208, suggesting that PAI1 could be used to predict increased efficacy of this PIM inhibitor and may complement radiotherapy in rectal tumors. PAI1 belongs to the family of serine protease inhibitors, also known as serpins, and is the main regulator of the plasminogen activation system, acting by inhibition of tPA and uPA. PAI1 deregulation has been associated with cardiovascular diseases, obesity, metabolic syndrome and various types of cancer. We found that PAI1 gene expression is upregulated in several CRC public datasets and validated this with our own cohort of patients with advanced rectal cancer (HUVR-IBIS). Patients with higher PAI1 expression show a decrease in the survival probability in our patient database, as compared to patients with lower PAI1 expression. This suggests that PAI1 expression could be a potential independent biomarker for survival. Accordingly, we found that patients with higher PAI1 expression showed a tendency to respond worse to treatment in the HUVR-IBIS cohort. Moreover, PAI1 expression is also correlated with poor outcome in several other cancer subtypes, particularly in ovarian serous carcinoma and node-negative breast cancer. In fact, PAI1 has been validated clinically in breast cancer as a biomarker. Furthermore, PAI1 expression is significantly enhanced in some other cancers, including stomach adenocarcinoma, head and neck squamous cell carcinoma, esophageal carcinoma or thymoma. Additionally, an increase in the plasma levels of PAI1 has been shown in acute leukemia, breast cancer, hepatocarcinoma and colon cancer. However, bladder urothelial carcinoma and testicular germ cell tumors do not show differences in PAI1 expression between normal and tumoral tissue, indicating that PAI1 is a common but not universal biomarker of cancer progression.
We have observed that high expression levels of PAI1 are associated with increased metastasis and invasion in rectal tumors of the HUVR-IBIS and TCGA cohorts, respectively, and that, accordingly, it is highly significantly correlated with EMT-associated gene expression in the latter database and in the GSE35452. In agreement with our results, overexpression of PAI1 was significantly associated with metastasis and invasion in lung, head and neck and breast cancers, as well as osteosarcoma. In contrast, PAI1 has been shown to have an inhibitory effect over cell migration and invasion in other types of cancer, as it is the case in pancreatic cancer, glioma or melanoma. Therefore, it seems that the outcomes of PAI1 overexpression depends on the cellular context. In line with the pro-metastatic role of PAI1, it has been shown to facilitate invasion and lung metastasis in osteosarcoma cells by promoting MMP13 expression and secretion. Moreover, it has been described that PAI1 and CCL5 signaling in endothelial cells leads to increased metastasis in EMT-induced triple-negative breast cancer cells. Finally, clinical data showed a significant increase in PAI1 levels in plasma of CRC patients with liver metastasis and infiltration.
Our analyses using rectal cancer patient public databases showed that non-responder rectal cancer patients had higher expression levels of PAI1 in the HUVR-IBIS cohort; GSE35452;. Moreover, we showed that PAI1 expression positively correlates with the expression of several drug targets genes, including PDGFRa, PDGFRb, FYN, PIM1 and BRAF in the TCGA database. A high number of preclinical studies have demonstrated the ability of Src inhibitors like dasatinib to inhibit most of these drug target genes, as PDGFRa, PDGFRb or FYN, and to overcome chemoresistance in CRC. However, clinical trials with dasatinib combined either with conventional chemotherapy or with suppression of EGFR by cetuximab and with FOLFOX, failed to show any meaningful clinical response. A recent study showed that this failure in clinical trials is possibly due to the fact that Desatinib reduces apoptosis triggered by 5-FU in colon carcinoma. However, data from our own laboratory suggest that dasatinib may be active in combination with oxaliplatin (but not with 5-FU) only in patients whose tumors show high p-Src levels.
Our data suggest that treatment with PIM inhibitors may be beneficial in combination with radiotherapy in patients with tumors showing PAI1 expression. Therefore, PAI1 can be used to select patients for whom this combination therapy could be beneficial. Overexpression of PIM kinases is common in many types of tumors, including CRCs. Indeed, both in vitro and in vivo studies have shown that individual PIM kinases are weak oncogenes that can become stronger ones in cooperation with c-MYC in some tumors. These precedents suggest that PIM kinases may be good targets for drug development. The pan-PIM kinase inhibitor AZD1208 acts over the serine/threonine kinases PIM1, 2 and 3, which may result in cell cycle arrest and apoptosis when those kinases are overexpressed. We used it as a tool compound to prove that one of the drug target paths correlating with PAI1 expression may be suitable to compensate tumor survival after radiotherapy and poor prognosis. We demonstrated that PAI1 expression leads to sensitivity to AZD1208 PIM inhibitor, suggesting that PAI1 expression could be used as a prognosis marker to stratify patients with bad prognosis of the rectal tumors and predict increased efficacy of this PIM inhibitor.
# Conclusions
Our study shows that PAI1 expression is upregulated and correlates with decreasing survival and increasing metastasis and invasion. Moreover, we propose PAI1 as a new predictive marker to stratify patients according to their response to standard treatment in rectal cancer. We demonstrate using a panel of CRC cell lines the sensitivity to AZD1208 Pim inhibitors of cells with PAI1 expression, suggesting that PAI1 could be used as a prognosis marker to stratify rectal cancer patients and predict increased efficacy of this Pim kinase inhibitor in those with a bad prognosis.
Supplementary Materials: The following are available online at http://www.mdpi.com/2073-4409/9/5/1071/s1. Table S1. Patient cohort characteristics; . Mutational analysis of the CRC cell lines used in this study; . Characteristics of patient public databases used in this study. . Determination of the IC50 value to AZD1208, Dasatinib and Vemurafenib; Supplementary Dataset. Complete Gene Ontology enrichment analysis of genes whose expression is correlated with PAI1. |
ICU Pad Project: application of modern computer technology in pediatric postoperative cardiac intensive care. Pilot study
## Quality in medicine
After cardiac surgery procedures, all vital functions of the patient must be monitored carefully with particular focus on hemodynamic parameters. Optimizing the function of the circulatory system is the primary task of an intensive care physician as low cardiac output syndrome is one of the fundamental challenges of providing medical care to both children and adults after cardiac procedures. After a cardiac procedure, the cardiac output and, thereby, patient condition and prognosis, can be influenced by many factors, including arterial blood pressure, circulating fluid volume, arrhythmia as well as myocardial dysfunction resulting from open heart surgery using extracorporeal circulation [bib_ref] The hemodynamic status of cardiac surgical patients in the initial 2-hrecovery period, Currey [/bib_ref] [bib_ref] Common complications after cardiacsurgery in adults, Foot [/bib_ref]. The evolution of hemodynamic parameters shortly after the procedure is usually difficult to predict. The situation may be additionally complicated by the presence of concomitant diseases or perioperative complications, specific properties of the procedure (palliative vs. total repair) and the heart defect, and the age of the patient (e.g., compensatory mechanisms are exhausted quicker in neonates and infants). ICU Pad Project: application of modern computer technology in pediatric postoperative cardiac intensive care. Pilot study errors associated with the necessity of alternately gaining access to different systems. Additionally, the variety of information sources precludes uniform data presentation and introduction of mechanisms for detecting variables that fall outside of the norms set for a given patient [bib_ref] Model-based data integration in clinical environments, Heldt [/bib_ref]. Using a clinical system to integrate access to various data sources describing the patient's condition and enable clear analysis (on a current and retrospective basis) of all registered variables would surely facilitate the work of intensivists treating hemodynamically unstable patients after cardiac procedures. Such a system should be simple and intuitive to use, mobile, and easy to disinfect. An additional advantage comes with the ability to couple a visualization system with an expert system or a clinical decision support system (CDSS), which are able to analyze the incoming stream of medical data on a current basis, pinpointing pathological trends and associations between all variables from the whole period of hospitalization, thus aiding the intensive care physician [bib_ref] A hybrid hierarchical decision support system for cardiac surgical intensive care patients...., Denaï [/bib_ref] [bib_ref] A hybrid hierarchical decision support system for cardiac surgical intensive care patients...., Ross [/bib_ref].
## Icu pad project
This paper describes a prototype of a system for integrating medical data, created on the basis of the foregoing rationale as part of the pilot program ICU Pad Project (IPP). The main component enabling communication between the user and the system is a mobile touch screen tablet. The system owes its development to intercollegiate cooperation with the Department of Information Technology of the Poznań University of Technology.
The ICU Pad Project was approved by the Bioethical Committee of the Poznań University of Medical Sciences (approval no. 837/12, date: September 13, 2012). Testing the system involved gathering sensitive information; for this reason, the system was isolated from the "external world" and the hospital medical network. Each time when the gathered data had to be exported outside of the hospital, sensitive data were anonymized. Work on the project involved the participation of physicians, nurses, as well as individuals who did not take part in the therapeutic process. Non-medical employees working on the project were required to sign confidentiality agreements and were pro-hibited from disclosing the medical data they accessed. Encrypted wireless transmission between authorized devices only was used during system testing.
The system consists of a server based on the Windows Vista operating system with the following software installed: Postgress SQL database (v. 9.2, http://www.postgresql. org) for data storage, Apache Tomcat http server (v. 7.0.37, http://httpd.apache.org) for handling data transmission to the mobile component, and a "stationary" software component enabling the management of the system (admitting and discharging the patients, adding users, printing observation cards). The element serving the function of joining the server and mobile components was a WiFi router ensuring encrypted transmission and access control on the level of trusted MAC addresses (Media Access Control addresses).
The mobile component consisted in tablet devices (ASUS EeePAD Transformer, 32GB RAM, Android 4.1, https:// www.android.com).
In order to test the system, it was used to gather the data of 28 patients in stable clinical condition; during the same time, conventional medical documentation, standard for postoperative wards, was maintained.
## System operation
After turning on the tablet, the user is presented with the login screen [fig_ref] Figure 2: Login screen [/fig_ref].
After the user's credentials are verified, the system displays the list of patients on the postoperative ward, their birthdates, gender, and duration of hospitalization [fig_ref] Figure 3: The list of admitted patients [/fig_ref].
Selecting one of the patient fields opens a window with the patient's card.
The heading of the patient card includes their first and last name, gender, PESEL number, and birthdate, as well as additional information: body mass, blood type, date of admission, and day of hospitalization. This part of the patient card remains unchanged: it is always visible on the screen after opening the patient card, regardless of which parameters are currently analyzed. The information gathered on the patient card is grouped under tabs -the tabs available at the present stage of the system's development include: Hemodynamics, Fluid balance, Arterial Blood Gas, Complete Blood Count, Biochemistry, and Coagulation. Touching any of them allows the user to switch between the desired data sets.
The Hemodynamics tab presents data in the form of a chart. Each parameter is marked with a different color, which facilitates reading and interpreting the data. The chart may be moved in any direction, changing the time of origin of the presented data. Additionally, the chart's scale can be adjusted using touch gestures. The history of a selected parameter can also be displayed in a digital form.
The Arterial Blood Gas, Biochemistry, Complete Blood Count, and Coagulation tabs present observations in the form of movable tables (Figs. 5 and 6).
The first row of the table contains the hour at which the measurement was made. Depending on the number of parameters in the rows, horizontal scrolling is enabled. Days on which the measurements were made are marked with a distinctive gray bar containing the date (day and month). The results table includes an automated alert mechanism, notifying the user when the norms for individual parameters are exceeded. Yellow is used to mark values below the norm, while red marks values above the norm. The norms are in accordance with the norms used in the hospital laboratory, defined for the current age and gender of the patient. This solution is meant to highlight parameter values of special note for the physician during therapy and to prevent potential mistakes when reading the results of laboratory examinations.
The FLUID BALANCE tab enables precise monitoring of the volume of fluids administered to and expelled by the patient during each day of hospitalization at the postoperative ward. When the data are entered into the system, the administered volume is divided into categories depending on the character of the fluid (colloid or crystalloid) and the method of administration (boluses, Continous infusions, nasogastric tube). A similar division has been employed in order to determine the amount of fluids lost by a patient during each day (drainage volume divided into mediastinal drainage, left and right pleural drainage, volume of diuresis or hemodiafiltration, and volume of fluid contents evacuated through a gastric tube). The tab works in a manner similar to that of the BIOCHEMISTRY, COAGULATION, and Complete Blood Count tabs -in the form of a table. Also, observations may be conveniently added and viewed in groups (Colloids, Drainage, Tube-supply, Tube-retention, boluses, continous infusions). The momentary value of the balance for the current day of hospitalization at the postoperative ward is calculated on a current basis for each group.
The individual groups in the Fluid Balance tab can be closed [fig_ref] Figure 7: Fluid balance tab -general view [/fig_ref] and opened [fig_ref] Figure 8: Fluid balance tab -detailed view [/fig_ref] by changing the selection of the 'Details' field available next to each group. . Hemodynamics tab. This tab presents observations in the form of a chart. Each color is assigned to a specific parameter (e.g., white -heart rate, green -saturation, red -arteria blood pressure). On the right side of the screen, there are fields with names of parameters; selecting one of them enables the user to see the history of the parameter in a digital form. Above the parameter fields there is the 'Add observations' field; selecting it opens a window which enables adding observations to this tab In order to add an observation, the user must activate the tab to which they wish to add new information and push the 'Add observations' button. At this point, a window for adding observations will be displayed . New data should be entered into the fields corresponding to the measured parameters; the data are saved using the 'Add observations' button. When entering the data, some of the fields can be left blank; this has no bearing on the saving of the remaining data. When the observations are successfully added, an appropriate message will be displayed, the adding window will be closed, and the tab will be refreshed. The default registered hour and date of an added observation is the time of entry, but this can be changed for each observation. The frequency of entering the data is at the user's discretion.
The IPP system enables the user to select parameters of interest only; viewing all available data is not necessary. The window presenting all parameters available in the system is divided in accordance with the names of the tabs relevant for the observation types . Only the selected parameters are available in the window for adding observations. This solution enables the physician to focus on the parameters that are especially important to be monitored for a given patient.
After logging into the system, the user may end the session at any time. After ending the session, the current user will be logged out, and the login screen will be displayed.
The system has undergone its first tests conducted in clinical conditions. The testing focused primarily on processing the hemodynamic variables and the results of selected laboratory examinations. The test was conducted by . Selecting observed parameters. The system enables selection of observed parameters. The 'Parameters monitored on admission' window contains a list of the parameters and shows to which tab they belong. The user can select the parameters that will be added and monitored for a given patient. The observed parameters can be changed at any time Colloids, Drainage, Diuresis, Tube-supply, Tube-retention, boluses, continous infusions). Next to each category there is a 'Details' button enabling presentation of detailed information for each category as well as the balance for the current day for each category Additional information includes the hour of the measurement and the amount administered to or expelled by the patient. Clicking for more details allows the user to display not only volume changes noted for the current day, but also the values noted on previous days, separated by columns containing the dates on which the information was entered . Adding observations. Each tab features an 'Add observations' option. After clicking the button, the system displays a list of parameters for which values can be added. The top of the window features fields in which the day and hour of the observation is automatically entered. After the added observations are confirmed, the results are saved by the system QUALITY IN MEDICINE 6 nurses during fourteen 12-hour shifts. The patients included in the test had been operated on due to the following types of heart defects: ASD II (atrial septal defect type 2), AS (aortic stenosis), DORV (double outlet right ventricle), VSD (ventricular septal defect), ToF (tetralogy of Fallot), PA (pulmonary atresia), BWG (Bland-White-Garland syndrome), AVSD (atrioventricular septal defect), CoA (coarctation of the aorta), d-TGA (transposition of the great arteries), ASD sinus venosus (atrial septal defect sinus venosus), ASD I (atrial septal defect type 1). The following parameters were registered: body temperature, heart rate, oxygen saturation, arterial blood pressure, pulmonary artery pressure, cardiac output, central venous pressure , arterial blood gas (pH, blood saturation, partial pressure of carbon dioxide, partial pressure of oxygen, base excess, concentrations of potassium, sodium, and calcium ions, lactate concentration, and glucose concentration) [fig_ref] Figure 5: Arterial Blood Gas tab [/fig_ref] , coagulation parameters (fibrinogen concentration, prothrombin ratio, INR, prothrombin time, partial thromboplastin time, D-dimer concentration, antithrombin III concentration), complete blood count (white blood cell count, red blood cell count, and blood platelet count in blood volume, as well as hematocrit and hemoglobin concentration in blood) [fig_ref] Figure 6: Complete Blood Count tab [/fig_ref]. Information concerning the fluid balance was also recorded (measurement of the volume drained from the right and left pleura, diuresis value and/or the volume expelled through hemodiafiltration, the volume of contents evacuated from the gastric tube, the volume of administered colloid solutions, the volume of fluids administered through the gastric tube, the volume of fluids provided to the patients through the administration of agents in boluses and infusions) [fig_ref] Figure 7: Fluid balance tab -general view [/fig_ref].
Before receiving the tablets, the nurses received approx. 30 minutes of hands-on training and were able to practice all possible scenarios of entering and accessing data. The IPP tablet system was well received by the medical personnel. The ease and intuitiveness of the touch interface enabled its uncomplicated use as a tool for gathering and visualizing hemodynamic data and results of biochemical analyses in selected patients. At the same time, suggestions for improvements, as well as information about glitches, were recorded.
# Discussion
The prototype of the IPP system presented above appears to be one of the first attempts to employ mobile touchsensitive devices in the postoperative care provided to children after congenital heart defect procedures. The impulse behind the creation of the system was the need to improve the conventional system of documenting the postoperative course which has been on paper observation cards. There are many unquestionable drawbacks of the traditional system: the data can be incomplete; the paper cards can be easily lost; their size is cumbersome; mistakes can be made when the results are copied (there are no mechanisms of error control); the cards are unhygienic and can be a habitat for hospital microorganisms; it is difficult to process the gathered data; using the cards consumes paper, and archiving is cumbersome. The electronic IPP system eliminates all of the abovementioned problems: - the data can be gathered and integrated using one medium, - the relationships between many variables can be visualized in a straightforward manner, - the data can be accessed quickly, - the system automatically visualizes the parameters whose norms are exceeded, - the data can be analyzed on a current basis using the tablet by one or multiple individuals (video access, potentially with an option of teleconsultation), - entering data is expedited, and the system checks for errors when the data are entered, - the data are stored on the server in real time (data security),
- the system saves space (no paper documentation), - the tablets are more hygienic.
Other advantages include the fact that the system can be integrated with existing information systems and treatment quality control mechanisms can be introduced.
The Intensive Care Unit of the Nicosia General Hospital has employed software similar to IPP (but without the use of tablets) called "Intensive Care Window" (ICW) [bib_ref] Intensive care window: real-time monitoring and analysis in the intensive care environment, Stylianides [/bib_ref]. The system consists in an ICW application installed on standard computers and a bedside controller collecting data from the cardiac monitor, respirator, and gasometric measurement equipment. The ICW application enables real-time presentation of the gathered data in a text and graphic form. Its users can highlight and mark variable changes of interest and monitor the clinical parameters using alerts that notify the user when any unfavorable changes occur. The software also enables offline data analysis. This function may be used during normal system operation and does not interfere with the process of gathering and storing data. Each piece of information gathered from a patient is identified (unambiguously assigned to a given patient) and immediately entered into the database. Helpfully, the application includes an efficient search system allowing the user to search for specific information in the gathered data [bib_ref] Intensive care window: real-time monitoring and analysis in the intensive care environment, Stylianides [/bib_ref]. Similarly, the presented IPP systems enables the graphic and numeric presentation of results, their retrospective analysis, and quick access to each moment of postoperative observation. Multiple individuals can access the data remotely, potentially from any hospital location with network access.
One of the primary goals of introducing the IPP system is to integrate and visualize medical data originating from different sources. This concerns mostly parameters registered by cardiac monitors and respirators, results of laboratory investigations, as well as information obtained during examinations performed by nurses and physicians. Integrating and visualizing the aforementioned data using one mobile medium undoubtedly allows intensive care physicians to better perceive the stream of information they are faced with when providing care to hemodynamically unstable patients. Apart from this aspect, integrating medi-ICU Pad Project: application of modern computer technology in pediatric postoperative cardiac intensive care. Pilot study cal data has other advantages as the information gathered and made available by the system can include information obtained on admission, results of preoperative diagnostic examinations, as well as descriptions of the procedure, the postoperative course, and the patient's condition on discharge. The ability to share information concerning the postoperative course in an orderly fashion with medical personnel from other wards when the patient is discharged from the postoperative ward for further treatment appears to be especially advantageous [bib_ref] Data integration in cardiac surgery health care institution: Experience at G. Pasquinucci..., Taddei [/bib_ref] [bib_ref] Multimedia abstract generation of intensive care data: the automation of clinical processes..., Jordan [/bib_ref].
Apart from the practical benefits, employing the IPP system offers the following scientific advantages: - in terms of information technology: development of new user-friendly interfaces for entering data along with mechanisms for teleconsultation as well as visualization and analysis of medical data generated by the touch screen tablet, - in terms of medical sciences: development of clinical decision support systems [6, 10], - intelligent system alerts (elimination of false alarms) [bib_ref] Reducing false alarm rates for critical arrhythmias using the arterial blood pressure..., Aboukhalil [/bib_ref] , - the ability to analyze the postoperative course of selected heart defects with consideration to previously unstudied parameters or to identify new risk factors for adverse events, - an opportunity to optimize postoperative treatment of selected heart defects based on the analysis of the data gathered by the system [bib_ref] Prediction of mortality in intensive care unit cardiac surgery patients, Hekmat [/bib_ref]. Systems that integrate medical data, such as the presented IPP or the aforementioned ICW systems, are able to generate large quantities of data, especially if the registered parameters include direct recordings of hemodynamic curves and ECG evolution, registered automatically, or the results of imaging examinations. As the cost of storing such information is no longer a limitation, considering the current technological capabilities, creating high definition medical databases describing the condition of patients undergoing intensive medical care is becoming a reality. This is one of important aspects of implementing and using IPP, ICW, and other such systems [bib_ref] Multiparameter Intelligent Monitoring in Intensive Care II: a public-access intensive care unit..., Saeed [/bib_ref] [bib_ref] Intensive care window: real-time monitoring and analysis in the intensive care environment, Stylianides [/bib_ref].
A flagship example of a high definition medical database is the MIMIC-II project (Multiparameter Intelligent Monitoring in Intensive Care II) [bib_ref] Multiparameter Intelligent Monitoring in Intensive Care II: a public-access intensive care unit..., Saeed [/bib_ref] [bib_ref] Open-access MIMIC-II database for intensive care research, Lee [/bib_ref] [bib_ref] MIMIC II: a massive temporal ICU patient database to support research in..., Saeed [/bib_ref]. Apart from its practical advantages, MIMIC-II can be employed in various scientific endeavors, owing to the fact that all the information in its database can be shared after the removal of sensitive data (anonymization) [bib_ref] Computer-assisted de-identification of the free text in the MIMIC II database, Douglass [/bib_ref]. The literature features reports in which data from the MIMIC-II project were analyzed offline. The published studies regarded, e.g., determining ARDS risk in patients undergoing mechanical ventilation longer than 48 hours [bib_ref] Risk factors for ARDS in patients receiving mechanical ventilation for > 48..., Jia [/bib_ref] , hypertension as a risk factor for acute kidney injury [bib_ref] Hypertension as a risk factor for acute kidney injury in ICU patients, Lehman [/bib_ref] , the estimation of cardiac output values based on the arterial blood pressure curve [bib_ref] Estimating cardiac output from arterial blood pressure waveforms: a critical evaluation using..., Sun [/bib_ref] , the identification of mortality risk factors, the possibility of discontinuing pressor agents within the first 12 hours after cardiac surgery procedures, and the assessment of factors affecting the development of septic shock. These examples demonstrate the enormous scientific potential, enabling the generation of new knowledge, which stems from the ability to analyze data whose storage and assessment would not be possible without systems designed for the integration of medical information.
The experiences accumulated during the design and creation of the IPP system as well as literature reports point to the significant number of challenges that must be overcome in order to create an efficient high definition medical database. One of the main challenges is the fact that the data, originating from various sources, must be properly attributed to a given patient, e.g., by using a marking system based on personal data [bib_ref] Multiparameter Intelligent Monitoring in Intensive Care II: a public-access intensive care unit..., Saeed [/bib_ref]. At the current stage of IPP development, the automation of this process has been unnecessary as these tasks are performed by tablet operators.
Apart from offline data mining, the information gathered by a medical data integration system can also be used to create expert systems and clinical decision support systems [bib_ref] A hybrid hierarchical decision support system for cardiac surgical intensive care patients...., Denaï [/bib_ref] [bib_ref] A hybrid hierarchical decision support system for cardiac surgical intensive care patients...., Ross [/bib_ref]. Such a system could recognize certain problems typical for intensive treatment after cardiac surgery procedures (e.g., hypotension, hypovolemia, vasodilatation, or SIRS) and introduce appropriate treatment. Studies have shown that in each clinical scenario, the tested clinical decision support system was able to select therapeutic procedures similar to those suggested by intensive care physicians. As a result, the system was able to control hemodynamic parameters and maintain their values within the norm [bib_ref] A hybrid hierarchical decision support system for cardiac surgical intensive care patients...., Denaï [/bib_ref] [bib_ref] A hybrid hierarchical decision support system for cardiac surgical intensive care patients...., Ross [/bib_ref].
# Limitations
The presented prototype of the IPP system constitutes, in its current form, evidence that the concept and premises assumed during its design are correct: a proof of concept. The fact that the nurses received the product well and are eager to use it and to contribute to its development stimulates its authors to further efforts. More work is required to tackle software glitches on an ongoing basis and eliminate the limitations related to the system's current isolation from the hospital network. As a pilot system, the IPP project was designed in isolation from the existing hospital network and the external world (intranet/ internet) in order to protect patient data and avoid potential conflicts with the current IT systems during these early stages of IPP development. This isolation precludes the automatic gathering of data (e.g., laboratory results) from the hospital system or autonomous data generating equipment (e.g., cardiac monitors, respirators). Although integration with the hospital IT system based on open communication protocols (HL-7) is planned, transferring data from the aforementioned autonomous equipment may pose a significant challenge due to the large number of restricted data exchange protocols [bib_ref] Intensive care window: real-time monitoring and analysis in the intensive care environment, Stylianides [/bib_ref]. The system would also be significantly improved if it was provided with capabilities to integrate information concerning drug interactions, suggested dosage dependent on age, gender, and body mass, as well as the patient's condition and concomitant diseases.
## Quality in medicine
## Perspectives
The development of the IPP system may improve the comfort and safety of work in intensive care units and contribute to the improvement of treatment outcomes in patients after congenital heart defect procedures. At the same time, IPP can be a unique tool for gathering data that will be useful in further clinical studies as well as a platform for the development of medical expert systems.
# Disclosure
Authors report no conflict of interest.
[fig] Figure 2: Login screen. On the left side of the screen is the user list, on the right -the 'User' and 'Password' fields. After selecting the user, entering the password, and clicking the 'Log in' button, the user's credentials are verified, and the list of admitted patients is displayed on the screen [/fig]
[fig] Figure 3: The list of admitted patients. This view consists of rectangular tiles; each one is assigned to one patient. Each 'tile' contains basic information about the patient: first and last name, birthday, PESEL number, gender, and length of stay QUALITY IN MEDICINE [/fig]
[fig] Figure 5: Arterial Blood Gas tab. This tab presents measurement results in the form of a table. The first row of the table contains the hour at which the measurement was made. The results from individual days of the patient's stay are divided by columns with dates. Values below the norm are marked in yellow, while values above the norm are marked in red. In the right upper corner of the tab there is the 'Add observation' field which allows the user to add measurement results to the table [/fig]
[fig] Figure 6: Complete Blood Count tab. The data are presented in the form of a table; this tab also features an alert system notifying the user when the norms of specific values are exceeded ICU Pad Project: application of modern computer technology in pediatric postoperative cardiac intensive care. Pilot study [/fig]
[fig] Figure 7: Fluid balance tab -general view. The figure presents the Fluid Balance tab; in this mode, details concerning specific groups are hidden. At the top of the tab there is the fluid balance value for a given day. Below, specific categories are displayed, presenting the volumes administered to or expelled by the patient (from the top: [/fig]
[fig] Figure 8: Fluid balance tab -detailed view. The figure presents the tab showing details for each parameter group. [/fig]
|
Mutational landscapes of tongue carcinoma reveal recurrent mutations in genes of therapeutic and prognostic relevance
Background: Carcinoma of the oral tongue (OTSCC) is the most common malignancy of the oral cavity, characterized by frequent recurrence and poor survival. The last three decades has witnessed a change in the OTSCC epidemiological profile, with increasing incidence in younger patients, females and never-smokers. Here, we sought to characterize the OTSCC genomic landscape and to determine factors that may delineate the genetic basis of this disease, inform prognosis and identify targets for therapeutic intervention. Methods: Seventy-eight cases were subjected to whole-exome (n = 18) and targeted deep sequencing (n = 60).Results: While the most common mutation was in TP53, the OTSCC genetic landscape differed from previously described cohorts of patients with head and neck tumors: OTSCCs demonstrated frequent mutations in DST and RNF213, while alterations in CDKN2A and NOTCH1 were significantly less frequent. Despite a lack of previously reported NOTCH1 mutations, integrated analysis showed enrichments of alterations affecting Notch signaling in OTSCC. Importantly, these Notch pathway alterations were prognostic on multivariate analyses. A high proportion of OTSCCs also presented with alterations in drug targetable and chromatin remodeling genes. Patients harboring mutations in actionable pathways were more likely to succumb from recurrent disease compared with those who did not, suggesting that the former should be considered for treatment with targeted compounds in future trials. Conclusions: Our study defines the Asian OTSCC mutational landscape, highlighting the key role of Notch signaling in oral tongue tumorigenesis. We also observed somatic mutations in multiple therapeutically relevant genes, which may represent candidate drug targets in this highly lethal tumor type.
# Background
Oral squamous cell carcinoma (OSCC) accounts for 270,000 new cancer cases and 145,000 deaths annually, the majority of which occur in developing countries [bib_ref] Cancer incidence and mortality worldwide: Sources, methods and major patterns in GLOBOCAN..., Ferlay [/bib_ref]. Of these, two-thirds are oral tongue squamous cell carcinoma (OTSCC) [bib_ref] The epidemiology of tongue cancer: a review of global incidence, Moore [/bib_ref] [bib_ref] Global epidemiology of oral and oropharyngeal cancer, Warnakulasuriya [/bib_ref]. Current treatment strategies involve a multimodality approach involving surgery, chemotherapy, and radiotherapy. Unfortunately, despite advances in detection and treatment, 5-year overall survival (OS) rates for OTSCC remain alarmingly low [bib_ref] Metastasis of squamous cell carcinoma of the oral tongue, Sano [/bib_ref]. Currently, for all cancers arising in the head and neck [head and neck squamous cell carcinoma (HNSCC)], cancers of the tongue remain among the worst in terms of prognosis [bib_ref] Metastasis of squamous cell carcinoma of the oral tongue, Sano [/bib_ref].
OTSCC traditionally was thought to be a cancer afflicting older males with longstanding exposure to tobacco and alcohol. In recent decades, however, there appears to have been a shift in epidemiology towards a higher proportion of never-smokers and younger patients, with an increase in incidence among women [bib_ref] Worldwide trends in incidence rates for oral cavity and oropharyngeal cancers, Chaturvedi [/bib_ref] [bib_ref] Increasing incidence of oral tongue squamous cell carcinoma in young white women,..., Patel [/bib_ref] [bib_ref] Head and neck squamous cell carcinoma in the young: a spectrum or..., Toner [/bib_ref] [bib_ref] Increasing incidence of oral cancer amongst young persons: what is the aetiology?, Mackenzie [/bib_ref] [bib_ref] Changing trends in smoking and alcohol consumption in patients with oral cancer..., Montero [/bib_ref]. Although the basis of this epidemiological trend remains elusive, what is definitely worrisome is that the cancers arising in this 'atypical population' tend to be more aggressive and resistant to conventional treatment [bib_ref] Increasing incidence of oral tongue squamous cell carcinoma in young white women,..., Patel [/bib_ref] [bib_ref] Increasing incidence of oral cancer amongst young persons: what is the aetiology?, Mackenzie [/bib_ref] [bib_ref] Changing trends in smoking and alcohol consumption in patients with oral cancer..., Montero [/bib_ref].
The expanding application of next-generation sequencing technology to cancer has revolutionized our ability to identify genetic alterations essential for carcinogenesis and to uncover therapeutic vulnerabilities. Coupled with the rapidly expanding array of therapeutic compounds targeting key somatic oncogenic 'driver' alterations, these technologies promise a new era for targeted therapy [bib_ref] Biomarker-driven early clinical trials in oncology: a paradigm shift in drug development, Tan [/bib_ref]. This approach has made significant headway in a number of cancers, such as breast and lung carcinoma, where alterations in the EGFR/ERBB pathways proved to be the main drivers that could be targeted with specific small molecules [bib_ref] Rational, biologically based treatment of EGFR-mutant non-small-cell lung cancer, Pao [/bib_ref] [bib_ref] Trastuzumab for patients with HER2 positive breast cancer: delivery, duration and combination..., Pinto [/bib_ref].
Unfortunately, it has been difficult to translate the same success across to HNSCC. Recent large-scale genome sequencing efforts have validated TP53 as the commonest mutation in HNSCC, but have failed to identify any single 'oncogenic driver' gene or family that is conventionally targetable, although some studies are examining the possibility of targeting tumor suppressors using synthetic lethal approaches [bib_ref] Exome sequencing of head and neck squamous cell carcinoma reveals inactivating mutations..., Agrawal [/bib_ref] [bib_ref] The mutational landscape of head and neck squamous cell carcinoma, Stransky [/bib_ref] [bib_ref] Genetic alterations in head and neck squamous cell carcinoma: The next-gen sequencing..., Nguyen [/bib_ref] [bib_ref] Applying synthetic lethality for the selective targeting of cancer, Mclornan [/bib_ref] [bib_ref] Integrative genomic characterization of oral squamous cell carcinoma identifies frequent somatic drivers, Pickering [/bib_ref]. Notably, the majority of these publications, including The Cancer Genome Atlas (TCGA), have focused on HNSCC as a single entity. However, epidemiological evidence suggests that biological differences and clinical peculiarities do exist between different head and neck sites, subsites and populations affected by this disease [bib_ref] Cancer incidence and mortality worldwide: Sources, methods and major patterns in GLOBOCAN..., Ferlay [/bib_ref] [bib_ref] Epidemiology of oral cancer in Asia in the past decade-an update, Rao [/bib_ref]. For example, certain molecular alterations appear to be present in higher frequencies or unique to certain sites/ subsites: these include PIK3CA mutations that are more prevalent in human papillomavirus (HPV)-induced oropharyngeal cancer, and alterations in USP9X, MLL4, ARID2, UNC13C, and TRPM3 in gingiva-buccal cancers. Even within the oral cavity alone, the distinct clinical behaviors of tumors arising from the oral tongue, buccal, alveolar, and hard palate suggests that these should be examined as different and separate entities [bib_ref] Changing trends in smoking and alcohol consumption in patients with oral cancer..., Montero [/bib_ref]. Specifically, in the oral cavity, tumors arising from the anterior tongue are known to be more infiltrative and metastatic and hence have the poorest outcome compared with tumors in other subsites [bib_ref] Metastasis of squamous cell carcinoma of the oral tongue, Sano [/bib_ref] [bib_ref] Frequency and therapeutic implications of "skip metastases" in the neck from squamous..., Byers [/bib_ref] [bib_ref] Prognostic significance of invasion depth in oral tongue squamous cell carcinoma, Tan [/bib_ref]. Furthermore, it is now emerging that the mutation spectrum of many malignancies can be strongly influenced by ethnic background, encompassing population genomic variations, environmental exposures, and lifestyle practices [bib_ref] Recurrent gain-of-function mutations of RHOA in diffuse-type gastric carcinoma, Kakiuchi [/bib_ref] [bib_ref] Whole-genome sequencing and comprehensive molecular profiling identify new driver mutations in gastric..., Wang [/bib_ref] [bib_ref] Exome sequencing identifies distinct mutational patterns in liver fluke-related and non-infection-related bile..., Chan-On [/bib_ref]. For example, in a recent International Cancer Genome Consortium (ICGC) analysis of gingiva-buccal cancers, specific genetic changes appeared to be unique to patient ethnicity (Indian) or disease etiology (betel nut chewing). Taken collectively, these findings raise the important need to consider HNSCC not as one condition but as a spectrum of different diseases, thus requiring genomic analysis of specific cohorts that are well-defined with regards to disease subsite and the at-risk patient population [bib_ref] Squamous cell carcinoma of the oral tongue in young non-smokers is genomically..., Pickering [/bib_ref].
To date, the majority of previous genetic studies focused on OTSCC have been confined to single gene analysis or used limited gene panels, and thus have not provided a comprehensive landscape [bib_ref] Common oncogenic mutations are infrequent in oral squamous cell carcinoma of Asian..., Zanaruddin [/bib_ref] [bib_ref] Tongue carcinoma infrequently harbor common actionable genetic alterations, Tan [/bib_ref]. A recent publication that examined the mutational spectra and copy number variations of oral tongue tumors showed that tumors from young patients and non-smokers appear to be genomically similar to those of older patients and smokers [bib_ref] Integrative genomic characterization of oral squamous cell carcinoma identifies frequent somatic drivers, Pickering [/bib_ref] [bib_ref] Squamous cell carcinoma of the oral tongue in young non-smokers is genomically..., Pickering [/bib_ref]. These data suggest that genomic alterations are unable to account for the changing demographics in OTSCC. In this study, we set out to define the mutational landscape of OTSCC in an Asian population, with a secondary focus on identifying driver mutations with a potential for therapeutic targeting. Our objectives were to determine whether tumors arising from the tongue had a distinct genetic landscape, if these differences were population specific or correlated with any epidemiological factors (e.g., gender), and to determine the proportion of alterations with therapeutic potential.
# Methods
## Patients and bio-specimens
Patients were identified from an institutional database of a consecutive series of patients treated at the National Cancer Centre Singapore between 1998 and 2012. Included patients were confirmed to have a histological diagnosis of squamous cell carcinoma involving the anterior tongue, with complete clinico-pathologic data and fresh frozen tumor samples available, with primary surgical treatment at the National Cancer Centre Singapore. Only patients aged 21 years or above diagnosed with primary OTSCC and not previously treated were included in the study. Treatment decisions were made in weekly multi-disciplinary meetings and recorded prospectively. All patients were treated with primary surgery followed by adjuvant therapy, if applicable. Samples for 78 patients were retrieved from the Singhealth Tissue Repository and tumor content confirmed to be greater than 50 % by a board-certified pathologist (TKL). For the 'discovery set' , germline DNA from blood was used as a reference for detecting somatic alterations. Medical records were reviewed for standard demographic data, risk factors, clinical, radiological, and histo-pathologic information, treatment parameters, and outcome of the disease. Smoking was defined as use of tobacco, chewable or smoked, for at least 1 year continuously. Patient characteristics are shown in Additional file 1.
# Ethical issues
In accordance with the Helsinki Declaration of 1975, standardized written consent for use of clinical material (which covers tumor tissue, blood, or other clinical specimens) and clinico-pathologic data for research was obtained at the time of surgical excision. Both this study and the tissue collection/consent protocols have been approved by the Singhealth Centralized Institutional Review Board (CIRB 2007/438/B).
DNA extraction, library construction, exome capture, and sequencing Genomic DNA was isolated from tumor specimens and blood samples using the DNeasy Blood and Tissue kit (Qiagen, Crawley, UK) according to the manufacturer's instructions. Two micrograms of genomic DNA was sheared to 200 bp using a Covaris E Series ultrasonic solubilizer (Covaris Inc., Woburn, MA, USA) and after adapter ligation the resultant library was subjected to DNA capture using either the Agilent SureSelect Human All Exon 50 MB V3 (Agilent, Santa Clara, CA, USA) or a custom-designed capture library (SureSelect XT2 Target Enrichment System, Agilent) designed to capture the coding sequences of 465 genes. Library construction and DNA capture were carried out according to the manufacturer's instructions. The custom 465-gene Sure-Select capture probes were designed using the Agilent SureDesign package. The concentration and size distribution of the libraries were determined on an Agilent Bioanalyzer DNA 1000 chip. Captured libraries were sequenced on a HiSeq2000 platform (Illumina, San Diego, CA, USA). The image analysis and base calling were done using the Illumina pipeline (v.1.6) with default settings.
## Sequence processing and variant calling
Read pairs were aligned to the National Center for Biotechnology Information (NCBI) human reference genome GRCh37 (hg19) using Burrows-Wheeler Aligner (BWA) software [bib_ref] Fast and accurate short read alignment with Burrows-Wheeler transform, Li [/bib_ref]. PCR duplicates removal was done using SAMTools [bib_ref] The Sequence Alignment/Map format and SAMtools, Li [/bib_ref]. After duplicate removal, the Genome Analysis Tool Kit (GATK) [bib_ref] The Genome Analysis Toolkit: a MapReduce framework for analyzing next-generation DNA sequencing..., Mckenna [/bib_ref] was used for base quality score recalibration, local realignment around indels, and variant calling. Partek Genomics Suite v.6.6 (Partek Inc., St. Louis, MO, USA) was also used to identify somatic single nucleotide mutations in targeted exons.
To exclude common germline polymorphisms, the detected variants were checked against dbSNP 138 [bib_ref] dbSNP: the NCBI database of genetic variation, Sherry [/bib_ref] , 1000 Genomes [bib_ref] A map of human genome variation from population-scale sequencing, Project [/bib_ref] , and ExACdatabases. In the discovery set, the variants detected in the matched normal controls were subtracted from variants detected in the tumors. In the 'prevalence set' , due to the absence of matched normal controls, rare germline SNVs were discarded using an in-house Southeast Asian normal exome dataset, including 523 exomes (sgSNPG). For further analysis, we restricted our attention to candidate somatic SNVs that were in the capture target and present inside exons or at canonical splice sites, with a variant depth ≥5 and depth in the normal sample ≥10.
Variants that passed this filtering were annotated using several gene-transcript databases (Consensus CDS, RefSeq, Ensembl and UCSC. Amino acid changes were determined based on the longest gene transcript. Somatic nonsynonymous single-nucleotide mutations were analyzed using PolyPhen2 [bib_ref] A method and server for predicting damaging missense mutations, Adzhubei [/bib_ref] in order to predict their impact on the protein function. To identify significantly mutated genes, the SMG test in the genome MuSiC (Mutational Significance in Cancer) suite was used with default parameters [bib_ref] MuSiC: identifying mutational significance in cancer genomes, Dees [/bib_ref]. Microindels were detected using GATK IndelGenotyperV2. For further analysis we only considered microindels that were in a target and in an exon or at a canonical splice site. Several filters were applied to reduce false positives: we retained microindels that (1) were not in a simple repeat, (2) had variant depth ≥5 with at least one read on each strand, (3) had depth in the normal sample ≥10, (4) were contained in ≥5 % of the reads, (5) had average base quality of all bases inside the microindel and ±5 bases around the microindel ≥25, (6) had an average number of mismatches in the reads containing the microindel ≤4, and (7) had an average fraction of mismatched bases in reads containing the microindel ≤0.2. Candidate somatic alterations were also assessed by visualizing sequencing data for each alteration using the Integrative Genomics Viewer (IGV; version 2.3.25) [bib_ref] Integrative genomics viewer, Robinson [/bib_ref]. The COSMIC database was used to annotate known cancer variants. Raw sequence data have been deposited at the European Genome-phenome Archive, accession number [EGA:EGAS00001001329].
## Sanger sequencing validation
To validate somatic mutations by Sanger sequencing, PCR amplicons encompassing the candidate mutation sites were sequenced in cancer tissues and their normal tissue counterparts using the ABI PRISM BigDye Terminator Cycle Sequencing Ready Reaction kit (Applied Biosystems, Foster City, CA, USA) on an ABI 3730 xl DNA Analyzer (Applied Biosystems). PCR primers for amplification and sequencing were designed by targeting regions immediately flanking the predicted variant using Primer3 software [bib_ref] Primer3-new capabilities and interfaces, Untergasser [/bib_ref]. The chromatograms were analyzed using the CodonCode Aligner softwareand manual review.
## Gene selection for targeted sequencing
For the target-exome sequencing, 465 protein-coding genes were selected according to the following criteria: (a) genes identified as somatically altered in the discovery set based on occurrence in at least two samples (28 genes); (b) genes identified as somatically altered in the discovery set based on occurrence in at least one sample, but harboring single-nucleotide variations (SNVs) called by both caller algorithms, GATK and PGS (164 genes); (c) genes described as frequently altered in head and neck tumors in earlier published literature [bib_ref] Exome sequencing of head and neck squamous cell carcinoma reveals inactivating mutations..., Agrawal [/bib_ref] [bib_ref] The mutational landscape of head and neck squamous cell carcinoma, Stransky [/bib_ref] (45 genes); (d) genes with significant mutation frequency across 279 HNSCC tumors present in TCGA dataset (42 genes); and (e) recognized cancer genes included in the Ion AmpliSeq™ Comprehensive Cancer Panel (Life Technologies) (186 genes) (Additional file 2).
## Statistical analyses
Statistical analysis was performed using the statistical software SPSS 19.0 for Windows (IBM, New York, NY, USA). In order to assess the association between the anatomical and pathological variables of patients and the molecular findings in the tumor samples, the chi-square or Fisher's exact tests were performed, as appropriate. Mutational frequency differences between the OTSCC and TCGA data were assessed using chi-square test, followed by the Benjamini-Hochberg procedure for multiple testing corrections. Survival curves were calculated by the Kaplan-Meier method and differences between groups were compared using the log-rank test. OS was measured as the time interval between the date of the initial treatment for the primary tumor and the date of the last follow-up or death, while disease-free survival (DFS) was defined as the time interval between the date of initial treatment and the date of the diagnosis of the first recurrence. For evaluation of the independent contribution of significant clinical and molecular variables on DFS or OS, all factors with significance in Kaplan-Meier analysis (P ≤ 0.2) were tested in the multivariate analyses using the Cox proportional hazard model. Results were calculated with 95 % confidence intervals (CIs). For all analysis we considered statistical significance when P value < 0.05. In the pathway analysis, the P values were calculated by Gene Set Enrichment Analysis (GSEA) algorithm by estimating the hypergeometric distribution of overlapping genes over all genes in the gene universe.
URLs UCSC Genome Bioinformatics, CodonCode Aligner software, dbSNP, 1000 Genomes Project [46], Exome Variant Server, COSMIC, ExAC database, cBioPortal for Cancer Genomics [49], MSigDB pathways database, Genomics of Drug Sensitivity in Cancer, Clinical trials database, Drugs approved by US Food and Drug Administration (FDA) [53], European Genome-phenome Archive.
# Results
## Patient characteristics
Eighteen OTSCC patient tumor and matched blood samples were used for the discovery set, and 60 tumor samples were used in the prevalence set (clinico-pathologic data for both cohorts is shown in Additional file 1). The prevalence set was used for all subsequent clinical correlations. In the latter cohort, approximately half (51.7 %; n = 31) were smokers, 63.3 % (n = 38) were males, with a median age of 55.5 years (range 21-89 years). All patients in the prevalence cohort underwent surgery as the primary modality of treatment, and 26 (43.3 %) received adjuvant radiation or chemo-radiation therapy. Approximately half presented with locally advanced disease (pT3/ T4, 48.3 %, n = 29) and had pathologic evidence of nodal metastasis (pN+, 46.7 %, n = 28) at presentation. Patients were followed up for a median of 36 months (range 1-190 months). During this time period, there were 17 (28.3 %) recurrences and 13 (21.7 %) deaths.
## Overview of otscc somatic mutation profile
To identify somatic mutations, we performed wholeexome capture and massively parallel DNA sequencing on paired DNA samples (tumor and matched white blood cells) from 18 OTSCCs (discovery set). From each sample, we sequenced around 16 billion bases, resulting in an average 93-fold coverage of each base in the targeted regions (range 70-115), with 94 % of DNA sequences being covered sufficiently (≥10×) for confident variant calling (Additional file 3).
Overall, 1,092 nonsynonymous somatic variations affecting 750 genes were detected, of which 992 were somatic SNVs and the remaining were somatic insertions or deletions (indels) (Additional file 4). The co-mutation plot for the discovery set is provided in Additional file 5. Of the single-nucleotide substitutions, 888 (81.3 %) were predicted to be missense mutations, 68 (6.2 %) nonsense mutations, and 36 (3.4 %) splice site mutations. The number of variants (SNVs and indels) per patient ranged from 11 to 105, with a mean of 61 mutations per sample, similar to previous oral cavity whole-exome sequencing studies (ranging from 29 to 85) [bib_ref] Squamous cell carcinoma of the oral tongue in young non-smokers is genomically..., Pickering [/bib_ref]. The alterations were verified manually using IGV. Subsequent technical validation by Sanger sequencing confirmed the presence of mutations in 28 genes (28/29, true-positive rate of 96.6 %), demonstrating that the algorithms used (GATK and PGS) for identifying mutations were reliable. Overall, the whole-exome sequencing of the discovery set indicated 28 genes harboring somatic coding mutations in two or more oral tongue tumors (a list of these 28 genes is presented in Additional file 6).
Nucleotide sequence contexts of somatic mutations can be indicative of specific mutagenesis mechanisms occurring in tumor cells [bib_ref] Signatures of mutational processes in human cancer, Alexandrov [/bib_ref] [bib_ref] Genome-wide mutational signatures of aristolochic acid and its application as a screening..., Poon [/bib_ref] [bib_ref] Mutational heterogeneity in cancer and the search for new cancer-associated genes, Lawrence [/bib_ref]. Analyzing synonymous and non-synonymous alterations, we observed elevated rates of C > T transitions, particularly at CpG positions, consistent with deamination of 5-methyl-cytosine to uracil by APOBEC or other endogenous processes. C > G mutations were also common (Additional file 7). We visually examined the somatic mutation spectra of the 18 exome-sequenced tumors and performed principal component analysis to search for possible systematic differences in spectra according to several clinical factors: age, smoking status, gender, HPV status, ethnicity, and recurrence. These analyses did not suggest any systematic differences according to any of these categories (Additional file 6; see "Discussion").
Copy number variant analyses were performed as previously described [bib_ref] Exome-wide sequencing shows low mutation rates and identifies novel mutated genes in..., Cutcutache [/bib_ref]. The results are presented in Additional files 8, 9, and 10. We identified 30 amplified regions and 56 regions with recurrent deletions in the discovery set. Focal amplifications in chromosome 7 harboring the MET oncogene were seen in 7 of 18 samples, while deletions were commonly seen on chromosome 3 (9/18 samples), involving VHL, SETD2, PBRM1, MLH1, and BAP1. All regions of gains and losses, with their corresponding tumor-associated genes are indicated in Additional files 8 and 9. There were no correlations between copy number variants and smoking status, age, gender, ethnicity, or recurrence.
## Evaluating the frequency and distribution of somatic mutations by target sequencing
To further investigate the precise frequency and distribution of somatic mutations in an expanded OTSCC cohort, we designed a panel with 465 genes including (a) the 28 candidate genes frequently mutated in the discovery set, (b) genes previously described as frequently mutated in HNSCC [bib_ref] Exome sequencing of head and neck squamous cell carcinoma reveals inactivating mutations..., Agrawal [/bib_ref] [bib_ref] The mutational landscape of head and neck squamous cell carcinoma, Stransky [/bib_ref] , and (c) recognized cancer genes that were manually curated (a list of all genes included in the target panel is presented in Additional file 10). These genes were sequenced in a prevalence set of 60 treatmentnaive snap-frozen OTSCC specimens by targeted sequencing with an average depth of 136× (Additional file 11).
For the prevalence set, we used a consecutive patient series treated in a standardized manner in order to make meaningful conclusions on epidemiology and outcome, and only tumor samples were available for this cohort [bib_ref] The rise of big clinical databases, Cook [/bib_ref]. Therefore, an alternative strategy to discriminate germline SNVs in the absence of normal tissue DNA was adopted, following previously described studies [bib_ref] Clinical and biological implications of driver mutations in myelodysplastic syndromes, Papaemmanuil [/bib_ref] [bib_ref] Identification and characterization of cancer mutations in Japanese lung adenocarcinoma without sequencing..., Suzuki [/bib_ref]. We created a high-fidelity tumor-only analysis pipeline that relied on two steps. First, we excluded variants found in either public databases of germline SNVs [dbSNP (build 132), 1000 Genomes, and EVS (6500-SI-V2)] [bib_ref] dbSNP: the NCBI database of genetic variation, Sherry [/bib_ref] [bib_ref] A map of human genome variation from population-scale sequencing, Project [/bib_ref] or in an in-house database of germline SNVs drawn from 523 normal Asian exomes. We initially also planned to exclude variants described in the beta version of the ExAC database, but preliminary tests showed that this filtering steps removes a significant proportion of the TP53 alterations in our cohorts (23 % discovery set and 22 % prevalence set). Furthermore, the examination of the TCGA data for HNSCC also revealed that 22 % of the TP53 mutations described there were also present in the ExAC database, with a significant proportion of damaging missense mutations. Given these findings, we did not utilize the ExAC database in the final analysis. Variants present in the COSMIC database (v.52) were kept regardless of their match in the other databases, because dbSNP in particular contains some known oncogenic somatic mutations. Second, we excluded variants with allele frequencies >0.41. The rationale for this criterion is that somatic mutations usually present with variant read proportions substantially less than 0.5 in next-generation sequencing data, due to admixture of genomes from non-malignant cells in a tumor sample. By contrast, germline variants have read proportions of either 1 (for homozygous variants) or~0.5 (for heterozygous variants) in regions where the tumor genome is diploid. We note that copy number gains or losses in tumor genomes may alter the expected variant read proportions of germline variants, so this two-step approach will not always exclude germline variants. Conversely, genomic copy number gains may sometimes cause somatic mutation to have read proportions greater than 0.5, so this two-step approach sometimes incorrectly excludes somatic mutations. As discussed below, we estimated the accuracy of this approach in a set of HNSCCs with matched normal DNA.
To validate this two-step tumor-only filtering protocol and to verify the extent to which it would retain rare germline alterations amidst bona fide somatic mutations, we applied it to both the discovery set and an independent cohort of 16 HNSCC exomes for which normal matched controls were available. For the discovery set, this algorithm showed a high degree of reliability and accuracy with a positive predictive value (PPV) of 92 % and negative predictive value (NPV) of 96 % (with 89 % sensitivity and 97 % specificity). Further validation of this tumor-only protocol in an independent cohort of 16 non-subsite specific HNSCC samples also showed good reproducibility with PPV of 94 % and NPV of 89 % (with 73 % sensitivity and 98 % specificity). These control analyses support the notion that our tumor-only filtering protocol does not lead to excessive numbers of germline variants miscalled as somatic mutations.
After applying the tumor-only filtering protocol to the prevalence set, 380 genes were detected as harboring SNVs or indels. All mutations identified in the prevalence set are shown in Additional file 12. Unsurprisingly, the most prevalent genetic alteration detected in this cohort comprised a broad spectrum of [fig_ref] Figure 1: Mutation plot summary of 60 oral tongue squamous cell carcinoma patients showing... [/fig_ref]. We examined expression levels of these ten genes using HNSCC RNA-seq data from TCGA and our own previously published microarray data [bib_ref] An eleven gene molecular signature for extra-capsular spread in oral squamous cell..., Wang [/bib_ref] , both of which confirm that the genes listed here are known to be expressed in OTSCC (data not shown).
This pattern of Asian OTSCC genetic mutations was compared with the profile of 172 exomes from oral cavity tumors available from TCGA (OC-TCGA). One major caveat to this comparison is that there are fundamental differences between the two cohorts examined here in terms of tumor site (tongue versus oral cavity), disease stage, ethnicity (Asian versus North American) and environmental exposure. Single gene comparisons showed that mutation frequencies of the following genes were higher in Asian OTSCCs compared with OC-TCGA: DST (P < 0.001), RNF213 (P = 0.012), COL6A6 (P = 0.049) and ZFHX4 (P = 0.031). Even after multiple hypothesis correction, mutation frequencies remained higher in Asian OTSCCs for DST (P = 0.001) and RNF213 (P = 0.033) . In contrast, OC-TCGAs exhibited more frequent mutations in TP53 (P < 0.001), CDKN2A (P = 0.005), and NOTCH1 (P = 0.019). Critically, no significant differences were observed in the mutation rates of many other genes frequently mutated in both cohorts (CASP8, CDH10, EP300, FAT1, MLL2/KMT2D, MLL3/KMT2C, NSD1, PIK3CA, PLEC, SYNE1, and USH2A). These similarities in mutation frequency in the latter genes provide a sound validation of and positive control for our methodology, for both the experimental approach (tumor content, library creation, and sequencing) and bioinformatics pipeline.
## Analysis of significantly altered pathways
Pathway analysis was performed to examine the association between mutation profiles and core canonical pathways in Asian OTSCC, using either the MSigDB pathways database [bib_ref] Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles, Subramanian [/bib_ref] or compared with manually curated cancer-related pathways [bib_ref] Exome sequencing identifies a spectrum of mutation frequencies in advanced and lethal..., Kumar [/bib_ref] [bib_ref] Whole-genome methylation analysis of benign and malignant colorectal tumours, Beggs [/bib_ref]. Putative cancerrelated pathways achieving statistical significance, as determined by GSEA, are listed in Additional file 13. Among the most prominently mutated pathways were the epidermal growth factor receptor (EGFR), ERRB, mitogen-activated protein kinase (MAPK), and NOTCH signaling pathways, as well as adherens junction and focal adhesion pathways [fig_ref] Figure 1: Mutation plot summary of 60 oral tongue squamous cell carcinoma patients showing... [/fig_ref]. Overall, 55 % of the OTSCC samples had mutations in at least one gene affecting the MAPK pathway, while EGFR, Notch, and ERBB signaling families where modified in 52 %, 32 % and 23 % of cases, respectively. Genes involved in cell junction organization and focal adhesion pathways were mutated in 22 % and 53 % of OTSCC cases, respectively. Other core canonical signaling pathways frequently found mutated in OTSCC included
## Notch pathway mutations
As stated above, we found the Notch pathway to be altered in a high proportion of the Asian OTSCCs.
However, classical single-gene NOTCH1 mutations, previously reported in other studies, were notably absent from our cohort (only three cases have NOTCH1 mutations). Nevertheless, pathway analyses demonstrated that the Notch signaling pathway was altered in OTSCC in the same proportion as previously reported for HNSCC [bib_ref] The mutational landscape of head and neck squamous cell carcinoma, Stransky [/bib_ref] or detected in oral cavity tumors and HNSCC by TCGA [fig_ref] Figure 4 a: Co-mutation map showing alterations in genes involved in the Notch pathway in... [/fig_ref]. Genes recurrently altered in this pathway include AR, ARNT, EP300, CREBBP, JAK2, JAK3 NCOA1, NOTCH2, NOTCH3, and PARP1. In total, the A B The frequency of commonly mutated genes in the current cohort of oral tongue squamous cell carcinoma (OTSCC) compared with oral cavity squamous cell carcinoma from TCGA (OC-TCGA). a The most frequently mutated genes in the OTSCC cohort. b The most frequently mutated genes in the OC-TCGA cohort. Significance is determined by multiple hypothesis testing using chi-square test, followed by the Benjamini-Hochberg procedure. NS not significant Notch signaling pathway was altered in 32 % (19/60) of the OTSCC cases, with all somatic events following a mutually exclusive pattern [fig_ref] Figure 4 a: Co-mutation map showing alterations in genes involved in the Notch pathway in... [/fig_ref].
Interestingly, mutations in the Notch signaling pathway were enriched in women compared with men (P = 0.04; Additional file 15). More importantly, a striking difference in DFS was observed between OTSCC cases with mutations in the Notch pathway compared with those without, where the former had significantly poorer outcomes (P = 0.005; [fig_ref] Figure 4 a: Co-mutation map showing alterations in genes involved in the Notch pathway in... [/fig_ref] ; Additional file 13). Based on multivariate models that included clinico-pathologic correlates associated with prognosis (age, tobacco consumption, N stage and extra-capsular spread), modifications in the Notch pathway remained an independent predictor for DFS. Patients with Notch pathway mutations were 3.3 times more likely to die with recurrent disease compared with those who did not have these alterations [p = 0.016, hazard ratio (HR) = 3.3, 95 % CI = 1.24-8.94; [fig_ref] Figure 4 a: Co-mutation map showing alterations in genes involved in the Notch pathway in... [/fig_ref] ]. Analysis of TCGA data for oral cavity suggests a similar trend towards poorer OS in patients with alterations in this pathway, however this was not statistically significant (p = 0.13; data not shown).
Gene expression microarray data were available for a limited number of these samples: Notch pathway altered (n = 9) versus Notch pathway wild type (n = 21) [bib_ref] An eleven gene molecular signature for extra-capsular spread in oral squamous cell..., Wang [/bib_ref].
GSEA analysis of gene expression confirmed that the presence or absence of these somatic events resulted in significant alteration of two Notch associated pathways Targets DN, p = 0.004; PID Notch pathway, p = 0.011), confirming that these alterations truly modulate Notch signaling (Additional file 16). These orthogonal data suggest that the Notch pathway is modulated as predicted in the cases with mutations in Notch pathway.
## Chromatin remodeling gene family
There has been significant interest in the role of chromatin remodeling enzymes in a range of cancers, and examination of this gene set revealed alterations in 42 % of the OTSCC cohort, similar to data from TCGA for the oral cavity (42 %) and for all HNSCC (47 %) [fig_ref] Figure 5 a: Co-mutation map showing alterations in genes involved in chromatin remodeling in 25 [/fig_ref]. Chromatin remodeling genes frequently altered in OTSCC included CREBBP, EP300, KAT6A, MLL/KMT2A, MLL2/ KMT2D, MLL3/KMT2C, NCOA1, NSD1, SETD2, and WHSC1. Moreover, most of the mutations in this gene family were mutually exclusive, especially mutations affecting the MLL family of histone methyltransferases [fig_ref] Figure 5 a: Co-mutation map showing alterations in genes involved in chromatin remodeling in 25 [/fig_ref] [fig_ref] Figure 5 a: Co-mutation map showing alterations in genes involved in chromatin remodeling in 25 [/fig_ref]. Based on multivariate models, modifications in this gene set remained an independent predictor for DFS. Patients with mutations in this pathway were 3.3 times more likely to die with recurrent disease compared with those who did not have these alterations (p = 0.03, HR = 3.3, 95 % CI = 1.12-9.55; [fig_ref] Figure 5 a: Co-mutation map showing alterations in genes involved in chromatin remodeling in 25 [/fig_ref]. Again, analysis of TCGA data for tongue cancers suggests a similar trend towards poorer OS in patients with alterations in this pathway, although this was not statistically significant (P = 0.11; data not shown).
Remarkably, multivariate analysis including both the Notch pathway and chromatin remodeling genes as variables showed that alterations in the Notch pathway remained as an independent predictor of recurrence. Patients harboring mutations in at least one member of this signaling pathway were 2.9 times more likely to develop recurrence and die (P = 0.037, HR = 2.9, 95 % CI = 1.07-7.73; [fig_ref] Figure 5 a: Co-mutation map showing alterations in genes involved in chromatin remodeling in 25 [/fig_ref]. Furthermore, this same multivariate model also revealed a tendency towards to significance for the presence of mutations in chromatin remodeling genes (P = 0.057, HR = 2.9, 95 % CI = 0.97-8.91; [fig_ref] Figure 5 a: Co-mutation map showing alterations in genes involved in chromatin remodeling in 25 [/fig_ref].
## Analysis of significantly altered actionable mutations
Finally, the mutation data were also used to characterize alterations in drug targetable genes. In this analysis, manually curated gene sets were assembled based on genes known to have similar functions/pathways, gene families, or overlapping cross-sensitivity to available targeted compounds [53, [bib_ref] Serglycin is a theranostic target in nasopharyngeal carcinoma that promotes metastasis, Li [/bib_ref]. This group of 20 genes included receptor tyrosine kinases, ERBB, JAK-STAT, MAPK-AKT, and phosphoinositide 3-kinase (PI3K) signaling, for which therapeutic agents are currently in clinical use or available in the trial setting. Based on this list, 50 % of OTSCC samples harbored modifications in at least one of these genes, and the majority were mutually exclusive [fig_ref] Figure 6 a: Co-mutation map showing alterations in drug targetable genes in 30 [/fig_ref]. The majority of the PIK3CA mutations (four out of five) are known to be activating mutations, while about half the mutations in the ERBB family of genes (three out of six) occur in regions that affect signaling (extracellular domain or tyrosine kinase domain) [bib_ref] Tongue carcinoma infrequently harbor common actionable genetic alterations, Tan [/bib_ref] [bib_ref] Whole-exome and targeted gene sequencing of gallbladder carcinoma identifies recurrent mutations in..., Li [/bib_ref]. Based on Polyphen2 analysis, a significant proportion of the other mutations in this set of 'druggable' genes are predicted to be damaging to the respective protein function. Of note, the frequency of these 'druggable mutations' is significantly higher in OTSCC compared with that detected in the oral cavity tumor exomes analyzed by TCGA (P = 0.020; [fig_ref] Figure 6 a: Co-mutation map showing alterations in drug targetable genes in 30 [/fig_ref].
# Discussion
To date, most cancer genome studies on OSCC have regarded this entity as a homogenous disease. However, differences in biological behavior already hint that OSCC is not a uniform condition [bib_ref] Metastasis of squamous cell carcinoma of the oral tongue, Sano [/bib_ref] [bib_ref] Changing trends in smoking and alcohol consumption in patients with oral cancer..., Montero [/bib_ref] [bib_ref] Epidemiology of oral cancer in Asia in the past decade-an update, Rao [/bib_ref] , likely requiring dissection by populations affected, anatomic subsite and disease etiology [bib_ref] Recurrent gain-of-function mutations of RHOA in diffuse-type gastric carcinoma, Kakiuchi [/bib_ref] [bib_ref] Exome sequencing identifies distinct mutational patterns in liver fluke-related and non-infection-related bile..., Chan-On [/bib_ref] [bib_ref] Tongue carcinoma infrequently harbor common actionable genetic alterations, Tan [/bib_ref]. This population/subsite/etiologybased approach requires rigorous sample collection and exclusion of a significant number of cases that do not fit the selection criteria. One of the first papers that used this approach was the Indian ICGC group, who used exome analysis for gingiva-buccal oral squamous cell carcinoma (OSCC-GB). In this study, apart from detecting mutations in genes previously associated with HNSCC (TP53, NOTCH1, FAT1, and CASP8), they also identified recurring alterations in genes specific to this disease (USP9X, MLL4/KMT2B, ARID2, UNC13C, and TRPM3). In a similar fashion, our study focused on the mutational landscape of squamous cell cancers arising from the single subsite of oral (or anterior) tongue in a South-East Asian population (OTSCC). More importantly, as this cohort of patients was treated in a uniform manner at a single tertiary institution, data from this study can also be used to draw meaningful conclusions based on clinical correlates. Our patient cohort exhibits several features that are representative of the evolution of this disease in Asia. First, there is an almost equal proportion of smokers and non-smokers. Second, the male:female ratio is approaching equivalence. Third, the median age of disease has been declining over the past three decades and is now around 55 years. Reassuringly, mutational analysis of our OTSCC cohort identified frequent nucleotide alterations in SYNE1, FAT1, MLL2/KMT2A, MLL3/KMT2C, PIK3CA, and CASP8 at similar frequencies reported by TCGA and ICGC. However, there were significantly fewer mutations in TP53, NOTCH1, and CDKN2A. Mutations in the RAS genes were extremely infrequent (HRAS, n = 1/60, with no mutations in KRAS and NRAS) and mutations in EGFR notably absent. In contrast, several less well-characterized genes (DST and RNF213) were mutated at higher rates in Asian OTSCC than previously reported. Most of the unique genes detected here were also not seen in the ICGC study, which focused on an Indian population with tobacco-and betel nut-associated OSCC-GB, reinforcing the need to carry out similar studies in different populations. Indeed, we hypothesize that the unique genetic alterations in both studies can be attributable to the etiology of disease, differences in the (Asian) population examined or the distinct biological features of the disease subsites (in this case oral tongue). As predicted, these genetic alterations also differed from HPV-driven cancers involving the base of tongue, which had a higher frequency of activating mutations in PIK3CA, with loss of TRAF3 and amplification of E2F1. In a similar vein, patterns attributable to etiology and specific populations have been observed for other cancer types [bib_ref] Exome sequencing identifies distinct mutational patterns in liver fluke-related and non-infection-related bile..., Chan-On [/bib_ref].
The reason for this changing epidemiology in a disease (OTSCC) classically seen in older, male smokers remains unknown, and one of the major objectives of this study was to determine the genetic basis for this 'emerging phenotype'. Our analysis of OTSCC found no specific mutations unique to younger patients or non-smokers, although there was a tendency towards more frequent mutations in MLL2/KMT2A and PIK3CA in older patients. Moreover, the mutational signature in our cohort did not reveal a tobacco-induced signature in OTSCC cases grouped by age (young versus old) or by smoking status, which is similar to findings by Pickering et al. [bib_ref] Squamous cell carcinoma of the oral tongue in young non-smokers is genomically..., Pickering [/bib_ref]. The latter recently reported whole-exome sequencing analysis of two independent cohorts of tongue tumors and found that mutational spectrum were similar between young and old patients. Further, they also observed that the types of base changes observed in the young cohort were similar to those in the old cohort even though they differed in smoking history, and no smoking signature could also be found in OTSCC. These findings, now supported by two independent studies, suggests that tobacco smoke, despite being an established risk factor for OTSCC, may act primarily as a tumor promoter while having only a minor impact on the initiation of tongue cancers [bib_ref] Tobacco smoke promotes lung tumorigenesis by triggering IKKbeta-and JNK1-dependent inflammation, Takahashi [/bib_ref]. Analysis of the OTSCC data at the gene set level provided valuable insights into pathways that potentially drive OTSCC carcinogenesis. We were initially surprised that NOTCH1 mutations were relatively infrequent in this cohort, although one previous study in esophageal squamous cell cancer had shown that NOTCH1 mutations are uncommon in Asian patients compared with matched Caucasian populations [bib_ref] Comparative genomic analysis of esophageal adenocarcinoma and squamous cell carcinoma, Agrawal [/bib_ref]. However, pathway analysis confirmed that the Notch pathway was affected in a third of our OTSCC cases by mutations that were mutually exclusive, distributed across 11 distinct genes along the pathway. These alterations were more common in women, and of concern, they conferred a poorer prognosis in these patients regardless of other factors or mutations, even on multivariate analysis. The role of Notch as a tumor suppressor pathway in OSCC has been suggested by several studies [bib_ref] Exome sequencing of head and neck squamous cell carcinoma reveals inactivating mutations..., Agrawal [/bib_ref] [bib_ref] The mutational landscape of head and neck squamous cell carcinoma, Stransky [/bib_ref] [bib_ref] Loss-offunction mutations in Notch receptors in cutaneous and lung squamous cell carcinoma, Wang [/bib_ref]. Activated NOTCH1 has an antiproliferative effect in tongue tumor cells through downregulation of Wnt/β-catenin signaling, inducing apoptosis and cell cycle arrest [bib_ref] Submit your next manuscript to BioMed Central and take full advantage of:..., Duan [/bib_ref]. Furthermore, in a tongue cancer model using mouse xenografts, the presence of NOTCH1 caused a dramatic reduction in tumor size compared with tumors in which this gene was absent [bib_ref] Integrative genomic characterization of oral squamous cell carcinoma identifies frequent somatic drivers, Pickering [/bib_ref]. This apparent discordance between infrequent NOTCH1 mutations per se and a still defective Notch pathway through other gene mutations in 32 % of OTSCC patients highlights the importance of deregulating the Notch-driven epithelial cell differentiation program in OTSCC through a range of convergent targets.
Analyses of the chromatin remodeling gene family led to similar conclusions. Interestingly, we observed a significantly higher frequency of MLL3/KMT2C mutations (16 %) in our cohort compared with mutations in the MLL homologs MLL2/KMT2A in OC-TCGA (Western/predominantly Caucasian population)and in MLL4/KMT2D in OSCC-GB. Damaging mutations in MLL2/KMT2A were more frequently identified in smokers in our cohort, while the mutations in MLL4/KMT2D seem to be unique to the Indian population studied by the ICGC. The higher frequency of MLL3/KMT2C mutations in our cohort may therefore be a phenomenon associated with never-smoker OTSCC populations, and represents another example of convergent evolution on a critical pathway in OTSCC carcinogenesis.
Despite advances in technology, clinical outcomes have not changed significantly over the past three decades for OSCC. However, analysis of our cohort showed that at least half of the patients harbored potentially targetable mutations in signal transduction pathways, where targeted compounds are already available. These results have several important implications. Firstly, the mutation frequencies of these genes in our population were significantly higher than in previously reported studies, and this pattern should therefore be taken into consideration in future trial designs. Second, previous pessimism regarding the low occurrence of sensitizing mutations in HNSCC should not be translated to any cohort until similar studies have been conducted in this population/subsite/etiology-based approach. Interestingly, there was a similar conclusion by the ICGC group regarding OSCC-GB, who also similarly identified higher rates of potentially actionable mutations.
Beyond OTSCC, our study is also notable from a technical standpoint. To date, most cancer sequencing studies have utilized tumor and normal matched pairs to distinguish de novo somatic mutations from existing germline variants. In this study, we needed to obtain tumor representation from a consecutive series of patients presenting with tongue cancer, treated in a standardized manner with clinic-pathologic and outcome data available, with no bias in patient selection (prevalence set), to enable robust clinico-pathological analysis. In this process, however, it was not possible to obtain matched normal blood from every patient in this prevalence set. This is a very common scenario encountered in the translational research domain, motivating the need to develop approaches to identify somatic mutations in the absence of matched normal controls. In our study, we addressed this challenge by using, in addition to commonly used pipelines (e.g., 1000 Genomes, dbSNP, EVS), a locally assembled 523-patient normal sample dataset to discriminate the germline SNVs. One major concern was that this approach may not remove all rare germline alterations, erroneously classifying these as somatic variations. To test the validity of this approach, we determined the PPV, NPV, specificity, and sensitivity of the filtering algorithm by applying the tumor-only filtering protocol on the discovery set as well as an independent cohort of HNSCC cases (comprising 16 tumor-normal matched pairs). The data for the discovery set show that the methodology was extremely reliable, with PPV and NPV over 90 %. Even after correcting for these accuracy data, the differences observed between our cohort and TCGA remains significant. Further validation of this approach can be inferred from the fact that we found similar frequencies of mutations in a number of genes previously detected by TCGA, including USH2A, FAT1, PLEC, SYNE1, PIK3CA, MLL2/KMT2A, MLL3/ KMT2C, CASP8 COL6A6, and ZFHX4. There are a few other studies in the literature where 'tumor-only' protocols have been used. In a study conducted at the Sanger Centre, investigators were able to identify base substitutions and small insertions/deletions in the exome data from 738 patients with myelodysplastic syndrome by comparison against one unmatched normal sample [bib_ref] Clinical and biological implications of driver mutations in myelodysplastic syndromes, Papaemmanuil [/bib_ref]. However, no data were presented on the accuracy of this approach. Along the same line, a Japanese study identified mutations in a cohort of 97 Japanese lung adenocarcinoma patients by filtering out germline variants present in an external exome dataset composite by 217 normal control exomes from individuals with the same ethnic background [bib_ref] Identification and characterization of cancer mutations in Japanese lung adenocarcinoma without sequencing..., Suzuki [/bib_ref]. The latter study reported accuracy rates of 64 %, establishing the validity of this approach. It is likely that the accuracy rates for any tumor-only pipeline will improve with increasing numbers of control/normal sample data from which germline alterations can be subtracted, and more importantly represent the same population distribution in which the primary analysis is carried out. The latter approach likely accounts for the accuracy of our approach, which focused on the same patient population. It is likely that with increasing numbers of normal samples deposited into 'open source' databases like ExAC, these algorithms can be further refined. Furthermore, the availability of these databases and algorithms allow future studies to proceed with retrospective sample collections, regardless of matched normal samples, at half the cost for sequencing. This approach also obviates the issue of dealing with patient germline data and the inherent ethical dilemma that this generates.
[fig] Figure 1: Mutation plot summary of 60 oral tongue squamous cell carcinoma patients showing frequently mutated genes. The top plot shows the key clinical parameters, below which the mutation status of the recurrently mutated genes for each tumor is indicated. Somatic mutations are colored according to functional class according to the legend below the plot. Prevalence is indicated as number of mutations in the graph on the right and mutational frequency is given on the left of the mutation plot mammalian target of rapamycin (mTOR), transforming growth factor β-SMAD, chromatin remodeling genes, KIT, β-catenin and apoptosis, many of which may have therapeutic implications. Correlations between significantly altered pathways and clinical characteristic/outcomes are presented in Additional file 15. [/fig]
[fig] Figure 3: Mutations affecting chromatin remodeling genes were significantly more frequent in older patients (p = 0.03; Additional file 15) and patients with mutations have significantly poorer DFS compared with those who Pathway map showing molecular circuits implicated in OTSCC. a EGFR and HER2. b Adherens junction. c NOTCH. d Focal adhesion. Numbers in each box represent the percentage of tumors harboring mutations in the respective gene did not (p = 0.01; [/fig]
[fig] Figure 4 a: Co-mutation map showing alterations in genes involved in the Notch pathway in 19 (32 %) oral tongue squamous cell carcinoma (OTSCC) samples. Gene names and mutation rates are as indicated on either side of the plot, while the type of mutation follows the legend below. b, c Comparison between OTSCC, oral cavity cohort from TCGA (OC-TCGA) and entire head and neck cohort of TCGA (HN-TCGA) for NOTCH1 mutation frequency and alterations in the entire Notch pathway (11 genes), respectively. P value based on t-test as indicated if significant or NS if not significant. d Kaplan-Meier plot showing DFS estimates according to the mutational status of the Notch pathway. P value is based on log-rank test. e Results of multivariate analysis of selected prognostic factors for DFS in OTSCC by Cox proportional hazards model. HR hazard ratio [/fig]
[fig] Figure 5 a: Co-mutation map showing alterations in genes involved in chromatin remodeling in 25 (42 %) oral tongue squamous cell carcinoma (OTSCC) samples. Gene names and mutation rates are as indicated on either side of the plot, while the type of mutation follows the legend below. b Comparison between OTSCC, the oral cavity cohort from TCGA (OC-TCGA) and the entire head and neck cohort of TCGA (HN-TCGA) for frequency of genetic alterations in ten chromatin remodeling genes. P value based on t-test; NS not significant. c Kaplan-Meier plot showing DFS estimates according to the mutational status of the chromatin remodeling family. P value is based on log-rank test. d, e Results of multivariate analysis of selected prognostic factors for DFS in OTSCC by Cox proportional hazards model. HR hazard ratio [/fig]
[fig] Figure 6 a: Co-mutation map showing alterations in drug targetable genes in 30 (50 %) oral tongue squamous cell carcinoma (OTSCC) samples with gene names indicated on the left. b Comparison between OTSCC and oral cavity cohort from TCGA (OC-TCGA) for frequency of genetic alterations in 20 drug targetable genes. P value based on t-test [/fig]
|
Increasing physical activity levels following treatment for cervical cancer: an intervention mapping approach
Purpose The purpose of this study was to utilise the intervention mapping (IM) protocol as a framework with which to develop an intervention underpinned by relevant behaviour change theory to promote physical activity (PA) following treatment for cervical cancer. Methods The six steps of the IM protocol were followed. A qualitative semi-structured interview study and a rapid review of the literature were conducted along with the development of a logic model of the problem and a logic model of change to inform intervention development.Results An intervention was developed which aims to increase PA levels following treatment for cervical cancer, tailored to address key findings from the IM needs assessment. These include embedding behavioural and social strategies that help participants to overcome perceived barriers to PA participation; goal setting strategies to gradually increase PA levels with a view of reaching relevant PA guidelines for cancer survivors and feedback to encourage self-assessment of well-being and PA capability. Conclusion This study maps the development of a novel PA intervention for those who have been treated for cervical cancer. The use of a systematic development framework was necessary as little insight exists regarding PA preferences after treatment for cervical cancer. Implications for Cancer Survivors PA behaviour is associated with positive physical and psychological health outcomes for cancer survivors. Optimising targeted promotion of PA behaviour following treatment for cervical cancer may result in an enhanced survivorship experience through increased PA behaviour and improved quality of life (QOL).
# Background
The most recent data suggests that there are on average 3200 cases of cervical cancer diagnosed each year. Survival rate has improved over the last 40 years with documented increases in the number of cervical cancer survivors beyond 5 years or more (60%). Despite survivors being at risk of experiencing several health and lifestyle related issues [bib_ref] Assessment of quality of life and urinary and sexual function after radical..., Plotti [/bib_ref] , very few lifestyle interventions exist which are specifically aimed at improving QOL following cervical cancer and its associated treatments. Therefore, an increased focus on survivorship and interventions to enhance recovery after cervical cancer are warranted.
Navigating the recovery period to regain a sense of normality following treatment for cervical cancer is perceived as challenging [bib_ref] A peculiar time in my life": making sense of illness and recovery..., Phillips [/bib_ref]. This can be due to a myriad of psychological, physical and social impacts of treatment-related changes [bib_ref] Women's experiences after a radical vaginal trachelectomy for early stage cervical cancer, Lloyd [/bib_ref]. Research suggests that individuals treated for cervical cancer, particularly with radiotherapy, tend to exhibit worse QOL scores compared to those treated for other gynaecological tumours [bib_ref] Cervical cancer survivorship: long-term quality of life and social support, Pfaendler [/bib_ref]. Premature menopause is a particularly disabling after-effect of treatment, predominantly for those treated at a young age due to the comorbid negative symptomology that may be experienced as a result (e.g. osteopenia and vasomotor symptoms [bib_ref] Quality of life and emotional distress in early stage and locally advanced..., Ferrandina [/bib_ref]. Additionally, long-term psychological distress is common [bib_ref] Comparison of quality of life and sexuality between cervical cancer survivors and..., Lee [/bib_ref] , often due to sexual dysfunction, fear and anxiety related to cancer recurrence and a loss of fertility [bib_ref] Iatrogenic menopause after treatment for cervical cancer, Moss [/bib_ref]. Previous explorations have demonstrated a need for psychosocial intervention to improve the survivorship experience [bib_ref] Women's experiences after a radical vaginal trachelectomy for early stage cervical cancer, Lloyd [/bib_ref].
PA has been consistently shown to have positive psychological benefits after cancer [bib_ref] Physical activity for cancer survivors: meta-analysis of randomised controlled trials, Fong [/bib_ref] , yet studies frequently highlight that survivors report low levels of PA up to 3 years after treatment for gynaecological cancers [bib_ref] The impact of gynaecological cancer treatment on physical activity levels: a systematic..., Lin [/bib_ref]. Meta-analysis data indicates that PA interventions are widely investigated and are a safe way to improve QOL [bib_ref] Effects and moderators of exercise on quality of life and physical function..., Buffart [/bib_ref]. However, the body of literature in this area lacks diversity in tumour type with most studies conducted among breast cancer survivors [bib_ref] Interventions for promoting habitual exercise in people living with and beyond cancer, Turner [/bib_ref]. The limited existing evidence following cervical cancer suggests that PA can result in improvements in fatigue [bib_ref] A randomised controlled trial testing the feasibility and efficacy of a physical..., Donnelly [/bib_ref] and sexual functioning [bib_ref] Effect of a pelvic floor muscle training program on gynecologic cancer survivors..., Yang [/bib_ref]. Further support is needed to strengthen findings and to determine preferable modalities and durations specifically in this population. Therefore, the current study aimed to design a behaviour change intervention to increase PA levels following treatment for cervical cancer.
The most effective behaviour change interventions are those which explicitly link theory to the components of an intervention [bib_ref] Developing and evaluating complex interventions: the new Medical Research Council guidance, Craig [/bib_ref]. It has been strongly suggested that this becomes common practice in order to maximise the effectiveness of interventions and the potential for refinement and progression through robust evaluation [bib_ref] DEBATE: Do interventions based on behavioral theory work in the real world?, Hagger [/bib_ref]. Considering this, the development of the current intervention followed the IM protocol which is a six-step systematic framework for intervention developmentand which has been previously used in cancer populations [bib_ref] Use of intervention mapping to adapt a health behavior change intervention for..., Koutoukidis [/bib_ref]. The IM protocol allows the integration of experience and insight from the target population whilst also prioritising relevant behaviour change theories to inform the intervention components (e.g. [bib_ref] Use of intervention mapping to adapt a health behavior change intervention for..., Koutoukidis [/bib_ref].
# Method
Firstly, a planning group with relevant experience was established. The planning group consisted of a PhD student specialising in exercise psychology with 2 years of experience in PA intervention design (NM). Three of the authors (FM, HM, CE) together have at least 15 years of experience in intervention design and implementation research. Three of the authors have also worked specifically with cancer groups in research and practice (EM, 18 years; FM; HM, at least 15 years). Secondly, in line with guidance by the National Institute for Health Research (NIHR) [bib_ref] Patient and public involvement in health and social care research: a handbook..., Nihr [/bib_ref] , two patient and public involvement (PPI) groups were established over the course of the research. The initial discussion with a group of six patients treated for cervical cancer enabled the identification of unmet needs and guided initial research aims. A second, smaller group of three patients treated for cervical cancer met with NM as a group every 6 months and individually to act as advisors throughout and to ensure that aspects of the research were user-friendly. The planning group worked through the six steps of the IM protocol as follows, which are presented in line with relevant previous research [bib_ref] theory and context: using intervention mapping to develop a worksite physical activity..., Mceachan [/bib_ref] :
Step 1: Needs assessment First, the planning group developed the logic model of the problem (i.e. low PA participation in the target group) based on their expertise and their knowledge of the relevant scientific literature. The logic model was then refined after undertaking a needs assessment which consisted of: (1) a rapid review of the relevant literature and (2) semi-structured interviews with women focusing on their experience following treatment for cervical cancer and their PA preferences.
## Rapid review of the literature
A scientific literature review was conducted by NM to identify published studies on cervical cancer survivorship and PA. Given the scarcity of literature in the area representing cervical cancer survivors, research in other gynaecological cancer populations was also considered; however, only studies which recruited gynaecological cancer survivors as their sole population were reported. The review utilised several search engines (Google Scholar, PubMed, Web of Science) to identify (a) PA intervention studies in gynaecological cancer survivors (including some representation of cervical cancer) and PA preferences of gynaecological cancer survivors (b) theoretical underpinnings of relevant PA interventions. The search terms used can be found in Appendix A.
Interviews with women who have been treated for cervical cancer Ethical approval was gained from the Loughborough University Ethics Sub-Committee to conduct interviews with those treated for cervical cancer aged between 18 and 65 years. Participants were purposefully recruited via charities, recruitment posters and social media. Those treated for preinvasive cervical lesions only and those who were unable to partake in PA due to an issue unrelated to their cancer were excluded. Informed consent was obtained prior to participation. Ten participants were interviewed either face to face or over the telephone. A semi-structured interview schedule was followed, which asked questions on what types of movements participants perceived themselves capable of doing since treatment, if, and how much, they were typically active and their perceived challenges and facilitators to taking part in PA (e.g. "If you are someone who would like to be more physically active, what is stopping you now?"). Interviews were digitally recorded and transcribed verbatim manually. Interviews were analysed manually using template analysis, a form of thematic analysis [bib_ref] The utility of template analysis in qualitative psychology research, Brooks [/bib_ref] which includes a large degree of structure in the analysis process by utilising a coding template (Appendix B).
Step 2: Identification of outcomes, performance objectives and change objectives In Step 2, a logic model of change was developed. This involved specifying the desired behavioural and environmental outcomes from the intervention in the target group (e.g. to take part in 120 min of purposeful walking per week). Next, the performance objectives (a description of what is required of the target group to perform the behavioural outcome or how environmental conditions will be modified) for each desired outcome were specified in a list-wise fashion (e.g. developing the intention to walk). This leads to the creation of a behaviour change matrix which specifies change objectives (i.e. what needs to change in order for the performance objectives to be met). Using the previous example, if a performance objective is "to develop the intention to walk", a change objective might be "to know the health benefits of walking". The matrix combines the evidence-and theory-based determinants of the desired outcomes that were identified in the needs assessments. The change objectives allow for the mapping of the performance objectives to practical strategies that encourage behaviour change.
Step 3: Selecting theoretical methods and practical strategies
Step 3 involved choosing theoretical change methods on which to design the intervention and, from this, translating change objectives into practical strategies, which can underpin intervention components. Theoretical methods and practical strategies were chosen after a process of consulting with the relevant literature and findings from interviews in Step 1, engaging with the guidance given by Bartholomew et al.and discussions with the planning group and PPI group.
Step 4: Intervention programme production In Step 4, the intervention materials and protocols were designed which included defining the scope and limitations of the intervention programme production and delivery. Materials were presented to and discussed with the PPI group to ensure that they could be refined appropriately and were user-friendly. The members of the PPI group were asked to complete a short questionnaire (Appendix C) exploring potential facilitators and barriers to intervention participation and what the indicators of a successful intervention might be. Based on their feedback, key changes that were feasible were made to the intervention components and materials.
Step 5: Implementation and adoption plan
Step 5 required the planning group to address how the intervention could be transitioned from its theoretical state to the real world, thus considering how it would be implemented and by whom. In this case, the intervention was to be implemented within a feasibility trial. Therefore, intervention execution was discussed in terms of a short-term adoption plan, as it was likely that the intervention would be refined for future long-term adoption.
Step 6: Evaluation Finally, the planning group developed a plan for evaluating the feasibility and acceptability of the intervention, based on the IM protocol and feasibility evaluation frameworks provided by the NIHR. In turn, this plan will evaluate the results of the IM protocol steps.
# Results
Step 1: Needs assessment
## Rapid review of the literature
The small body of literature surrounding PA in the context of cervical cancer survivorship mainly focuses on whether survivors take part in PA [bib_ref] Functional impairment and physical activity adherence among gynecologic cancer survivors: a population-based..., Nayak [/bib_ref] , self-reported PA levels (e.g. [bib_ref] Quality of life, lifestyle behavior and employment experience: a comparison between young..., Bifulco [/bib_ref] and the relationship between PA and other outcomes such as QOL, fatigue and long-term survival (e.g. [bib_ref] Health-related quality of life and sociodemographic characteristics as prognostic indicators of long-term..., Kim [/bib_ref]. Two studies investigating the effect of a PA intervention in gynaecological cancer survivors, with some representation of cervical cancer, were found and are presented in [fig_ref] Table 1: Physical activity intervention studies in gynaecological cancer, with some representation of cervical... [/fig_ref]. Nine studies were found which provide insight regarding PA preferences of gynaecological cancer survivors (endometrial, uterine, ovarian, cervical), which used either a cross sectional or qualitative study design.
## Pa preferences
The common types of PA performed by gynaecological cancer survivors are walking, gym-based activities and swimming [bib_ref] Physical activity and lower limb lymphedema among uterine cancer survivors, Brown [/bib_ref] [bib_ref] Cancer survivors' exercise barriers, facilitators and preferences in the context of fatigue,..., Blaney [/bib_ref]. Moderate-intensity walking was found to be a preference for those not meeting the current PA guidelines and who reported lower-income levels [bib_ref] Physical activity preferences of ovarian cancer survivors, Stevinson [/bib_ref] [bib_ref] Physical activity correlates, barriers, and preferences for women with gynecological cancer, Farrokhzadi [/bib_ref] [bib_ref] Exercise preferences of endometrial cancer survivors: a population-based study, Karvinen [/bib_ref] [bib_ref] Attitudes, challenges and needs about diet and physical activity in endometrial cancer..., Koutoukidis [/bib_ref]. The evaluation of a PA programme for socio-culturally diverse endometrial cancer survivors found that women particularly enjoy exercise programmes which provide opportunities for social interaction and which lead to physical benefits in terms of pain attenuation. Programmes which gradually increase walking; incorporate PA into the daily routine of participants; provide a variety of exercises; incorporate regular goal setting; deliver regular counselling and provide the opportunity to socialise with similar others are also favourable [bib_ref] Cancer survivors' exercise barriers, facilitators and preferences in the context of fatigue,..., Blaney [/bib_ref] [bib_ref] Attitudes, challenges and needs about diet and physical activity in endometrial cancer..., Koutoukidis [/bib_ref] [bib_ref] A focus group study exploring gynecological cancer survivors' experiences and perceptions of..., Donnelly [/bib_ref] [bib_ref] The physical activity preferences of gynecologic cancer survivors, Tyrrell [/bib_ref].
Interviews with women who have been treated for cervical cancer Participant characteristics, treatment details and details of PA participation are included in [fig_ref] Table 1: Physical activity intervention studies in gynaecological cancer, with some representation of cervical... [/fig_ref]. Participants enjoyed walking alone or with others, in contrast to class-based activities, which they perceived to be challenging to replicate alone.
Challenges and facilitators to PA most commonly spoken about are shown in [fig_ref] Table 2: Participant characteristics, treatment type and physical activity [/fig_ref]. Challenges tended to be unique for each participant. Physical after-effects of treatment act as challenges either physically or by contributing to a psychological factor which hindered participation. For example, neuropathy affected balance, which in turn impacted one's competence and confidence to try certain activities (e.g. cycling). A lack of knowledge of safe activities to undertake and environmental factors (e.g. requiring access to toilet facilities) were also challenges. Regarding facilitators, women were more likely to be active if they were aware of the benefits of being active and could incorporate being active into their daily routine. Enjoyment of PA was important and could be enhanced by setting small goals, tracking self-improvement, a sense of competition and being active with others who have similar experiences.
Step 2: Identification of outcomes, performance objectives and change objectives
The National PA recommendations for adults are at least 150 min of moderate-intensity aerobic PA per week in combination with strength training [bib_ref] UK chief medical officers' physical activity guidelines, Davies [/bib_ref] , whilst PA recommendations for cancer survivors as suggested by the American College of Sports Medicine are at least 30 min of moderateintensity activity, 3 times per week for at least 8-12 weeks [bib_ref] Exercise guidelines for cancer survivors: consensus statement from international multidisciplinary roundtable, Campbell [/bib_ref]. Based on the data collected in Step 1, it was decided that the desired behavioural outcomes of the intervention were the adoption and increase of PA levels among the target group (those previously treated for cervical cancer and currently not taking part in 150 min of moderate-intensity activity per week) with a view of reaching the recommended weekly PA guidelines for aerobic activity. The chosen avenue for PA was walking as it is an activity which is specifically recommended for cancer survivors [bib_ref] Exercise guidelines for cancer survivors: consensus statement from international multidisciplinary roundtable, Campbell [/bib_ref] , with numerous health benefits documented such as reduced fatigue [bib_ref] A randomised controlled trial testing the feasibility and efficacy of a physical..., Donnelly [/bib_ref] and reduced cardio vascular risk [bib_ref] The cardiac rehabilitation model improves fitness, quality of life, and depression in..., Dolan [/bib_ref]. It is an activity which is flexible and overcomes the many PA challenges faced by cancer survivors. For example, it places little pressure on the pelvic muscles which can be weakened as a result of treatment (e.g. [bib_ref] Physical activity and the pelvic floor, Nygaard [/bib_ref]. Additionally, its intensity and duration can be adjusted to suit individual preferences and can promote social interaction by "walking and talking".
Next, performance objectives and the behaviour change matrix were constructed, and the theoretical determinants involved in changing behaviour for each objective and the resulting performance objectives were identified (e.g. self-efficacy, knowledge, intrinsic motivation). Examples of performance objectives, theoretical determinants and change objectives can be found in Appendix D.
Step 3: Selecting theoretical change methods and practical strategies
The social cognitive theory (SCT)was selected as relevant for this intervention as many of the theory's determinants (e.g. self-efficacy, outcome expectations, outcome expectancies, self-efficacy, behavioural capability and observational learning)hold relevance for this intervention and have established theoretical change methods. The health belief model (HBM) and theories of self-regulation also informed intervention development.
During several planning group meetings, theoretical change methods were translated into practical strategies to directly manipulate elements of behaviour and the environment. Strategies were created based on what would be deemed feasible, given the population characteristics and previous successful strategies from relevant literature [bib_ref] RiseTx : testing the feasibility of a web application for reducing sedentary..., Trinh [/bib_ref]. Examples of how theoretical methods were translated into practical strategies for this intervention can be found in Appendix E.
Step 4: Producing an intervention programme plan In Step 4, a structured intervention programme was developed by the planning group and incorporated suggestions from the PPI groups on the content, materials and time scales for the intervention. Scope and limitations of the intervention were identified. These were budget restrictions and the challenges as a result of COVID-19 restrictions on social distancing and household mixing. Whilst group walking was seen as the most effective strategy to encourage PA participation after cervical cancer treatment, it was not possible to make this a condition of the intervention as it needed to be flexible in order to be continued during possible restrictions imposed in the future.
The intervention programme to increase walking following treatment for cervical cancer contains structured components, which also provide flexibility and the opportunity for individualisation. The 12-week intervention will target individual determinants of behaviour change, with a focus on the practical strategies outlined in [fig_ref] Table 3: Challenges and facilitators to PA participation identified in qualitative interviews [/fig_ref]. The strategies are delivered in six related components of the intervention:
Intervention launch and brief education First, all participants will attend a virtual intervention launch session designed to increase knowledge on the benefits of walking for those diagnosed and treated for cervical cancer (education provision), facilitate group discussions on addressing perceived challenges to taking part in PA and consider the benefits of goal setting (and how to set goals) to gradually increase PA.
Physical activity self-monitoring tool Each participant will be provided with a Fitbit activity monitor to support selfmonitoring of their PA and to provide PA prompts and prompts to review goals throughout the intervention. Selfmonitoring, via feedback, is associated with PA behaviour through promoting self-efficacy for PA [bib_ref] Behavior change techniques implemented in electronic lifestyle activity monitors: a systematic content..., Lyons [/bib_ref] and is accepted after gynaecological cancer [bib_ref] Fitbit wear-time and patterns of activity in cancer survivors throughout a physical..., Hardcastle [/bib_ref].
An intervention diary A handwritten diary with daily and weekly inputs will facilitate assessment and evaluation of physical and psychological well-being. The diary is an Health coaching Health coaching will be offered to participants fortnightly via telephone or video call. Health coaching sessions will be delivered by NM who has MSc level training in exercise psychology and informal training supported by all authors who have various health coaching qualifications, of which FM has used in her cancer research. The sessions will follow a schedule, based on the GROW modelto guide individualised barrier identification and problem solving, goal review and goal setting to increase PA.
A messaging platform Group messaging between participants on the Fitbit community or via WhatsApp will act as a platform to encourage peer support and to organise a time and a place for group walks (if possible), whereby participants can state their intention to be active and hold themselves accountable (implementation intentions).
Group walking Providing the opportunity for participants to be active together will create a supportive environment which fosters feelings of relatedness and identification with similar others which have been linked with enhanced PA enjoyment [bib_ref] Peer support for physical activity adoption among breast cancer survivors: do the..., Demello [/bib_ref]. It was chosen to deliver intervention components on virtual and technology-based platforms where possible to allow for individualisation and to enhance adherence. For example, education and health coaching will be delivered on the Microsoft Teams platform. The intervention will be delivered in groups of up to six participants who live in relatively close proximity to each other, to facilitate group walking.
## Step 5: implementation and adoption
The planning group agreed that the intervention should be implemented within a feasibility trial. Discussions around the intervention programme identity with the PPI groups led to the name ACCEPTANCE (Acceptability in Cervical Cancer of an Exercise-based Programme delivered Through An Online Community Environment).
Six months post-cervical cancer treatment was chosen as an appropriate criterion for inclusion in the intervention study. The qualitative findings suggest that it is a time when patients feel ready to resume normal activities. Previous research suggests that gynaecological cancer survivors have a preference for PA after treatment rather than at diagnosis or during treatment [bib_ref] Physical activity preferences of ovarian cancer survivors, Stevinson [/bib_ref] and that between 6 and 9 months post-treatment is a period of heightened motivation for positive change, therefore presenting a critical time for intervention [bib_ref] Cancer survivors' exercise barriers, facilitators and preferences in the context of fatigue,..., Blaney [/bib_ref]. In terms of reach, it was decided that clinical nurse specialists would be engaged and trained to offer the intervention programme to patients who attend clinic. For those who have been discharged, relevant cancer charities would be engaged to promote the intervention programme.
Step 6: Evaluation plan An evaluation plan to test the feasibility of the PA intervention programme was created to gain insight into the acceptability of the intervention components and to determine whether refinement was needed in order for the intervention to be suitable as a full-scale pilot trial. Evaluation time points will be at baseline, 6-week, 12-week and 3-month post-intervention (24 weeks). The process evaluation methods will be a set of 3 online questionnaires designed to obtain feedback on participant experiences of attending the programme launch session, of using the Fitbit and of participating on the messaging platform; follow-up interviews to gain insight into how participants experienced the intervention; device assessed PA data via 8 days of wearing an accelerometer at all 4 time points; and a record of participant feedback throughout the intervention.
# Discussion
The IM protocol provided the planning group with a structured and systematic approach to intervention development. The planning group also benefitted from the feedback given by the PPI group who represent the target population for the proposed intervention. The integration of theory is at the centre of this protocol which is in line with the recommendations offered by the Medical Research Council [bib_ref] Developing and evaluating complex interventions: the new Medical Research Council guidance, Craig [/bib_ref] and scholars specialising in behaviour change [bib_ref] Advancing the science of behaviour change: a plea for scientific reporting, Michie [/bib_ref]. Based on previous literature and findings from the interview study, SCT was chosen as a main theory which incorporates self-efficacy along with input from the HBM and theories of self-regulation. Among cancer survivors, perceived self-efficacy in undertaking PA is positively associated with increased PA behaviour [bib_ref] A systematic review and meta-analysis of social cognitive theorybased physical activity and/or..., Stacey [/bib_ref]. More specifically, SCT has been used to design and develop feasibility and longitudinal PA interventions for endometrial cancer survivors [bib_ref] Acceptability and feasibility of a Fitbit physical activity monitor for endometrial cancer..., Rossi [/bib_ref]. Practical strategies were also theory-driven; for example, technology was chosen as a key strategy to promote self-monitoring behaviour, self -efficacy for PA, implementation intentions and to provide choice and flexibility within the intervention. Technologies, such as the Fitbit, are a widely accepted and integrated facet of society and have been utilised to increase PA behaviour after cancer [bib_ref] Acceptability and feasibility of a Fitbit physical activity monitor for endometrial cancer..., Rossi [/bib_ref]. Cancer survivors have previously accepted technologybased methods for enhancing social support, instruction via coaching and self-regulation behaviours [bib_ref] Attitudes, challenges and needs about diet and physical activity in endometrial cancer..., Koutoukidis [/bib_ref]. Given that exploring cervical cancer survivorship is a novel avenue in the PA research landscape, assembling a PPI group to advise the development of this intervention along with findings from the interview study was highly valuable. A key assumption of this research is that behaviour change techniques need to be specific to those treated for cervical cancer. Regular contact with the PPI group enhanced this focus and ensured that all decisions were made with respect to those who would receive the intervention. Such insights allowed the planning group to build a unique knowledge base regarding the experience of recovery which was essential when choosing theoretical determinants to target behaviour.
The qualitative findings echo previous data suggesting that gynaecological cancer survivors do associate PA with cancerspecific health benefits; however, treatment side effects act as a significant barrier [bib_ref] Attitudes, challenges and needs about diet and physical activity in endometrial cancer..., Koutoukidis [/bib_ref]. In this study, we found that participants enjoyed home-based walking and gym activities and welcomed exercising with other survivors who had similar experiences. Such findings are both supported and contested in the literature with some research suggesting that cancer survivors do not enjoy partaking in lifestyle interventions with other survivors [bib_ref] Physical activity preferences of ovarian cancer survivors, Stevinson [/bib_ref] , hence further highlighting the need for interventions tailored by cancer type, particularly when cervical cancer survivors generally represent a younger demographic compared to more commonly researched populations (e.g. breast cancer survivors). A limitation of the qualitative study was that it lacked diversity in terms of PA participation. All interviewees identified as taking part in some physical activity, which contradicts the current literature [bib_ref] Population estimates of meeting strength training and aerobic guidelines, by gender and..., Ottenbacher [/bib_ref]. However, this may be due to the study explicitly stating its interest in PA during recruitment. The intervention strategies aim to increase PA behaviour in those who are not physically active, and so it would have been meaningful to also speak with women who were not active to gain a greater insight into further possible barriers experienced. Greater diversity in recruitment could have been achieved had it been possible to recruit via hospital clinics.
Whilst the protocol provides a structured and systematic framework to base intervention development, the level of detail required to complete each step leads to an onerous process. Previous researchers have commented on the time-consuming nature of the IM protocol [bib_ref] theory and context: using intervention mapping to develop a worksite physical activity..., Mceachan [/bib_ref]. The protocol may not always be a feasible option when developing clinical interventions as such studies require more time-consuming ethical clearances and funding. It took the planning group 8 months to complete the IM process, and a further 7 months to obtain ethical and regulatory approval. These two processes cannot happen simultaneously because the exact protocol needs to be confirmed before the study can be submitted to the NHS ethics committee.
The protocol was suited to the development of a feasibility trial. However, adoption and implementation of the intervention (Step 5) will have to be revisited should a full-scale trial be implemented as the considerations made here will differ to those made in preparation for a feasibility trial. In terms of feasibility evaluation, it was also necessary to combine another framework in order to develop an appropriate evaluation plan; in this case, the NIHR framework was usedas the IM protocol does not specifically deal with the evaluation of feasibility aspects.
# Conclusion
This paper maps the development of a physical activity intervention programme suitable for those recovering from cervical cancer treatment. The initial protocol steps informed what needed to be targeted by the intervention, whilst the latter steps allowed development of strategies and protocols to achieve this. Although it is an onerous process, intervention mapping is worthwhile if adequate resources are in place.
## Supplementary information
The online version contains supplementary material available at https://doi.org/10.1007/s11764-021-01058-y.
Acknowledgements We would like to thank the members of the Patient and the Public Involvement group for their contributions throughout the intervention development stages.
Availability of data and materials The logic model of the problem and the logic model of change are available on request from the corresponding author. The qualitative transcripts generated and analysed during the current study are not publicly available in order to maintain participant anonymity.
Code availability N/A.
Author contributions All authors contributed to the study conception and design. Data collection and analysis were performed by NM. The first draft of the manuscript was written by NM and FM. All authors commented on the manuscript. All authors read and approved the final manuscript.
Funding Support and funding for this research were provided by the Economic, Social and Research Council (ES/P000711/1), Loughborough University (ES/P000711/1), and the Leicester Hospital's charity, University Hospitals of Leicester (Q843). This research is supported by the National Institute for Health Research (NIHR) Leicester Biomedical Research Centre, which is a partnership between University Hospitals of Leicester NHS Trust, Loughborough University, and the University of Leicester, UK.
# Declarations
Ethics approval This study was performed in line with the principles of the Declaration of Helsinki. This research was approved by the Loughborough University ethics committee (REF, R19-P085).
Consent to participate Informed consent was obtained from all individual participants included in the study.
## Consent for publication
The authors affirm that participants provided informed consent for publication of data generated from participant interviews.
Conflict of interest ELM has received research grants from the Intuitive Surgical and Hope Against Cancer for unrelated work. CLE has received lecture fees for GSK and is on the clinical advisory board for Inivata.
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://creativecommons.org/licenses/by/4.0/.
[table] Table 1: Physical activity intervention studies in gynaecological cancer, with some representation of cervical cancer [/table]
[table] Table 2: Participant characteristics, treatment type and physical activity [/table]
[table] Table 3: Challenges and facilitators to PA participation identified in qualitative interviews [/table]
|
Third Line Eribulin for Triple-negative Metastatic Breast Ductal Carcinoma Resulting in Extended Progression-free Survival of 57 Months
# Introduction
Triple-negative breast cancer (TNBC) is a term that has been applied to breast cancers that lack expression of the estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER2). TNBC tends to behave more aggressively than other types of breast cancer. TNBC accounts for approximately 15% of breast cancers diagnosed worldwide and is more commonly diagnosed in women younger than 40 years. Eribulin is a non-taxane microtubule inhibitor approved for the treatment of metastatic breast cancer after two prior chemotherapeutic regimens. We report a patient with extended progression-free survival (PFS) of more than 57 months with metastatic breast cancer treated with eribulin in the third-line setting.
## Case presentation
A 48-year-old lady was diagnosed with stage IIA (pT2N0M0), high-grade, triple-negative, invasive ductal carcinoma (IDC) of the left breast. No germline breast cancer type 1 (BRCA1) or type 2 (BRCA2) mutation was detected. She underwent neoadjuvant chemotherapy with adriamycin and cyclophosphamide (AC) followed by a negative sentinel lymph node biopsy. At mastectomy, a 2.5 cm tumor, high grade, triplenegative IDC with three additional lymph nodes negative for metastatic carcinoma was noted. She subsequently pursued further chemotherapy and was treated with two more cycles of AC followed by six cycles of cyclophosphamide, methotrexate, and 5-fluorouracil (CMF) and then transferred care to our cancer center. Eight months into a surveillance program, she developed a 2.8 cm right lower lobe (RLL) lung mass with a standard uptake value (SUV) of 27 along with multiple smaller nodules within the lungs noted on positron emission tomography-computed tomography (PET-CT). A core needle biopsy of the RLL lung mass was consistent with metastatic TNBC with sheets of poorly differentiated carcinoma similar in morphology and immunohistochemical studies to the initial diagnosis of breast pathology. Approximately 35 months later, a follow-up PET-CT showed increasing cluster nodularity in RLL measuring 4.5 x 2.6 cm with SUV of 7.9 for which a repeated biopsy showed benign lung parenchyma with non-necrotizing granuloma and no evidence of malignancy. She has maintained no evidence of disease (NED) on scans over 57 months.Currently, she is clinically stable, and her serum tumor markers have plateaued. She has required eribulin 20% dose reduction (1.12 mg/m2 days 1 and 8 every 21 days) on account of neuropathy.
# Discussion
TNBC is a term referred to as breast cancers that lack the expression of ER, PR, and HER2 on immunohistochemical or molecular studies. TNBC accounts for approximately 15% of all breast cancers diagnosed worldwide that accounts for almost 200,000 cases each year. TNBC is more commonly diagnosed in women younger than 40 years of age and appears to be relatively more common among black women compared with white women. The triple-negative clinical phenotype mostly comprises the basal-like molecular subtype, although there is substantial heterogeneity within TNBCs. In one study of utilizing DNA and RNA profiling of TNBCs, four stable subtypes were identified including luminal androgen receptor, mesenchymal, basal-like immunosuppressed, and basal-like immune-activated. These molecular features may have implications for chemotherapy selection in the future but currently, no such guidelines exist.
TNBC tends to behave more aggressively than other types of breast cancer. The risk of distant recurrence and death peaks approximately three years after diagnosis and declines rapidly thereafter. Chemotherapy has been the mainstay of systemic treatment for TNBC, as endocrine therapy and HER2-directed therapies, are ineffective. There are no prospective data that show combination chemotherapy improves overall survival (OS) compared with singleagent sequential cytotoxic chemotherapy. Combination chemotherapy may be appropriate for those with rapidly progressive visceral disease or crisis (symptomatic lymphangitic lung metastases, bone marrow replacement with associated cytopenias, leptomeningeal carcinomatous, significant liver metastases with associated liver dysfunction), in which the high chance of response is thought to outweigh the risk of toxicity, due to concerns about impending organ dysfunction.
Anthracyclines and taxanes are considered the most active chemotherapy agents for metastatic breast cancer, although the increased use of both drugs in adjuvant treatment has prompted the development of other non-cross-reacting agents. Unlike other breast cancer subtypes (i.e., ERpositive, HER2-positive subtypes), there are no approved targeted treatments available for TNBC. Although immunotherapy -in combination with chemotherapy is available for advanced TNBC that expresses PD-L1. For patients without germline BRCA mutation and PD-L1 positive metastatic tumors, the US Food and Drug Administration (FDA) has now approved the use of nab-paclitaxel and atezolizumab. In this randomized trial planned subset analysis of outcomes according to PD-L1-expressing immune effector cells within the tumors, atezolizumab improved both PFS (7.5 versus 5 months), and, importantly, OS (25 versus 15.5 months). Our patient's PD-L1 expressing tumor-infiltrating cells were 0% suggesting no benefit of immunotherapy in combination with chemotherapy if she progresses on current third-line chemotherapy.
In patients with no germline mutation and tumors that do not express PD-L1, preferred regimens are anthracyclines (doxorubicin, liposomal doxorubicin), taxanes (paclitaxel), antimetabolites (capecitabine, gemcitabine), microtubule inhibitors (vinorelbine or eribulin).
Eribulin mesylate (1.4 mg/m2 days 1 and 8 every 21 days) inhibits the polymerization of tubulin and microtubules. In a phase III trial of 762 heavily pretreated patients randomly assigned to treatment with eribulin or other chemotherapy (based on physician's and patient's choice), eribulin improved PFS (3.7 versus 2.2 months) and significantly improved OS (median, 13.1 versus 10.6 months). The objective response rate was 12% with only a 1% complete response (CR) reported in the literature. The second phase III trial was published in 2015 with 1102 patients, who had previously received anthracyclines and taxanes, were randomized in a 1:1 ratio to eribulin or capecitabine. The trial showed eribulin and capecitabine to be equally effective for both PFS and OS. Median PFS was 4.1 months in the eribulin arm in this study. Our patient has shown excellent response and tolerance to eribulin with ongoing PFS of over 57 months (nineteen times the norm) which is rare.
Case reports with eribulin for heavily pre-treated metastatic breast cancer have been published in the literature. In one series, 15% of patients were TNBC (8/53 patients). In this series median time to progression in TNBC patients was 4.7 months and median OS of 7.43 months. To our knowledge, no cases with such lengthy PFS on eribulin chemotherapy, as seen in our patient in a metastatic setting, were reported in the literature. One possible explanation for such a durable response could be low volume disease to start.
# Conclusions
Historically, eribulin was added to the armamentarium of drugs after being studied in patients who progressed on two or more chemotherapy regimens and PFS was approximately 3.7 to 4.1 months in phase III trials. It is currently unclear if a specific subset of TNBC patients gains maximal benefit. Identification of predictive biomarkers and the concurrent development of diagnostics for these biomarkers are needed. Herein, we present a case of a woman with recurrent triple-negative metastatic breast carcinoma with an extraordinarily lengthy PFS on eribulin chemotherapy.
# Additional information disclosures
Human subjects: Consent was obtained by all participants in this study.
## Conflicts of interest:
In compliance with the ICMJE uniform disclosure form, all authors declare the following: Payment/services info: All authors have declared that no financial support was received from any organization for the submitted work. Financial relationships: All authors have declared that they have no financial relationships at present or within the previous three years with any |
Feeding Difficulty in an Infant with Stickler‘s Syndrome
Stickler syndrome or hereditary progressive arthro-ophthalmopathy is an autosomal dominant condition characterized by ocular manifestations, arthritic changes, orofacial features, and deafness, in variable degrees. This is a case report of a 48-day-old infant who presented with severe feeding difficulty. It also details the clinical, diagnostic features, and management of an infant with Stickler syndrome.
# Introduction
Stickler's syndrome also known as hereditary arthro-ophthalmopathy is a dominantly inherited disorder of collagen connective tissues, resulting in an abnormal vitreous of the eye, variable degrees of myopia, the risk of retinal detachment, cataract, and glaucoma. [bib_ref] Clinical features of hereditary progressive arthro-ophthalmopathy (Stickler syndrome): A survey, Stickler [/bib_ref] The disease was first described by Stickler et al. and subsequently reported in two families with Pierre-Robin sequel as a connective tissue disorder that may affect the formation of the eyes, ears, palate mandible, and joints.Stickler's syndrome Affects both males and females. Prevalence rates have been estimated at 1-3/1000 births and at 1/7500 births. Most investigators believe that the disorder is highly under-diagnosed making it difficult to determine the true prevalence of Stickler's syndrome in the general population. [bib_ref] Stickler's syndrome (hereditary progressive arthro-ophthalmopathy), Popkin [/bib_ref] Early recognition of the syndrome is important, not only for genetic counseling but also to offer a more precise prognosis and improved treatment of the many serious disorders that may occur in affected children.Case Report A 48-day-old infant had reported to our tertiary care hospital with the complaint of feeding difficulty. Clinical examination
This is an open access journal, and articles are distributed under the terms of the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 License, which allows others to remix, tweak, and build upon the work non-commercially, as long as appropriate credit is given and the new creations are licensed under the identical terms.
For reprints contact: [email protected] revealed syndromic facies with retrognathia, micrognathia, cleft palate, and increased work on breathing. Intra-oral examination revealed a secondary cleft palate involving the hard and soft palate [ ].
Detailed case history and genetic evaluation were done which suggested features of syndromic phenotype (Stickler's syndrome). Medical records revealed a large patent ductus arteriosus continuous left-to-right shunt and a refractive error on the myopic side in both eyes. Ear, nose, and throat evaluation revealed decreased hearing with respect to the right ear. Flexible endoscopy showed stridor due to tongue fall and retrognathia. However, she was referred to the Department of Pedodontics due to a severe feeding difficulty.
A putty-based impression was used to record the anatomy of the cleft using a special tray customized for this child. The light-body impression recorded the cleft margins distinctly. A dental stone cast was made out of the impression received [ ].
The cleft margins and area were covered by a wax spacer and an acrylic-based feeding plate was then fabricated on the cast. Once the feeding plate was finished and polished, the wax spacer was removed. Two holes were made on either side of the mesiobuccal and distobuccal alveolar ridge, to insert a dental floss [ ]. This was done to prevent the plate from being aspirated. Excess margins on the palatal aspect were reduced till the child could retain the plate in the mouth without triggering a gag reflex. The mother was then asked to feed the child immediately to assess if the child tolerated the feeds well []. As the feeding turned out hassle-free, the mother was instructed to regularly use the plate before every feed. The patient was reviewed periodically every week thereafter for 2 months, and no complaints were reported.
# Discussion
Stickler et al. described an autosomal dominant connective tissue disorder with hereditary progressive arthro-ohpthalmopathy. Features of the Stickler syndrome include premature osteoarthritis; ocular involvement; sensorineural hearing loss; a characteristic facies with maxillary hypoplasia, midface hypoplasia, long philtrum, and micrognathia; and cleft palate (Pierre-Robin sequence). [bib_ref] The stickler syndrome: Evidence for close linkage to the structural gene for..., Francomano [/bib_ref] [bib_ref] Case report of a family affected by stickler syndrome in which rhegmatogenous..., Kimura [/bib_ref] However, most of these features become less distinct as the child reaches adulthood. The facial features are so variable that in isolation they are unreliable for making a diagnosis. The infant, in this particular case, had distinctive facial features such as retrognathia, retroglossoptosis, flattened nasal bridge, hypoplastic maxilla, and shape anomaly of the external ear. A quarter of patients (25% of cases) have some evidence of midline clefting. This can range from the extreme of the Pierre-Robin sequence, through clefting of the hard/soft palate, to the mildest manifestation of the bifid uvula. [bib_ref] Clinical and molecular genetics of stickler syndrome, Snead [/bib_ref] Affected children may present with speech defects and may require speech therapy. The infant presented here with secondary cleft involving the hard and soft palate. Early diagnosis of such infants can only result in a better prognosis.
Of babies born with Pierre-Robin sequence, 30%-44% are subsequently diagnosed with Stickler's syndrome. For the same reason, if a child presents with Pierre-Robin sequence, the ophthalmic investigation should be performed to rule out Stickler's syndrome. [bib_ref] The diagnosis and consequences of stickler syndrome, Webb [/bib_ref] Feeding problems leading to failure to thrive in babies with cleft conditions were recorded as early as the 1600s. It has been accepted that babies with nonsyndromic oral clefts have feeding difficulties related to the structural malformation, which causes abnormal suction and compression during nutritive sucking. Glass and Wolf in 1999 suggested that in cases where the cleft occurs with multiple congenital anomalies, the feeding difficulty is more complex and thought to relate to the neurological status of the baby and/or the presenting medical condition as well as the structural anomaly. [bib_ref] Sucking efficiency of early orthopaedic plate and teats in infants with cleft..., Choi [/bib_ref] For optimal suction, while feeding, a baby must have intact oral structures, especially lip and palatal structures, and functional competence of relevant musculature including the lips, cheeks, tongue, velum, and pharyngeal walls. Given their abnormal oral anatomy, it is not surprising that babies with CL/P are reported to have difficulty creating the oral pressure gradients necessary for the bottle-or breast-feeding. [bib_ref] Sucking efficiency of early orthopaedic plate and teats in infants with cleft..., Choi [/bib_ref] documented that the degree of impairment expected as a consequence of particular cleft malformations remains controversial, but the extent, place, and width of the defect may be important factors in influencing the outcome. [bib_ref] Sucking efficiency of early orthopaedic plate and teats in infants with cleft..., Choi [/bib_ref] The infant presented in this case had secondary cleft palate which failed to create the necessary oral seal to help the child be able to feed, which in turn failed to thrive. Thus, the main objective here was to fabricate a feeding plate, so the infant can have proper nourishment and subsequent weight gain to proceed with other surgeries. [bib_ref] Sucking performance of babies with cleft conditions, Reid [/bib_ref] The management of patients with Stickler's syndrome is more complex than a simple cleft repair or airway Contemporary Clinical Dentistry | Volume 10 | Issue 3 | July-September 2019 management. Successful long-term outcomes depend on individualized, fully integrated, long-term treatment provided in an effective and coordinated manner by a multidisciplinary team of experts in the field from early infancy, through adolescence. [bib_ref] Keys to a successful cleft lip and palate team, Wellens [/bib_ref] Declaration of patient consent
The authors certify that they have obtained all appropriate patient consent forms. In the form the patient(s) has/have given his/her/their consent for his/her/their images and other clinical information to be reported in the journal. The patients understand that their names and initials will not be published and due efforts will be made to conceal their identity, but anonymity cannot be guaranteed.
[fig] Figure 2, Figure 1: (a-d)Step-wise procedure of fabrication of feeding plate (a-d) Facial features in Stickler syndrome with secondary cleft involving hard and soft palate [/fig]
[fig] Figure 3: (a and b) Dental floss knotted for ease of handle b a [/fig]
|
Identifying clusters of leprosy patients in India: A comparison of methods
BackgroundPreventive interventions with post-exposure prophylaxis (PEP) are needed in leprosy highendemic areas to interrupt the transmission of Mycobacterium leprae. Program managers intend to use Geographic Information Systems (GIS) to target preventive interventions considering efficient use of public health resources. Statistical GIS analyses are commonly used to identify clusters of disease without accounting for the local context. Therefore, we propose a contextualized spatial approach that includes expert consultation to identify clusters and compare it with a standard statistical approach. OPEN ACCESS Citation: Taal AT, Garg A, Lisam S, Agarwal A, Barreto JG, van Brakel WH, et al. (2022) Identifying clusters of leprosy patients in India: A comparison of methods. PLoS Negl Trop Dis 16(12): e0010972. https://doi.org/10.Leprosy is chronic infectious disease characterized by skin and peripheral nerve lesions. Despite the efforts to eliminate leprosy, around 210,000 new cases are still found annually of which 60% is reported by India alone. To reduce the incidence significantly, new active case finding approaches in combination with preventive treatment to at-risk populations are needed in high-endemic areas in India. Geospatial methods can support program managers and policy makers to identify clusters of leprosy patients and target at-risk populations for preventive interventions. However, often standard spatial methods do not account sufficiently for the local context (i.e., local barriers and social determinants). In this study, we describe a contextualized spatial approach that includes expert consultation to identify context specific clusters and compared it with a standard approach. Overall, our results show that the contextualized approach is able to identify more clusters precisely and covers a larger proportion of the population in clusters that would need to be targeted for preventive interventions. For program managers and policy makers, the contextualized approach can be useful to target at-risk populations in high-endemic areas while ensuring efficient use of public health resources. Further research is needed to test the scalability in different endemic settings and to apply to other Neglected Tropical Diseases.
# Methodology/principal findings
We included all leprosy patients registered from 2014 to 2020 at the Health Centers in Fatehpur and Chandauli districts, Uttar Pradesh State, India (n = 3,855). Our contextualized spatial approach included expert consultation determining criteria and definition for the identification of clusters using Density Based Spatial Clustering Algorithm with Noise, followed by creating cluster maps considering natural boundaries and the local context. We compared this approach with the commonly used Anselin Local Moran's I statistic to identify high-risk villages. In the contextualized approach, 374 clusters were identified in Chandauli and 512 in Fatehpur. In total, 75% and 57% of all cases were captured by the identified clusters in Chandauli and Fatehpur, respectively. If 100 individuals per case were targeted for PEP, 33% and 11% of the total cluster population would receive PEP, respectively. In the statistical approach, more clusters in Chandauli and fewer clusters in Fatehpur (508 and 193) and lower proportions of cases in clusters (66% and 43%) were identified, and lower proportions of population targeted for PEP was calculated compared to the contextualized approach (11% and 11%).
# Conclusion
A contextualized spatial approach could identify clusters in high-endemic districts more precisely than a standard statistical approach. Therefore, it can be a useful alternative to detect preventive intervention targets in high-endemic areas.
## Plos neglected tropical diseases
# Introduction
Leprosy is caused by an infection with Mycobacterium leprae. It affects the peripheral nerves and skin, and if left untreated, can progress to lifelong disabilities [bib_ref] Leprosy-an overview of clinical features, diagnosis, and treatment, Fischer [/bib_ref] [bib_ref] The clinical and immunological features of leprosy, Walker [/bib_ref]. Therefore, patients affected by leprosy, and especially those who develop disabilities, often suffer from social stigma and discrimination leading to social exclusion, economic loss, and depression [bib_ref] Disability in people affected by leprosy: the role of impairment, activity, social..., Van Brakel [/bib_ref]. The World Health Organization (WHO) still reports 200,000 new cases of leprosy annually, of which 80% are registered in India, Brazil, and Indonesia. India alone reported 114,451 new cases in 2019, which are unevenly distributed within the country. In many districts, high endemic pockets with ongoing transmission exist. Despite the many efforts to eliminate leprosy, new cases are still found in India during past and current leprosy case detection campaigns . Therefore, new active case finding approaches in combination with prophylactic treatment to at-risk populations are needed to reduce the incidence considerably in these high endemic areas in India [bib_ref] Current situation of leprosy in India and its future implications, Rao [/bib_ref].
Progress have been made in the field of prophylactic interventions to prevent leprosy among at risk populations [bib_ref] The state of affairs in post-exposure leprosy prevention: a descriptive meta-analysis on..., Schoenmakers [/bib_ref]. The COLEP study demonstrated that a single dose of rifampicin provided as post-exposure prophylaxis (SDR-PEP) to contacts of leprosy patients reduces the risk of developing leprosy by, on average, 57% [bib_ref] Effectiveness of single dose rifampicin in preventing leprosy in close contacts of..., Moet [/bib_ref]. SDR-PEP has now been taken up in the WHO leprosy guidelines to be used as preventive treatment for contacts of leprosy patients after excluding leprosy and tuberculosis disease. Recently, new studies have started to assess the effectiveness of novel PEP strategies and/or PEP regimens. For example, the PEP+ + trial compares the effectiveness of an enhanced PEP regimen, a combination of two antibiotics, with SDR-PEP administered to close contacts of leprosy patients in high-endemic areas in five countries. In this trial, enhanced PEP is provided in combination with a cluster-based door-to-door campaign. People within identified clusters will be screened for leprosy signs and symptoms and provided with PEP.
Geographic Information System (GIS) or spatial analysis are frequently used to identify clusters of disease and to measure spatial patterns that are not random. Among these clustering tools, many statistical methods have been developed to verify if disease clusters are of sufficient geographic size to have not occurred by chance [bib_ref] A review of spatial methods in epidemiology, Auchincloss [/bib_ref]. Also in leprosy, spatial techniques are perceived as an important tool to identify clusters [bib_ref] The use of GIS in leprosy control, Bakker [/bib_ref] [bib_ref] Geographic Information Systems and applied spatial statistics are efficient tools to study..., Queiroz [/bib_ref]. Global clustering methods such as the Global Moran's I statistic are used to indicate the occurrence of spatial clustering in the complete area and are often followed by local clustering methods [bib_ref] Geographic Information Systems and applied spatial statistics are efficient tools to study..., Queiroz [/bib_ref] [bib_ref] Epidemiological, temporal and spatial dynamics of leprosy in a municipality in northeastern..., Albuquerque [/bib_ref]. Local clustering methods such as Local Moran's I statistic (Local Index of Spatial Association [LISA]) and Getis Ord Gi � statistic are commonly used to identify areas of significant high leprosy incidence and to find correlations with high grade 2 disability 2 (G2D) rates and low socioeconomic status [bib_ref] Epidemiological, temporal and spatial dynamics of leprosy in a municipality in northeastern..., Albuquerque [/bib_ref] [bib_ref] Spatial distribution of leprosy in Nigeria, Daniel [/bib_ref] [bib_ref] Leprosy in urban space, areas of risk for disability and worsening of..., De Assis [/bib_ref]. With Kulldorff's Spatial and Space Time Scan Statistics (SaTScan) [bib_ref] A spatial scan statistic, Kulldorff [/bib_ref] , a space-time local clustering technique that can adjust for heterogeneous background population densities and confounding variables [bib_ref] A space-time permutation scan statistic for disease outbreak detection, Kulldorff [/bib_ref] , the most likely cluster(s) of disease are identified controlled for gender, age, type of leprosy and deformity, and over a period of time [bib_ref] Geospatial epidemiology of leprosy in northwest Bangladesh: a 20-year retrospective observational study, Bulstra [/bib_ref] [bib_ref] Trends and spatial clustering of leprosy cases over a decade in a..., Kuruwa [/bib_ref]. Non-statistical spatial methods such as the Kernel Density Estimation which estimates the probability density function of a random variable, are applied to visualize clusters or so-called hotspots of disease and disability [bib_ref] Geographic Information Systems and applied spatial statistics are efficient tools to study..., Queiroz [/bib_ref] [bib_ref] Misdiagnosis of leprosy in Brazil in the period 2003-2017: spatial pattern and..., Neves [/bib_ref] [bib_ref] Determining target populations for leprosy prophylactic interventions: a hotspot analysis in Indonesia, Taal [/bib_ref].
Statistical spatial methods are mainly used to test a hypothesis and identify clusters in leprosy research context. These methods can control for possible confounding and adjust for population densities, but do not account for the local barriers or social determinants. Involvement of local communities and stakeholders for identifying high risk areas (intervention target) is beneficial as they provide valuable information on the local context such as the natural boundaries, presence of stigma or travel and social behavior of the population in a specific area. Program managers and policy makers intend to use spatial tools to identify the at-risk populations for preventive interventions while considering efficient use of public health resources [bib_ref] The use of GIS in leprosy control, Bakker [/bib_ref]. They often rely on simple maps that present the leprosy incidence, grade 2 disability or child rate per district only. Therefore, to guide program managers and policy makers for identification of at-risk clusters, a contextualized spatial approach that allows them to identify clusters while considering the local context would be valuable.
In the PEP++ trial, a contextualized spatial approach was developed to identify high-risk areas for door-to-door PEP interventions, and not necessarily to find statistical clusters. This approach includes GIS spatial methods and expert consultation to account for the local context. Non-statistical methods were required to identify clustering at a local level using individual level data (i.e., location of the house of the leprosy patient) and can include considering of local barriers. This study describes the contextualized spatial approach that has been used in the PEP++ trial to identify clusters of leprosy patients targeted for active case finding and preventive intervention and compares it with a standard statistical spatial analysis method.
# Methods
# Ethics statement
The PEP++ trial received ethical clearance from the Vardhman Mahavir Medical College & Safdarjung Hospital (IEC/VMMC/SJH/Project/2019-09/40) and is approved by the Indian Council of Medical Research (ICMR) in New Delhi, India on September 17, 2018. For this study, ethical clearance has been obtained for collecting patient information from the national register and the GPS coordinates of patients registered from 2014 to 2020 at the public health centers in Chandauli and Fatehpur. A copy of the dataset has been anonymized to perform the analysis.
## Study area
This study used spatial and demographic data of registered leprosy patients in Chandauli and Fatehpur districts, Uttar Pradesh (UP) State, India. UP is one of the most endemic states of India with an annual new case detection rate (ANCDR) of 65.8 per 1,000,000 population in 2019. Chandauli district is in southeast of UP and borders Bihar State. It has an area of 2,484 km 2 and population of 2.34 million in 2020 [ in S1 Text]. Fatehpur district is located in central south of UP and has an area of 4,152 km 2 and population of around 3.10 million in 2020 [ in S1 Text]. Both districts are considered high endemic for leprosy with an ANCDR of 54.6 per 1,000,000 population in Chandauli and of 76.6 per 1,000,000 population in Fatehpur in 2021 [ in S1 Text]. [fig_ref] Fig 1: Overview of the study area in Uttar Pradesh, India, which covers Fatehpur... [/fig_ref] the geographic location of the two districts.
## Data collection
We collected the medical records of all leprosy index cases registered from April 2014 to March 2020 across 10 Primary Health Centers (PHCs) and 17 PHCs in Chandauli and Fatehpur, respectively. Between January 2018 and June 2020, we conducted five mapping surveys to collect the spatial data based on information on patient' names and addresses with village name. Project staff were trained in data collection using the mobile application MapIt software (version 7.6.0, https://mapitgis.com/), a tool for Geographic Positioning Systems (GPS) data collection and management, in January 2018. They were assigned to one of the blocks and worked together with Indian government staff, such as Medical Officers (MOs), Non-Medical Assistants (NMAs), Paramedical workers (PMWs) of the PHC and Accredited Social Health Activists (ASHAs) as community volunteers. Since we only had the village and patient name, the support of PHC staff and ASHAs including the chief of village (i.e., pradhans) was required to find the patients' houses. Once a patient's house was located, the GPS coordinates of the house were recorded offline with MapIt. All data points were uploaded to the server in the NLR India office in New Delhi to be aggregated and combined with demographic information for analysis. in S1 Text shows an overview of the five mapping surveys. All data were imported in the open-source Quantum Geographic Information System (QGIS) version 3.4.1 (QGIS Developer team, Madeira 2018)) for validation. Incorrect data points (e.g., situated in a lake or field) were removed and re-collected by project staff.
## Spatial analysis: contextualized spatial approach
We developed a contextualized spatial approach for the PEP++ trial to identify clusters of leprosy cases that account for the local barriers and social determinants. It comprises of the nonstatistical Density-Based Spatial Clustering of Applications with Noise (DBSCAN) tool and an expert consultation to identify clusters of leprosy cases (high-transmission areas) at individual level. The approach was divided into three steps: a preliminary spatial analysis, an expert consultation to decide on criteria and a cluster definition, and the development of context specific cluster maps that can be used in the field. [fig_ref] Fig 2: Overview of the different steps in the contextualized spatial approach with the... [/fig_ref] an overview of the steps, and S1 Text describes the complete methodology used in this approach.
Step 1 Preliminary spatial analysis. In the preliminary analysis, we analyzed the data on strength of clustering, then visualized clusters, and finally identified cases that were part of each main cluster. The strength of clustering of leprosy patients was calculated using Global Moran's Index statistic [bib_ref] Notes on continuous stochastic phenomena, Moran [/bib_ref] in GeoDa (https://geodacenter.github.io/) version 1.18 (Anselin, Santa Barbara, CA, USA). It calculates the autocorrelation coefficient (degree of similarity) between spatial points (and takes values ranging between -1 and 1), where -1 indicates dispersed distribution, 0 no clustering, and 1 strong clustering. In this approach, the Global Moran's I was used to determine reasonable cluster sizes. Strong clustering indicates that spatial points are located close together and therefore a cluster can contain many patients. Weak clustering indicates that only a few spatial points are close together and therefore a cluster should contain a few patients.
Clustering of cases was visualized using the heatmap tool (Heatmap plugin, QGIS). The heatmap tool draws a circle with a specified radius (e.g., 1000 m) around each data point and creates a raster file as output. Then, it uses Kernel Density Estimation to calculate the density of points for each raster cell. Raster cells that are close to data points will have a higher density (higher value) compared to raster cells that are further away which will have a low density (low value). The resulting map showed the density distribution of clusters and non-clusters. In this approach, the heatmap radius was used to select the maximum distance between patients that would be part of a cluster. A heatmap that shows many individual density spots can be selected as the maximum distance.
To identify cases that are part of a cluster, we used the DBSCAN tool [bib_ref] A density-based clustering methods for discovering clusters in large spatial databases with..., Ester [/bib_ref] of QGIS. This tool focuses on the proximity and density of points to form (arbitrary shaped) clusters. A minimum cluster size (minimum points; MinPts) and maximum distance to a nearest neighbor (radius; ε) needs to be selected beforehand. In the DBSCAN tool, points can be classified as core points, reachable points and outliers. A point is identified as a core point if it has the specified minimum number of points in its radius-neighborhood (including the core point). A reachable point is a point within the radius-neighborhood of the core point and is identified as part of the cluster. All points outside the radius-neighborhood are 'outliers' or 'noise points' and will not be identified as cluster points. The parameters of the DBSCAN were determined by the strength of clustering (Global Moran's I) and density of points (Heatmap). We used the DBSCAN tool with 12 different combinations of minimum cluster size (i.e., 2, 3, 4) and maximum distance (i.e., 200 m, 300 m, 400 m, and 500 m). For each combination, we calculated the proportion of the total cases georeferenced that would be part of a cluster.
Step 2 Expert consultation. An expert consultation was organized with the aim to i) discuss the preliminary results of the clustering analysis, ii) decide which cluster definition to use in the concerned setting and iii) discuss the implications of this choice for the door-to-door campaigns. Among the participants were representatives of the state and district government health department, health care workers of both districts, and project staff that collected the data (n = 60). Their knowledge and experience of the Chandauli and Fatehpur context was important to determine the definition of a cluster.
To get a better understanding of the Chandauli and Fatehpur context, additional information about i) transmission of M. leprae in the two districts, including MB leprosy, child, and grade 2 disability proportions, ii) the travel and social behavior of leprosy infected persons, and iii) specifications of the environment (i.e., districts), including rural or urban areas, natural and social boundaries, population counts, poverty, presence of violence or stigma and access to health centers, were gathered and discussed. The experts recommended using: a small cluster size (Minpt) and small maximum distance (ε, radius) for the cluster analysis and that 'at least 60% of the leprosy cases should be part of a detected cluster' (S1 Text). Based on this input the appropriate DBSCAN parameter combination was selected, which included a minimum cluster size of 2 and maximum distance of 500 m (i.e., at least 2 cases living within 500 m of each other; in S1 Text).
Step 3 Develop context specific cluster maps. The data points that are within 500 m of another point were identified as a cluster case and used for further analysis. Cases that are part of the same cluster were grouped and we drew a polygon shape around them using Google Satellite Imagery. While shaping the areas, we considered the village, street and natural boundaries, and capturing on average 20 households (i.e., 100 individuals) around each cluster case as was decided during the expert consultation.
## Spatial analysis: statistical approach
Statistical tools are often used by academics to identify significant clusters with risk ratios. We used the Local Moran's Index (Local Indicator of Spatial Association [LISA]) [bib_ref] Local Indicators of Spatial Association-LISA, Anselin [/bib_ref] because it is a commonly used statistical method to test for local clustering and can be used to identify the highest risk areas of occurrence of leprosy in Fatehpur and Chandauli district. As the areas of interest, we selected village level since this is the smallest area with known population data.
Similar as the Global Moran's I, the Local Moran's Index measures the correlation coefficient (degree of similarity) between neighbouring spatial points. The difference is that the Global Moran's I calculates one summarizing I value for the complete study area and the Local Moran's calculates the I value for each spatial point (local level). In this study, we calculated the correlation coefficient between villages with or without leprosy cases to identify clustering of villages. The algorithm removes the village from its neighborhood and determines if the neighborhood is significantly different from the study area, then it determines if each village is significantly different from its neighborhood. The tool calculates the Moran's Index value and both a z-score and p-value to evaluate the significance. A Moran's I value of zero indicates homogenous distribution (no clustering). A value between 0 and 1 indicates that a village has neighboring villages with similarly high or low number of leprosy cases; this village is part of a cluster. A value between -1 and 0 indicates that a village has neighboring villages with dissimilar number of cases in the village; this village is an outlier.
We checked the data for outliers using the histogram function in GeoDa version 1.18. Since the data showed a non-normal distribution with extreme high values, we decided to use the 'Univariate Median Local Moran's I' tool in GeoDa to identify the high-risk cluster villages. This is an extension of the Univariate Local Moran's I tool that can correct for the variance caused by extreme high or low values. We appointed the total number of cases per village as 'event' variable. The queen contiguity with first order was selected to determine the neighbors, which includes all neighboring villages that directly share a spatial border with the village of interest. To construct the reference distribution, we selected the Monte Carlo replication of data sets of 999 permutations (i.e., pseudo p value of 0.001) to ensure adequate power for defining clusters. As the output, GeoDa provides a significance map and a cluster map. The significance map shows the villages with a significant local I statistic (p value � 0.05). The thematic cluster map shows the clusters high-high (i.e., village with high number of cases surrounded by villages with high number of cases) and low-low (i.e., village with low number of cases surrounded by villages with low number of cases), the outliers high-low (i.e., village with high number of cases surrounded by villages with low number of cases) and low-high (i.e., village with low number of cases surrounded by villages with high number of cases), and the villages that are not significant. In this study, the high-high and high-low villages (villages with leprosy cases and a p value lower than 0.05) were considered as leprosy cluster.
## Outcome measures
To compare the contextualized spatial and statistical approach, the proportion of cases in clusters, the number of clusters, the cluster area, the total number of people in clusters, and the proportion of contextualized clusters that overlap with statistical clusters are calculated for urban and rural areas in both districts and for each approach. In the contextualized spatial approach, the proportion of cases in clusters was calculated as the number of leprosy cases in clusters divided by the total number of leprosy cases. The total population in clusters is the sum of populations as provided by the pradhans of the villages with leprosy clusters in 2021. In the statistical approach, the proportion of cases in clusters was calculated as the number of leprosy cases in high-high and high-low villages divided by the total number of leprosy cases. The total population in clusters is the sum of the populations in high-high and high-low villages. The proportion of cluster area that overlap will be calculated as the total area overlap divided by the total cluster area of the contextualized spatial approach.
The implications for the door-to-door active case finding and PEP interventions in the clusters were also compared for the two approaches. The estimated number of people to be targeted with PEP was calculated for both approaches as the number of cases in clusters multiplied by 100 individuals. Following the PEP++ strategy, we assumed that 20 households which is about 100 individuals were targeted per leprosy case in a detected cluster. The proportion of total population in the clusters to be targeted for PEP was also calculated for both approaches. The results and the implications for PEP interventions of both approaches for Chandauli and Fatehpur are compared and presented in [fig_ref] Table 1: The results of two approaches compared [/fig_ref]. Overall, the indicators were lower for urban compared to rural clusters and villages. The proportion of cases in clusters for urban and rural areas together is 75.0% and 57.2% in Chandauli and Fatehpur, respectively, in the contextualized spatial approach, and 66.2% and 43.0% in the statistical approach [fig_ref] Table 1: The results of two approaches compared [/fig_ref]. The proportion of total population in clusters that would be targeted for PEP (20 households per leprosy cluster case) is 31.3% in Chandauli and 11.4% in Fatehpur in the contextualized spatial approach, and 10.8% and 10.8%, respectively in the statistical approach [fig_ref] Table 2: The implications for PEP interventions of two approaches compared [/fig_ref]. The proportion of contextualized clusters that overlap with statistical clusters is for Chandauli 41.3% and 62.0% in urban and rural areas, respectively, and for Fatehpur, this is 60.6% and 39.8% [fig_ref] Table 1: The results of two approaches compared [/fig_ref].
# Results
## Distribution of cases
## Comparison of the contextualized spatial approach with standard statistical approach
## Plos neglected tropical diseases
Identifying clusters of leprosy patients in India: A comparison of methods
# Discussion
Our study described a contextualized spatial approach with expert consultation to identify contextualized clusters of leprosy cases to target door-to-door PEP interventions in two highendemic districts in India and compared this approach with a standard statistical approach. Both approaches could identify clusters of cases in the two high-endemic districts. The contextualized spatial approach, however, identified smaller clusters and a higher proportion of cases in clusters in both districts. This may imply that the contextualized approach identified clusters more precisely in high-endemic districts. Also, the total population in clusters targeted for PEP were smaller in urban settings of both districts in the contextualized approach compared to the statistical approach.
As program managers often operate with limited budgets and resources, it is necessary to target PEP interventions only to those that are most likely to be infected with M. leprae and are at risk of developing leprosy. Our findings showed that the contextualized spatial approach covers a higher proportion of cases in clusters of rural and urban areas combined and therefore, results in larger number of individuals targeted for PEP intervention. A contextualized spatial approach is very suitable if one aims to target the complete cluster or target the at-risk population in smaller but more precise clusters that account for the local context. Especially with door-to-door campaigns in urban areas, targeting smaller and more precise clusters within a city can be more efficient in terms of resources and time compared to large urban � Chandauli has 20 urban areas (total population of 264,000) and 1405 rural areas (total population of 1,939,000), Fatehpur has 56 urban areas (total population of 356,4,000) and 1296 rural areas (total population of 2,641,000)a Number of cluster cases divided by total number of cases for the contextualized spatial approach and the number of cases in red and pink villages divided by the total number of cases for the statistical approach b Calculated as the sum of populations in each cluster as estimated by the head of village in the contextualized spatial approach and as the sum of populations in cluster villages for the statistical approachc Calculated as the total cluster area that overlap with each other divided by the total cluster area of the contextualized spatial approach.
https://doi.org/10.1371/journal.pntd.0010972.t001
## Plos neglected tropical diseases
Identifying clusters of leprosy patients in India: A comparison of methods clusters as identified in the statistical approach. The statistical spatial approach should be considered if the local context is not considered relevant and if targeting larger clusters is acceptable, or if the interest is to target only the clusters that are identified as statistical significant (i.e., areas where cases are not appearing randomly). Generally, this approach would result in fewer and larger urban clusters. Moreover, the proportion of total cases in clusters (urban and rural) is lower resulting in a lower number of individuals targeted for the PEP intervention. This implies that a (significant) part of the at-risk population may be missed for preventive interventions.
Additional considerations for choosing a suitable and feasible spatial approach for a particular district could be: i) population density, ii) granularity of geographic data, and iii) time and resources available for the data collection and analysis. Our contextualized spatial approach can be applied in areas with both low and high population densities, where preferably individual level data is available. However, in our study this approach required at least four weeks to collect the data, including, GPS coordinates of patients and information about local barriers, epidemiological indicators, social determinants, presence of stigma, migration and an estimation of the population, and another two weeks for the data manager to construct the context specific maps. Door-to-door campaigns (with PEP) would therefore require additional time and resources for planning but may save resources (precise clusters and less travelling) for the actual implementation. Nevertheless, this approach would only be feasible when enough time and resources are available to carry this out. Otherwise, we recommend using a standard statistical approach. A statistical approach is also recommended if the population is distributed heterogeneously and if data is only available at village or health center level.
The expert consultation in our study was the added value of the contextualized approach. Expert consultations or key stakeholder meetings are often used in leprosy to gain support for new ideas or strategies. For example, at global level to develop leprosy elimination strategies or guidelines for treatment, at national level to improve leprosy control activities [bib_ref] Stakeholders perspectives on perceived needs and priorities for leprosy control and care, Jaeggi [/bib_ref] , or at research level to develop innovative tools or a research agenda [bib_ref] An enhanced regimen as post-exposure chemoprophylaxis for leprosy: PEP++, Mieras [/bib_ref] [bib_ref] A comprehensive research agenda for zero leprosy, Steinmann [/bib_ref]. The success of our expert consultation and development of context specific cluster maps depended on relevant questions related to the i) transmission of M. leprae in the two districts, including MB leprosy, child, and grade 2 disability proportions, ii) the travel and social behavior of leprosy infected persons, and iii) specifications of the environment (i.e., districts), including rural or urban areas, natural and social boundaries, population counts, poverty, presence of violence or stigma and access to health centers. This resulted in an understanding of the clustering of leprosy cases, a cluster definition and methodology based on the travel behavior of population in the concerned setting, context specific cluster maps presenting more precise target areas for PEP, and a PEP door-to-door strategy considering available resources in the two districts.
The transmission of M. leprae and clustering of cases can be different in rural and urban areas as a result of migration to cities, higher awareness among urban population, and closer distance to an infected person [bib_ref] The spatial distribution of leprosy cases during 15 years of a leprosy..., Fischer [/bib_ref] [bib_ref] Social distance and spatial distance are not the same, observations on the..., Hoeven [/bib_ref]. found an increase in number of child cases in urban areas compared to rural areas indicating that urban areas become an important source of transmission and therefore, may need a different clustering approach and possible PEP strategy [bib_ref] Differential trend of leprosy in rural and urban area of Western Maharashtra, Mohite [/bib_ref]. Currently, urban clusters are mainly identified using statistical tests, including Kulldorff's spatial scan statistic [bib_ref] Socio-economic and environmental effects influencing the development of leprosy in Bahia, north-eastern..., Cabral-Miranda [/bib_ref] [bib_ref] Identification of urban leprosy clusters, Paschoal [/bib_ref]. In our study, we did not use a different spatial approach for urban or rural areas. However, we observed that the contextualized spatial approach is able to identify more precise and smaller clusters in urban areas than the statistical spatial approach, resulting in a higher proportion of total population in clusters targeted for PEP. An example is Fatehpur city [fig_ref] Fig 3: shows the contextualized clusters identified by the contextualized spatial approach [/fig_ref]. The city is identified as a cluster village by the statistical spatial approach and therefore, for each leprosy case in the city 100 individuals will be targeted for door-to-door PEP interventions. In the contextualized spatial approach, we could identify small individual clusters that are mainly located in the center of Fatehpur. This implies that more transmission is taking place in the center than at the outskirts of the city. As a result, the leprosy cases that are not within a distance of 500 m from another case would not be targeted for door-to-door PEP interventions and therefore, the proportion of total population in clusters targeted for PEP is higher compared to the statistical approach. Considering more efficient use of public health resources during active case finding and PEP intervention and limiting exposure to stigma, our contextualized spatial approach could be very useful to identify more precise clusters, especially in urban areas.
The maximum proportion of cluster overlap was 60% indicating that the two approaches can identify the same clusters but also have 40% of the clusters not in common. In clusters identified by the statistical approach only, leprosy cases were located more than 500 meters from each other, and therefore not identified as a cluster in the contextualized approach. In clusters identified by our contextualized approach only, the number of leprosy cases in a village was considerably lower (or equally low) compared to neighboring villages, and therefore not identified as a cluster in the statistical approach. To increase the accuracy of identifying clusters, especially in the urban areas, the program manager can consider using a different DBSCAN setting for urban areas or a different statistical approach that corrects for population densities. As a result, the proportion of cases in clusters may increase for both approaches. The drawback of increasing the proportion of cases in clusters is the more individuals need to be targeted for PEP and therefore more resources are needed.
# Strengths and limitations
Our contextualized spatial approach is a first step towards a feasible spatial approach that accounts for the local context. This approach has worked well in our study setting, and could therefore, be considered a useful method for program managers and policy makers if time and resources can be made available for collecting coordinates at individual level and contextual information. In addition, the output of the expert consultation would likely be relevant for the coming years as the local barriers don't change much over time. The return of investment in time and resources can be significant because changes in, for example, the distribution of new leprosy cases in the studied areas, would require minimal updates to this approach. Further research in different settings is recommended to i) test and, if necessary, adapt the approach, ii) scale up to high-endemic leprosy districts, iii) apply to other neglected tropical diseases such as Buruli Ulcer, and iv) measure the cost-effectiveness of both approaches.
The DBSCAN clustering tool used in the contextualized spatial approach differs from other clustering algorithms by i) not using a pre-set number of clusters such as in the K-Means tool, but instead using a selected Minpt and maximum distance [bib_ref] A density-based clustering methods for discovering clusters in large spatial databases with..., Ester [/bib_ref] , and ii) not correcting for population densities. Selecting a maximum distance demands a proper understanding of the leprosy transmission and the population at risk in the study area. This was discussed in the expert consultation and a decision was made on the maximum distance for Chandauli and Fatehpur. A limitation in our study, however, is that we didn't differentiate between rural and urban settings or correct for population densities and used the same DBSCAN parameters (Minpt of 2 and maximum distance of 500 m). For urban settings with a higher population density, program managers can consider selecting a higher Minpt and smaller maximum distance for the DBSCAN.
In the statistical spatial approach, we used the total number of cases as the event while other leprosy studies selected new case detection rate (NCDR) as event to identify clusters [bib_ref] Spatial distribution of leprosy in Nigeria, Daniel [/bib_ref]. As a result, a village with a high number of cases but low NCDR due to large population can be identified as a cluster, while a village with a low number of cases but high NCDR due to a small population can be missed as a cluster. If we used the NCDR as the event, however, we would have included many rural villages with one leprosy case in both Chandauli and Fatehpur. Even though the NCDR is often used as measure of risk, a single leprosy case in a rural village may also be a sporadic case.
It should be noted that the use of spatial analysis to guide intervention targets and efficient use of public health resources is challenging, especially in high endemic districts. Clustering of leprosy cases can only be measured if the locations of registered cases are known. Cases that are not in the national health registers or from which the GPS coordinates are not collected are automatically excluded from the analysis. Such 'hidden cases' may be due to poor performance or absence of leprosy services at the health center level. These areas where hidden cases may exist will not be identified as high risk areas in the clustering analysis and not be targeted with (contact) screening interventions. While in these areas, public health resources may be used in an efficient way: detecting as many new cases of leprosy as possible while screening few individuals as possible. If a statistical approach is used, these areas will be identified as non-significant and will not be of interest for a program manager to target interventions resulting in continuous transmission of M. leprae. If a contextualized spatial approach is used, however, these results and areas will be discussed during the expert consultation and can be explained according to the local context or history. If it is expected that hidden cases may exist in these areas, the program manager can decide to implement or increase active case finding efforts. This can be followed by door-to-door PEP interventions or other PEP strategies to target the at risk populations and interrupt the transmission of M. leprae.
# Conclusion
This study showed that our contextualized spatial approach could identify clusters in highendemic districts more precisely than a statistical spatial approach. It captures a larger proportion of cases in the identified clusters and covers a larger proportion of the population in the clusters (at risk) that would need to be targeted for a door-to-door approach with active case finding and PEP to support interruption of transmission of M. leprae. Therefore, this approach appears to be a useful approach to detect at-risk populations for preventive interventions in high-endemic areas. In other endemic settings, further research is needed to test the scalability.
Supporting information S1 Text. Text document containing 1) the contextualized spatial approach, 2) number of new patients registered in the five mapping surveys conducted from January 2018 until June 2020 in Chandauli and Fatehpur , 3) block-wise epidemiological status of leprosy during the year 2020 to 2021 reported by the District Leprosy Office in Chandauli and Fatehpur , 4) the new case detection rate per 1,000,000 population per village for Chandauli and Fatehpur , 5) heatmaps of Chandauli and Fatehpur district using a 100 m, 500 m, 1000 m, 2000 m and 3000 m radius , and 6) DBSCAN clustering results of preliminary analysis (2014-2017) . supporting the PEP++ team with data collection. We also like to acknowledge the support of Dr. Anil Kumar, DDG (L) and study PI (program) and Dr. Jugal Kishore, PI and Dr. M.A. Arif, former Country Director, NLR India, and also former Co-PI for their overall contributions, and guidance in conceptualization and preparatory phases of this PEP++ trial. Lastly, we would like to express our gratitude to the participants in the expert meeting for providing their invaluable input in this study.
# Author contributions
Conceptualization: Anneke T. Taal, Suchitra Lisam, Josafá G. Barreto, Wim H. van Brakel, Jan Hendrik Richardus, David J. Blok.
[fig] Fig 1: Overview of the study area in Uttar Pradesh, India, which covers Fatehpur and Chandauli district. The red dots represent the distribution of the mapped leprosy index cases registered from 2014 to 2020. Base layers from https://sedac.ciesin.columbia.edu/data/set/india-india-village-level-geospatial-socioecon-1991-2001/data-download. https://doi.org/10.1371/journal.pntd.0010972.g001 [/fig]
[fig] Fig 2: Overview of the different steps in the contextualized spatial approach with the input and output by step.Step 1 is the preliminary analysis using the Global Moran's I, Heatmap and the DBSCAN tool with the data of 2014-2017. The output of the Global Moran's I and Heatmap are used for the selection of parameters of the DBSCAN (arrows).Step 2 is the expert consultation during which the preliminary results are discussed followed by decisions that have been taken as criteria on cluster definition. Step 3 is the development of context specific cluster maps for the PEP interventions. Minpt: minimum number of points in a cluster; PEP: post-exposure prophylaxis; HH: household.https://doi.org/10.1371/journal.pntd.0010972.g002 [/fig]
[fig] From: April 2014 to March 2020, in total, 4,039 new leprosy cases were registered at the PHCs, 1,710 in Chandauli district and 2,329 in Fatehpur district. The locations of 1,647 (96.3%) and 2,208 (94.8%) leprosy patient' houses were georeferenced during the five mapping studies in Chandauli and Fatehpur, respectively. The New Case Detection Rate (NCDR) per village was calculated as the total number of leprosy cases registered from April 2014 to March 2020 in the village divided by the population of the village in 2021. Fig A in S1 Text shows the NCDR per 1,000,000 population for Chandauli and Fatehpur. A Moran's I statistic of 0.030 (z score of 2.078; p value of 0.031) and 0.028 (z score of 2.624; p value of 0.018) in Chandauli and Fatehpur respectively, indicates weak clustering of cases. The heatmaps of the two districts showed many medium density areas and a few hotspots of leprosy cases (Fig B in S1 Text). [/fig]
[fig] Fig 3: shows the contextualized clusters identified by the contextualized spatial approach (dark red) and the cluster villages identified by the statistical approach (light pink). In the contextualized spatial approach, 374 clusters are identified in Chandauli and 512 cluster in Fatehpur. These clusters are distributed throughout the districts. In the statistical approach, 508 cluster villages are identified in Chandauli and 193 cluster villages in Fatehpur. The cluster villages in Chandauli are distributed throughout the district and the cluster villages in Fatehpur, are identified in the center and North part of the district. [/fig]
[table] Table 1: The results of two approaches compared. [/table]
[table] Table 2: The implications for PEP interventions of two approaches compared. [/table]
|
The Spatial Differentiation and Driving Forces of Ecological Welfare Performance in the Yangtze River Economic Belt
Ecological welfare performance contributes directly to human well-being and regional sustainable development. Improving the regional ecological welfare performance in the process of pursuing green and sustainable development demands theoretical innovation and empirical exploration. Based on the super-efficiency SBM model, this study evaluated the ecological welfare performance of 108 cities during the period of 2009 to 2019. The Dagum Gini coefficient decomposition and spatial convergence model were employed to analyze the differences in ecological welfare performance across and within the study area and explore the underlining causes of such spatial differentiation in the Yangtze River Economic Belt and the upper, middle and lower reaches. It can be seen from the results that: (1) the overall difference in the ecological welfare performance of the Yangtze River Economic Belt is associated with a fluctuating downward trend during the study period. Regional and inter-regional differences were revealed and hypervariable density was identified as the main source of the differences. (2) The ecological welfare performance of the Yangtze River Economic Belt has absolute and conditional β convergence, and the ecological welfare performance of each city-region and surrounding urban areas has a positive impact on each other.(3) The difference in the spatial-temporal differentiation trend is manifested by the difference in the convergence rate. The cities in the middle reaches of the Yangtze River have the fastest convergence rate, followed by the cities in the upper reaches, and the cities in the lower reaches are the slowest. This geographic difference is mainly driven by the combined effects of industrial structure, urban characteristics, environmental regulation, foreign direct investment, and transportation accessibility. Finally, it is proposed that future policies should focus on the imbalanced regional development in the study area, and each region needs to explore ways to improve local ecological welfare performance according to local conditions, and ultimately promote the overall green, coordinated and high-quality development in the Yangtze River Economic Belt.
# Introduction
Coupled with the acceleration of urbanization and industrialization, the rapid economic and social development has also brought about a series of ecological and environmental problems in China, such as environmental pollution and resource shortages [bib_ref] Have Chinese cities achieved the win-win between environmental protection and economic development?..., Wang [/bib_ref] [bib_ref] Policy instruments for addressing construction equipment emission-A research review from a global..., Huang [/bib_ref] , seriously reducing the quality of life and happiness of residents and thus restricting the high-quality economic and social development [bib_ref] Spatial effects of economic performance on the carbon intensity of human well-being:..., Li [/bib_ref]. In the report of the 19th National Congress of the Communist Party of China, it was proposed that "We will continue to ensure and improve people's well-being through development; improving people's livelihood and well-being is the fundamental purpose of development". (https://www.12371.cn/2017/10/27/ARTI1509103656574313.shtml (accessed on 23 September 2022)). This statement means that the previously adopted two-dimensional concept of sustainable development as economic growth and green ecology needs to be broadened to include welfare enhancement as a third dimension of sustainable development strategy in China. Accordingly, in order to promote regional sustainable development, there is a need to establish a comprehensive regional performance evaluation framework that simultaneously considers economic, social and environmental dimensions under the conditions of resource constraints, environmental protection requirements, and the improvement of people's well-being.
Ecological efficiency is a widely adopted index for sustainable development evaluation, but it focuses only on reducing environmental pollution and maximizing regional economic output [bib_ref] Eco-efficiency, eco-technology innovation and eco-well-being performance to improve global sustainable development, Zhang [/bib_ref]. Ecological welfare performance (EWP) index is different because it includes social, economic, environmental dimensions with an emphasis on maximizing human well-being with a minimum amount of ecological resource consumption [bib_ref] Assessment of ecological well-being performance and its spatial correlation analysis in the..., Xia [/bib_ref]. Furthermore, because it estimates the efficiency of converting ecological input into social welfare output, EWP index addresses the issues of efficiency and fairness emphasized by green economic development and, consequently, the quality of economy growth [bib_ref] Sustainable urban development of the π-shaped Curve Area in the Yellow River..., Sun [/bib_ref]. Therefore, the EWP index has the potential to be used to evaluate human well-being and sustainable regional development. However, the applicability of the current EWP index remains problematic. First, there is no agreeable indicators that constitute each dimension of the current EWP index. Secondly, how to empirically measure EWP index indicators is uncertain. Third, there are challenges in the existing data collection and analysis methods in implementing the EWP index for sustainable regional development evaluation. This study attempted to tackle these deficiencies in the literature to accelerate to green and sustainable economic development and to achieve the coordinated development of economy, environment and people's well-being.
As a strategically important region for the national economic development in the new era and a driving force for high-quality development, the Yangtze River Economic Belt (YREB) contributes to more than forty percent of China's population and GDP with merely 21.4% of the nation's land. Due to cumulative environmental effects of long-term high-intensity progress and a lack of sustainable development measures and controls, the YREB is faced with pressing social, economic and environmental consequences in recent years including uncoordinated economic development, lagging environmental protection, severe smog pollution and water eutrophication [bib_ref] Does China's Yangtze River Economic Belt policy impact on local ecosystem services?, Luo [/bib_ref]. For instance, its wastewater discharge already accounts for more than 40% of the national discharge. Its unit emission intensity of sulfur dioxide and ammonia nitrogen is more than 1.5 times of the national average emission intensity. These environmental problems have inevitably hindered the improvement of regional EWP and further retarded sustainable development in the YREB. Therefore, research is needed to inform policy makers about how to improve and balance the relationship among economy, society and the ecological environment and achieve the high-quality coordinated and integrated development in the YREB.
Existing research demonstrates that there is a geographic gradient differentiation pattern in industrial ecological efficiency, green economic efficiency and factor agglomeration capacity across the YREB [bib_ref] Industrial eco-efficiency evaluation and regional differences of Yangtze River Economic Belt, Ren [/bib_ref] [bib_ref] Spatial-temporal evolution of green total factor productivity and identification of area problems..., Li [/bib_ref] , but few scholars have probed the temporal and spatial differentiation of the EWP at an urban scale from the visual angle of human well-being. While ecological welfare studies are abundant, most focus on the national and provincial scales in analyzing spatial and temporal patterns. In reasoning the influencing factors of regional ecological welfare, fewer studies have employed spatial panel econometric models to investigate spatial effects [bib_ref] Applying the Super-EBM model and spatial Durbin model to examining total-factor ecological..., Chengyu [/bib_ref]. In fact, the differences in EWP across different regions are not only constrained by their own resource endowments, but also are jointly affected by multiple effects of external factors, such as factor flow, technology diffusion and differential policy implementation. These effects can often be explained by means of a spatial panel Durbin convergence model in which the spatial interaction effect of "local-neighbor" is considered [bib_ref] The spatial convergence and drivers of environmental efficiency under haze constraints-Evidence from..., Guo [/bib_ref].
This research develops an all-round EWP index to assess the YREB's EWP at a city scale. It attempts to reveal the spatial differentiation in the EWP and explore its driving forces using a spatial panel model. This study contributes to the theorization of comprehensive evaluation of regional EWP by incorporating three dimensions of economy, society and ecological environment into building a comprehensive EWP index and its associated analytical indicators. The developed EWP index improves the current index evaluation system for regional sustainable development evaluation. Methodologically, this research utilizes the spatial convergence model for estimating the spatio-temporal convergence of EWP across the cities in the YREB. The proposed model enables one to investigate the spatial effects of driving forces and their spatial interaction effects of "local-neighbors" in configuring the spatial patterns of regional ecological welfare, which are rarely examined in the existing studies. The empirical evidence generated in this research will contribute to figure out the relationship between regional economy, society and environment, and also provide a scientific basis for formulating policies aiming for the high-quality coordinated regional development.
## Literature review
Due to pressing problems associated with environmental degradation, economic growth and social welfare fluctuation, there is a growing interest in understanding how to promote sustainable regional development that simultaneously considers economy, society and ecological environment. Numerous studies have been published on EWP. Three broad categories of relevant studies can be identified: (1) the measurement of EWP; (2) the regional difference analysis of EWP; (3) the analysis of influential factors on the spatial-temporal changes of EWP.
Daly first introduced the concept of EWP when evaluating the sustainable development in various countries, but it was not widely used due to its difficulty in quantification and low practicability of the proposed indices [bib_ref] The world dynamics of economic growth: The economics of the steady state, Daly [/bib_ref]. Over time, the idea of EWP was reconceptualized to include economic dimensions, which led to the development of comprehensive welfare indicators. Among these indicators, the human development index (HDI) is one of the widely adopted indicators in evaluating EWP. As a comprehensive index, the HDI takes into account the basic essentials of human development by considering life expectancy at birth, education status and income level in its quantification. It embraces both economic benefits on account of national revenue and non-economic benefits on account of social choices. This composite approach using a limited number of variables is convenient for calculation and analysis [bib_ref] Linking nuclear energy, human development and carbon emission in BRICS region: Do..., Sadiq [/bib_ref] [bib_ref] Does nuclear energy consumption contribute to human development? Modeling the effects of..., Sadiq [/bib_ref]. However, HDI still has two shortcomings. First, HDI focuses too much on economic growth and fails to consider the influence of environmental factors on human welfare. The index does not pay enough attention to ecology and ignores information about environmental sustainability. Second, HDI only considers the "output" of human welfare but fails to include the cost invested for this purpose [bib_ref] The environmental efficiency of well-being: A cross-national analysis, Knight [/bib_ref]. To address the deficiencies of the HDI, some scholars such as Yew, Abdallah and Caillon et al. employed the ratio method to build the ecological performance evaluation index system [bib_ref] Environmentally responsible happy nation index: Towards an internationally acceptable national success indicator, Yew [/bib_ref] [bib_ref] Estimating worldwide life satisfaction, Abdallah [/bib_ref] , where the numerator is the social welfare indicators that are excluded by the HDI, and the denominator measures the consumption of ecological resources. However, none of the above methods measure the level of regional EWP comprehensively.
The measurement method of EWP has gradually evolved into a more acceptable efficiency measurement method based on the input-output analysis framework [bib_ref] Environmentally efficient well-being: Rethinking sustainability as the relationship between human well-being and..., Dietz [/bib_ref] [bib_ref] A two-stage data envelopment analysis of efficiency of social-ecological systems: Inference from..., Ibrahim [/bib_ref]. There are two categories of efficiency measurement methods. The first is the non-parametric approach of Data Envelopment Analysis (DEA), and the second is the parametric approach of Stochastic Frontier Analysis (SFA). Although the influence of interference items is considered in SFA, the functional relationship between input and output needs to be assumed in advance, which has a certain degree of subjectivity [bib_ref] Trajectories of efficiency measurement: A bibliometric analysis of DEA and SFA, Lampe [/bib_ref]. The DEA model has become one of the most widely used models in the literature due to its advantages of requiring fewer indicator variables and retaining completely the original indicator information. DEA models include CRS (Constant Return to Scale), VRS (Variable Return to Scale), SBM (Slack-Based-Measure), three-stage DEA, Malmquist index, and the super-efficiency DEA model. Among them, the SBM model is widely used in the literature, because it can better deal with the negative environmental pollution output in the consumption of resource elements [bib_ref] A comprehensive eco-efficiency model and dynamics of regional eco-efficiency in China, Huang [/bib_ref] [bib_ref] Green economic efficiency and its influencing factors in China from 2008 to..., Zhao [/bib_ref].
The current input-output based evaluation index system and measurement methods of regional EWP are however limited by indicator selection, data collection and analysis methods. First, the existing researches only consider resource consumption as an input indicator, ignoring the capital and technological factors that transform resource consumption into outputs. Second, the current studies tend only to include industrial wastes as undesirable outputs for measurement, but exclude household waste discharge, PM 2.5 and other indicators which are intimately linked to human well-being. Third, the current measure of people's living standards only considers a single economic dimension. Indicators on education, medical care and health need to be integrated into the existing input-output analysis framework. To address these deficiencies, this paper proposes a "3C" (Classification, Coordination, Collaboration) analysis framework for the EWP [fig_ref] Figure 1: "Classification-Coordination-Collaboration" Analysis Framework of EWP [/fig_ref] based on the "3C" theorization for sustainable development in the literature [bib_ref] Classification-coordination-collaboration: A systems approach for advancing Sustainable Development Goals, Fu [/bib_ref]. The framework provides a basis for developing a comprehensive evaluation index system and selecting appropriate evaluation indicators and measurements. In the framework, input indicators include resource consumption, ecological capital and technological capital, and output indicators consist of desired outputs and non-desired outputs. The goals of achieving ecological protection and green development are multi-dimensional and need to be coordinated. A multi-objective collaborative regional development model in the evaluation index system needs to be established for promoting the overall realization of the high-quality sustainable development of economy, society and ecological environment. Scale), SBM (Slack-Based-Measure), three-stage DEA, Malmquist index, and the superefficiency DEA model. Among them, the SBM model is widely used in the literature, because it can better deal with the negative environmental pollution output in the consumption of resource elements [bib_ref] A comprehensive eco-efficiency model and dynamics of regional eco-efficiency in China, Huang [/bib_ref] [bib_ref] Green economic efficiency and its influencing factors in China from 2008 to..., Zhao [/bib_ref]. The current input-output based evaluation index system and measurement methods of regional EWP are however limited by indicator selection, data collection and analysis methods. First, the existing researches only consider resource consumption as an input indicator, ignoring the capital and technological factors that transform resource consumption into outputs. Second, the current studies tend only to include industrial wastes as undesirable outputs for measurement, but exclude household waste discharge, PM2.5 and other indicators which are intimately linked to human well-being. Third, the current measure of people's living standards only considers a single economic dimension. Indicators on education, medical care and health need to be integrated into the existing inputoutput analysis framework. To address these deficiencies, this paper proposes a "3C" (Classification, Coordination, Collaboration) analysis framework for the EWP [fig_ref] Figure 1: "Classification-Coordination-Collaboration" Analysis Framework of EWP [/fig_ref] based on the "3C" theorization for sustainable development in the literature [bib_ref] Classification-coordination-collaboration: A systems approach for advancing Sustainable Development Goals, Fu [/bib_ref]. The framework provides a basis for developing a comprehensive evaluation index system and selecting appropriate evaluation indicators and measurements. In the framework, input indicators include resource consumption, ecological capital and technological capital, and output indicators consist of desired outputs and non-desired outputs. The goals of achieving ecological protection and green development are multi-dimensional and need to be coordinated. A multi-objective collaborative regional development model in the evaluation index system needs to be established for promoting the overall realization of the highquality sustainable development of economy, society and ecological environment. The literature on the regional differences in the EWP has so far focused on the national and provincial scales, and paid less attention to the city scale [bib_ref] Have cities effectively improved ecological well-being performance? Empirical analysis of 278 Chinese..., Bian [/bib_ref]. For example, Zhang et al. compared and analyzed the comprehensive evaluation values of EWP in 82 countries. It was found that developed countries and G20 countries performed relatively worse [bib_ref] Which countries are more ecologically efficient in improving human well-being? An application..., Zhang [/bib_ref]. Long used cross-sectional data of 42 countries to conduct a horizontal international comparison, and found that the EWP was the highest in OECD countries, followed by G20 countries, and the lowest in BRICS [bib_ref] Evaluation of urban ecological well-being performance of Chinese major cities based on..., Long [/bib_ref]. Xu et al. measured the EWP of 30 provinces and cities in mainland China from 2005 to 2014. They found that the EWP generally showed an "east > central > west" gradient, and revealed the pattern of high-high or lowlow agglomeration in the local space, indicating that China's EWP presented a certain degree of spatial heterogeneity and spatial autocorrelation [bib_ref] Spatial-temporal differentiation of Chinese provincial ecological well-being performance. Reg, Xu [/bib_ref]. Fang and Xiao, Wang et The literature on the regional differences in the EWP has so far focused on the national and provincial scales, and paid less attention to the city scale [bib_ref] Have cities effectively improved ecological well-being performance? Empirical analysis of 278 Chinese..., Bian [/bib_ref]. For example, Zhang et al. compared and analyzed the comprehensive evaluation values of EWP in 82 countries. It was found that developed countries and G20 countries performed relatively worse [bib_ref] Which countries are more ecologically efficient in improving human well-being? An application..., Zhang [/bib_ref]. Long used cross-sectional data of 42 countries to conduct a horizontal international comparison, and found that the EWP was the highest in OECD countries, followed by G20 countries, and the lowest in BRICS [bib_ref] Evaluation of urban ecological well-being performance of Chinese major cities based on..., Long [/bib_ref]. Xu et al. measured the EWP of 30 provinces and cities in mainland China from 2005 to 2014. They found that the EWP generally showed an "east > central > west" gradient, and revealed the pattern of high-high or low-low agglomeration in the local space, indicating that China's EWP presented a certain degree of spatial heterogeneity and spatial autocorrelation [bib_ref] Spatial-temporal differentiation of Chinese provincial ecological well-being performance. Reg, Xu [/bib_ref]. Xu et al. [bib_ref] Research on regional ecological well-being performance and spatial effect in China, Fang [/bib_ref] [bib_ref] Research on Spatial Unbalance and Influencing Factors of Ecological Well-Being Performance in..., Wang [/bib_ref]. Most of the above studies attempt to understand the spatial differences in EWP at the national, provincial and other macro scales. Little is known about how such differences manifest across various cities and other mesoscale units. In fact, the conclusions on EWP differences and causal mechanisms drawn from the macro scale analysis may not be applicable to the mesoscale, making it difficult to provide an effective reference for the policy-making of prefecture-level cities at the mesoscale. The YREB, which spans eastern and western China and connects its north and south, is a significant strategic support belt for national development in the new era, and it is also an ecologically sensitive area. However, the existing studies pay little attention to the temporal and spatial pattern of EWP at the urban scale in the belt. This paper, therefore, attempts to supply gaps in literature and will investigate the regional differentiation and dynamic trends in the YREB at a city scale.
Regarding the research on the influencing factors of EWP, scholars have mainly investigated the effects of economic growth, industrial structure, urbanization, technological progress, and foreign direct investment [bib_ref] Spatial and Temporal Evolution and Driving Factors of Urban Ecological Well-Being Performance..., Bian [/bib_ref]. However, the knowledge about the role of causal factors in influencing EWP patterns remains to be limited. For example, the relationship between economic growth and the EWP has been studied extensively in the literature, but findings seem inconclusive. Common used the 2001 cross-sectional data to find that the richer countries have lower EWP, and the EWP is negatively associated with economic growth [bib_ref] Measuring national economic performance without using prices, Common [/bib_ref]. Dietz et al. conducted a cross-country panel data analysis and reached a similar conclusion [bib_ref] Environmentally efficient well-being: Is there a Kuznets curve?, Dietz [/bib_ref]. However, Jorgenson et al. used transnational panel data in their study and found that economic growth may not cause the EWP to decline [bib_ref] Economic development and the carbon intensity of human well-being, Jorgenson [/bib_ref]. The effects of other influencing factors on the EWP seem mixed. For example, Xiao and Zhang discovered that urbanization and industrialization have significant negative spatial spillover effects on the EWP [bib_ref] Spatio-temporal characteristics of coupling coordination between green innovation efficiency and ecological welfare..., Xiao [/bib_ref]. Ma et al. found that urban growth, industrial structure and government investment have a promoting effect on the EWP, while foreign direct investment has an inhibitory effect [bib_ref] Can Green Economy and Ecological Welfare Achieve Synergistic Development? The Perspective of..., Ma [/bib_ref].
While many factors may affect the EWP, the effects of influencing factors may vary significantly. The mixed findings on the effects of influencing factors may result from the data and analytical methods employed in various studies. The existing analysis methods investigating the influencing factors of EWP include cross section regression, panel Tobit, and LMDI. These analytical methods are often employed to explore the relations between EWP and social variables and are not suitable for detecting spatial effects. In fact, the estimation results that ignore spatial effects are often biased [bib_ref] Specification and estimation of spatial panel data models, Elhorst [/bib_ref]. Further, most of these methods are considered static models as they are incapable of revealing temporal trends of the associations between EWP and causal factors. The spatial convergence model (SCM) is used on the basis of spatial angles for investigating the differentiated causes of the associations among different regions and their convergence trends [bib_ref] What drives technology convergence? Exploring the influence of technological and resource allocation..., Jeong [/bib_ref] [bib_ref] Convergence of green total factor productivity in China's service industry, Wu [/bib_ref] [bib_ref] Spatial Convergence of Carbon Productivity: Theoretical Analysis and Chinese Experience, Sun [/bib_ref]. Therefore, this research builds a spatial Durbin convergence model which incorporates the spatial impacts. The study helps reveal the dynamic driving forces of regional differentiation in the EWP of the YREB.
# Methods and data
# Methods
## Measurement of ewp
Charns first presented the Data Envelopment Analysis (DEA). It is used to measure the efficiency of a multi-input-output DMU [bib_ref] Measuring the efficiency of decision making units, Charnes [/bib_ref]. The initial DEA model is based on either radial or angular measurement. The radial method requires that inputs and outputs must change in the same proportion when evaluating efficiency, so it cannot address the slackness in inputs and outputs and the resulting measurement errors. The angular method requires that the evaluation of efficiency must be based on the perspective of input or output, and cannot give consideration to both. To solve the above problems, Tone presented a nonradial and non-angular DEA model on the basis of slackness variables, namely the SBM (Slack-Based Measure) model [bib_ref] A slacks-based measure of efficiency in data envelopment analysis, Tone [/bib_ref]. However, plural DMUs will be effective simultaneously when using the SBM model to measure efficiency, so it is unable to evaluate and rank DMUs effectively. The super-efficiency SBM model is an improvement on the SBM model, which can further evaluate the effective units (the efficiency value is 1), so as to obtain more accurate measurement results. The model is as follows:
[formula] ρ se = min 1− 1 m m ∑ i=1 s − i x ik 1+ 1 q 1 +q 2 ( q 1 ∑ r=1 s g r y g rk + q 2 ∑ r=1 s b r y b rk ) s.t. x k = Xλ + s − y g k = Y g λ − s g y b k = Y b λ + s b λ ≥ 0, s − ≥ 0, s g ≥ 0, s b ≥ 0 (1) [/formula]
where, x k , y g k and y b k are the input, desirable output and undesirable output vectors of the DMUs, respectively. X k , Y g and Y b are input, desirable output and undesirable output matrices. s − , s g , s b are the slackness of inputs, desirable outputs and non-desirable outputs. λ is the weight column vector. ρ se is the measured value of efficiency. When 0 < ρ se < 1, there is redundancy in the DMUs, and the efficiency can be raised by optimizing the input structure. When 1 ≤ ρ se , the DMU is valid.
## Dagum gini coefficient
The Gini coefficient and subgroup decomposition method is presented by Dagum [bib_ref] Decomposition and interpretation of Gini and the generalized entropy inequality measures, Dagum [/bib_ref]. Compared with the traditional imbalance measurement indexes, it can decompose the indexes into regional differences, inter-regional differences and hypervariable density, and solve the problems of subsample distribution, overlap and regional differences, so that it is easier to obtain indexes with economic implications. The calculation formula is as follows:
[formula] G = k ∑ e=1 k ∑ f =1 n e ∑ i=1 n f ∑ j=1 x ei − x f j 2n 2 x (2) [/formula]
where, G denotes the Dagum Gini coefficient, x denotes the average value of the EWP in the YREB, n denotes the total number of research units in the study area (108 in this paper), k denotes the number of areas, n e (n f ) denotes the number of research units in area e (f ), and x ei (x fj ) is the EWP of no.i (j) research unit in the region e (f ). G ee and G ef are shown as follows:
[formula] G ee = n e ∑ i=1 n f ∑ j=1 x ei − x f j 2n e 2 x (3) G e f = n e ∑ i=1 n f ∑ j=1 x ei − x f j n e n f (x e − x f )(4) [/formula]
Based on the decomposition of Dagum Gini coefficient, the total Gini coefficient is broken down into regional difference (G rd ), inter-regional difference (G ird ) and hypervariable density (G t ), and G = G rd + G nb + G t . The expression is as follows:
[formula] G rd = k ∑ e=1 G ee p e s e(5)G ird = k ∑ e=2 e−1 ∑ f =1 D e f G e f (p e s f + p f s e ) (6) G t = k ∑ e=2 e−1 ∑ f =1 (1 − D e f )G e f (p e s f + p f s e )(7) [/formula]
where, p e = n e /n, s e = n e x e /nx, . D ef is the relative influence of EWP between the regions e and f. d ef is the performance difference of ecological welfare in the region, representing the mathematical expectation of the sum of all samples (x ei − x f j > 0) in the regions e and f, and p ef is the hypervariable first-order moment, representing the mathematical expectation of the sum of all samples (x f j − x e f > 0) in the areas e and f. For details of each coefficient, refer to Stéphane and Patrick [bib_ref] Linking Yitzhaki's and Dagum's Gini decompositions, Mussard [/bib_ref].
## Verification of spatial correlation
For investigating the spatial distribution of the EWP in the YREB comprehensively, the global spatial correlation index in ESDA was employed in this research. Moran's Index was commonly employed for measuring the global spatial correlation. The results reflect the concentration level of spatial distribution of the EWP in the YREB, and embody the similarity of adjacent cities. The expression is as follows:
[formula] Moran s I = n n ∑ i=1 n ∑ j=1 w ij (y i − y)(y j − y) n ∑ i=1 n ∑ j=1 w ij n ∑ i=1 (y i − y) 2 (8) [/formula]
where n represents 108 cities in the YREB, w ij denotes the spatial weight metric, and y and y denote the EWP and its mean value, respectively. For exploring systematically the spatial relevance characteristics of the EWP, three spatial weight matrices were set and standardized based on the Queen Adjacent Matrix (w xy ), Geographic Distance Matrix (w d ) and Economic Distance Matrix (w e ), as shown in [fig_ref] Table 1: Spatial weight matrix types [/fig_ref].
[formula] x = y 0 x = y Economic distance 1/|g x −g y | n ∑ x,y 1/|g x −g y | x = y 0 x = y [/formula]
Note: d xy denotes the distance between city x and city y; g x , g y represent the per capita GDP of city x and city y respectively, n denotes the total number of cities.
## Spatial panel durbin convergence model
β-convergence was first used to analyze whether the economic development of a region will change with time in a steady state. In this paper, β-convergence means that regions with a low EWP are able to overtake regions with a high EWP, with a higher growth rate over time, and the gap between the two gradually narrows, eventually reaching the same stable level. β-convergence includes absolute and conditional β-convergence. Absolute βconvergence refers to all cities in YREB having identical economic and social characteristics, and the EWP of each prefecture-level city will tend to reach the equivalent level over time. The spatial dependence of different regions is not considered in the traditional standard β-convergence model, but the spatial panel Durbin convergence model can explain the influence of the spatial interaction effect of "local-neighbor" on the dependent variable, so this model was selected for investigating the convergence of EWP in the YREB. The expression is as follows:
[formula] Ln(EWP i,t+1 /EWP i,t ) = α +ρ n ∑ j=1 W ij Ln(EWP i,t+1 /EWP i,t ) + βLnEWP i,t +λ n ∑ j=1 W ij LnEWP i,t + δ t + µ i,t(9) [/formula]
where, EWP i,t , and EWP i,t+1 are the EWP of the city i in year t and t + 1 respectively, W ij refers to the spatial weight matrix, and ρ refers to the spatial autoregression coefficient, indicating the impact of the growth rate of EWP in adjacent cities on the local city. λ refers to the impact of EWP in adjacent cities on the local city. δ t denotes time fixed effect. µ i,t denotes a random perturbation term. β is the convergence coefficient. If β is significantly negative, it means that regions with a lower EWP have faster growth rates than the regions with higher EWP, and there is β-convergence. On the contrary, if it is positive, there is no convergence.
Secondly, the conditional β-convergence of EWP is not just affected by the initial factor endowment of each region, but also by the surrounding external social-economic factors, which will lead to differences in its steady-state values. Therefore, the spatial lag term and the influence factors of surrounding areas (WX) were selected to construct the spatial panel Durbin model. The convergence expression is:
[formula] Ln(EWP i,t+1 /EWP i,t ) = α + ρ n ∑ j=1 W ij Ln(EWP i,t+1 /EWP i,t ) + βLnEWP i,t + n ∑ k=1 θ k X i,t +λ n ∑ k=1 W ij EWP i,t + ϕ n ∑ k=1 W ij X i,t + δ t + µ i,t(10) [/formula]
In this expression, θ k denotes the coefficient of the influencing factors and ϕ refers to the impact of influencing factors in adjacent cities on the local city. Other coefficients are the same as above.
## Establishment of indicator system
## Establishment of indicator system
According to the "3C" analysis framework of EWP, resource consumption, ecological capital and technological capital are considered to be the ecological inputs with reference to ecological welfare input-output indicator selection proposed by Zhang et al. and Li et al. [bib_ref] China's multi-dimensional ecological well-being performance evaluation: A new method based on coupling..., Li [/bib_ref] [bib_ref] How is the environmental efficiency in the process of dramatic economic development..., Zhang [/bib_ref] [bib_ref] Mining Eco-Efficiency Measurement and Driving Factors Identification Based on Meta-US-SBM in Guangxi..., Li [/bib_ref]. In addition to the traditional resource consumption indicators, ecological capital and technology input are also included and measured by urban environmental infrastructure investment, per capita science and technology and education expenditure. It includes a multi-dimensional comprehensive appraising indicator system of both wellbeing outputs and non-well-being outputs. Among them, the non-well-being output is an indicator of environmental pollution. In addition to the three industrial wastes commonly used by scholars, it also includes domestic waste removal and transportation volume and PM 2.5 particles that are closely related to the life quality of residents. In terms of desirable outputs, relevant indicators reflecting economic benefits and education level [bib_ref] Sustainable development of China's regions from the perspective of ecological welfare performance:..., Song [/bib_ref] were selected, and environmental, medical and health aspects were also taken into consideration. They are represented by the park land area per capita and the number of doctors per 10,000 inhabitants, respectively. The comprehensive appraising indicator system of the EWP along with multi-source data is listed as follows.
This research employs the statistical data from China Urban Statistical Yearbook, China Urban Construction Statistical Yearbook and statistical yearbooks of provinces and cities from 2009 to 2019. The data of PM 2.5 in the undesirable outputs comes from the grid data of PM 2.5 concentrations released by the Social Economic Data and Application Center of Columbia University and obtained by satellite monitoring (https://sedac.ciesin.columbia. edu/data/set/sdei-global-annual-gwr-pm2-5-modis-misr-seawifs-aod-v4-gl-03 (accessed on 7 September 2022). The above data are described in [fig_ref] Table 2: Appraising indicator system of the EWP in the YREB [/fig_ref].
## Selection of influence variables
When constructing the conditional β convergence model, this paper considers the influence of industrial structure, technological input, urbanization development level, environmental regulation, foreign capital utilization, government intervention and transportation accessibility on the convergence of EWP. Considering the availability of data, the specific indicators are as follows(see [fig_ref] Table 3: Control variables and their descriptive statistics [/fig_ref] : Industrial structure (Ind). Industrial upgrading and shifting service-oriented economy can effectively promote carbon emission reduction, improve environmental quality, and thus improve the EWP of residents [bib_ref] Does the expansion of a motorway network lead to economic agglomeration? Evidence..., Yu [/bib_ref]. The proportion of tertiary industry in GDP is used for measuring the industrial structure.
Technology input (Tec). Chen and Cheng argued that advances in technology are the main means for improving China's extensive economic growth mode with high consumption and emissions [bib_ref] Haze pollution and economic green transformation assessment of Chinese cities, Chen [/bib_ref]. Therefore, it is measured based on the sum of science and technology expenditure and education expenditure per capita.
City. Urbanization is one of the most influential human activities. On the one hand, urban growth will increase energy consumption, thereby reducing environmental quality [bib_ref] How does urbanization affect carbon emission efficiency? Evidence from, Sun [/bib_ref]. On the other hand, the increasing level of urbanization contributes to the utilization efficiency of infrastructure and improves people's living standards [bib_ref] Scale, composition, and technique effects through which the economic growth, foreign direct..., Liobikienė [/bib_ref]. The proportion of the population in municipal districts of a prefectural city-region is used to characterize the urbanization process.
Environmental regulation (Er). On the one hand, environmental regulation may force some heavy pollution enterprises to withdraw from the market, thereby improving the environmental quality. Another aspect, environmental regulation could lead to higher production costs and environmental governance costs of enterprises, resulting in less pro-ductive investment and lower profits and consequently competitiveness of enterprises. Environmental regulation may thereby limit the improvement of enterprise production efficiency and affect economic development [bib_ref] Impact of environmental regulations on carbon emissions of transportation infrastructure: China's evidence, Zhang [/bib_ref]. Comprehensive indicators such as domestic sewage treatment rate, industrial sulfur dioxide removal rate, industrial solid waste comprehensive utilization rate, and industrial dust removal rate are selected to illustrate the strength of environmental supervision.
Foreign direct investment (Fdi). The "pollution refuge" hypothesis believes that foreign direct investment will aggravate the environmental pollution of the host country under the conditions of loose environmental regulations [bib_ref] Trade, environmental regulations and industrial mobility: An industry-level study of Japan, Cole [/bib_ref]. The "pollution halo" hypothesis believes that the adoption of standardized management methods and advanced production technologies by foreign-funded enterprises can promote domestic enterprises to optimize their environmental management system through technology spillover effects, thus helping to improve the environmental quality of host countries [bib_ref] Is free trade good for the environment?, Antweiler [/bib_ref]. The ratio of the actual amount of foreign investment used in a year to GDP is used for measuring foreign direct investment.
Government intervention (Gov). Government intervention determines the regional institutional environment to a large extent, thus affecting the efficiency of resource allocation and economic development [bib_ref] The impact of environmental regulation, shadow economy, and corruption on environmental quality:..., Chen [/bib_ref]. The share of fiscal expenditure in GDP is used for measuring the governmental intervention.
Transportation accessibility (Trans). Convenient transportation can shorten the space and time distance between regions, promote the flow and exchange of regional resources, and thus improve the living standard of residents. This research employs per capita urban road area for measuring the transportation convenience in each city. [fig_ref] Figure 2: Distribution map of EWP in 2019 [/fig_ref] displays the spatial distribution and temporal evolution of YREB's EWP. The higher the EWP, the healthier urban development and greater people's well-being will be. From the angle of spatial distribution, most of the high-value regions are concentrated in Jiangsu and Zhejiang in the downstream of Yangtze River, however, Wuhan, Changsha, Zhuzhou and Xiangtan in midstream, and provincial capital cities in the upstream are also associated with a high EWP, reflecting the high level of urban development in these regions and the effective improvement of residents' well-being. Low-value areas are mostly located in non-capital cities in the upstream, including western Hubei and Hunan, and southern Jiangxi. Macroscopically, it presents a ladder-shaped distribution pattern of east-high and west-low, showing a gradual decline gradient. From the angle of temporal evolution, the EWP for each region in YREB shows a fluctuating but upward trend. The gap in EWP between different regions still exists, so further analysis of differences in the EWP between different regions in the YREB is essential. will be. From the angle of spatial distribution, most of the high-value regions are concentrated in Jiangsu and Zhejiang in the downstream of Yangtze River, however, Wuhan, Changsha, Zhuzhou and Xiangtan in midstream, and provincial capital cities in the upstream are also associated with a high EWP, reflecting the high level of urban development in these regions and the effective improvement of residents' well-being. Low-value areas are mostly located in non-capital cities in the upstream, including western Hubei and Hunan, and southern Jiangxi. Macroscopically, it presents a ladder-shaped distribution pattern of east-high and west-low, showing a gradual decline gradient. From the angle of temporal evolution, the EWP for each region in YREB shows a fluctuating but upward trend. The gap in EWP between different regions still exists, so further analysis of differences in the EWP between different regions in the YREB is essential.
## Spatial variation of ewp in yreb
## Spatial and temporal evolution of ewp
## Regional differences of ewp
The Dagum Gini coefficient was adopted for measuring the regional variations about EWP in YREB [fig_ref] Figure 3: Regional differences of the EWP in YREB [/fig_ref]. As a whole, the Gini coefficient of the YREB's EWP is on the high side, but it fluctuates and declines in the sample period, indicating that there is a promi-
## Regional differences of ewp
The Dagum Gini coefficient was adopted for measuring the regional variations about EWP in YREB [fig_ref] Figure 3: Regional differences of the EWP in YREB [/fig_ref]. As a whole, the Gini coefficient of the YREB's EWP is on the high side, but it fluctuates and declines in the sample period, indicating that there is a prominent geographic variation in the EWP among the prefecture level cities in the YREB, but the degree of spatial variation has been reduced. Specifically, the Gini coefficient in the upstream region of the YREB ranks first in all regions, because the upstream region of the YREB covers a wide area and has complex terrain. Furthermore, different cities have different resource endowments, varying degrees of industrialization, and different paces of economic development. The problem of unbalanced and insufficient high-quality development in the region is prominent. The EWP gap among cities in the midstream YREB is comparatively small and more balanced. The mean regional difference in the lower reaches of the YREB is the lowest, and it has maintained a lower level from 2009 to 2019, reflecting the positive impact of integrative and inclusive development in the YREB. Moreover, its internal central cities can generate positive spatial-spillover effects on surrounding cities, helping them improve their EWP. [fig_ref] Figure 4: Inter-regional differences of the EWP in the YREB [/fig_ref] refers to the time-series evolution of the inter-regional variance of EWP between the downstream, midstream and upstream YREB during the sample observation period. From the overall trend, the Gini coefficient between paired regions is fluctuating and declining. Among them, the Gini coefficient between the midstream and the downstream is the smallest, and shows a further downward trend. The possible reason is that the midstream region may have benefitted from industrial diffusion out from the downstream region, optimized the industrial structure, and improved its own economic development, thus narrowing cross regional differences. In contrast, although the Gini coefficient between the downstream and upstream regions, and between the midstream and upstream regions, declined significantly during the sample observation period, it was still at a high level. It can be seen from the figure that the inter-regional difference in the EWP is the greatest between the downstream and upstream regions, followed by the difference between the midstream and upstream regions. [fig_ref] Figure 4: Inter-regional differences of the EWP in the YREB [/fig_ref] refers to the time-series evolution of the inter-regional variance of EWP between the downstream, midstream and upstream YREB during the sample observation period. From the overall trend, the Gini coefficient between paired regions is fluctuating and declining. Among them, the Gini coefficient between the midstream and the downstream is the smallest, and shows a further downward trend. The possible reason is that the midstream region may have benefitted from industrial diffusion out from the downstream region, optimized the industrial structure, and improved its own economic development, thus narrowing cross regional differences. In contrast, although the Gini coefficient between the downstream and upstream regions, and between the midstream and upstream regions, declined significantly during the sample observation period, it was still at a high level. It can be seen from the figure that the inter-regional difference in the EWP is the greatest between the downstream and upstream regions, followed by the difference between the midstream and upstream regions. opment, thus narrowing cross regional differences. In contrast, although the Gini coefficient between the downstream and upstream regions, and between the midstream and upstream regions, declined significantly during the sample observation period, it was still at a high level. It can be seen from the figure that the inter-regional difference in the EWP is the greatest between the downstream and upstream regions, followed by the difference between the midstream and upstream regions. [fig_ref] Figure 5: Spatial differentiation [/fig_ref] shows the contribution rate of regional variance, inter-regional variance and hypervariable density in the EWP of the YREB. As shown in [fig_ref] Figure 5: Spatial differentiation [/fig_ref] the whole variance of the EWP shows a fluctuating decline, which specifically exhibits an inverted "N" downward trend of decrease-increase-decrease, but the difference remains at a high level. In addition, the Gini coefficient between regions has decreased significantly, indicating that the gap in EWP among regions in the YREB has narrowed. [fig_ref] Figure 5: Spatial differentiation [/fig_ref] shows the contribution rate of regional variance, inter-regional variance and hypervariable density in the EWP of the YREB. As shown in [fig_ref] Figure 5: Spatial differentiation [/fig_ref] the whole variance of the EWP shows a fluctuating decline, which specifically exhibits an inverted "N" downward trend of decrease-increase-decrease, but the difference remains at a high level. In addition, the Gini coefficient between regions has decreased significantly, indicating that the gap in EWP among regions in the YREB has narrowed.
## Inter-regional differences of ewp
## Inter-regional differences of ewp
## Overall difference and decomposition of ewp
## Overall difference and decomposition of ewp
From the contribution rate of the variance [fig_ref] Figure 5: Spatial differentiation [/fig_ref] , the hypervariable density is the major source of the whole variance during the sample study period, and it is always higher than the regional and inter-regional variance, which denotes that the overlapping phenomenon between different regions of YREB is an important reason for the poor EWP. The hypervariable density reflects the spatial imbalance of the development of the EWP. While the developed regions have improved their EWP, the underdeveloped cities have "left behind" and failed to keep up with the pace of development. Therefore, we should not only pay attention to the underdeveloped regions and give them corresponding fiscal and tax subsidy policies, but also find out the reasons why the underdeveloped cities in the developed regions fail to develop, and take corresponding measures according to local conditions for narrowing the gap and promoting the coordinated development of the YREB.
## Analysis on spatial differentiation and drives of ewp in the yreb
## Spatial autocorrelation of ewp
Moran's I of the EWP in the YREB during 2009-2019 was calculated by using ArcGIS software, and the results are presented in [fig_ref] Table 4: Moran's I of overall EWP during 2009~2019 [/fig_ref]. It denotes that Moran's I in all years has passed the significance test and is larger than zero, denoting that the spatial distribution of YREB's EWP during this period has an obvious spillover effect. The prefecture level cities with a high (or low) EWP are often adjacent and have spatial dependence, which may be mainly due to the regional economic integration. Each city is not an "island"; its EWP is no longer limited to its own development, and the influence of surrounding areas cannot be ignored. From the contribution rate of the variance [fig_ref] Figure 5: Spatial differentiation [/fig_ref] , the hypervariable density is the major source of the whole variance during the sample study period, and it is always higher than the regional and inter-regional variance, which denotes that the overlapping phenomenon between different regions of YREB is an important reason for the poor EWP. The hypervariable density reflects the spatial imbalance of the development of the EWP. While the developed regions have improved their EWP, the underdeveloped cities have "left behind" and failed to keep up with the pace of development. Therefore, we should not only pay attention to the underdeveloped regions and give them corresponding fiscal and tax subsidy policies, but also find out the reasons why the underdeveloped cities in the developed regions fail to develop, and take corresponding measures according to local conditions for narrowing the gap and promoting the coordinated development of the YREB.
## Analysis on spatial differentiation and drives of ewp in the yreb
## Spatial autocorrelation of ewp
Moran's I of the EWP in the YREB during 2009-2019 was calculated by using ArcGIS software, and the results are presented in [fig_ref] Table 4: Moran's I of overall EWP during 2009~2019 [/fig_ref]. It denotes that Moran's I in all years has passed the significance test and is larger than zero, denoting that the spatial distribution of YREB's EWP during this period has an obvious spillover effect. The prefecture level cities with a high (or low) EWP are often adjacent and have spatial dependence, which may be mainly due to the regional economic integration. Each city is not an "island"; its EWP is no longer limited to its own development, and the influence of surrounding areas cannot be ignored.
## Spatial convergence model test model selection
It can be seen from the previous analysis that the EWP has a strong spatial correlation. If the inherent spatial spillover effect is ignored when analyzing its spatial convergence trend, it is likely to obtain erroneous empirical results. In addition, not only the explained variable itself, but also the explanatory variable and error term will cause spatial autocorrelation. The spatial Durbin model (SPDM) could well recognize the spatial effects of different sources [bib_ref] Applied Spatial Econometrics: Raising the Bar, Elhorst [/bib_ref]. Thus, this research starts from the SPDM. First, LR tests and Wald tests were adopted for determining whether SPDM could degenerate to SAR or SEM. Results passed the 5% significance level test and the original hypothesis was rejected (see [fig_ref] Table 5: Spatial panel econometric model test [/fig_ref]. The SPDM was employed to conduct empirical test. Then the existence form of spatial dependence was tested, and the results show that the spatial dependency of YREB's EWP exists in the form of error. In addition, Hausman test results show that the log likelihood value of the time fixed effect model (1055.3) is greater than those of the space fixed effect model and the spatial-temporal fixed effect model (935.39, 900.2), so this research selected the time-fixed SPDM. [fig_ref] Table 6: Absolute β-convergence estimation results [/fig_ref] shows the estimation results of the absolute β-convergence model for YREB's EWP under the three spatial matrices. It can be seen as follows. First, under the three matrix conditions, the β-convergence coefficients are all less than zero at the 0.01 significance level, indicating the gap has narrowed without considering the impact of other factors. Cities with relatively low EWP have a "catching-up effect" to the cities with a relatively high EWP, eventually converging to a steady state level. Second, although the significance is slightly different, ρ and λ are all positive, indicating that the change rate of the EWP of this city is affected by the change rate and the positive spatial spillover of other cities. Under the proximity matrix and geographical distance matrix, both are significantly positive, indicating that its change rate is more vulnerable to the impact of neighboring cities. Therefore, the local government shouldn't only take into account its own development, but also think about the influence of neighboring cities when formulating relevant policies to improve the EWP of each city. Note: *** refer to significant levels of 0.01, with t-values in parentheses. denotes the conditional β-convergence analysis for the EWP of 108 prefecture level cities in YREB. First, the results show there is significant conditional β-convergence in the YREB. Second, the conditional β-convergence rate has increased compared to absolute β-convergence, indicating the convergent rate is driven by the industrial structure, city, environmental regulation, foreign investment, transportation accessibility and other factors. . Conditional β convergence estimation results under adjacent space weight matrixes. According to the regression results, this research obtains several discoveries. First, the overall spatial spillover effect of the YREB has not changed, and both ρ and λ are positive, indicating that the change rate of EWP of each city is subject to the impact of neighboring cities. Second, from the perspective of economics, the convergence of each control variable on the EWP of the YREB is significantly different. Specifically, the adjustment of industrial structure (Ind) and technological innovation (Tec) have a generally positive effect, which can significantly promote the overall EWP to converge to a high value, but the spatial spillover effects of both are significantly negative, possibly because the development of the tertiary industry, especially the modern service industry, can improve the efficiency of resource utilization, and the improvement of technological level reduces the pollution emissions of unit products, thus promoting the improvement of EWP. However, on the one hand, the transformation and upgrading of the industrial structure of the surrounding cities will enhance their competitiveness; on the other hand, they will transfer their outdated production capacity to other regions, thus preventing them from converging to high values. Secondly, the acceleration of urbanization will not only promote regional growth, but also improve the infrastructure construction, health care, employment, education and income of residents in the surrounding areas, so to facilitate EWP convergence to a higher value. Thirdly, environmental regulation (Er) inhibits the convergence of EWP, and its spatial spillover effect is significantly negative. The reason may be that the current environmental regulation is still in the "passive governance" stage. With the deepening of pollution, the governance will be strengthened, so the overall level of environmental regulation shows a certain inhibition. Fourthly, foreign direct investment (Fdi) has significantly inhibited the convergence of the YREB's EWP, but its spatial spillover effect is positive. The possible reason is that there are huge differences among cities in the YREB. Some developed cities have attracted foreign capital early, and the role of foreign capital has gone beyond the stage of advanced management concepts and technologies to accelerate the convergence of EWP to a high value. However, for some underdeveloped regions, the advanced technology and management level brought by foreign investment can continuously spread from developed regions with the flow of personnel and resource elements, thus promoting the EWP of underdeveloped regions to converge to a high value, so its spatial spillover effect is positive. Fifthly, transportation accessibility (Trans) has a generally positive effect on the overall EWP in the YREB. The improvement of transportation will help guide the efficient distribution and flow of various elements and expand the living space of residents. Under the effect of agglomeration and scale, social security and public service facilities will develop in a convenient and reasonable direction, and this sharing will greatly enhance people's sense of happiness. Finally, the impact of governmental intervention (Gov) on the convergence of the YREB's EWP is not significant, which does not mean that the governmental intervention has no impact on the EWP. Rather, government intervention cannot promote the EWP in the YREB to converge to high or low values.
## Absolute β-convergence
# Conditional β-convergence analysis
## Variables
## Robustness test
For testing the authority of the regression outcomes, this research replaces the different spatial matrices. From the regression outcomes in [fig_ref] Table 8: Conditional β convergence estimation results of two spatial weight matrices [/fig_ref] , the YREB's EWP has a significant conditional convergence trend under the different spatial matrices. At the same time, the significance level of other control variables on the convergence of EWP is not much different from the results in [fig_ref] Table 8: Conditional β convergence estimation results of two spatial weight matrices [/fig_ref] , which shows that the spatial econometric model that was designed is robust. Note: ***, ** and * refer to significant levels of 0.01, 0.05 and 0.1 respectively, with t-values in parentheses.
# Heterogeneity analysis
The YREB has a vast territory, with different conditions of urban development in the upstream, midstream and downstream, and the problem of unbalanced and insufficient regional development is also very prominent. At the same time, there are obvious differences in economic development, regional advantages, factor endowments and industrial structures among each region, which makes the influencing factors of EWP heterogeneous across different city-region in the YREB. Therefore, this paper further investigated the convergence of EWP in the upstream, midstream and downstream of the Yangtze River. The estimated results are shown in . Several points need to be noted. First, significant conditional β-convergences are presented in the YREB as well as in its lower, middle and upper reaches. However, there are different rates of convergence. The difference in spatial differentiation of the EWP is manifested through different convergence rates. Cities in the midstream of the Yangtze River have the fastest convergence rate, followed by upstream cities, and downstream cities are the slowest. Second, the spatial effects of different regions in YREB are different. ρ and λ in the midstream and upstream of the YREB are both negative, showing a negative spatial spillover effect, because the development of cities in midstream and upstream is extremely uneven, and most of those with a high EWP are concentrated in the developed provincial capital and central cities where most of the regional education, medical care, science and technology resources have been accumulated, and the siphon effect is significant. ρ and λ in the lower reaches of the YREB are positive, showing a positive spillover effect of "high interdependence", because the integrated regional development is actively promoted in YREB's downstream, especially the Yangtze River Delta. Talents, resources and technology in developed cities with a high EWP can be circulated to the developing cities, thus driving their EWP to converge to high values. Third, the spatial differentiation in the upstream, midstream and downstream of YREB is manifested through different convergence rates, and the differences in convergence rates in different regions are mainly driven by the combined effects of industrial structure, technology input, urbanization, environmental regulation, foreign direct investment, transportation accessibility and other factors. Different driving factors have significant heterogeneous impacts on the EWP's convergence . For instance, technology input inhibits the convergence of the EWP in midstream regions, the city has significant negative effects on the EWP in upstream regions, and transportation accessibility has significant heterogeneous effects on EWP. Obviously, the regional heterogeneity is caused by the different resource endowment, the social and economic development and the characteristics of the main stages of development in different regions. . Analysis results of heterogeneity of the EWP based on regional differences.
## Variable
Lower
# Discussion and conclusions
Based on the "3C" analysis framework of "classification, coordination and collaboration", this research developed a comprehensive appraising indicator system of inputs and outputs that considers economy, environment and society simultaneously for measuring the efficiency of regional EWP. Multi-source data were integrated and the super-efficiency SBM model was used for measuring EWP of 108 cities in the YREB from 2009 to 2019. Then, the Dagum Gini coefficient decomposition method was used for exploring the integral and regional variations in EWP, and the SPDM of convergence was employed to explore driving factors of their spatial differentiation.
The study identifies that there are gradient differences in the EWP of the YREB from the perspective of regional differences, and this difference has gradually narrowed over time, confirming the finding by Bian et al. that there are regional variations in China's EWP [bib_ref] Have cities effectively improved ecological well-being performance? Empirical analysis of 278 Chinese..., Bian [/bib_ref]. In addition, it reveals the existence of regional differences in the EWP, and finds that hypervariable density has always played a primary role in the integral difference of the YREB's EWP, indicating that underdeveloped cities in developed regions have lagged behind. From the angle of inter-regional differences, Wang et al. studied the differences in EWP based on provincial panel data, and found that the internal variations of the EWP in the eastern region were the maximum [bib_ref] Research on Spatial Unbalance and Influencing Factors of Ecological Well-Being Performance in..., Wang [/bib_ref]. This paper offers new insight into regional differences at the urban scale. EWP differences in the upper reaches are the largest among all regions, reflecting the prominent imbalance of the development of their internal EWP. The difference of the lower reaches is the smallest, which indicates that the steady progress of the integration process in the Yangtze River Delta will help narrow the gap in EWP across the cities.
Second, the spatial-clustering features of the YREB's EWP have been analyzed through ESDA. During the sample study period, the Global Moran's I in all years was greater than zero. It is concluded that the spatial distribution of the YREB's EWP has obvious spillover effects, and its distribution shows "H-H" and "L-L" agglomeration features. This finding confirms the spatial distribution characteristics of EWP observed at the national scale by Bian et al. [bib_ref] Spatial and Temporal Evolution and Driving Factors of Urban Ecological Well-Being Performance..., Bian [/bib_ref] , indicating a cross-scale similarity in spatial clustering of the EWP.
Third, some scholars have investigated the evolution of resource allocation efficiency, green total factor productivity and carbon emission rate from the perspective of spatial convergence [bib_ref] What drives technology convergence? Exploring the influence of technological and resource allocation..., Jeong [/bib_ref] [bib_ref] Convergence of green total factor productivity in China's service industry, Wu [/bib_ref] [bib_ref] Spatial Convergence of Carbon Productivity: Theoretical Analysis and Chinese Experience, Sun [/bib_ref] , but few scholars use spatial convergence models to study the differences in EWP. By constructing a spatial panel Durbin convergence model, this study proposes and verifies the difference in the convergence rate of the EWP and the main influencing factors in the upstream, midstream and downstream of the YREB. It also supplements and improves the research contents of EWP from the angle of spatio-temporal evolution, offering a novel method to relevant scholars to continue to engage in this research.
Fourth, in explaining the influence of causal factors on EWP, the extant studies often employ traditional non-spatial econometric paradigms and the spatial auto-correlation is often ignored. In fact, the evident spatial agglomeration and spatially correlated distribution of EWP require us to consider the spatial spillover effect in analyzing its changes and driving forces; otherwise, it will lead to biased analytical results. Ma et al. [bib_ref] Can Green Economy and Ecological Welfare Achieve Synergistic Development? The Perspective of..., Ma [/bib_ref] pointed out that industrial structure, urbanization and government regulation are conducive to improving the EWP, but the use of foreign capital will have a negative effect. The findings of this study confirm that these factors not only influence the EWP but also how it converges over time. In particular, its spatial effect cannot be ignored in modeling spatial and temporal characteristics of the EWP. Specifically, our study indicates industrial structure and technology investment can promote the convergence of the EWP to a high value, but their spatial dispersion effects play an opposite role. Further, foreign capital and environmental regulation can have a significant restraining effect on the convergence of EWP, but the spatial effect of neighboring foreign capital is positive and the spatial effect of environmental regulation of neighboring cities is negative. These results provide an enhanced understanding of how spatial interactive effects of the influencing factors on EWP may behave differently from their non-spatial effects, indicating the necessity of using spatial models in future research.
Fifth, the heterogeneity analysis is another highlight of this paper. The existing research on EWP convergence tends to treat the study area as a unified homogeneous whole and its internal differences are often ignored [bib_ref] A new class of composite indicators for measuring well-being at the local..., Ciommi [/bib_ref]. This study reveals that the impact size and direction of influencing factors vary across different regions within the study area of the YREB. The heterogeneity of the YREB is identified by dividing it into three parts of upstream, midstream and downstream. The results confirm that the direct and spatial effects of different influencing factors are heterogeneous, and this heterogeneity is also evident in the EWP convergence speed. The cities in the midstream of the Yangtze River converge fastest, followed by the cities in the upstream, and the cities in the downstream are the slowest. From the angle of heterogeneity, this paper offers a novel angle to the study of the EWP.
The present research still has some limitations. First, objective indicators were selected based on the availability of data in the selection of indicators related to EWP. This study does not include subjective well-being indicators that should be considered in future studies. Second, the convergence of EWP was the outcome of the interactive impacts of many factors. This study only considered the direct impact of the influencing factors. The indirect and interactive effects of these factors need to be further studied. Third, while this study reveals the association of the influencing factors with the convergence of EWP, future studies may need to explore the causal mechanism of these factors using a system dynamics model.
## Policy implications
There are several policy implications from the findings of this research. First, the regional imbalance of urban development in the YREB should be focused on. It is necessary to strengthen the spreading effects from central cities to surrounding cities, reduce the backwash effect, and stop the spatially polarizing development pattern in the YREB. Global Moran's I shows that the distribution of EWP is characterized by a remarkable regional interdependence. Therefore, cities or regions in geographical proximity and complementary development modes should be encouraged to formulate coordinated development plans and strive to realize the free flow of production factors, product outputs, technologies, and labor forces within the region. Because the spatial spillover effect cannot be ignored, the local government must consider the impact of surrounding areas, especially in the developed regions. In formulating relevant policies, the impact on the surrounding under-developed areas must be considered. Furthermore, it is required to establish a long-term cooperation for coordinated development strategies between the developed and the developing regions in order to promote coordinated sustainable regional development. Second, due to the differences in the convergence rate of the EWP in the upstream, midstream and downstream of the YREB and the different influencing factors, the local government should propose and follow a targeted development path according to the local characteristics of each region. As far as the lower reaches are concerned, social inequity, income gaps and other aspects of human and ecological wellbeing should be focused on achieving coordinated regional sustainable development. For the middle reaches, it is urgent to speed up the upgrading of industrial structure, eliminate backwards production capacity, increase attraction to talent, provide corresponding support policies, and focus on people's living standards in the development process. For the upper reaches, it is necessary to further expand foreign direct investment, increase infrastructure construction, and make full use of their own ecological advantages to transform them into ecological and economic advantages.
[fig] Figure 1: "Classification-Coordination-Collaboration" Analysis Framework of EWP. [/fig]
[fig] Figure 2: Distribution map of EWP in 2019 (a) and its dynamic changes (b) in the YREB. [/fig]
[fig] Figure 3: Regional differences of the EWP in YREB. [/fig]
[fig] Figure 4: Inter-regional differences of the EWP in the YREB. [/fig]
[fig] Figure 5: Spatial differentiation (a) and contribution rate (b) of EWP in the YREB. [/fig]
[table] Table 1: Spatial weight matrix types. [/table]
[table] Table 2: Appraising indicator system of the EWP in the YREB. [/table]
[table] Table 3: Control variables and their descriptive statistics. [/table]
[table] Table 4: Moran's I of overall EWP during 2009~2019. [/table]
[table] Table 5: Spatial panel econometric model test. [/table]
[table] Table 6: Absolute β-convergence estimation results. [/table]
[table] Table 8: Conditional β convergence estimation results of two spatial weight matrices. [/table]
|
Applications of water molecules for analysis of macromolecule properties
a b s t r a c tWater molecules maintain proteins' structures, functions, stabilities and dynamics. They can occupy certain positions or pass quickly via a protein's interior. Regardless of their behaviour, water molecules can be used for the analysis of proteins' structural features and biochemical properties. Here, we present a list of several software programs that use the information provided by water molecules to: i) analyse protein structures and provide the optimal positions of water molecules for protein hydration, ii) identify highoccupancy water sites in order to analyse ligand binding modes, and iii) detect and describe tunnels and cavities. The analysis of water molecules' distribution and trajectories sheds a light on proteins' interactions with small molecules, on the dynamics of tunnels and cavities, on protein composition and also on the functionality, transportation network and location of functionally relevant residues. Finally, the correct placement of water molecules in protein crystal structures can significantly improve the reliability of molecular dynamics simulations.
# Introduction
Life began to evolve in an aqueous milieu, and the unique properties of water determine the chemistry of all living organisms. Water is a ubiquitous and essential substance in cells, accounting for about 70% of their mass. It is not only the environment for biological processes, but also an integral part of them. At a macromolecular level, water contributes to biomolecules' formation and their stability, dynamics and functions. Water serves as a reaction reagent or mediates ligand-protein and protein-protein interactions. Water molecules are small enough to penetrate a macromolecule's core, to stabilise its native structure and also to participate in processes occurring in the protein's core. X-rayand neutron diffractioncrystallography provide an insight into the spatial distribution of water molecules in the vicinity of biomolecular surfaces and confined regions such as active sites, pockets and cavities. Depending on the crystal quality, atomic resolutions can be achieved. Protein structures deposited in the Protein Data Bank (PDB)contain an abundance of information, i.e., alternative conformations of amino acid side chains and potential rearrangements of protein compartments. about water molecules' positions is usually incomplete or can be strongly influenced by experimental conditions. Therefore, it is unclear how closely the distribution of crystal water molecules resembles the native conditions of the biomolecule. Nuclear Magnetic Resonance (NMR) is useful for discovering the hydration properties of proteins, especially their dynamics. Unfortunately, this technique cannot provide information about the threedimensional structure of the hydration sites, and its time scales are shorter by an order of magnitude than the residence times of water molecules.
The limitations of experimental methods can be overcome by computational techniques. Ab initio and DFT (Density Functional Theory) methods can be used for a precise description of a reaction mechanism, including the contribution of water. Molecular dynamics (MD) and Monte Carlo (MC) simulations provide a detailed atomic description of a biomolecule and a solvent, along with their dynamics. These simulations, however, do not tackle many equilibrium and long-time-scale kinetic properties.
The increasing awareness of the significant role of water molecules has given rise to a range of software focused on the analysis of water molecules' behaviour. Recent reviews focused on virtual screening strategies describe several docking software applications that are capable of utilising information related to water molecules. This paper provides a review of the available computational methods that employ water molecules for the analysis of macromolecules' properties and structure dynamics. In the first part, we provide an overview of the techniques used for the prediction of water molecules' locations. The following chapter describes the water sites that may participate in ligand binding. Next, water molecules are analysed in terms of ligand transportation and the detection of tunnels and cavities. For all three chapters, a list of software along with their functionality and/or their characteristics and principles is provided. The last chapter comprises conclusions and general remarks, as well as perspectives on the further development of software.
## Software for protein hydration analysis
Water molecules not only maintain the functions of proteins but also stabilise their native structure. The presence of water molecules in proteins' internal cavities is conserved among homologous proteins families, as well as the key residues are. It was shown, by reducing the amount of water during crystallisationor by using mutants of particular proteins, that internal water molecules contribute to the structural folding and the stability of enzymes, ion channels, proton pumps and other macromolecules. However, as we pointed out above, the experimental results are insufficient and can be inconsistent with each other. The water molecules inside a protein's structure may also be distorted, and their position may depend on the orientation of a particular side chain. They may also be trapped inside a protein's cavity due to a process of large conformational changes.
The residence time of a water molecule buried in an internal cavity or trapped in a narrow cleft depends on its location and connectivity with the bulk solvent. Fast minimisation and short equilibrium stages can provide insufficient or inaccurate solvation of the protein interior and can bias the results. Therefore, it is important to fill the internal cavities with water molecules precisely, prior to running MD simulations. Lengthening the minimisation and equilibration procedure can provide sufficient exchange of water molecules between the surroundings and the protein interior; however, it is strongly system-dependent. Application of software developed to place water molecules into a protein's cavities and its surroundings is proposed as an alternative strategy, especially for systems with large interior volumes, homology-modelled proteins or proteins with mutations introduced inside their cores .
Three different methodshave been implemented for the placement of water molecules in a protein's interior: i) based on the docking of water molecules, such as Dowserand Water-Dock, ii) based on the reference interaction site model (RISM), such as 3D-RISM, GAsoland Placevent, and iii) based List of available software to predict water molecules' positions and orientation.
## Software
Testing set Accuracy* Remarks
Docking-based Dowser14 crystal structures of OppA; D-and K-channels of cytochrome c oxidase; Photosystem II 63% Not available Dowser+74% Not available Dowser++85% Dowser++ standalone link WaterDock14 on the assumption that internal water molecules are conserved among similar proteins' structures (PyWATERand ProBiS H2O. The docking-based methods assume that the protein structure is the target and the water molecule is the ligand. Both Dowser and WaterDock utilise the commonly available docking software -AutoDock Vina. These methods are fast and provide accurate positioning of the water molecules determined by crystallography. The water molecule docking algorithms have improved over new software versions, and for the Dowser software 'generations', the average accuracy of their predictions have increased from 63% in Dowser, to 74% in Dowser+ and up to 85% of the water molecules in Dowser++, when compared to high-resolution crystallographic structures. WaterDock software presented a higher accuracy of crystallographic water molecules' prediction than Dowser ++: it was 88% for the original WaterDock and 91% for WaterDock 2.0; however, it should be kept in mind that there were some differences in the details of the accuracy measurements described in the original publications. Along with their ability to predict water molecules' positions, both WaterDock software releases are also able to determine if water molecules are displaced or ordered. WaterDock 2.0 comes with an easy-to-use PyMOL plugin.
The RISM theory is used for calculating the distribution of solvent molecules around a solute and has its roots in statistical, mechanical integral equation theories (IET) of liquids. Due to the fact that the distribution calculated by 3D-RISM theory is continuous, it is difficult to directly examine specific solvent interactions, especially when they are numerous. 3D-RISM has been successfully used to locate water molecules in proteins as compared to experimentand simulation, to calculate hydration free energiesand to predict fragment and drug positions. The Placevent algorithm gave an average error for water molecules' positions of about 0.5 Å. The GAsol software, in which the 3D-RISM theory was combined with a genetic algorithm and a desirability function, showed the highest accuracy, with 93.4% of the predicted water molecules within 2 Å from their crystallographic positions. Generally, RISM-based methods for water molecules' prediction are slower than docking-based ones, and the computational time is system-size-dependent. However, they can be more accurate, especially for complex systems (e.g., metalloproteins, proteins equipped with large cavities or in complexes with nucleic acids). Moreover, it was shown that the RISM theory may break down in larger systems and systematically underestimates the partial molar volume (PMV) of amino acids.
As an alternative to the methods based on the physicochemical properties comes a simple similarity-based approach, implemented in PyWATERand ProBiS H2O, which both superimpose crystallographic structures similar to the target protein and cluster the positions of conserved water molecules inside the protein cavities. ProBiS H2O is the first software that utilises the Pro-BiS algorithmto perform local superimposition of the detected conserved water molecules. It also reduces the bias introduced by comparing similar protein structures or structures in different conformations than the query protein. PyWATER searches for similar structures using the PDB database. The accuracy of such an approach strongly depends on the number, similarity and quality of related structures. Generally, ProBiS H2O gives fewer clusters with more tightly packed water molecules in comparison to PyWATER due to the clustering algorithms used (PyWATER uses hierarchical clustering, while ProBis H2O uses a Python implementation of 3D-DBSCAN (Three Dimensional Density-Based Spatial Clustering of Applications with Noise)). In addition, PyWATER stores information on the degree of conservation of each water molecule cluster with related atom numbers of water oxygen atoms from the superimposed protein structures.
All the tools mentioned above provide relatively fast, intuitive and accurate modelling of the water molecules in low-quality crystal structures and thus provide a more accurate starting point for an MD or MC study. Their usage can be also recommended for in silico prepared mutant structures, where substitution of residues enlarges or significantly reshapes internal cavities. However, the user should keep in mind that none of these methods assume flexibility or large conformational changes in the target structure. Also, in the case of preparing a very demanding protein structure, i.e., an unrefined homology model or a sole representative of a particular protein family, the user should be encouraged to avoid similaritybased or docking-based methods and focus on the RISM-based software to properly sample the positions of water molecules.
## Software for water site detection and ligand binding analysis
Studies of protein-ligand interactions are crucial for a better understanding of the mechanisms of biological processes and their regulation. Water-mediated interactions were found in 85% of 392 analysed protein-ligand complexes. Structural and thermodynamic data indicate that water molecules mediate interactions between proteins and ions, substrates, cofactors, inhibitors and other macromolecules. Water molecules are placed methodically within the surroundings and inside the protein, and display a particular structure characterised by the presence of hydration or water sites -regions of high-water density. They act as locations that attract water and can be used to describe water behaviour around chemical molecules. The hydration/dehydration balance is relevant for protein-ligand formation and binding affinity, involving both entropic and enthalpic contributions. During the binding process, water molecules can either be displaced or conserved, bridging the protein-ligand interactions in the latter case. The presence of water molecules in protein binding sites may imply different effects on the energy, entropy and enthalpy of the system. Depending on the situation, such effects may be favourable or unfavourable. For example, in a case where water molecules are trapped in a hydrophobic cavity filled by residues that cannot make appropriate hydrogen bonds, the enthalpic contribution is unfavourable. An opposite situation occurs when water molecules are engaged in forming hydrogen bonds to hydrophilic residues, and here the enthalpic effect may be favourable. The displace-ment of such water molecules can contribute to the binding free energy, impact affinity during ligands' association and govern enthalpy and entropy partitioning, according to the properties of the individual water molecules compared to those in the bulk phase. Developing a ligand with a high binding affinity towards a specific target is one of the most important steps during the entire drug design process. Thus, a lot of effort is focused on the prediction of whether a water site should be displaced and whether this would cause an increase in a ligand's affinity.
Different approaches, both experimentally-based (i.e., location in crystal structures and B-factors) and knowledge-based (i.e., free energy, water's contribution to binding free energy, entropic contribution), have been reported to assess information about water sites. One of the very first experimentally-based software programs, GRID, uses a regular array of 'grid points', established throughout and around the protein, to calculate the energetics of water probes inside a macromolecular binding site. GRID places a chemical probe and calculates an empirical interaction energy at all grid points. An approach using crystallographic B-factors to determine which water molecules in a protein's structure are likely to be displaced has been implemented in Consolvand WaterScoresoftware. Using geometric criteria can also indicate the positions of water molecules mediating proteinligand interactions. Such a procedure has been included in the AcquaAlta programfor estimating the propensities of ligand hydration. In HINT software, the Gibbs free energy of noncovalent interactions is based on van der Waals interactions and partial atomic partition coefficients. A knowledge-based approach has been implemented, e.g., in AQUARIUSor AQUARIUS2software. The probable positions for hydration sites are predicted based on solvent distributions surrounding particular amino acids derived from the analysis of a protein's structure. However, most of the software applications mentioned above are not currently used or are used very rarely. This is due to the fact that another class of methods, describing the thermodynamic properties of water by analysing data from MD and MC simulations, became very popular and easy to use. Most of the recently developed tools are based on Inhomogeneous Fluid Solvation Theory (IFST) derived in 1998 by Lazardis. IFST is a statistical mechanical method that calculates free energy differences from short MD or MC simulations by quantifying the effect of a solute acting as a perturbation to bulk water. The solute may be different molecules, such as proteins, peptides or other molecules. One of the major advantages of IFST is that the system is spatially decomposed to consider the contribution of specific regions to the total solvation free energies. The contributions of each individual water molecule to the enthalpy are calculated by computing the average interaction energies, whereas the contributions to the entropy are calculated from intermolecular correlations. The Gibbs free energy equation can be used to calculate the contribution to the free energy from the enthalpy and List of software for analysis of ligand binding and drug design with respect to the water molecules in the binding cavity.
## Software
Testing set Functionality Remarks
Input -a single structure AcquaAltatrypsin entropy. The result is then compared with the contribution of one water molecule to the free energy of bulk water, again calculated using IFST. The results of IFST depend on the accuracy of the force field and the water model that was used: a detailed discussion can be found elsewhere. IFST is implemented in WaterMap, GIST, WATCLUST, STOWand WATsitesoftware. Methods based on IFST are limited to the analysis of highoccupancy hydration sites and therefore omit solvent molecules found in lower density regions. Some of the software listed in perform calculations based on static input (WATsite, WATGEN, FLAP, Water-FLAP, SZMAP, WRAPPA, WaterScore, AcquaAlta, JAWSand WScore, while some of them rely on data obtained from MC or MD simulations (AQUA-DUCT, GIST, WATCLUST, STOW, SSTMap, WaterMap, Water-swap, WatMD, SPAMand AquaMMapS. In static structure-based software, hydration and water sites are determined by investigating potential binding sites through the placement of water probes. The only exception is WATsite, which conducts MD simulations based on the static input structure. In simulation-based software, the system is simulated with explicit water molecules, which are free to explore the system's space. These water molecules are then clustered in hydration sites, and their thermodynamic properties are calculated. Some of the software only identifies the water sites, without any further information, while some can estimate the binding free energy and the corresponding thermodynamic components for water molecules in the binding sites. In a very recent paper, the authors combined WATsite data with neural networks and deep learning to significantly improve the speed of water site identification and the calculation of the free energy contributions. The authors claim that such an approach will allow the inclusion of solvation components, such as water-mediated interactions or enthalpically stable hydration networks in proximity to the protein-ligand complex, in structure-based ligand design.
## Software for tunnel detection and transportation phenomena analysis
The intramolecular voids inside a protein structure, such as cavities, tunnels, channels and pores, are often important for protein functions. While we have already shown the importance of cavities, this section focuses on the function of tunnels and channels. For proteins with a buried active site, tunnels facilitate substrate entry and enable product egress. Tunnels, as well as the whole protein structure, should not be seen as rigid bodies. In fact, a reasonable degree of flexibility is often required to maintain the catalytic reaction. The geometry and amino acid composition of a particular tunnel determine the shape and chemical properties of a potential ligand. Tunnels are also equipped with a much more sophisticated mechanism of small molecule discrimination -gates. Gates are capable of controlling substrate access to the active site, preventing solvent access to particular protein regions and synchronising processes occurring in distant parts of the protein. Tunnels, pores, gates and cavities constitute a dynamic network inside a protein. Therefore, for proper tunnel detection, a single crystal structure of a protein may be insufficient. MD simulations provide a picture of a protein's movements in an aqueous solution. Reasonably long simulations give insights into the dynamics of the tunnel network. Well-defined tunnels allow fast water exchange over a time of about 10 À9 s, while transient tunnels extend the required time up to 10 À3 s. In comparison, the exchange time of water molecules at the protein surface with bulk ones is in the sub-nanosecond range. Therefore, from the computational point of view, the lengths of the required molecular dynamics simulations depend on the studied system. In the case of buried active sites linked with the solvent via a network of tunnels, hundreds of nanoseconds are enough to provide good sampling, whereas to observe the exchange of deeply buried water molecules with the bulk solvent up to as much as tens of milliseconds are necessary. Shorter simulations can provide information about the potential pathways of such an exchange and can suggest mutations that can open an alternative tunnel. The second parameter which might influence the required length of simulations is the frequency of gating phenomena. Gates defined by a single amino acid's rotation require shorter experiments than those defined by, e.g., loops or controlled by proteins' breathing motions.
The first tunnel detection software used a geometry-based approach to identify 'empty spaces' inside protein structures. The most successful ones, such as Mole 2.0, CAVER 3.0, and CAVER Analyst 1.0 and 2.0, are widely used by the scientific community, predominantly to describe tunnels identified in crystal structures. The most successful strategy employs the construction of a Voronoi diagram to detect and describe voids within the macromolecule. Using a defined probe radius and internal cavity identification, the software is able to detect tunnels providing access from the selected area to the surroundings. Such a strategy assumes that the tunnels are a summation of connected cavities and is very often used for the analysis of single crystallographic structures. The structural information obtained on such a basis is mostly incomplete, due to tunnels' flexibility. Moreover, using spherical probes for tunnel exploration provides only an approximation of tunnels to tubes with symmetrical diameters and thus prevents analysis of tunnels' asymmetry. It is also difficult to analyse the regulation and direction of the solvent flow, as well as the contribution of tunnels to an enzyme's activity and selectivity. Some of these weaknesses were targeted in 2014 by a non-spherical approach by Benkaidali et al.; however, due to its complex implementation the tool was rarely used. Results provided by geometry-based tunnel detection software were unable to answer questions about solvent flow direction and how tunnels contribute to this.
To analyse the solvent flow direction and tunnels' contribution to this parameter, we need to concentrate on solvent/ligand analysis. Several different methods have been implemented, based on very diverse approaches. The first attempt was made in 2008 by Bidmon et al., who introduced the Visual Abstractions of Solvent Pathlines method. The pathways of solvent molecules passing through the particular region of interest (so-called ROI) were pre-processed and visualised as Bézier curves. The next attempt at such an analysis was carried out in 2010 by Vasiliev et al.. Their streamline tracing method was applied to photosystem II and was used to visualise water flux in particular regions of the protein. However, the results of the calculations are hard to interpret, and only a few applications of this method can be found in the literature. In 2014, Benson and Pleiss proposed a solvent flux method to study water influx in the Candida antarctica lipase B protein cavity from the triglyceride-water interface. They introduced a solvent concentration gradient and the reorientation and rescaling of the velocity vectors of selected water molecules in order to accelerate the influx and increase the probability of rare events in the study. Similarly to widely used strategies (aMD, REMD, SMD and RAMD), it was applied to investigate rare events in a reasonable computing time range (e.g., it overcame the significant energy barriers of slow biophysical events). In contrast to known methods, this technique allowed the flow of multiple molecules, including the selected solvent molecules, to be precisely investigated during a single simulation. Since artificial external forces are introduced to classic MD simulations, one could be concerned about misleading biases and the complicated proto-col that is dependent on the system. In 2014 another software program emerged as a GROMACS plugin, called trj_cavity. Trj_cavity is capable of cavity and tunnel identification together with time-dependent calculations of their volume and solvent capacity. In the vast majority of research papers, trj_cavity is used only for the identification of cavities and calculating their volumes and occupancy, while only one study was found where the authors used trj_cavity to actually trace ligands. The existing gap between tools searching for tunnels and pathways, and advanced tools for accelerated water flux investigations was filled in 2017 List of software and methods for tunnel detection and transportation phenomena observation.
## Software or method applicability remarks
System Functionality AQUA-DUCTMus musculus epoxide hydrolase water molecule tracking can be used also for cavity and hot-spot detection Standalone: link D-amino-acid oxidasetunnel and gating residue detection
Pyrococcus furiosus phosphoglucose isomerasewater molecule tracking and occupancy analysis in the internal cavity Claudin-2 ion channelion transportation pathways identification Solvent flux methodCandida antarctica lipase B identification of water access pathway; hot-spot identification based on an artificial gradient. Code not available. Streamline tracingphotosystem II fibre tracing; tunnel detection; gating residue (access control points) identification visual analysis only; code available on request squalene-hopene cyclasechanges in water flow after introducing amino acid substitution trj_cavity [108]* Der p 2 protein; TM pore; pullulanase generating the trajectory of discovered cavities, quantification of time-dependent cavity volume, solvent presence inside a particular cavity; tunnel detection implemented in GROMACS: link polydicyclopentadiene; herkinorin; glycidoxypropyltrimethoxy silane; mammalian translocator membrane protein; human G-protein coupled receptors; amyloid fibrils; sperm whale myoglobin; 07A metalloprotease; human erythrocyte anion exchanger 1 (Band 3 protein); amorphous silica; full-length TLR4 dimer; profilin; human serum albumin; laccase; dengue capsid protein (C protein); horseradish peroxidase; lactoperoxidase; OmpC-MlaA complex; MATE transportercavity analysis cholesteryl ester transfer protein; acyl carrier proteinstime-dependent cavity analysis mouse myoglobinanalysis of the movement of ligands, movements within the cavities and tunnels of proteins Visual Abstractions of Solvent PathlinesTEM b-lactamaseidentification of the role of water in gating loop flexibility visual analysis only; code not available Watergatehaloalkane dehalogenase mutants visualisation of water molecule trajectories visual analysis only; code available on request *For trj_cavity software only recent applications are presented (2017-2019). Information about currently unavailable software is in italics. by AQUA-DUCT, an easy-to-use tool facilitating analysis of the behaviour of water (and, if necessary, other solvent molecules) penetrating any selected region in a protein. AQUA-DUCT comprises a Valve module, which is capable of tracking water molecules, clustering their trajectories and enabling visualisation in PyMOL. The Valve module was used to investigate relatively small proteins, such as D-amino-acid oxidase (DAAO)and Pyrococcus furiosus phosphoglucose isomerase, as well as ion channels such as claudins. In contrast to the geometrybased approach, water tracking analysis provides information about tunnels' functionality, allows their permeability to be compared and facilitates the detection of the gating residues controlling access to the binding cavity. At the same time, Watergate, a software application for statistical overview of the overall solvent flow, water trajectory clustering, and visualisation was developed. The software programs using water molecules for tunnel detection are listed in . Tools based on water molecules as a molecular probe for tunnel detection (listed in can provide much more complex information about proteins than simple geometry-based methods. Since they are focused on the information provided by the solvent itself, they also take into account the physicochemical properties of the solute. Such information is useful for examining the effects of the introduced mutation on the solvent flow and thereby the enzyme's activity. By using the pathways of the solvent molecules, the user is able to identify the key residues important for the enzyme's activity and selectivity, and the amino acids that contribute to gating residues and control small ligands' entry/egress. These tools can additionally facilitate the description of cavity shape evolution during simulation time, which can be used for inhibitor design or hot-spot detection for substrate specificity modification, and also the identification of residues distant from the active site which contribute to the activity and selectivity, and thus can be considered as a safe alternative to smart mutant library design. However, it should be kept in mind that to properly sample events such as substrate entrance, product release or the exchange of the trapped solvent molecules with the bulk solvent, the analysed simulations must be of reasonable length and conducted in physiological-like conditions (please see the Summary and outlook section for more details).
## Summary and outlook
The important role of water molecules in structural biology is reflected by a large number of different software programs dedicated to various types of water-molecule-based analysis. Most of the software presented here is focused only on particular aspects of water's presence in a macromolecular structure, such as its contribution to protein stability, ligand binding and drug design, or cavity and tunnel description.
Among all the described software, the role of software in optimising water placement inside proteins' cavities is probably the most underestimated, although placing water molecules is not a trivial task. Three different strategies, RISM theory, the docking of water molecules, and the analysis of conserved water molecules among similar proteins, are used and complement each other. The RISM-based software probably provides the most accurate model; however, it is time-consuming. Docking-based methods are the fastest; however, they may provide biased results for systems, such as metalloproteins, proteins with large cavities and protein-nucleic acids complexes, that are problematic for such software. Both methods can be considered for the prediction of differences in water rearrangement when mutant structures are designed. The third strategy requires a collection of similar structures and may not be sensitive enough to provide correct predictions when a single mutation occurs. Therefore, depending on the investigated system, different strategies are optimal for water placement and can provide reliable starting points for molecular dynamics studies. As already stated, the water placement method should be considered as a standard approach for homology models or structures with introduced amino acid modifications.
Concerning the role of water in the description of ligand binding, the most commonly used methods are based on IFST. Given a water site, the software can predict how much free energy is gained (or lost) by displacing the water molecules that occupied the potential ligand-binding site. The solvent contribution cannot be neglected, as was shown in several excellent papers, and therefore continuous progress in both the accuracy and parallel analysis of alternative states is highly desired. Binding enthalpies and entropies of water molecules may also be calculated based on Grid Cell Theory (GCT). This is a recently developed method for investigating hydration thermodynamics from a molecular dynamics or Monte Carlo standpoint. In the GCT approach, the density, enthalpy, entropy and free energy of water are evaluated for an arbitrary region of space around a system of interest. These parameters refer to the water molecules which enter a particular hydration site of the protein(s) from the bulk concentration. However, this theory has not yet been implemented in known software. The recent AQUA-DUCT version provides an approach combining information on water and cosolvent high-occupancy sites. It can be used for pharmacophore design and suggests directions for future software development. All of the abovementioned methods can provide additional support to drug design and provide more accurate results in comparison to methods neglecting water molecules' contribution.
In the third group of software, water molecules are used as a molecular probe to sample the 'empty spaces' in proteins during molecular dynamics simulations to detect tunnels and cavities. This field is so far monopolised by a widely used geometry-based approach which on the one hand is very simple, but provides rather approximate results. It neglects the tunnels' asymmetry and the physicochemical properties of the tunnel-lining amino acids. Therefore, it is difficult to use such software for analysis of tunnels' functionality. The alternative approaches presented in our review are a most diverse group of software utilising water as a molecular probe. Depending on the implemented algorithms, they provide information about local water flow changes (such as the streamline tracing method), changes in cavity volumes (trj_cavity) or can provide a holistic picture of water flow via tunnel networks and an approximation of the energy profiles of particular pathways (AQUA-DUCT). The utilisation of water for cavity and tunnel description removes most of the limitations of the standard approach. It seems that using water molecules as a molecular probe enables more sophisticated analysis of the substrate transportation network provided by tunnels, handling tunnels and cavities together and describing the protein interior in a holistic way as a single entity. However, so far it is hard to provide an estimation of how accurate they are. This problem is caused not only by difficulties in experimental verification of their findings, nor the question of how accurately the hydrophobic cavities can be described, but also due to the lack of benchmarks for the performance inside protein structures of the different water models used in MD simulations.
Protein engineering is one of the most promising, but still largely unexplored, fields of application for software focused on the analysis of water as a molecular probe. So far, most of the examples of such studies are focused on understanding the changes introduced by mutant proteins' construction. However, there are papers showing the potential applicability of the water-based approach for hot-spot detectionand mutant library design. Successful verification of such a strategy can greatly facilitate protein engineering and provide an interesting and easy-to-apply technique. Also, using the information on the water molecules' (or small ligands, or other types of solvent molecules) tracking, the user gains knowledge on a protein's internal architecture, which might be used to develop a successful strategy for further modifications; for instance, to search for more potent inhibitors which will explore previously unused cavities, or to improve the protein's activity and/or selectivity by adjusting the pathways leading to and from the active site.
Since the analysis of water-mediated interactions has become of greater interest, we hope that the number and quality of software programs using water molecules to analyse macromolecules' properties will only increase. However, progress in this promising area cannot be achieved without the further joint efforts of theoreticians and experimentalists. One needs to consider that the majority of the tools described above depend on water models and force fields (e.g., tools based on Inhomogeneous Fluid Solvation Theory or benefiting from MD simulations). Both force fields and water models are being constantly upgraded to provide more accurate descriptions of studied systems. For example, the most recent papers of Huang, et al. provide force fields which can be used for both ordered and disordered proteins. Recent four-point water models have improved the description of its thermodynamic properties; however, water molecules' non-bonded interactions still require validation. Nevertheless, the majority of computational studies employ simple non-polarizable models of water (e.g., TIP3P, SPC/E, TIP4P) and assume that they will describe water molecules in macromolecular surroundings equally well as in the bulk water. Unfortunately, there is no study that can confirm such a presupposition, simply due to the limited access to experimental data providing insight into water's behaviour inside a protein's core. Moreover, even the benchmark analysis of a particular software data's dependency on the used parameters is very limited. As we mentioned above, the comparison of the IFST results obtained with different water models suggests that the quantitative application of IFST to biological systems is strictly modeldependent and has to be carefully analysed. Fortunately, several successful verifications of the findings guided by the software developed to analyse the behaviour and/or properties of water molecules have accelerated research in the field of protein research and each year bring to the scientific community new, optimised, versatile and reliable tools which greatly improve our understanding of nature. |
Interactions between Microbial Food Safety and Environmental Sustainability in the Fresh Produce Supply Chain
# Introduction
Fresh produce is being increasingly recognized as a source of foodborne outbreaks. The burden of such outbreaks includes economic losses, healthcare costs, loss of productivity, reductions in the quality of life, and mortality [bib_ref] Economics applied to food safety, Focker [/bib_ref]. Using data from foodborne outbreaks that occurred between 1998 and 2018 in the USA, the Centers for Disease Control and Prevention linked a significant proportion of the illnesses caused by three key pathogens (Salmonella, Escherichia coli O157, and Listeria monocytogenes) to produce (including fruits, sprouts, vegetable row crops, and seeded vegetables). According to a report from World Health Organization (WHO) and Food and Agriculture Organization (FAO), fruits and vegetables are among the main identified vehicles of foodborne Shiga-toxin-producing E. coli (STEC) illness. Regarding the economic losses, the study by The potential conflicts between microbial food safety and environmental sustainability at the pre-harvest level will be illustrated using mainly examples of food safety measures undertaken in the USA in the last two decades and their consequences. Particularly since the E. coli O157:H7 outbreak linked to California-grown spinach from 2006, fresh produce growers in the USA have faced conflicting demands regarding food safety and environmental preservation, with priority being given to the former [bib_ref] Constrained choice and ethical dilemmas in land management: Environmental quality and food..., Stuart [/bib_ref] [bib_ref] Comanaging fresh produce for nature conservation and food safety, Karp [/bib_ref] [bib_ref] When farmers are pulled in too many directions: Comparing institutional drivers of..., Baur [/bib_ref]. The measures taken for the sake of fresh produce microbial safety included the elimination of natural vegetation, reduction of the presence of wild animals in agricultural areas, avoidance of the use of manure-based amendments, and, more recently, the disinfection of irrigation water [bib_ref] Food safety and environmental quality impose conflicting demands on Central Coast growers, Beretti [/bib_ref] [bib_ref] Farm practices for food safety: An emerging threat to floodplain and riparian..., Gennet [/bib_ref] [bib_ref] Evolving food safety pressures in California's Central Coast Region, Olimpi [/bib_ref].
The following paragraphs provide details on the food safety measures applied and the conflicts with sustainability and detail current opinions regarding the outcomes and the way forward.
Farmers have been encouraged to remove natural vegetation to reduce the risks of wildlife intrusion [bib_ref] The new food safety, Leib [/bib_ref]. However, it has not been confirmed whether the elimination of non-crop vegetation reduces the presence of pathogenic bacteria in the crops. For example, al. [bib_ref] Comanaging fresh produce for nature conservation and food safety, Karp [/bib_ref] did not find an increased prevalence of pathogenic (enterohemorrhagic E. coli, Salmonella spp.) and indicator bacteria (generic E. coli) on leafy green vegetables grown near areas with non-grazed non-crop vegetation. Sellers et al. [bib_ref] Impact of field-edge habitat on mammalian wildlife abundance, distribution, and vectored foodborne..., Sellers [/bib_ref] , analyzing fecal samples from wildlife intruders (rodents), did not observe a higher risk of the presence of pathogenic microorganisms in agricultural fields (walnut, tomato) surrounded by hedgerows, compared to fields with controlled field edge vegetation. Smith et al. [bib_ref] Agricultural intensification heightens food safety risks posed by wild birds, Smith [/bib_ref] detected a higher presence of Campylobacter spp. in avian fecal samples from crop farms (brassica plants) located in landscapes with high mammalian livestock densities, compared to farms located in landscapes with larger areas of natural habitat. Fonseca et al. [bib_ref] Assessing the food safety risk posed by birds entering leafy greens fields..., Fonseca [/bib_ref] , analyzing local birds that inhabited near leafy green growing fields in the U.S. southwest, reported the absence of Salmonella spp. and E. coli O157:H7. The presence of vegetation barriers surrounding agricultural fields can have benefits for the environment and also limit the wind dispersion of pathogenic bacteria [bib_ref] The effect of vegetation barriers at reducing the transmission of Salmonella and..., Glaize [/bib_ref]. Therefore, recent studies raise doubts as to whether limiting the presence of non-crop vegetation in farmlands does lead to safer fresh produce.
The use of organic amendments has positive effects on the health of agricultural soils (e.g., on soil microbiota functional diversity) [bib_ref] Changes in bacterial community structure of agricultural land due to long-term organic..., Chaudhry [/bib_ref] [bib_ref] Long-term application of manures plus chemical fertilizers sustained high rice yield and..., Chen [/bib_ref]. On the other hand, the use of raw manure has been linked to a higher prevalence of pathogenic microorganisms in agricultural soils compared to the use of synthetic fertilizers [bib_ref] Agricultural practices influence Salmonella contamination and survival in pre-harvest tomato production, Gu [/bib_ref]. Avoiding the use of biological soil amendments is one of the preventive strategies that has been proposed and used to reduce the food safety risk of fresh produce [bib_ref] Evolving food safety pressures in California's Central Coast Region, Olimpi [/bib_ref]. However, recent studies suggest that using properly treated animal-based manure should be reconsidered. Devarajan et al. [bib_ref] Cascading effects of composts and cover crops on soil chemistry, bacterial communities..., Devarajan [/bib_ref] , analyzing corn growing fields, concluded that the application of appropriately managed poultry litter could lead to a lower risk of the presence of Salmonella spp. and Listeria monocytogenes in farms. Those authors suggest that this organic manure would stimulate the presence of a thriving pathogen-inhibiting microbiota in the soil. Gu et al. [bib_ref] Agricultural practices influence Salmonella contamination and survival in pre-harvest tomato production, Gu [/bib_ref] did not detect Salmonella spp. in samples from tomato fields fertilized using poultry litter ash. Further research should confirm whether the use of properly managed animal-based manure can enable the combination of waste reuse, the preservation of soil health, and an adequate level of food safety.
Current evidence does not demonstrate that organic farming provides produce that is less safe than that grown using conventional practices. However, the higher environmental sustainability of organic farming is questioned. Although it can achieve a local reduction in the environmental impact, the economic and social aspects (e.g., food security) make organic farming an inadequate alternative to the predominant conventional farming systems [bib_ref] Moving beyond organic-A food system approach to assessing sustainable and resilient farming, Röös [/bib_ref]. To maintain current produce supply levels using mainly organic agriculture would demand a substantial increase in the area of land devoted to agricultural activities, with the risk of leading to an increase in greenhouse gas emissions [bib_ref] The greenhouse gas impacts of converting food production in England and Wales..., Smith [/bib_ref].
Irrigation water is an important vector for the contamination of fresh produce (e.g., leafy crops) with pathogenic microorganisms [bib_ref] Irrigation water quality for leafy crops: A perspective of risks and potential..., Allende [/bib_ref]. Disinfection of irrigation water can be used as a preventive measure in those settings in which water presents a higher microbiological risk (e.g., when reclaimed urban wastewater is used for irrigation) [bib_ref] Pre-and postharvest preventive measures and intervention strategies to control microbial food safety..., Gil [/bib_ref] [bib_ref] Microbiological reduction strategies of irrigation water for fresh produce, Banach [/bib_ref] [bib_ref] Disinfection options for irrigation water: Reducing the risk of fresh produce contamination..., Dandie [/bib_ref]. However, risk-benefit assessments should consider potential negative effects regarding the presence of disinfection by-products (DBPs) in the crop. For example, the presence of chlorate (ClO 3− ) has been reported in leafy greens irrigated with water treated with chlorine [37], chlorine dioxide , and electrolyzed water . In the mentioned studies, the levels of ClO 3− in the crop were above the current maximum residue levels (0.7 mg/kg for leaf vegetables) allowed in the European Union [40] only when the irrigation water was treated with chlorine dioxide. The accumulation of disinfection residues in the soil and the potential alterations in the soil microbiota should also be considered. , for example, observed no relevant changes in the crop and soil microbiota in a baby spinach field irrigated with water treated using chlorine dioxide, suggesting that this treatment (as applied in their study) would be eco-compatible. Martínez-Sánchez and on the quality of capsicum seedlings grown in the nursery and found a reduced microbial load in the water (E. coli and total Enterobacteria) as well as a decrease in the mesophilic load of capsicum seedlings. [fig_ref] Figure 1: Potential environmental impacts [/fig_ref] summarizes the topics covered in this section. In conclusion, the effectiveness and the side-effects of the microbial food safety measures taken in the primary production step should be carefully assessed to determine opportunities for co-management of microbial safety and environmental sustainability.
irrigation water was treated with chlorine dioxide. The accumulation of disinfection residues in the soil and the potential alterations in the soil microbiota should also be considered. , for example, observed no relevant changes in the crop and soil microbiota in a baby spinach field irrigated with water treated using chlorine dioxide, suggesting that this treatment (as applied in their study) would be ecocompatible. Martínez-Sánchez and Aguayo [42] studied the effect of irrigation with ozonated water (0.35-0.40 mg/L of O3) on the quality of capsicum seedlings grown in the nursery and found a reduced microbial load in the water (E. coli and total Enterobacteria) as well as a decrease in the mesophilic load of capsicum seedlings. [fig_ref] Figure 1: Potential environmental impacts [/fig_ref] summarizes the topics covered in this section. In conclusion, the effectiveness and the side-effects of the microbial food safety measures taken in the primary production step should be carefully assessed to determine opportunities for comanagement of microbial safety and environmental sustainability.
## Post-harvest management in the fresh produce supply chain and interactions between safety and sustainability
## Structure of the fresh produce supply chain
Some studies assign an important share of the environmental impact of fresh produce (e.g., tomato, apple) to the transport stage . Current food transport systems cause the emission of greenhouse gases (e.g., methane, carbon dioxide, nitrous oxide), which are implicated in global warming. Transportation is also involved in other environmental impacts such as non-renewable energy use, terrestrial acidification, and freshwater eutrophication . Consequently, shortening the food supply chains has been proposed as a strategy to increase the sustainability of the food industry . However, while the positive effects of short food supply chains in social sustainability are clear, the impacts on economic and environmental sustainability are questionable . Furthermore, concerns have also been raised about the safety standards of local supply chains . According to Schmitt et al. , food safety is more closely monitored in products managed in large quantities in the global supply chains compared to that in local
## Post-harvest management in the fresh produce supply chain and interactions between safety and sustainability
## Structure of the fresh produce supply chain
Some studies assign an important share of the environmental impact of fresh produce (e.g., tomato, apple) to the transport stage . Current food transport systems cause the emission of greenhouse gases (e.g., methane, carbon dioxide, nitrous oxide), which are implicated in global warming. Transportation is also involved in other environmental impacts such as non-renewable energy use, terrestrial acidification, and freshwater eutrophication . Consequently, shortening the food supply chains has been proposed as a strategy to increase the sustainability of the food industry . However, while the positive effects of short food supply chains in social sustainability are clear, the impacts on economic and environmental sustainability are questionable . Furthermore, concerns have also been raised about the safety standards of local supply chains . According to Schmitt et al. , food safety is more closely monitored in products managed in large quantities in the global supply chains compared to that in local products. In the fresh fruit and vegetable supply chain, the size of the customer affects the safety management, with major retailers putting pressure on wholesalers regarding private certifications, thereby leading to greater safety . In contrast, some consumers attribute higher food safety to short supply chains . In developing countries, fresh produce companies oriented to the export market (therefore, involved in global supply chains) have more advanced management regarding food safety issues, as compared to smallholdings, which are oriented to . In many countries around the world, farmers' markets or local food markets are popular settings that facilitate consumer access to local fresh produce. Despite concerns regarding the food safety procedures in such markets [52], the potential higher prevalence of microbial contamination in the products sold in farmers' markets compared with other retailers remains controversial [47,53].
## Water reuse and food safety in the fresh produce industry
The industrial handling, conditioning, and processing of fresh produce have an important water footprint, due to their considerable water demand and the generation of large quantities of wastewater . Water can be used for cleaning, washing, disinfection and rinsing, transportation, blanching, cooling, or even heating the products [56-60]. Apart from water consumption, the use of water also involves the consumption of energy for cooling, heating, or pumping [61]. Mundi et al. indicated that one kilogram of processed fruit and vegetables entails the generation of 5 L of wastewater; its characteristics depend on factors such as the type of processed product, and the configuration and management of the processing lines. In the case of fresh-cut produce processing plants, 2 to 11 m 3 of good-quality water is consumed per ton of product [63], although a significant part of this water is commonly reconditioned and reused to reduce water consumption and wastewater generation [bib_ref] Water reconditioning and reuse in the food processing industry: Current situation and..., Meneses [/bib_ref]. However, environmentally beneficial water reuse can have consequences from the microbial food safety point of view, due to potential crosscontamination between batches [61]. Furthermore, it can also lead to the accumulation of disinfection by-products in the wash water [bib_ref] Microbial and chemical characterization of commercial washing lines of fresh produce highlights..., López-Gálvez [/bib_ref]. Water treatment needs to be optimized to enable water reuse, whilst reducing the microbial and chemical safety risks [bib_ref] Association between bacterial survival and free chlorine concentration during commercial fresh-cut produce..., Luo [/bib_ref] [bib_ref] Chlorination management in commercial fresh produce processing lines, Tudela [/bib_ref]. In particular, reconditioning the water using physical treatments can be a sustainable alternative . In any case, the presence of chemical antimicrobials is required to reduce the risk of cross-contamination by maintaining continuous disinfection processes in the washing tanks [bib_ref] Guidelines to validate control of cross-contamination during washing of fresh-cut leafy vegetables, Gombas [/bib_ref]. Apart from water reuse, the recovery of useful compounds from the fresh fruit and vegetable processing wastewater has been suggested as an approach to increase the sustainability of the industry [bib_ref] Valorization of artichoke wastewaters by integrated membrane process, Conidi [/bib_ref]. However, the implementation of such a strategy faces different obstacles, including safety issues such as the potential presence of pathogenic microorganisms in the material recovered [bib_ref] Recovery of High Value-Added Nutrients from Fruit and Vegetable Industrial Wastewater, Chen [/bib_ref].
## Packaging of fresh produce
Packaging is another aspect of the fresh produce supply chain in which conflicts between food safety and sustainability can appear. Guaranteeing food safety is one of the benefits of fresh produce packaging. Nevertheless, packaging can also have drawbacks regarding microbial safety, such as the increased survival of pathogens in highmoisture environments (e.g., bagged lettuce) [bib_ref] A comparison of Listeria monocytogenes contamination in bagged and un-bagged lettuce in..., Kyere [/bib_ref]. Regarding the packaging/sustainability interaction, the recommendations and regulations aimed at guaranteeing food safety tend to promote the utilization of single-use packaging [bib_ref] The new food safety, Leib [/bib_ref] , although multiple-use containers can be more sustainable. For example, the utilization of reusable plastic containers (RPCs) for the handling, transport, and commercialization of fruits and vegetables has the potential to improve the sustainability of the fresh produce supply chain. However, although the use of these RPCs has never been linked directly with any fresh produce outbreak, a lack of hygiene can lead to unwanted risks [bib_ref] Microbiological status of RPCs in commercial grower/packer operations and risk of Salmonella..., Zhu [/bib_ref]. On the other hand, at the consumer level, Barbosa et al. [bib_ref] Microbiological contamination of reusable plastic bags for food transportation, Barbosa [/bib_ref] detected diverse microorganisms, including pathogens, in multiple-use plastic bags utilized for food transportation. In the specific case of the controversial plastic packaging, it helps to reduce food waste in the fresh produce supply chain, thus improving sustainability. However, the current life cycle of plastic packaging does not fit a circular economy approach. Consequently, policymakers should promote the utilization of alternative and sustainable packaging options, due to the importance of the packaging sector in the sustainability of the food supply chain [79,80].
## Temperature control
Control of storage temperatures during preservation is essential for maintaining the quality of fresh produce and, thereby, avoiding food waste and the associated impact on sustainability . It also helps in guaranteeing the microbial safety of such products (e.g., leafy greens) [bib_ref] Quantitative assessment of the microbial risk of leafy greens from farm to..., Danyluk [/bib_ref]. However, storage temperature control entails a cost in energy consumption that affects the sustainability of the supply chain. Tools for the optimization of temperature control in the fresh produce supply chain must consider safety, spoilage, and energy consumption [bib_ref] Linking food waste prevention, energy consumption and microbial food safety: The next..., Guillier [/bib_ref]. New technological developments (e.g., Internet of Things, Artificial Intelligence, Big Data, Blockchains, etc.) are expected to improve the control of food cold-chain logistics in the coming years, with positive impacts on the safety and quality of fresh produce, reducing food waste and the environmental impact [bib_ref] A comprehensive review of cold chain logistics for fresh agricultural products: Current..., Han [/bib_ref] [bib_ref] How might broad adoption of blockchain-based traceability impact the U.S. fresh produce..., Collart [/bib_ref]. Wu et al. [bib_ref] Environmental trade-offs in fresh-fruit cold chains by combining virtual cold chains with..., Wu [/bib_ref] suggested that a holistic approach combining life cycle assessment with virtual cold chains could help to design more sustainable fresh fruit cold chains. However, simpler changes could also have a significant impact. For example, the study by Xie et al. [bib_ref] Improving temperature management and retaining quality of fresh-cut leafy greens by retrofitting..., Xie [/bib_ref] suggests the use of closed displays for refrigerated fresh-cut leafy greens in retail shops as a solution that combines keeping the quality (avoiding food waste), microbial safety, and energy savings.
## Relationship between food loss/waste and food safety
Food loss is the decrease in the quantity or quality of food resulting from decisions and actions by food suppliers in the chain, excluding retailers, food service providers, and consumers. Food waste refers to the decrease in the quantity or quality of food resulting from decisions and actions by retailers, food service providers, and consumers. The reduction of food loss/waste is included in the United Nations Sustainable Development Goals from the 2030 Agenda for Sustainable Development (SDG 12, Target 12.3). Food is lost/wasted in many ways: fresh produce that deviates from what is considered optimal, for example in terms of shape, size, and color, is often removed from the supply chain during sorting operations. Foods that are close to, at, or beyond the "best before" date are often discarded by retailers and consumers. Large quantities of wholesome edible foodstuffs are often unused or left over and discarded from household kitchens and eating establishments. Around one-third of the world's food is lost or wasted every year. This 2011 estimate by the FAO is in the process of being replaced by two separate indices: the Food Loss Index (FLI) and the Food Waste Index (FWI). The FLI provides new loss estimates from post-harvest up to, but not including, the retail stage.
The significant amount of the food produced that is lost or wasted entails an unnecessary environmental impact (greenhouse gas emissions, use of land and water resources). Food loss and waste includes not only the organic material but also the water and energy utilized for the production and the components of the packages [bib_ref] Global missions and the critical needs of food science and technology, Lillford [/bib_ref]. A reduction in food loss/waste could improve the sustainability of fresh food, adding a sizable quantity to the global food supply, thereby reducing the need to intensify production in the future. Less food loss and waste would lead to more efficient land use and better water resource management, with positive impacts on climate change and livelihoods (http://www.fao.org/food-loss-and-food-waste/flw-data accessed on 16 July 2021).
The recommendations to reduce food loss/waste should consider their viability (technical and economic), the constraints due to food quality and safety requisites, the point of view of society, and the environmental impacts. Fresh produce supply chain stakeholders are receiving messages on the topics of food loss/waste and food safety that are, to a certain extent, contradictory [bib_ref] Linking food waste prevention, energy consumption and microbial food safety: The next..., Guillier [/bib_ref] [bib_ref] Food, waste and safety: Negotiating conflicting social anxieties into the practices of..., Watson [/bib_ref]. On the one hand, measures that help to reduce food loss/waste could increase the food safety risk for the consumers [bib_ref] Balancing the desire to decrease food waste with requirements of food safety, Kasza [/bib_ref]. On the other hand, the food safety policy should be well calibrated to avoid unnecessary food loss/waste [bib_ref] The new food safety, Leib [/bib_ref]. For example, the confusion of consumers regarding the relationship between food safety and food date labels can lead to food waste. In the European Union, perishable foods after the "use by" date shall be deemed to be unsafe and, consequently, their marketing is prohibited [bib_ref] on the Provision of Food Information to Consumers, Parliament [/bib_ref]. Conversely, the "best before" label informs about quality, not safety, but up to a quarter of the population thinks that food should not be eaten after that date.
Food waste management methods include (in order of priority) prevention (e.g., prediction of demand by consumers, planned food shopping), redistribution (for human consumption), valorization (e.g., industrial recycling of waste, its use to produce animal feed), and food waste treatment (e.g., composting, incineration) [bib_ref] Balancing the desire to decrease food waste with requirements of food safety, Kasza [/bib_ref]. The redistribution of excess food for human consumption (e.g., donation) or its use to produce animal feed can help to reduce food waste. Decentralization, lack of professionalization, insufficient or non-existent regulation, and lack of monitoring by authorities have been identified as important problems that make the optimization of food safety within food donation/acceptation chains difficult [bib_ref] Ensuring food safety in food donations: Case study of the Belgian donation/acceptation..., Boeck [/bib_ref] [bib_ref] Food donation and food safety: Challenges, current practices, and the road ahead, Bierma [/bib_ref]. Hecht and Neffindicated that future studies aimed at performing a risk-benefit assessment of food redistribution interventions need to include the effects on health, the environment, and the economy. Safety requirements are also essential when assessing the feasibility of the valorization of food waste [bib_ref] Food waste valorization, Plazzotta [/bib_ref]. To prevent microbial safety issues, food waste could require treatment (e.g., pasteurization) to enable its valorization as a food-grade ingredient [bib_ref] The Impact of Pasteurisation and Sterilisation on Bioactive Compounds of Onion By-products, Benítez [/bib_ref]. The option of food waste treatment is appropriate for a wider range of food waste categories than redistribution and valorization [bib_ref] Are waste hierarchies effective in reducing environmental impacts from food waste? A..., Redlingshöfer [/bib_ref] , because fruits and vegetables recovered and reintroduced into the supply chain for human consumption constitute higher safety risks [bib_ref] Recovery of wasted fruit and vegetables for improving sustainable diets, Augustin [/bib_ref].
In the case of fresh produce, the use of different pre-and post-harvest tools can help to decrease food loss/waste. The effects of proper management, such as storage in well-ventilated rooms, storage in a controlled atmosphere, modified atmosphere, ethylene scavengers, proper temperature and relative humidity (RH), heat treatment, and others, plus different sustainable pre-and post-harvest treatments (i.e., natural compounds, ozone, ultraviolet irradiation, biocontrol agents), and their combinations are sustainable treatment methods that help to reduce the decay of fresh fruits and vegetables (e.g., carrot, spinach, peach, nectarine) [42, [bib_ref] Ethylene scavengers for the preservation of fruits and vegetables: A review, Wei [/bib_ref] [bib_ref] Postharvest environmentally and human-friendly pre-treatments to minimize carrot waste in the supply..., Papoutsis [/bib_ref] [bib_ref] Inducing fungal resistance of spinach treated with preharvest hormetic doses of UV-C, Martínez-Sánchez [/bib_ref] [bib_ref] Deficit irrigation strategies combined with controlled atmosphere preserve quality in early peaches, Falagán [/bib_ref] [bib_ref] Natural additives to preserve quality and improve nutritional value of fresh-cut nectarine, Falagán [/bib_ref] [bib_ref] Heat treatment as postharvest tool for improving quality in extra-early nectarines, Falagán [/bib_ref].
## Climate change and fresh produce safety
Modifications in the Earth's atmospheric composition caused by human activities are driving climate change. Increases in the mean air temperature and the frequency of extreme weather events are among the expected consequences of climate change. These changes are associated with a potential risk of an increased presence of certain pathogenic microorganisms and toxins in food, including vegetables [bib_ref] Impacts of climate change on the microbial safety of pre-harvest leafy green..., Liu [/bib_ref]. The work by Liu et al. [bib_ref] Impacts of climate change on the microbial safety of pre-harvest leafy green..., Liu [/bib_ref] focusing on pre-harvest leafy green vegetables concluded that the rise in temperatures and modifications in precipitation patterns will affect the contamination sources and the pathways of pathogens, likely leading to an increase in the contamination of these products with pathogenic microorganisms. Foodborne pathogens are among the most climate-sensitive human pathogenic microorganisms [bib_ref] Systematic Assessment of the Climate Sensitivity of Important Human and Domestic Animals..., Mcintyre [/bib_ref]. Holvoet et al. [bib_ref] Relationships among hygiene indicators and enteric pathogens in irrigation water, soil and..., Holvoet [/bib_ref] observed a positive correlation between the presence of pathogenic microorganisms in lettuce irrigation water and temperature. Extreme precipitation can cause flooding in agricultural fields, and the risks of using open-air areas after a flood event, where potential exposure to infective microbial contamination exists, must be evaluated. Flooding has been associated with pathogenic contamination of leafy green vegetables [bib_ref] Microbial safety considerations of flooding in primary production of leafy greens: A..., Castro-Ibáñez [/bib_ref]. Droughts are also expected to be more frequent and intense in the future in some areas of the planet that are currently affected by that phenomenon, and drought-stressed plants (e.g., lettuce) could be more susceptible to the internalization of pathogenic bacteria [bib_ref] Effects of soil texture and drought stress on the uptake of antibiotics..., Zhang [/bib_ref]. The potential impacts of climate change on the contamination of food with pathogenic microorganisms are complex, and knowledge gaps are numerous [bib_ref] Effects of climate change on the persistence and dispersal of foodborne bacterial..., Hellberg [/bib_ref]. In any case, adaptation and mitigation strategies will need to be implemented to reduce the negative impacts of climate change on fresh produce safety [bib_ref] Towards strategies to adapt to pressures on safety of fresh produce due..., Kirezieva [/bib_ref] [bib_ref] A risk modeling framework to evaluate the impacts of climate change and..., Smith [/bib_ref]. The work by Kirezieva et al. [bib_ref] Towards strategies to adapt to pressures on safety of fresh produce due..., Kirezieva [/bib_ref] used experts' opinions to evaluate potential responses to the impacts of climate change on fresh produce safety, concluding that strengthened control activities (e.g., water microbial quality monitoring, personal hygiene requirements) and improved guidance and training for farmers will be needed.
## Approaches, strategies and solutions to solve conflicts between fresh produce microbial safety and environmental sustainability
Co-management at the farm level comprises balancing environmental protection with food safety and productivity goals. Certain types of agricultural management, such as organic agriculture and, particularly, biodynamic agriculture, which fosters the diversity of plant and animal life, increase the health and resilience of the organism farm. Biodynamic farms aspire to generate their fertility through composting, integrating animals, cover cropping, and crop rotation.
The work by Crohn and Bianchi [bib_ref] Research priorities for coordinating management of food safety and water quality, Crohn [/bib_ref] identified the assessment of the fate of pathogenic microorganisms in farmlands as the most urgent research topic regarding the co-management of food safety and surface water quality. In the last two decades, the information available on the behavior of pathogenic microorganisms in agricultural settings has increased significantly, mainly based on controlled tests with the inoculation of lab-prepared pathogens [bib_ref] Effect of weather on the die-off of Escherichia coli and attenuated Salmonella..., Belias [/bib_ref]. The development of methods for the large-scale affordable detection of pathogenic microorganisms in the agricultural environment would be of enormous help in tracking the sources of produce-borne outbreaks, as well as in the assessment of the fate of pathogens [bib_ref] Quantifying the sensitivity of scent detection dogs to identify fecal contamination on..., Partyka [/bib_ref]. Information obtained directly from the environment would provide us with a more reliable picture of the situation, which could then be used as a background for the development of recommendations and legislation. Currently, one of the tools assessed to detect fecal contamination in fresh produce in the growing fields is hyperspectral imaging. Cho et al. evaluated this technique for the on-site detection of fecal contamination in romaine lettuce, with positive results [bib_ref] Hyperspectral determination of fluorescence wavebands for multispectral imaging detection of multiple animal..., Cho [/bib_ref]. Until more detailed and complete information on this topic becomes available, a conservative approach is likely to be taken by the competent authorities and by supply-chain stakeholders to avoid outbreaks and the subsequent consequences on public health and the agrifood sector.
In many cases, the experimental studies focus on a topic with a narrow approach, without considering interactions with other aspects. For example, over the years, numerous studies have assessed the efficacy of antimicrobial treatments that are applicable to fresh produce, without considering aspects such as the economic and environmental sustainability of their usage. However, in the last decade, studies with a more global perspective have been performed. Vigil et al.assessed sanitation and decontamination techniques for fresh-cut produce using a life cycle approach. Papoutsis and Edelenbos [bib_ref] Postharvest environmentally and human-friendly pre-treatments to minimize carrot waste in the supply..., Papoutsis [/bib_ref] reviewed different sustainable post-harvest treatments for carrots (considering both human health and the environment). On the topic of food-waste reduction, Tromp et al. [bib_ref] Reusing salad from salad bars-simulating the effects on product loss, microbial safety..., Tromp [/bib_ref] assessed the potential reuse of salads in salad bars considering safety and quality. Yam and Takhistov [bib_ref] Sustainable packaging technology to improve food safety, Yam [/bib_ref] also considered microbial safety, as well as economic and environmental sustainability, when assessing an alternative packaging technology for fresh produce. Looking to the future, both available and innovative technologies (e.g., nanotechnology) will help to make fresh produce safety and sustainability compatible [bib_ref] Nanobiotechnology for agricultural sustainability, and food and environmental safety, Madanayake [/bib_ref].
Proper traceability is crucial in the management of the fresh produce supply chain to ensure safety and avoid loss and waste [bib_ref] Pre-and postharvest preventive measures and intervention strategies to control microbial food safety..., Gil [/bib_ref]. Being able to quickly trace back any contamination to its source can reduce food loss/waste by defining precisely which lots should be disposed of, thereby avoiding the unnecessary elimination of uncontaminated batches [bib_ref] Modeling disruption in a fresh produce supply chain, Mackenzie [/bib_ref]. The suitability of the use of blockchain technology to enhance traceability (and therefore the safety and sustainability) of fresh produce is being assessed [bib_ref] How might broad adoption of blockchain-based traceability impact the U.S. fresh produce..., Collart [/bib_ref].
In the conflicts between sustainability and safety, from the political organization's standpoint, there is a lack of a global perspective. Different departments, agencies, etc. have different and narrow-sighted ideas on how to deal with the issues of safety and sustainability (environmental, economic, and social) of the food supply chain [bib_ref] The new food safety, Leib [/bib_ref]. Multiple criteria decision analysis has been suggested as a structured tool for decision-making in this complicated framework [bib_ref] Linking food waste prevention, energy consumption and microbial food safety: The next..., Guillier [/bib_ref]. Regarding the concept of food safety, Leib and Pollans [bib_ref] The new food safety, Leib [/bib_ref] proposed a more global view that should include not only the current concept (acute risks linked to ingestion of pathogens or toxins) but also other issues such as the health risks associated with cumulative ingestion and the health risks linked to the life cycle of food from production to end of life. Furthermore, measures aimed at improving fresh produce safety should be adapted taking the diversity of agrifood systems into account [bib_ref] Agrifood systems and the microbial safety of fresh produce: Trade-offs in the..., Nguyen-The [/bib_ref]. The One Health approach, based on the concept of the interconnection between human beings, animals, and the environment, promotes the formation of multidisciplinary teams that can work to obtain solutions to challenges that involve health, social, and environmental issues [bib_ref] One Health for Food Safety, Food Security, and Sustainable Food Production, Garcia [/bib_ref]. Finally, to promote the safety, security, and sustainability of the produce supply chain, the training and education of all the stakeholders are crucial. summarizes the topics covered, the potential implications on fresh produce microbial safety and environmental sustainability, and the optimization options covered in this review. . Summary of topics covered, potential implications on fresh produce microbial safety and environmental sustainability discussed, and optimization options.
## Topic
## Sub-topic microbial safety environmental sustainability optimization options
Fresh produce safety at primary production
# Conclusions
Clear conflicts arise when analyzing the interactions between environmental sustainability and microbial safety of the fresh produce supply chain. Although the safety aspect has priority, the decisions taken for the sake of fresh fruit and vegetable safety ought to consider the potential impacts on the whole sustainability (environmental, economic, and social). The present work provides examples of frictions between microbial safety and environmental sustainability in the fresh produce supply chain. The stated issues are present throughout the supply chain (pre-and post-harvest) and affect all the different stakeholders (from primary producers to consumers). A global approach to deal with these safety/sustainability interactions is required. Widening the concept of food safety, co-management, multicriteria decision analysis, technological advances (e.g., cold chain management), working in multidisciplinary teams, and training the stakeholders are some of the strategies and approaches that will help to deal with sustainability/safety conflicts. In this context, the concept of One Health applied to the fresh produce supply chain appears as a correct approach to analyze and make decisions aimed at solving these challenges.
[fig] Figure 1: Potential environmental impacts (left) and fresh produce safety gains (right) of different measures applicable in primary production. DBPs: disinfection by-products. [/fig]
[fig] Author: Contributions: F.L.-G.: conceptualization, investigation, and writing-original draft. P.A.G., F.A. and F.A.-H.: writing-review and editing. E.A.: conceptualization, investigation, supervision, writing-review and editing, project administration, and funding acquisition. All authors have read and agreed to the published version of the manuscript. [/fig]
[fig] Funding: This research was funded by the Fondo Europeo de Desarrollo Regional/Ministerio de Ciencia e Innovación-Agencia Estatal de Investigación (FEDER/MICINN-AEI), project RTI2018-099139-B-C21 (https://www.karpolife.com accessed on 16 January 2020). Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. Conflicts of Interest: The authors declare no conflict of interest. The funders had no role in the design of the study; in the writing of the manuscript; or in the decision to publish it. 37. Garrido, Y.; Marín, A.; Tudela, J.A.; Truchado, P.; Allende, A.; Gil, M.I. Chlorate accumulation in commercial lettuce cultivated in open field and irrigated with reclaimed water. Food Control 2020, 114, 107283. [CrossRef] 38. Tombini-Decol, L.; López-Gálvez, F.; Truchado, P.; Tondo, E.C.; Gil, M.I.; Allende, A. Suitability of chlorine dioxide as a tertiary treatment for municipal wastewater and use of reclaimed water for overhead irrigation of baby lettuce. Food Control 2019, 96, 186-193. [CrossRef] 39. López-Gálvez, F.; Andújar, S.; Marín, A.; Tudela, J.A.; Allende, A.; Gil, M.I. Disinfection by-products in baby lettuce irrigated with electrolysed water. J. Sci. Food Agric. 2018, 98, 2981-2988. [CrossRef] 40. European Parliament and of the Council. Consolidated Text: Regulation (EC) No 396/2005 of the of 23 February 2005 on Maximum Residue Levels of Pesticides in or on Food and Feed of Plant and Animal Origin and Amending Council Directive 91/414/EEC. 2005. Available online: https://eur-lex.europa.eu/eli/reg/2005/396/2021-01-06 (accessed on 17 May 2021). 41. Truchado, P.; Gil, M.I.; Suslow, T.; Allende, A. Impact of chlorine dioxide disinfection of irrigation water on the epiphytic bacterial community of baby spinach and underlying soil. PLoS ONE 2018, 13, e199291. [CrossRef] 42. Martínez-Sánchez, A.; Aguayo, E. Effect of irrigation with ozonated water on the quality of capsicum seedlings grown in the nursery. Agric. Water Manag. 2019, 221, 547-555. [CrossRef] 43. Payen, S.; Basset-Mens, C.; Perret, S. LCA of local and imported tomato: An energy and water trade-off. J. Clean. Prod. 2015, 87, 139-148. [CrossRef] 44. Boschiero, M.; Zanotelli, D.; Ciarapica, F.E.; Fadanelli, L.; Tagliavini, M. Greenhouse gas emissions and energy consumption during the post-harvest life of apples as affected by storage type, packaging and transport. J. Clean. Prod. 2019, 220, 45-56. [CrossRef] 45. Kumar, V.; Wang, M.; Kumari, A.; Akkaranggoon, S.; Garza-Reyes, J.A.; Neutzling, D.M.; Tupa, J. Exploring short food supply chains from triple bottom line lens: A comprehensive systematic review. In Proceedings of the International Conference on Industrial Engineering and Operations Management 2019 (MAR), Bangkok, Thailand, 5-7 March 2018; pp. 728-738. 46. Vittersø, G.; Torjusen, H.; Laitala, K.; Tocco, B.; Biasini, B.; Csillag, P.; de Labarre, M.D.; Lecoeur, J.-L.; Maj, A.; Majewski, E.; et al. Short food supply chains and their contributions to sustainability: Participants' views and perceptions from 12 European cases. Sustainability 2019, 11, 4800. [CrossRef] 47. Chapman, B.; Gunter, C. Local food systems food safety concerns. Microbiol. Spectr. 2018, 6. [CrossRef] [PubMed] 48. Schmitt, E.; Galli, F.; Menozzi, D.; Maye, D.; Touzard, J.-M.; Marescotti, A.; Six, J.; Brunori, G. Comparing the sustainability of local and global food products in Europe. J. Clean. Prod. 2017, 165, 346-359. [CrossRef] 49. Hernández-Rubio, J.; Pérez-Mesa, J.C.; Piedra-Muñoz, L.; Galdeano-Gómez, E. Determinants of food safety level in fruit and vegetable wholesalers' supply chain: Evidence from Spain and France. Int. J. Environ. Res. Public Health 2018, 15, 2246. [CrossRef] [PubMed] 50. González-Azcárate, M.; Cruz Maceín, J.L.; Bardají, I. Why buying directly from producers is a valuable choice? Expanding the scope of short food supply chains in Spain. Sustain. Prod. Consum. 2021, 26, 911-920. [CrossRef] 51. Kirezieva, K.; Luning, P.A.; Jacxsens, L.; Uyttendaele, M. Status of food safety management activities in fresh produce companies in the European Union and beyond. Acta Hortic. 2015, 1103, 167-173. [CrossRef] 52. Wright, K.G.; Sirsat, S.A.; Neal, J.A.; Gibson, K.E. Growth of local food systems: A review of potential food safety implications. CAB Rev. Perspect. Agric. Vet. Sci. Nutr. Nat. Resour. 2015, 10, 025. [CrossRef] 53. Young, I.; Thaivalappil, A.; Reimer, D.; Greig, J. Food safety at farmers' markets: A knowledge synthesis of published research. J. Food Protect. 2017, 80, 2033-2047. [CrossRef] [PubMed] 54. Manzocco, L.; Ignat, A.; Anese, M.; Bot, F.; Calligaris, S.; Valoppi, F.; Nicoli, M.C. Efficient management of the water resource in the fresh-cut industry: Current status and perspectives. Trends Food Sci. Tech. 2015, 46, 286-294. [CrossRef] 55. Fusi, A.; Castellani, V.; Bacenetti, J.; Cocetta, G.; Fiala, M.; Guidetti, R. The environmental impact of the production of fresh cut salad: A case study in Italy. Int. J. Life Cycle Ass. 2016, 21, 162-175. [CrossRef] 56. Morris, J.R.; Brady, P.L. Temperature effects on produce degradation. In Produce Degradation: Pathways and Prevention; Lamikanra, O., Imam, S., Ukuku, D., Eds.; CRC Press: Boca Raton, FI, USA, 2005; pp. 599-647. 57. Artés, F.; Gómez, P.A.; Aguayo, E.; Escalona, V.H.; Artés-Hernández, F. Sustainable sanitation techniques for keeping quality and safety of fresh-cut plant commodities. Postharvest Biol. Technol. 2009, 51, 287-296. [CrossRef] 58. Silveira, A.C.; Aguayo, E.; Escalona, V.H.; Artés, F. Combined effect of hot water treatment and peracetic acid to maintain the overall quality in fresh-cut Galia melon. Innov. Food Sci. Emerg. 2011, 12, 569-576. [CrossRef] 59. Ölmez, H. Water Consumption, Reuse and Reduction Strategies in Food Processing. In Sustainable Food Processing; Tiwari, B.K., Norton, T., Holden, N.M., Eds.; Wiley Blackwell: Hoboken, NJ, USA, 2013; pp. 401-434. 60. López-Gálvez, F.; Gil, M.I. The importance of water in the fresh produce industry | [La importancia del agua en la industria de alimentos vegetales]. Arbor 2020, 196, 1-9. [CrossRef] 61. Weng, S.-C.; Jacangelo, J.G.; Schwab, K.J. Sustainable practice for the food industry: Assessment of selected treatment options for reclamation of washwater from vegetable processing. Int. J. Environ. Sci. Technol. 2019, 16, 1369-1378. [CrossRef] 62. Mundi, G.S.; Zytner, R.G.; Warriner, K. Fruit and vegetable wash-water characterization, treatment feasibility study and decision matrices. Can. J. Civil Eng. 2017, 44, 971-983. [CrossRef] 63. Nahim-Granados, S.; Sánchez-Pérez, J.A.; Polo-Lopez, M.I. Effective solar processes in fresh-cut wastewater disinfection: Inactivation of pathogenic E. coli O157:H7 and Salmonella enteritidis. Catal. Today 2018, 313, 79-85. [CrossRef] [/fig]
|
The beginning and the end of SNARE‐induced membrane fusion
Membrane fusion is not a spontaneous process. Physiologically, the formation of coiled-coil protein complexes, the SNAREpins, bridges the membrane of a vesicle and a target membrane, brings them in close contact, and provides the energy necessary for their fusion. In this review, we utilize results from in vitro experiments and simple physics and chemistry models to dissect the kinetics and energetics of the fusion process from the encounter of the two membranes to the full expansion of a fusion pore. We find three main energy barriers that oppose the fusion process: SNAREpin initiation, fusion pore opening, and expansion. SNAREpin initiation is inherent to the proteins and makes in vitro fusion kinetic experiments rather slow. The kinetics are physiologically accelerated by effectors. The energy barriers that precede pore opening and pore expansion can be overcome by several SNAREpins acting in concert.Biological membranes reliably separate two aqueous regions and delineate the contours of cells and of the organelles they contain[1,2]. Their integrity is ensured by their thin ∼ 3 nm hydrophobic core that prevents the crossing of any solute and sparsely allows water molecules to pass from one side to the other [3]. This spatial separation is indeed critical for them to separately accomplish their function[2]. Despite this individual specialization, organelles must work collectively. For instance, molecular exchanges must sometimes occur between them to share information and/or material. A major pathway for this molecular transport within or between cells is vesicular trafficking [4], which always follows the same steps. First, 50-200 nm vesicles containing selected cargo are formed from the membrane of the donor compartment. They travel to the target membrane where they fuse, thereby releasing both the encapsulated soluble cargo into the lumen and the membrane-bound molecules in the membrane of the target compartment. This last key step of the transport process does not occur spontaneously. A high energy barrier, typically 25-30 k B T over a couple of nanometer displacement, needs to be overcome[5][6][7][8][9]. This high activation cost is not surprising because, to prevent untimely vesicle fusion, two very cohesive membranes must be actively, cooperatively, and simultaneously disrupted and merged.It has been demonstrated more than a quarter of a century ago that the mechanical energy source of these mechanisms comes from proteins, the SNAREs [10-12].Abbreviations FRET, Förster resonance energy transfer; MTC, multisubunit tethering complex; SFA, surface force apparatus; SNARE, soluble N-ethylmaleimide-sensitive factor attachment protein receptor; SUV, small unilamellar vesicle; Syn1A, syntaxin 1A; TIRF, total internal reflection florescence microscopy; TMD, transmembrane domain; V c , C-terminal region of the SNARE domain of VAMP2 (typically residues 57-92).
Membrane fusion is not a spontaneous process. Physiologically, the formation of coiled-coil protein complexes, the SNAREpins, bridges the membrane of a vesicle and a target membrane, brings them in close contact, and provides the energy necessary for their fusion. In this review, we utilize results from in vitro experiments and simple physics and chemistry models to dissect the kinetics and energetics of the fusion process from the encounter of the two membranes to the full expansion of a fusion pore. We find three main energy barriers that oppose the fusion process: SNAREpin initiation, fusion pore opening, and expansion. SNAREpin initiation is inherent to the proteins and makes in vitro fusion kinetic experiments rather slow. The kinetics are physiologically accelerated by effectors. The energy barriers that precede pore opening and pore expansion can be overcome by several SNAREpins acting in concert.
Biological membranes reliably separate two aqueous regions and delineate the contours of cells and of the organelles they contain [bib_ref] The fluid mosaic model of the structure of cell membranes, Singer [/bib_ref] [bib_ref] Morphology and function of membrane-bound organelles, Heald [/bib_ref]. Their integrity is ensured by their thin ∼ 3 nm hydrophobic core that prevents the crossing of any solute and sparsely allows water molecules to pass from one side to the other [bib_ref] Permeability of lipid bilayers to water and ionic solutes, Deamer [/bib_ref]. This spatial separation is indeed critical for them to separately accomplish their function [bib_ref] Morphology and function of membrane-bound organelles, Heald [/bib_ref]. Despite this individual specialization, organelles must work collectively. For instance, molecular exchanges must sometimes occur between them to share information and/or material. A major pathway for this molecular transport within or between cells is vesicular trafficking [bib_ref] SNAREs-engines for membrane fusion, Jahn [/bib_ref] , which always follows the same steps. First, 50-200 nm vesicles containing selected cargo are formed from the membrane of the donor compartment. They travel to the target membrane where they fuse, thereby releasing both the encapsulated soluble cargo into the lumen and the membrane-bound molecules in the membrane of the target compartment. This last key step of the transport process does not occur spontaneously. A high energy barrier, typically 25-30 k B T over a couple of nanometer displacement, needs to be overcome [bib_ref] Low energy cost for optimal speed and control of membrane fusion, Francois-Martin [/bib_ref] [bib_ref] Cooperation of conical and polyunsaturated lipids to regulate initiation and processing of..., Francois-Martin [/bib_ref] [bib_ref] Mimicry and mechanism in phospholipid models of membrane-fusion, Rand [/bib_ref] [bib_ref] Hydration forces between phospholipid-bilayers, Rand [/bib_ref] [bib_ref] New phases of phospholipids and implications to the membrane fusion problem, Yang [/bib_ref]. This high activation cost is not surprising because, to prevent untimely vesicle fusion, two very cohesive membranes must be actively, cooperatively, and simultaneously disrupted and merged.
It has been demonstrated more than a quarter of a century ago that the mechanical energy source of these mechanisms comes from proteins, the SNAREs [bib_ref] A protein assembly-disassembly pathway in vitro that may correspond to sequential steps..., Sollner [/bib_ref] [bib_ref] SNAP receptors implicated in vesicle targeting and fusion, Sollner [/bib_ref] [bib_ref] SNAREpins: minimal machinery for membrane fusion, Weber [/bib_ref].
Abbreviations FRET, Förster resonance energy transfer; MTC, multisubunit tethering complex; SFA, surface force apparatus; SNARE, soluble N-ethylmaleimide-sensitive factor attachment protein receptor; SUV, small unilamellar vesicle; Syn1A, syntaxin 1A; TIRF, total internal reflection florescence microscopy; TMD, transmembrane domain; V c , C-terminal region of the SNARE domain of VAMP2 (typically residues .
These proteins form a complex between the two membranes, the SNAREpin. There is a whole family of SNAREpins [bib_ref] SNAREs-engines for membrane fusion, Jahn [/bib_ref]. They all contain a 'SNARE domain' characterized by four-coiled alpha helices [bib_ref] Crystal structure of a SNARE complex involved in synaptic exocytosis at 2.4..., Sutton [/bib_ref]. Each helix displays a heptad repeat, i.e., a hydrophobic residue every 3 and 4 alternating residues. In the coiled-coil, the hydrophobic residues are aligned, forming 'hydrophobic pocket' or 'hydrophobic layer' [bib_ref] Crystal structure of a SNARE complex involved in synaptic exocytosis at 2.4..., Sutton [/bib_ref] [bib_ref] SNARE complex formation is triggered by Ca2+ and drives membrane fusion, Chen [/bib_ref] [bib_ref] Helical extension of the neuronal SNARE complex into the membrane, Stein [/bib_ref]. Fifteen hydrophobic pockets in the SNARE domains are numbered from −7 at the membrane-distal N-terminus end to +8 at the membrane-proximal C-terminus end. The layer in the middle, referred to as layer 0, is hydrophilic and may help the correct register of the hydrophobic layers emanating from the four helices.
An archetypal example is the SNAREpin responsible for the fusion of synaptic vesicles and the neuronal presynaptic plasma membrane. Since the synaptic SNAREpin is among the most studied and best characterized, we will focus on this specific one. This SNAREpin forms a four helical bundle composed of VAMP2 (also known as synaptobrevin) that contains a single cytosolic helix and a transmembrane domain (TMD) embedded in the synaptic vesicle and the binary complex made of syntaxin1a (Syn1A, one helix, and TMD) and SNAP25 (two helices separated by a linker containing cysteine clusters to conjugate to palmitic acids) on the presynaptic plasma membrane. Short linker domains (∼ 10 residues) connect the helix and TMD of Syn1a and VAMP2. The four helices of VAMP2, Syn1A, and SNAP25 represent the SNARE domains of the proteins.
The SNARE-induced fusion process can be cut into six main stages. First, the membranes must meet (i). Then, SNAREs have to 'find each other' (ii) and initiate their assembly through their N-terminal regions (iii). Next the SNAREpin zippers in an effort to bring the membranes in close apposition (iv). When the intermembrane distance is small enough, the membranes merge and a fusion pore opens (v) and subsequently expands (vi). This dissection of the fusion process is valid in vitro and in vivo when no other factor is involved which is not the case, for instance, in evoked neurotransmitter release where many steps are bypassed or facilitated by scaffold proteins; this will be briefly discussed.
Even though SNAREpins were proven to be a minimal, necessary, and sufficient machinery for fusion [bib_ref] SNAREpins: minimal machinery for membrane fusion, Weber [/bib_ref] [bib_ref] Use of resonance energy transfer to monitor membrane fusion, Struck [/bib_ref] , this breakthrough result was set in question because of the apparently slow kinetics. Fusion occurred on the timescale of dozens of minutes whereas the order of magnitude in vivo is seconds or
[formula] (A) (B) Vamp2 # of layer -7 -6 -5 -4 -3 -2 -1 0 +1 +2 +3 +4 +5 +6 +7 +8 Syntaxin1a SNAP25-N SNAP25-C SNARE domain Linker TMD v i) Membranes meet [/formula]
iii) Assembly initiation iv) Zippering v) Pore opening vi) Pore expansion ii) SNAREs meet The synaptic SNAREpin is held together by 'SNARE domains' coming from the v-SNARE VAMP2 and from the t-SNAREs Syntaxin1a and SNAP25. The 53 residues presented here for each SNARE domain are distributed in a heptad repeat fashion: Every 3 and 4 alternating residues are hydrophobic. The resulting hydrophobic pockets, highlighted in yellow and numbered from −7 at the N-terminal end to +8 at the C-terminal end, ensure the stability of the SNAREpin. The middle layer, referred to as layer 0, is actually hydrophilic. (B) These cartoons depict the six stages of SNARE-induced fusion that occur when no other protein is involved, like in most in vitro experiments presented here but unlike evoked neurotransmitter release in which some of the stages are bypassed by chaperones. minutes and can be as low as milliseconds for synaptic SNAREs [bib_ref] The synaptic vesicle cycle, Sudhof [/bib_ref] [bib_ref] Distinct initial SNARE configurations underlying the diversity of exocytosis, Kasai [/bib_ref]. This surprising discrepancy is the starting point of our review: we will try to figure out and quantify the various kinetic and energetic hurdles during fusion induced by SNAREpins alone. We will not discuss stage (iv) that was exquisitely deciphered by optical tweezers [bib_ref] Single reconstituted neuronal SNARE complexes zipper in three distinct stages, Gao [/bib_ref] [bib_ref] Energetics, kinetics, and pathway of SNARE folding and assembly revealed by optical..., Zhang [/bib_ref] [bib_ref] Common intermediates and kinetics, but different energetics, in the assembly of SNARE..., Zorman [/bib_ref]. The remaining stages will be split into two parts: towards SNAREpin assembly (stages i, ii, and iii) and fusion pore opening and expansion (stages v and vi). We will combine experimental observations and simple physics/chemistry models to show that SNAREpin initiation (iii), fusion pore opening (v), and subsequent expansion (vi) are the energetically limiting steps of the fusion process.
Towards SNAREpin assembly: the beginning of the fusion process
The historic observation that SNAREpins are the minimal machinery for fusion was performed with an in vitro 'lipid mixing' fusion assay . In this first part, we will apply the common conditions (concentrations, vesicle dimensions, protein densities, . . .) used in this lipid mixing assay to quantitatively analyze and model the three stages that precede SNAREpin zippering.
Getting in touch: meeting of the two membranes Prior to fusion, two free-diffusing vesicles containing synaptic vesicle v-SNAREs (vSUV) and target plasma membrane t-SNAREs (tSUV) have to come in close proximity [fig_ref] Figure 3: Initiation of SNAREpin assembly [/fig_ref]. The encounter rate depends on the vesicle concentration: the less concentrated, the fewer the collisions and consequently the slower the kinetics. This concentration effect can be quantitatively predicted by the theory of collision developed mainly by . This theory provides equations to compute the initial collision rate, ν, of vSUVs with tSUVs. If we note R t (resp. R v ) and ρ t,∞ (resp. ρ v,∞ ) the radius and the initial concentration in tSUVs (resp. vSUVs), the collision rate can be expressed as:
[formula] ν ¼ 2 R t þ R v ð Þ 2 k B Tρ t,∞ 3ηR t R v(1) [/formula]
where k B is Boltzmann constant, T the temperature and η the viscosity of the solution (Eqn A7 in Appendix A). Under standard conditions of the lipid mixing assay, i.e., with 50 nm monodisperse SUVs mixed at 9 : 1 (mol%) of tSUV:vSUV for a total of 1 mM lipids, a vSUV will experience roughly 300 collisions with tSUV per second and there are about 2.10 14 collisions between v-and tSUVs per second in 100 {L of solution. Under conditions with one VAMP2 per ∼ 100 lipids in vSUV and one t-SNARE per ∼ 200 lipids in tSUV, it was found that the mean time for the first fusion event of a vSUV is typically 60 min [bib_ref] Rapid and efficient fusion of phospholipid vesicles by the alpha-helical core of..., Parlati [/bib_ref] [bib_ref] Actual fusion efficiency in the lipid mixing assay -comparison between nanodiscs and..., Francois-Martin [/bib_ref] , which corresponds to ∼ 10 6 collision events per vesicle. This is consistent with previous results that estimated only 1 fusion every 10 6 -10 7 collision events [bib_ref] Actual fusion efficiency in the lipid mixing assay -comparison between nanodiscs and..., Francois-Martin [/bib_ref] [bib_ref] Determinants of liposome fusion mediated by synaptic SNARE proteins, Schuette [/bib_ref]. This very low yield of successful fusion per collision shows that one or several subsequent stages of the fusion process are slower than the meeting of the two membranes in vitro. Nevertheless, it is worth noting that, the collision rate being proportional to the vesicle densities, it can quickly decrease and become a slow step in diluted situations. To ensure that the fluorescence increase is indeed due to fusion, additional controls are needed, the most common of which being the content mixing depicted in [fig_ref] Figure 4: Optical assays [/fig_ref]. The sudden rise at the end of the bottom cartoon depicts the addition of detergent, which maximizes the dequenching of the dyes and is used as a reference for analysis.
membranes will have to 'find each other' for fusion to proceed [fig_ref] Figure 3: Initiation of SNAREpin assembly [/fig_ref]. This encounter of the two SNAREs can occur when the membranes are less than a certain distance d apart. The mean square displacement of a freely diffusing particle predicts the approximate time during which two vesicles, with a relative diffusion coefficient D SUV , remain less than a distance d apart:
[formula] τ collision ¼ d 2 6D SUV(2) [/formula]
Surface Forces Apparatus (SFA) measurements hinted that cognate SNAREs can find each other when the membranes are less than 8-15 nm apart [bib_ref] Energetics and dynamics of SNAREpin folding across lipid bilayers, Li [/bib_ref] [bib_ref] Snapshot of sequential SNARE assembling states between membranes shows that Nterminal transient..., Wang [/bib_ref] , which seems reasonable considering that the fully assembled complex, from N-to C-terminal, has a size of 12 nm. For d ¼ 15 nm and a typical value of 20 {m 2 Ás −1 for D SUV , τ collision boils down to a few microseconds. Random movement of a protein on an artificial membrane is characterized by a diffusion coefficient, D protein . During the timescale τ collision , this leaves a SNARE time to explore an area
[formula] a exp ¼ 4D protein τ collision ¼ 2D protein 3D SUV d 2(3) [/formula]
D protein was measured to be about 5 {m 2 Ás −1 [bib_ref] FRAP to characterize molecular diffusion and interaction in various membrane environments, Pincet [/bib_ref]. With these experimental values, a exp scales as a few tens of nm 2 , which indicates that one SNARE covers the area of ∼ 100 lipids during a collision event. This threshold density, one SNARE per 100 lipids, is close to the standard SNARE density used in the lipid mixing assay and to the physiological density of v-SNARE on a synaptic vesicle [bib_ref] Molecular anatomy of a trafficking organelle, Takamori [/bib_ref]. Below this threshold, our simple model predicts that the fusion rate should vary linearly with the density of SNAREs. A new analysis of a study that systematically varied SNARE densities on both v-and tSUVs [bib_ref] Protein determinants of SNARE-mediated lipid mixing, Ji [/bib_ref] confirms this linear dependency (Appendix B). Above the threshold density, during the course of a collision, there is enough time for v-SNAREs to completely cover the surface of the vSUV and meet an opposing t-SNARE or any protein of similar dimension. Hence, at physiological SNARE densities, it seems that the encounter of the cognate SNAREs should occur systematically during the course of a collision.
Overcoming timidity. . . : difficult initiation of SNAREpin assembly
In the lipid mixing assay, a v-SNARE starts binding via its N-terminal end on a well-structured t-SNARE (SNAP25 + Syn1) N-terminal part [fig_ref] Figure 3: Initiation of SNAREpin assembly [/fig_ref] , left) [bib_ref] N-to C-terminal SNARE complex assembly promotes rapid membrane fusion, Pobbati [/bib_ref]. When the v-and t-SNAREs meet, their very few Nterminal residues weakly bind in a matter of at most a few seconds when membranes are 8 nm apart, as observed in the SFA [bib_ref] Snapshot of sequential SNARE assembling states between membranes shows that Nterminal transient..., Wang [/bib_ref]. However, it was also found that it takes up to half an hour to observe complete zippering. This long delay between initial contact was first attributed to the SFA geometry in which two macroscopic flat surfaces (∼ 1 cm 2 ) are placed a few nanometers apart. This nonphysiological confinement of the proteins was assumed to slow down the process. It turns out that, considering the collision and fusion In the course of their movement, a vSUV and tSUV collide. After most collisions, they move away from each other (first two yellow stars), but sometimes, they stay bound through a SNAREpin. (B) For a SNAREpin to form, a v-SNARE from the vSUV and a t-SNARE from the tSUV must meet. (C) Meeting of the two SNAREpins is not sufficient. There initial assembly of the SNAREpin is limited by an energy barrier. In the lipid mixing bulk assay, this energy barrier comes from the need to disrupt the preassembled t-SNAREs: Opening the 3-helix bundle formed by their N-terminal opens the groove for v-SNARE to bind. rates in the lipid mixing assay, this initial explanation of the long delay for the initiation of SNAREpin assembly is most likely incorrect. Indeed, it takes 1-10 million collisions for a vSUV to fuse with a tSUV under physiological concentrations of v-SNARE in the lipid mixing assay. Each collision takes 10 μs and involves ∼ 10 v-SNAREs that all have the opportunity to meet an opposing t-SNARE. Assuming that there is no cooperativity between the SNAREpins and the first SNAREpin forms on average after a cumulated contact time for all SNAREs in play, a v-SNARE needs to be in contact with a t-SNARE for 100-1000 s on average. This delay is commensurate with the upper limit of 1800 s measured in the SFA.
[formula] 'Collision' vSUV VAMP2 E E a d R t Syn1a SNAP25 (A) (B) (C) tSUV 2R v 2R t Top view Opened Closed R v [/formula]
Overall, this suggests that the initialization of SNAREpin zippering is the limiting step in the lipid mixing assay. From the 100-1000 s mean time required to start SNARE zippering, it is possible to estimate the activation energy of the process, E a , using Kramers reaction-rate theory [bib_ref] Brownian motion in a field of force and the diffusion model of..., Kramers [/bib_ref] [bib_ref] Reaction-rate theory -50 years after Kramers, Hanggi [/bib_ref]. The mean time, τ, can be expressed as:
[formula] τ ¼ τ 0 e Ea k B T(4) [/formula]
For reactions that occur over a couple of nanometers, as is the case here, the prefactor time, τ 0 , is between 0.1 and 10 ns [bib_ref] Reaction-rate theory -50 years after Kramers, Hanggi [/bib_ref] [bib_ref] Energy landscapes of biomolecular adhesion and receptor anchoring at interfaces explored with..., Evans [/bib_ref] , and E a would therefore be between 23 and 30 k B T. This value is consistent with bulk measurements of the binding rate of v-SNARE with t-SNARE [bib_ref] Kinetic barriers to SNAREpin assembly in the regulation of membrane docking/priming and..., Li [/bib_ref]. It was suggested that such a high energy value may be due to the necessity for the proteins to position and locally change structure to be able to bind [bib_ref] N-to C-terminal SNARE complex assembly promotes rapid membrane fusion, Pobbati [/bib_ref] [bib_ref] Kinetic barriers to SNAREpin assembly in the regulation of membrane docking/priming and..., Li [/bib_ref] [bib_ref] Stability, folding dynamics, and long-range conformational transition of the synaptic t-SNARE complex, Zhang [/bib_ref]. A possible pathway for this structural change is suggested by experiments showing that in the 1:1 Syntaxin:SNAP25 t-SNARE complex, i.e., the physiological stoichiometry, the N-terminal portion forms a three-helix coiled coil while the C-terminal region remains frayed [bib_ref] Stability, folding dynamics, and long-range conformational transition of the synaptic t-SNARE complex, Zhang [/bib_ref] [bib_ref] Folding intermediates of SNARE complex assembly, Fiebig [/bib_ref]. The N-terminal coiled coil would need to be opened for the v and t-SNAREs to bind [bib_ref] Stability, folding dynamics, and long-range conformational transition of the synaptic t-SNARE complex, Zhang [/bib_ref] [bib_ref] Folding intermediates of SNARE complex assembly, Fiebig [/bib_ref] [bib_ref] A transient N-terminal interaction of SNAP-25 and syntaxin nucleates SNARE assembly, Fasshauer [/bib_ref]. It has also been long known that structuring the C-terminal part of the t-SNARE by prebinding it with the soluble cognate v-SNARE region, V c , accelerates SNAREpin assembly [bib_ref] N-to C-terminal SNARE complex assembly promotes rapid membrane fusion, Pobbati [/bib_ref] [bib_ref] Regulation of membrane fusion by the membrane-proximal coil of the t-SNARE during..., Melia [/bib_ref]. A putative explanation to conciliate these observations is that V cbinding structures the four-helix bundle at the Cterminal part of the SNARE domain, and this structure propagates in the N-terminal region of the t-SNARE domain, thereby opening the groove where v-SNARE can directly bind. In vitro experiments suggest that this structural remodeling reduces the activation energy for SNAREpin initiation to 8 k B T [bib_ref] Kinetic barriers to SNAREpin assembly in the regulation of membrane docking/priming and..., Li [/bib_ref] , which would make the SNAREpin assembly extremely fast (0.3-30 {s according to Eqn 4).
## . . .with the help of chaperones
From the minimal initial assembly models presented above, two main hurdles to achieve fusion can be identified: vesicletarget membrane meeting and SNAREpin zippering initiation. These limitations may look like they would create difficulties in physiology by slowing down the fusion process, just as it was observed in the lipid mixing assay. Actually, they can be switched on or off by effectors and appear as assets used by cells and organelles to control and induce SNAREpin formation [bib_ref] SNARE regulatory proteins in synaptic vesicle fusion and recycling, Sauvola [/bib_ref].
To bring and maintain the vesicles at a distance compatible with SNAREpin formation, 10-20 nm, long tethers are used. They include the long bananashaped protein Munc13 that can form complexes with proteins from the RIM and Rab families [bib_ref] A Munc13/RIM/Rab3 tripartite complex: from priming to plasticity?, Dulubova [/bib_ref] [bib_ref] Heterodimerization of Munc13 C2A domain with RIM regulates synaptic vesicle docking and..., Camacho [/bib_ref] [bib_ref] Functional interaction of the active zone proteins Munc13-1 and RIM1 in synaptic..., Betz [/bib_ref] [bib_ref] RIM proteins activate vesicle priming by reversing autoinhibitory homodimerization of Munc13, Deng [/bib_ref] [bib_ref] RIM proteins tether Ca2+ channels to presynaptic active zones via a direct..., Kaeser [/bib_ref] [bib_ref] The RAB3-RIM pathway is essential for the release of neuromodulators, Persoon [/bib_ref] [bib_ref] Rab GTPases as coordinators of vesicle traffic, Stenmark [/bib_ref] [bib_ref] Mechanistic insights into neurotransmitter release and presynaptic plasticity from the crystal structure..., Xu [/bib_ref]. By extending these tethers away from the target membrane surface near calcium channels, these complexes are able to capture vesicles and position them at the location where calcium will enter during neurotransmission. A pool of vesicles can thus be permanently docked, thereby overpassing the difficulty for the vesicles to meet the target membrane [bib_ref] Vesicle capture by membrane-bound Munc13-1 requires selfassembly into discrete clusters, Li [/bib_ref]. This description corresponds to the synaptic SNAREpins. For other SNAREpins, other macromolecular complexes are used to tether the vesicles. They are often referred to as Multisubunit Tethering Complexes or MTCs [bib_ref] Structural basis for the binding of SNAREs to the multisubunit tethering complex..., Travis [/bib_ref] [bib_ref] Multisubunit tethering complexes and their role in membrane fusion, Brocker [/bib_ref].
Switching off the energy barrier for initial SNAREpin assembly requires another effector. This function is achieved by Sec1/Munc18-like proteins [bib_ref] Selective activation of cognate SNAREpins by Sec1/Munc18 proteins, Shen [/bib_ref] [bib_ref] Reconstitution of the vital functions of Munc18 and Munc13 in neurotransmitter release, Ma [/bib_ref] [bib_ref] Mechanism of neurotransmitter release coming into focus, Rizo [/bib_ref] [bib_ref] Munc18-1 is crucial to overcome the inhibition of synaptic vesicle fusion by..., Stepien [/bib_ref] [bib_ref] Synaptotagmin-1-, Munc18-1-, and Munc13-1-dependent liposome fusion with a few neuronal SNAREs, Stepien [/bib_ref]. To better understand the activation role of Munc18, a point needs to be clarified regarding t-SNAREs. In most in vitro experiments, the t-SNARE complex made of Syn1A and SNAP25 is preassembled. This is not the case in vivo. Hence, the activation energy for SNAREpin initiation presented above cannot quantitatively represent the physiological reality: for example, there is no need to open a groove in the t-SNARE for v-SNARE to bind. However, in neurons, the Nterminal part of the Syn1A SNARE domain forms a four-helix bundle with the so-called Habc N-terminal of Syn1A [bib_ref] Syntaxin-1 N-peptide and Habc-domain perform distinct essential functions in synaptic vesicle fusion, Zhou [/bib_ref] [bib_ref] Threedimensional structure of the neuronal-Sec1-syntaxin 1a complex, Misura [/bib_ref]. This coiled-coil needs to be disrupted to allow SNAP25 and VAMP2 binding. This disruption of the protein complex requires energy that will be a barrier to initial SNAREpin assembly. To our knowledge, the actual value of this energy barrier has not been measured but, because it entails disrupting more bonds than the opening of a groove in vitro, it is likely to be larger than the 23 and 30 k B T. At the molecular level, Munc13 is needed to open the Habc domain [bib_ref] Reconstitution of the vital functions of Munc18 and Munc13 in neurotransmitter release, Ma [/bib_ref] , Munc18 can grab the N-terminal of both Syn1A and VAMP2 SNARE domains [bib_ref] Membrane fusion: grappling with SNARE and SM proteins, Sudhof [/bib_ref] [bib_ref] Testing the SNARE/ SM protein model of membrane fusion, Bacaj [/bib_ref] and bring them together [bib_ref] A direct role for the Sec1/Munc18-family protein Vps33 as a template for..., Baker [/bib_ref]. SNAP25 binds to Munc13, which chaperones its assembly with Syn1A and VAMP2 to initiate SNAREpin formation [bib_ref] Munc13 binds and recruits SNAP25 to chaperone SNARE complex assembly, Sundaram [/bib_ref].
Finally, the membrane distribution of the t-SNARE may help chaperones to accelerate the initial SNAREpin assembly by increasing encounter probability between cognate SNAREs. For instance, Syn1A is known to form in vivo microdomains of different sizes in equilibrium with freely diffusing proteins. Superresolution techniques hint at clusters of diameter 50-80 nm with 30-90 copies of Syn1A, colocalizing with SNAP25 clusters having at least a similar number of copies [bib_ref] Anatomy and dynamics of a supramolecular membrane protein cluster, Sieber [/bib_ref] [bib_ref] T-SNARE protein conformations patterned by the lipid microenvironment, Rickman [/bib_ref] [bib_ref] Dynamical Organization of Syntaxin-1A at the presynaptic active zone, Ullrich [/bib_ref] [bib_ref] Syntaxin clusters assemble reversibly at sites of secretory granules in live cells, Barg [/bib_ref] [bib_ref] Single secretory granules of live cells recruit syntaxin-1 and synaptosomal associated protein..., Knowles [/bib_ref]. The evidence hence suggests the existence of domains with very high concentrations in t-SNAREs, scaling as tens of thousands of complexes per {m 2 , which probably improves the speed of the docking and priming process. The size, composition, structure, and organization of the clusters are not yet fully understood but might be controlled by lipid composition, protein-protein interactions [bib_ref] Anatomy and dynamics of a supramolecular membrane protein cluster, Sieber [/bib_ref] [bib_ref] T-SNARE protein conformations patterned by the lipid microenvironment, Rickman [/bib_ref] , the inclusion in an active zone [bib_ref] Dynamical Organization of Syntaxin-1A at the presynaptic active zone, Ullrich [/bib_ref] , and the presence of a primed vesicle [bib_ref] Syntaxin clusters assemble reversibly at sites of secretory granules in live cells, Barg [/bib_ref] [bib_ref] Single secretory granules of live cells recruit syntaxin-1 and synaptosomal associated protein..., Knowles [/bib_ref]. These possibilities of modulating clusters could as well provide more control over fusion.
## Pore opening and expansion: the end of the fusion process
In this second part, we will focus on the final action of the SNAREpins: the fusion process itself, i.e., the actual merging of the two membranes into a single entity. As the SNAREpin zippers, the apposed membranes come in close proximity. When the remaining water layer between them is 1-2 nm, depending on the membrane composition [bib_ref] Hydration forces between phospholipid-bilayers, Rand [/bib_ref] [bib_ref] New phases of phospholipids and implications to the membrane fusion problem, Yang [/bib_ref] , they are destabilized and a fusion pore opens. This short intermembrane separation at fusion suggests that only the C-terminal regions of the SNAREs, probably beyond layer +3, may play an active part in the actual fusion process. This hypothesis is consistent with experimental observations [bib_ref] Close is not enough: SNAREdependent membrane fusion requires an active mechanism that..., Mcnew [/bib_ref] [bib_ref] A conformational switch in complexin is required for synaptotagmin to trigger synaptic..., Krishnakumar [/bib_ref].
At the molecular level, the destabilization of the membranes towards the formation of the fusion pore is a complex process that has been the focus of many studies [bib_ref] Energetics of intermediates in membrane fusion: comparison of stalk and inverted micellar..., Siegel [/bib_ref] [bib_ref] Stalk mechanism of vesicle fusion. Intermixing of aqueous contents, Kozlov [/bib_ref] [bib_ref] A quantitative model for membrane fusion based on low-energy intermediates, Kuzmin [/bib_ref] [bib_ref] Membrane fusion: stalk model revisited, Markin [/bib_ref] [bib_ref] Lipid intermediates in membrane fusion: formation, structure, and decay of hemifusion diaphragm, Kozlovsky [/bib_ref] [bib_ref] Stalk model of membrane fusion: solution of energy crisis, Kozlovsky [/bib_ref] [bib_ref] Structure and energy of fusion stalks: the role of membrane edges, May [/bib_ref] [bib_ref] Stalk phase formation: effects of dehydration and saddle splay modulus, Kozlovsky [/bib_ref] [bib_ref] Solvent-exposed tails as prestalk transition states for membrane fusion at low hydration, Smirnova [/bib_ref] [bib_ref] Free energy analysis along the stalk mechanism of membrane fusion, Kawamoto [/bib_ref] [bib_ref] Calculating transition energy barriers and characterizing activation states for steps of fusion, Ryham [/bib_ref] [bib_ref] Secretory and viral fusion may share mechanistic events with fusion between curved..., Lee [/bib_ref] [bib_ref] Topological relaxation of a shear-induced lamellar phase to sponge equilibrium and the..., Porcar [/bib_ref] [bib_ref] Mechanics of membrane fusion, Chernomordik [/bib_ref]. SNAREpins may actually influence this molecular pathway, favor the formation of intermediate lipid and protein arrangements [bib_ref] A cascade of multiple proteins and lipids catalyzes membrane fusion, Wickner [/bib_ref] [bib_ref] Conflicting views on the membrane fusion machinery and the fusion pore, Sorensen [/bib_ref] , and affect the nature of the pore (see Box 1 for related discussion). Each fusion event will go through a different molecular pathway since hundreds of molecules are involved and the geometry of lipids will favor some fusion pathways over others [bib_ref] Cooperation of conical and polyunsaturated lipids to regulate initiation and processing of..., Francois-Martin [/bib_ref] [bib_ref] Non-bilayer lipids and biological fusion intermediates, Chernomordik [/bib_ref] [bib_ref] Bending membranes to the task: structural intermediates in bilayer fusion, Chernomordik [/bib_ref] [bib_ref] Lipid regulation of cell membrane structure and function, Yeagle [/bib_ref] [bib_ref] Polyunsaturated phospholipids facilitate membrane deformation and fission by endocytic proteins, Pinot [/bib_ref]. To circumvent this variability inherent to complex systems, we will envision the fusion process as a single reaction with a global activation energy barrier that needs to be passed to open the fusion pore. The main reason for this approach is that, experimentally, fusion is usually demonstrated by the actual opening of a fusion pore and not by the intermediate states. In any case, SNAREpins lead to the same final result: the formation of an extended fusion pore.
In this part devoted to the formation and expansion of the fusion pore, we will first describe the different types of observations, present the current view of pore opening and expansion, and model the energetics involved in each step.
Box 1: Nature of the fusion pore: Lipids, proteins, both?
Understanding the molecular nature of the fusion pore is a prerequisite to understand quantitatively how the fusion pore opens and expands. This is a difficult task because at 1 to 10 nanometer scale, molecules are very dynamic, and the timescale for movement is dozens of ns [bib_ref] Lipid simulations: a perspective on lipids in action, Vattulainen [/bib_ref]. This fast movement of the molecules always needs to be kept in mind: there is no such thing as a constant nature of a fusion pore. In any case, we will try to identify the molecular regions that are the most likely to be decorating the rim of the pore. Since the pore is aqueous it will always be energetically more favorable to have hydrophilic motives exposed to the inside of the pore. However, in lipid bilayers, it is well documented that hydrophobic parts are frequently facing the aqueous region. In the same way, there is no doubt that lipid chains and hydrophobic residues from the SNARE transmembrane domains can transiently be exposed at the rim of the pore.
The best picture of the typical molecular nature of the fusion pore is probably obtained by molecular dynamics simulations. They show that the pore is mainly decorated by polar heads of lipids and C-terminal regions of the SNAREs [bib_ref] Molecular mechanism of fusion pore formation driven by the neuronal SNARE complex, Sharma [/bib_ref]. Experimental observation suggests that the transmembrane domains can also be in contact with aqueous phases [bib_ref] Transmembrane segments of syntaxin line the fusion pore of Ca2+−triggered exocytosis, Han [/bib_ref]. It remains unclear whether they are in direct contact with the aqueous pore or with inverted micelles that may form during the fusion process [bib_ref] Exocytotic fusion pores are composed of both lipids and proteins, Bao [/bib_ref].
In summary, the pore appears to be mainly delineated by polar heads of lipids with a scarce presence of protein residues, primarily coming from the C-terminal region of the SNAREs and possibly also from the transmembrane domain [bib_ref] The fusion pore, 60 years after the first cartoon, Sharma [/bib_ref].
How to probe the fusion pore
Two main types of experimental measurements are performed: optical and electrical. Optical observations using fluorescence dequenching can be achieved in bulk or at the single fusion event level. Quenched fluorescent dyes are placed in the vesicle, either bound to the membrane as in the lipid mixing assay presented in the first part or in the lumen, referred to as 'content mixing' assays [fig_ref] Figure 4: Optical assays [/fig_ref] [bib_ref] Content mixing and membrane integrity during membrane fusion driven by pairing of..., Nickel [/bib_ref] [bib_ref] A fast, singlevesicle fusion assay mimics physiological SNARE requirements, Karatekin [/bib_ref] [bib_ref] Single molecule observation of liposome-bilayer fusion thermally induced by soluble N-ethyl maleimide..., Bowen [/bib_ref] [bib_ref] Imaging single membrane fusion events mediated by SNARE proteins, Fix [/bib_ref] [bib_ref] High-throughput monitoring of single vesicle fusion using freestanding membranes and automated analysis, Ramakrishnan [/bib_ref] [bib_ref] Dynamics and number of trans-SNARE complexes determine nascent fusion pore properties, Bao [/bib_ref] [bib_ref] Resolving kinetic intermediates during the regulated assembly and disassembly of fusion pores, Das [/bib_ref] [bib_ref] Nanodisc-cell fusion: control of fusion pore nucleation and lifetimes by SNARE protein..., Wu [/bib_ref] [bib_ref] Multiple intermediates in SNARE-induced membrane fusion, Yoon [/bib_ref] [bib_ref] A chemical controller of SNARE-driven membrane fusion that primes vesicles for ca(2+)-triggered..., Heo [/bib_ref] [bib_ref] In vitro system capable of differentiating fast Ca2+−triggered content mixing from lipid..., Kyoung [/bib_ref]. Upon fusion, the dyes diffuse away from the vesicle and their release is observed by the resulting increase in fluorescence due to the dequenching. The main limitation of lipid mixing assays is that they do not directly account for the opening of a fusion pore. For instance, a hemifusion state in which only the external leaflets of the two membranes have merged may be mistakenly confused with fusion. Lipid mixing also provides limited information on the fusion pore kinetics because the dyes are released extremely fast, typically in ms for a 1 nm diameter pore (see Appendix C). Conversely, the release of encapsulated fluorescent dyes through a fusion pore occurs on a slower time scale because the pore first needs to expand. The main difficulty of content release assays is to verify that the cargo does not diffuse away from the vesicle through leaks induced by the mechanical action of surface tension or by chemical modification of the membrane properties. Ideally, both 'lipid mixing' and 'content release' assays should be performed in parallel to ensure the validity of the results.
Electrical observations can also be used to monitor the kinetics of the fusion pore [bib_ref] Currents through the fusion pore that forms during exocytosis of a secretory..., Breckenridge [/bib_ref] [bib_ref] Nascent fusion pore opening monitored at single-SNAREpin resolution, Heo [/bib_ref]. They require to place at least one electrode on each side of the target membrane. In theory, monitoring the impedance of the vesicle/target membrane system during the fusion process allows the simultaneous characterization of the pore kinetics and the vesicle size by measuring the conductance and capacitance, respectively [fig_ref] Figure 5: Electrical assays [/fig_ref]. In reality, the conductance only provides transient information because the voltage difference between both sides of the fusion pore quickly vanishes to zero. This issue can be resolved by imposing a permanent voltage between the two sides of the target membrane and placing the lumen of the vesicle in electrolytic contact with the vesicle exterior [fig_ref] Figure 5: Electrical assays [/fig_ref]. Electrolytic contact can be achieved by either adding channels in the vesicle membrane [bib_ref] Nascent fusion pore opening monitored at single-SNAREpin resolution, Heo [/bib_ref] or replacing the vesicle with a small membrane patch, called nanodisc [bib_ref] Dynamics and number of trans-SNARE complexes determine nascent fusion pore properties, Bao [/bib_ref] [bib_ref] Resolving kinetic intermediates during the regulated assembly and disassembly of fusion pores, Das [/bib_ref] [bib_ref] Nanodisc-cell fusion: control of fusion pore nucleation and lifetimes by SNARE protein..., Wu [/bib_ref].
Finally, in the last years, with the increasing computational strength and the theoretical progress in the field, molecular dynamics simulations have proven to be a more insightful method to numerically probe the structural and functional properties of biological systems. Molecular dynamics simulations, by providing unique information on molecular remodeling and arrangement during fusion, nicely complement experimental observations.
We will now discuss the two main steps in the fusion process: the nucleation, i.e., the opening of a pore, and its subsequent growth.
## How to seed a fusion pore
Before discussing SNARE-induced fusion pore opening, it is important to understand the dimensions and energies involved. Here, we will present a model assuming that membrane interactions involved in the fusion process are purely associated with the physical and chemical properties of lipid bilayers; proteins may actually alter these interactions but are unlikely to significantly change the orders of magnitude (see Box 1). The analogy noticed almost 50 years ago between the fusion process and the transition from lamellar phases to other phases, e.g., hexagonal or rhombohedral phases [bib_ref] Mimicry and mechanism in phospholipid models of membrane-fusion, Rand [/bib_ref] [bib_ref] Hydration forces between phospholipid-bilayers, Rand [/bib_ref] [bib_ref] A rhombohedral phase of lipid containing a membrane fusion intermediate structure, Yang [/bib_ref] , provides quantitative insights. Merging the membranes and opening a fusion pore necessitates overcoming the sharp short-range hydration/ protrusion forces between membranes. These repulsive surface forces, F R , decay exponentially with the separation distance, d:
[formula] F R d ð Þ ¼ P 0 e À d λ(5) [/formula]
with a characteristic length, λ, of a few Angstroms and a prefactor, P 0 , of about 100 atm [bib_ref] Mimicry and mechanism in phospholipid models of membrane-fusion, Rand [/bib_ref] [bib_ref] Hydration forces between phospholipid-bilayers, Rand [/bib_ref]. This explains why merging the membranes and opening a fusion pore is energetically costly, ∼ 25 k B T [bib_ref] Low energy cost for optimal speed and control of membrane fusion, Francois-Martin [/bib_ref] [bib_ref] Cooperation of conical and polyunsaturated lipids to regulate initiation and processing of..., Francois-Martin [/bib_ref]. Since this energy must be provided over a very short distance, typically 1 nm [bib_ref] Mimicry and mechanism in phospholipid models of membrane-fusion, Rand [/bib_ref] [bib_ref] Hydration forces between phospholipid-bilayers, Rand [/bib_ref] , the average force is 100 pN.
Assuming that the pore opens at 100 atm pressure, this force should be applied to an area of 10 nm 2 , which is occupied by 15 lipids. Hence, the initial opening of a fusion pore probably involves ∼ 100 lipids when accounting for both leaflets.
## I i
## Membrane fusion
Membrane fusion A 25 k B T energy barrier is sufficient to prevent spontaneous fusion. Indeed, just as in the first part for the initiation of the SNAREpin, we can use Kramers reaction-rate theory (Eqn 4, [bib_ref] Brownian motion in a field of force and the diffusion model of..., Kramers [/bib_ref] [bib_ref] Reaction-rate theory -50 years after Kramers, Hanggi [/bib_ref] , to estimate the waiting time, τ w , before thermal fluctuations provide enough energy for passing the fusion barrier, E b :
[formula] V v V b V t V t V b V t V v V b V t V b V bτ w ¼ τ 0 exp E b k B T(6) [/formula]
Equation (6) indicates that the waiting time is in the minute scale for E b ¼ 25 k B T, which predicts that fusion will not spontaneously occur on an experimentally relevant time scale for neurotransmission, highlighting the physiological need for SNAREpins. Here, we will provide a simple model describing N SNAREpins temporarily clamped in a partly assembled state and simultaneously released, approximately mimicking the role of the calcium sensor Synaptotagmin-1. In this model, the acceleration of the fusion process by SNAREs can be quantitatively estimated by calculating the duration of two distinct phases in the SNAREpin: approaching the membranes and the actual opening of the fusion pore. First, the SNAREpins must reduce the vesicle-target membrane distance from their initial separation to the minimum of the energy landscape before the fusion barrier, i.e., 2 to 3 nm. This is achieved by the pulling force applied by each SNAREpin, F p . Because the system is overdamped, the speed of the vesicle, v, is driven by the drag force, i.e., the Stokes force, that opposes the N SNAREpins pulling force:
v ¼ NF p 3πdη (7) where η is the viscosity of the surrounding aqueous medium and d the vesicle diameter. Hence, the time, τ t , to travel a distance is:
[formula] τ t N ð Þ ¼ 3πldη NF p(8) [/formula]
where l is the total displacement of the vesicle. Once the vesicle has reached the minimum of the energy landscape, it faces the fusion barrier that must be overcome by thermal fluctuations for fusion to occur. Because the SNAREpins are pulling on the membranes they reduce the height of the fusion barrier. Hence, using Kramers' reaction-rate theory, the waiting time for N SNAREpins becomes:
[formula] τ w N ð Þ ¼ τ 0 exp E b ÀNδe k B T(9) [/formula]
where δe is the energy reduction in the fusion barrier due to a single SNAREpin. The fusion time to bring the vesicle from their initial separation distance to contact and subsequent fusion τ f N ð Þ is the sum of the travel and the waiting time:
[formula] τ f N ð Þ ¼ τ t N ð Þ þ τ w N ð Þ ¼ 3πldη NF p þ τ 0 exp E b ÀNδe k B T(10) [/formula]
Two regimes are predicted for τ f N ð Þ [fig_ref] Figure 6: Fusion time [/fig_ref]. In the first regime, at low N, the waiting time is limiting. Then, the fusion time decays exponentially with the number of SNAREpins. In the second regime, at higher N, the travel time is limiting and τ f N ð Þ is inversely proportional to N. Using the values in Appendix D, Eqn (8) predicts a threshold value of N = 4 SNAREpins for which the fusion time is dozens of ns, i.e., extremely fast [fig_ref] Figure 6: Fusion time [/fig_ref]. These predictions are in quantitative agreement with the experimental observations suggesting that it takes ∼ 1 s for a single SNAREpin to drive fusion [bib_ref] A programmable DNA origami platform to organize SNAREs for membrane fusion, Xu [/bib_ref] [bib_ref] Calcium secretion coupling at calyx of held governed by nonuniform channel-vesicle topography, Meinrenken [/bib_ref]. [fig_ref] Figure 6: Fusion time [/fig_ref] also implies that three or more SNAREpins must act simultaneously to achieve neurotransmitter release in less than 1 ms in synaptic transmission. The prediction of this simple model on the number of SNAREpins is consistent with experimental observations [bib_ref] One SNARE complex is sufficient for membrane fusion, Van Den Bogaart [/bib_ref] [bib_ref] SNARE proteins: one to fuse and three to keep the nascent fusion..., Shi [/bib_ref] and molecular dynamic simulations showing once the SNARE domains are almost fully zippered, the membranes are in such close apposition that the polar headgroups of the outer lipid leaflets are dehydrated to a level allowing fusion [bib_ref] Coarse-grained model of SNARE-mediated docking, Fortoul [/bib_ref].
The simple model we present here suggests a monotonic decrease in the fusion time with the number of SNAREpins. Intriguingly, several models suggest that there is an optimum SNAREpin number for fast fusion because using too many SNAREpins in the contact area of the vesicle and the target membrane would actually slow down the fusion process [bib_ref] Coarse-grained model of SNARE-mediated docking, Fortoul [/bib_ref] [bib_ref] SNARE machinery is optimized for ultrafast fusion, Manca [/bib_ref]. Two reasons for the existence of an optimum number have been proposed. First, steric repulsions increase the equilibrium intermembrane docking distance impeding efficient fusion. Molecular dynamics simulations predict a shift of the membrane separation from 2 to 3 nm when varying the number of active SNAREpins from 7 to 13 [bib_ref] Coarse-grained model of SNARE-mediated docking, Fortoul [/bib_ref].
Second, a mechanical model shows that the SNAREpins, which are not sufficiently zippered provide a force opposing fusion; the predicted optimal number of SNAREpins before this effect becomes dominant is 3 to 7 [bib_ref] SNARE machinery is optimized for ultrafast fusion, Manca [/bib_ref]. There is no experimental proof yet of the existence of such an optimum number of SNAREpins.
For fusion to actually occur, the zippering force applied by the SNAREpins to the apposing membranes needs to be transmitted by the linker and transmembrane domains [fig_ref] Figure 7: Role of the various parts of the C-terminal end of the SNAREpin... [/fig_ref]. To test the actual role of these domains, experiments and molecular dynamics simulations have been performed with specific mutations, deletions, or substitutions with lipid chains [bib_ref] Helical extension of the neuronal SNARE complex into the membrane, Stein [/bib_ref] [bib_ref] Molecular mechanism of fusion pore formation driven by the neuronal SNARE complex, Sharma [/bib_ref] [bib_ref] Caught in the act: visualization of SNARE-mediated fusion events in molecular detail, Risselada [/bib_ref] [bib_ref] Transmembrane-domain determinants for SNARE-mediated membrane fusion, Fdez [/bib_ref]. The assembly of the linker and transmembrane domains into coiled coils seems to provide energy to help pore opening and possibly subsequent expansion. However, there is an open question on the structure and rigidity of the linker domain that condition the efficiency of the force transmission.
The last element of the SNAREs that plays a significant part in the nucleation process is the very Cterminal end of VAMP2 [bib_ref] Role of the synaptobrevin C terminus in fusion pore formation, Ngatchou [/bib_ref]. VAMP2 has 2 hydrophilic uncharged residues after the transmembrane domains that are preserved across species, usually Ser-Thr or sometimes Ser-Ser. Several studies show that they play an important role in pore opening by inducing the deformation of the bilayer around the Cterminal leading to the nucleation of the pore by forcing the rearrangement of lipids. Because the vesicle has a high positive curvature in contrast to the nascent fusion pore characterized by a high negative curvature, a dramatic change in curvature occurs on the vesicle side. The two hydrophilic residues provide leverage for this transition. Intriguingly, Syn1A ends with the transmembrane domain without any subsequent hydrophilic residue. The curvature changes on the target membrane side are not as drastic and the strong anchorage of the t-SNARE in the hydrophobic core through Syn1A and SNAP25 seems to be sufficient to ensure optimal fusion [bib_ref] Close is not enough: SNAREdependent membrane fusion requires an active mechanism that..., Mcnew [/bib_ref] [bib_ref] Lipid-anchored SNAREs lacking transmembrane regions fully support membrane fusion during neurotransmitter release, Zhou [/bib_ref].
## How to grow a fusion pore
Opening a fusion pore is not sufficient to ensure full fusion. The importance of the subsequent expansion of the pore must not be underestimated because it is not a spontaneous process and also requires some energy. Expansion of the nascent fusion pore is associated with the energetically costly extension of the highly curved rim. Those curvatures will be continuously reduced as the pore extends (see Appendix E for explanations). Using a simple model based on curvature energy and crude torus-like geometry, there is a threshold pore diameter corresponding to an expansion barrier [fig_ref] Figure 8: Expansion barrier [/fig_ref]. If the pore expands above this threshold diameter, it spontaneously expands. Conversely, if there is not enough energy to pass the barrier, the pore ultimately reseals. Resealing is not a straightforward process either since, just like for opening the fusion pore, the two membranes that form the rim of the pore must merge to form fully distinct lumens. In this situation, the pore is trapped in a transiently open state and will eventually reseal when thermal fluctuations provide enough energy to overcome this resistance to resealing.
Considering each SNAREpin provides a constant force towards the expansion of the fusion pore, the energy landscape with one, two, or three SNAREpins can be computed in the crude torus-pore model. Using the energy landscape for pore expansion resulting from this model and the typical force applied by SNAREpins, 3 SNAREpins would start spontaneously expanding the pore [fig_ref] Figure 8: Expansion barrier [/fig_ref] and Appendix E). Several types of in vitro experiments with nanodiscs, vesicles, and suspended membranes have investigated the effect of the number of SNAREpins on the nascent fusion pore [bib_ref] Dynamics and number of trans-SNARE complexes determine nascent fusion pore properties, Bao [/bib_ref] [bib_ref] Nascent fusion pore opening monitored at single-SNAREpin resolution, Heo [/bib_ref] [bib_ref] One SNARE complex is sufficient for membrane fusion, Van Den Bogaart [/bib_ref] [bib_ref] SNARE proteins: one to fuse and three to keep the nascent fusion..., Shi [/bib_ref]. These studies consistently suggest that one or two SNAREpins are indeed able to open a fusion pore but cannot expand beyond the expansion barrier, making the fusion pore transient. The average apparent diameter of a fusion pore induced by a single (resp. two) SNAREpin(s) seems to be in the range of 0.3-0.4 nm (resp. 0.8-0.9 nm) [bib_ref] Nascent fusion pore opening monitored at single-SNAREpin resolution, Heo [/bib_ref]. These transient pores reseal after a few 100 ms when the SNAREpins run out of energy, i.e., when the transmembrane domains are fully zipped.
After fusion pore opening, the SNARE domains and at least part of the linker domains are already assembled. Hence, the energy for expanding the pore is expected to come from the zipping of the transmembrane domains. This is indeed what is experimentally observed. When the transmembrane domains are replaced by lipid chains or other noninteracting transmembrane domains, cargo release is reduced to the level induced by one or two SNAREpins [bib_ref] SNARE proteins: one to fuse and three to keep the nascent fusion..., Shi [/bib_ref].
## How to catalyze pore opening and expansion in vivo?
We saw that the formation of an expanded fusion pore is energetically opposed at two stages of the process: the fusion and the expansion barriers. Intriguingly, while a couple of SNAREpins are not sufficient to bypass these barriers, a handful of simultaneously acting SNAREpins provides enough energy to make them both disappear, making the fusion process spontaneous. Hence, for cells to precisely control the time of fusion, several SNAREpins must be Examples of energy landscapes of the pore expansion without SNAREpin (blue) and upon the action of one (orange), two (green), or three (red) SNAREpins. For these predictions, the vesicle radius, bending modulus, height of the vesicle distance, membrane thickness, and SNAREpin contribution were set at 25 nm, 10 k B T , 2 nm, 5 nm, and 10 pN, respectively. 10 pN means each SANREpin provides~2.5 k B T per nm increase in the pore radius. The quantitative details of the model used to obtain these landscapes are described in Appendix E.
synchronized. This synchronicity is primarily achieved by several copies of the calcium sensors Synaptotagmin-1 that clamp a few SNAREpins in a partly zipped state and synchronously release them upon calcium entry. Another potential protagonist of this synchronization on the synaptic vesicle, synaptophysin, forms hexameric structures necessary to make the synaptic vesicle functional and is able to bind VAMP2 [bib_ref] Architecture of the Synaptophysin/Synaptobrevin complex: structural evidence for an entropic clustering function..., Adams [/bib_ref] [bib_ref] Room for two: the Synaptophysin/Synaptobrevin complex, White [/bib_ref] [bib_ref] The synaptophysinsynaptobrevin complex: a hallmark of synaptic vesicle maturation, Becher [/bib_ref] [bib_ref] Synaptophysin I controls the targeting of VAMP2/ synaptobrevin II to synaptic vesicles, Pennuto [/bib_ref]. This organization may regulate the number of v-SNAREs presented to the target membrane [bib_ref] Symmetrical arrangement of proteins under release-ready vesicles in presynaptic terminals, Radhakrishnan [/bib_ref]. In this cryoelectron tomography study, it was proposed that each protein complex of the hexameric structure contains one partly assembled SNAREpin with VAMP2 emanating from the Synaptophysin:VAMP2 complex. To match this hexameric structure on the plasma membrane side, it has been proposed that Munc13, possibly helped by Synaptotagmin-1, oligomerizes in a ring-like structure, facilitating the assembly of exactly 6 SNAREpins [bib_ref] Hypothesis -buttressed rings assemble, clamp, and release SNAREpins for synaptic transmission, Rothman [/bib_ref]. The perfect matching of the symmetry between the two membranes is an appealing solution to guarantee that the optimum number of SNAREpins are acting together when fusion is triggered. However, these mechanisms still remain to be proven both structurally and functionally.
# Conclusion
The overall SNARE-induced fusion process is considerably slowed down by three main energy barriers: initial assembly of the SNAREpin, fusion pore opening, and pore expansion. Initial assembly occurs at the very N-terminal part of the SNARE domains and requires structural changes in the t-SNARE that are energetically costly. To open and expand the fusion pore, each part of the SNAREpins from layer +3/+4 of the SNARE domain to the very C-terminal plays a specific part. Zipping of the SNAREpin from layer +3 to +7 is responsible for bringing the two membranes into molecular contact. Zipping of layers +7, +8, and possibly part of the linker domain provides the energy for overcoming the fusion barrier. It is likely that 5 SNAREpins or more are necessary to make the fusion barrier disappear although thermal fluctuations are sufficient to overcome it in less than Appendix A
## Rate of collision of vesicles/nanodiscs in a lipid mixing assay
The rate of collision of vesicles or nanodiscs in the bulk can be computed with the standard Smoluchowski approach [bib_ref] Experiments on a mathematical theory of kinetic coagulation of colloid solutions, Smoluchowski [/bib_ref]. We consider two types of vesicles/nanodiscs:
- Those containing v-SNAREs, named 'v-particle' in the following, of hydrodynamic radius R v , bulk concentration ρ v,∞ , and diffusion coefficient D v . - Those containing t-SNAREs, named 't-particle' in the following, of hydrodynamic radius R t , bulk concentration ρ t,∞ , and diffusion coefficient D t .
Let us fix the coordinate system on the center of a v-particle. We want to know the flux of t-particles colliding with our v-particle because of diffusion processes. The spatiotemporal profile of the concentration in t-particles ρ t,∞ is given by Fick's second law:
[formula] ∂ρ t ∂t ¼ DΔρ t A1 [/formula]
where D is the diffusion coefficient of t-particles in the referential of v, which can be shown to be
[formula] D ¼ D t þ D v . [/formula]
## Stokes-einstein's equation then gives us
[formula] D ¼ k B T R t þ R v ð Þ 6πηR t R v (A2) [/formula]
with η the dynamic viscosity of the solution, which is approximately equal to that of water. In a steady-state regime, Eqn (A1) boils down to the Laplace equation:
[formula] Δρ t ¼ 0 (A3) [/formula]
which can be solved in spherical coordinates.
Assuming that, upon a collision, there is no aggregation (fusion or bouncing back are rapid events) the boundary condition around the v-particle is:
[formula] ρ t R t þ R v ð Þ¼0 (A4) [/formula]
The spatial concentration of the t-particle at a distance r from the considered v-particle can then be obtained from Eqns (A3,A4):
[formula] ρ tj r > R t þRv ¼ ρ t,∞ 1À R t þ R v r (A5) [/formula]
The flux of t-particles coming in collision with our vparticle can then be deduced from Fick's first law:
[formula] J ¼ ÀD dρ t dr jR t þRv ¼ k B Tρ t,∞ 6πηR t R v (A6) [/formula]
By integrating the flux over the sphere of radius R t + R v , we finally get the following collision rate:
[formula] ν ¼ 2 R t þ R v ð Þ 2 k B Tρ t,∞ 3ηR t R v (A7) [/formula]
What is remarkable here is that the collision rate only depends on the ratio between the two radii. In particular, if the two objects have the same radius:
[formula] ν ¼ 8k B Tρ t,∞ 3η A8 [/formula]
To find orders of magnitude that are consistent with what is usually done experimentally, we will take vand t-vesicles of monodisperse radius 25 nm, each of them with a final lipid concentration of 1 mM [bib_ref] SNAREpins: minimal machinery for membrane fusion, Weber [/bib_ref] [bib_ref] Actual fusion efficiency in the lipid mixing assay -comparison between nanodiscs and..., Francois-Martin [/bib_ref] [bib_ref] Protein determinants of SNARE-mediated lipid mixing, Ji [/bib_ref]. The molecular area of a lipid is typically 0.65 nm 2 [bib_ref] Hydration forces between phospholipid-bilayers, Rand [/bib_ref]. Thus, the number of lipids per SUV can be estimated to be 20 000. The resulting molar concentration in vesicles is 50 nM, i.e., the t-vesicle concentration is ρ t,∞ ¼ 2:5:10 19 vesiclesÁm −3 , which yields ν ∼ 300 collisionsÁs −1 . Given an initial slope of the corrected dequenching curve ∼ 1Á4000 s −1 for vesicle-vesicle, this means less than 1 collision out of 1 million is successful [bib_ref] Actual fusion efficiency in the lipid mixing assay -comparison between nanodiscs and..., Francois-Martin [/bib_ref] [bib_ref] Determinants of liposome fusion mediated by synaptic SNARE proteins, Schuette [/bib_ref].
# Appendix b
SNARE additivity and cooperativity in the lipid mixing bulk assay According to Eqn (3), during a collision, a SNARE covers the area occupied by ∼ 100 lipids. Assuming there is no cooperativity such as oligomerization between SNAREs, when the lipid to protein ratio is significantly larger than 100, each SNARE can be considered independent of the others. With this assumption, the probability that a SNAREpin starts assembling from a specific vSUV will vary linearly with the concentration of VAMP2 in the vSUV and with the concentration of t-SNARE in the tSUV. Hence, the fusion rate in the lipid mixing bulk fusion assay, ν f , should be inversely proportional to the lipid to protein ratios in both types of SUVs:
[formula] ν f / ν ref r v r t (B1) [/formula]
Where r v (resp. r t ) is the lipid to protein ratio in the vSUV (resp. tSUV) and ν ref a reference fusion rate.
The assumption that the SNAREs behave independently of each other and the existence of threshold lipid to protein ratios can be tested by comparing the fusion rates at various values of r v and r t . The fusion kinetics has previously been systematically measured at different, accurately measured lipid to SNARE ratios varying from ∼ 80 to ∼ 3000 lipids per outward-facing SNARE [bib_ref] Protein determinants of SNARE-mediated lipid mixing, Ji [/bib_ref]. We reanalyzed the data and considered the initial kinetics are well represented by the percentage increase in fluorescence at 80 min. This approach underestimates the kinetics at high protein density because the fusion rate will go down as more vSUV fuse with tSUV but is the most accurate in the lower concentrations.
To check Eqn (B1), we took r vref ¼ 649 lipids per VAMP2 as a reference and, for each r v , we averaged
[formula] ν f rv ð Þrv ν f r v ref ð Þrvref [/formula]
over all tested r t . We calculated the same parameter for the t-SNARE taking r vref = 362 lipids per t-SNARE. According to our assumptions that there is no cooperativity of the SNARE in the lipid fusion bulk assay and that the SNARE contributions are additive below a concentration threshold, the resulting parameters, p v and p t , should be equal to 1 above a certain lipid to protein ratio.confirms this prediction suggesting that the SNAREs do not exhibit any cooperativity and have additive contributions to fusion under these experimental conditions.
# Appendix c
## Kinetics of lipid mixing and content release
In this Appendix, we will consider the fusion of a vesicle with an infinitely large target membrane. The extracellular medium will also be considered infinitely large. The vesicle membrane initially contains membranebound molecules (resp. encapsulated cargo) at a concentration c m0 (resp. c c0 ). The fusion pore is assumed to have a fixed radius, r p , and a length, L p . We will estimate the concentration of membrane-bound molecules, c m , and encapsulated cargo, c c , remaining in the vesicle in time. At any time, the concentration of membrane molecules (resp. cargo) initially in the vesicle that diffused to the target membrane (resp. extracellular medium) is zero. Hence, the membrane molecule and cargo gradients in the pore can be written as c m L p and c c L p , respectively. Using Fick's first law, the variation of c m in time can be written as:
[formula] ∂c m t ð Þ ∂t ¼ À 2πr p 4πr 2 v D m c m t ð Þ L p ¼ À r p D m 2r 2 v L p c m t ð Þ (C1) [/formula]
where D m is the diffusion coefficient of the molecules and r v the vesicle radius. The situation is slightly more complex in volume because the pore radius needs to be larger than the hydration radius of the cargo, r c . Hence, the effective pore radius is rÀr c ð Þ. Fick's first law leads to:
[formula] ∂c c t ð Þ ∂t ¼ À 3π r p Àr c À Á 2 4πr 3 v D c c c t ð Þ L p ¼ À 3 r p Àr c À Á 2 D c 4r 3 v L p c c t ð Þ (C2) [/formula]
where D c is the diffusion coefficient of the cargo, which can be estimated from the Stokes-Einstein equation:
[formula] D c ¼ k B T 6πr c η (C3) [/formula]
Hence Eqn (C2) can be rewritten:
[formula] ∂c c t ð Þ ∂t ¼ À r p Àr c À Á 2 k B T 8πr 3 v L p r c η c c t ð Þ (C4) [/formula]
For a pore of fixed radius, Eqns (C1,C4) can be readily integrated:
[formula] c m t ð Þ ¼ c m0 e À t τm(C5) [/formula]
With the following expression for the characteristic time, τ m :
[formula] τ m ¼ 2r 2 v L p r p D m (C6) [/formula]
And:
[formula] c c t ð Þ ¼ c c0 e À t τc ð Þ if r p > r c C7ac c t ð Þ ¼ c c0 if r p < r c (C7b) [/formula]
With the following expression for the characteristic time, τ c : typical examples for the release time of 95% of the molecules presented inshow that encapsulated cargos will be released much slower than membranebound molecules for fusion pores up to ∼ 1 nm in diameter but will be faster released for larger pores.
[formula] τ c ¼ 8πr 3 v L p r c η r p Àr c À Á 2 k B T( [/formula]
Hence, for an efficient cargo release during fusion, the pore needs to expand beyond a few nm. The membrane-bound molecule will be released in ms even with a small pore. Physiologically, several SNAREs act on the membrane. The fusion pore radius increases in time with a speed v almost constant, typically 1 nmÁms −1 [bib_ref] Currents through the fusion pore that forms during exocytosis of a secretory..., Breckenridge [/bib_ref]. In that case, Eqns (C1,C4) become:
[formula] ∂c m t ð Þ ∂t ¼ À vD m 2r 2 v L p tc m t ð Þ (C9a) ∂c c t ð Þ ∂t ¼ À k B T 8πr 3 v L p r c η vtÀr c ð Þ 2 c c t ð Þ (C9b) [/formula]
which can be integrated:shows the release of membrane-bound and encapsulated cargo. Expanding the fusion pore ensures a fast and complete release of both types of contents in a couple of milliseconds, which is critical for neurotransmission.
[formula] c m t ð Þ ¼ c m0 e À vDm 4r 2 v Lp t 2 C10a c c t ð Þ ¼ c c0 e À k B T 24πvr 3 v Lp rc η vtÀr c ð Þ 3 ½ , for t > r c v C10b c c t ð Þ ¼ c c0 , for t < r c v C10b [/formula]
## Appendix d time for fusion
We will choose an energy reduction in the fusion barrier due to a single SNAREpin, δe, of 6 k B T, which is similar to that previously predicted [bib_ref] SNARE machinery is optimized for ultrafast fusion, Manca [/bib_ref]. For E b ¼ 25 k B T, 5 SNAREpins or more will completely abolish the fusion barrier. The force applied by a single SNAREpin is of the order of 30 pN in this range of intermembrane distance (∼ 30 k B T energy gain over a 4 nm displacement [bib_ref] Energetics, kinetics, and pathway of SNARE folding and assembly revealed by optical..., Zhang [/bib_ref]. With these values, the travel time of the vesicle from a 5 nm separation to 2 nm before fusion, the decreased fusion barrier, the waiting time, and the fusion time can be calculated from Eqns [bib_ref] New phases of phospholipids and implications to the membrane fusion problem, Yang [/bib_ref] [bib_ref] A protein assembly-disassembly pathway in vitro that may correspond to sequential steps..., Sollner [/bib_ref]. The results are presented in . The fusion time is plotted in [fig_ref] Figure 6: Fusion time [/fig_ref].
## Appendix e energy landscape of the fusion pore expansion
Preliminary description of the model Let us consider a fusion pore between a vesicle of external radius r v and a flat membrane that opens as the two objects are at a distance h that we will assume constant during the pore expansion process. The thickness of the lipid bilayers is noted t. The lipid bilayer will be considered as a continuous and differentiable (i.e., 'smooth') curve, which is a simplifying but also bold assumption at this scale. We will focus on the case in which the pore has a circular tore-like geometry. The radius (taken up to the middle of the bilayer) of the small circle forming the tore by rotation is noted r c and increases as the fusion pore expands. The portion of the tore spans from an angle À π 2 to a maximal angle called θ m , which will decrease. We will call θ the variable describing the portion of the circle between À π 2 and θ m . We can also choose a cylindrical coordinate system to describe the shape of the pore with the radial coordinate denoted r that describes the radius of the pore at a certain height, taken up to between the monolayers. The radius of the pore taken up to its rim will be noted, r p ; it is the minimum value of rÀ t 2 .Given our parameters, we have that:
[formula] r ¼ r p þ t 2 þ r c 1Àcos θ ð Þ ð Þ (E1) [/formula]
r c can be obtained from the Pythagorean theorem (see:
[formula] r c ¼ 2r v Àr p þ hÀt=2À [/formula]
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffi 2 2r 2 v Àht þ hr v À2r v tÀhr p À2r v r p À Á q (E2) Finally, we can compute θ m through:
[formula] cos θ m ð Þ ¼ r p þ t 2 þ r c r v þ r c À t 2 (E3a) sin θ m ð Þ ¼ r v Àr c þ t 2 þ h r v þ r c À t 2 (E3b) [/formula]
Computation of the curvature energy We will assume here that curvature is the sole driving force. Hypothesizing that the membranes are spontaneously flat, the curvature energy of the system is E c ¼ κ=2∬ VþT c h i 2 dS, where K is the membrane bending modulus, <c> is the mean curvature at the considered point, V is the surface of the vesicle and T that of the partial circular torus that forms the pore.
Curvature energy of the fusing vesicle The curvature energy of the fusing vesicle is that of the full vesicle minus that of the spherical cap of surface S ¼ 2π r v À t 2 À Á 2 1Àsin θ m ð Þ ð Þ that disappeared because of fusion. The curvature for the vesicle being constant equal to 2/(rv-t/2) at any point of the vesicle, the total curvature energy of the vesicle is: . A vesicle initially located 5 nm from the target membrane is pulled by N SNAREpins, brought in contact and fusion occurs subsequently. The travel time, reduction in the fusion barrier, waiting time, and fusion time are indicated for 1 to 6 SNAREpins.
[formula] E v ¼ K 2 4 r v À t 2 À Á 2 4π r v À t 2 2 ÀS ! ¼ 4K π 1 þ sin θ m ð Þ ½ (E4) [/formula]
Curvature energy of the tore The curvature on the torus is given by:
[formula] c h i ¼ À cos θ ð Þ r þ 1 r c (E5) [/formula]
The elementary surface dS is given by: dS ¼ 2πr r c dθ, so that the torus curvature energy is:
[formula] E t ¼ πK r c Z θ m À π 2 rÀr c cos θ ð Þ ð Þ 2 r dθ (E6) [/formula]
By substituting r by its expression and noting a≔ r p þ t 2 þr c rc > 1, we get: It is worth noticing that the vesicle curvature energy is exactly compensated by one of the terms of the energy of the pore, which is probably not innocuous.
[formula] E t ¼ πK Z θ m À [/formula]
Energy landscape of the pore In addition to the curvature energy, an energy barrier for resealing needs to be added to the pore energy landscape. This barrier is due to the merging of the rim upon pore closure. Hence, it should resemble that of the fusion barrier. A typical energy landscape of a fusion pore based on E9 is presented in [fig_ref] Figure 8: Expansion barrier [/fig_ref]. An expansion barrier as high and much larger than the resealing barrier is clearly observed. This resistance to expansion arises from the high curvature energies involved in the process. The next step is to add SNAREpins and observe how they affect the pore expansion energy landscape.
Assuming each SNAREpin contributes with a force f, the energy landscape of the pore upon the action of N SNAREpins is reduced by:
[formula] E s ¼ ÀNfr p (E10) [/formula]
Using reasonable values for all parameters predicts that the pore will spontaneously expand only when there are three SNAREpins acting simultaneously [fig_ref] Figure 8: Expansion barrier [/fig_ref]. The local minima for one and two SNAREpins are located at a pore radius of 0.6 and 0.9 nm, respectively. Overall, our model seems consistent with what was put forward experimentally [bib_ref] Nascent fusion pore opening monitored at single-SNAREpin resolution, Heo [/bib_ref].
r v r p. The expression of r c is obtained through the triangle presented here. The sides of the triangle are, respectively, r v À r c Àt=2Àh ð Þ , r p þ r c þ t 2 , and r v þ r c À t 2 for the purple, green/orange, and green/blue sides.
However, those theoretical results should be taken very cautiously for the crude approximations made are to be challenged, including:
- Continuum approach for the bilayer whereas scales are that of lipids. - The distance between membrane and vesicle taken fixed, whereas it probably decreases over time. - The vesicle shape assumed to remain spherical whereas it likely deforms during the process. - The circular torus-like geometry does not represent all possible membrane shapes.
[fig] Figure 1: (A) SNAREpin molecular organization. [/fig]
[fig] Figure 3: Initiation of SNAREpin assembly. (A) [/fig]
[fig] Figure 4: Optical assays (in addition to the lipid mixing assay presented in Fig. 2). (A) In the content release bulk assay, dyes are quenched in the vSUV. Upon fusion, these dyes are diluted and their fluorescence increases. Monitoring this increase in time provides a direct quantification of the content released during fusion. (B) Single vesicles immobilized on a surface can be monitored by total internal reflection fluorescence microscopy (TIRF). The fusion of a vesicle with cognate SNAREs can be observed either by Forster resonance energy transfer (FRET) for lipid mixing, left, or by dequenching of encapsulated dyes for content release (right). (C) Fusion of a vesicle with a flat membrane, supported (left) or suspended (right), can be observed by TIRF. Dequenching of membrane-bound or encapsulated dyes provides a direct observation of single fusion event at the level of lipid mixing and content release, respectively. [/fig]
[fig] Figure 5: Electrical assays. The fusion of a vesicle with a suspended membrane can be observed by placing two electrodes on either side of the membrane. (A) Impedance measurements. Upon fusion, the membrane of the vesicle is incorporated in the suspended membrane, which increases the capacitance. Also, initially, the potential in the lumen of the vesicle is different from the potentials on either side of the suspended membrane. Hence, upon fusion, there is a transient current to equilibrate the potentials of the vesicle, V v , and the lower side of the membrane, V b (right panel). This current is due to ions that flow through the pore and therefore provides a direct measure of the pore kinetics. (B) Conductance measurements. Because the current in the impedance measurement is transient, the kinetics of the pore can only be measured over a short period of time (~1 ms). To obtain longer kinetics, a constant voltage can be applied between the two sides of the suspended membrane. Using vesicles with embedded channels, thus at the same potential as the top side (V t ), or nanodiscs, ions flow continuously when the fusion pore opens; the whole kinetics of pore opening and expansion can then be monitored.FEBS Open Bio 12 (2022) 1958-1979 © 2022 The Authors. FEBS Open Bio published by John Wiley & Sons Ltd on behalf of Federation of European Biochemical Societies. [/fig]
[fig] Figure 6: Fusion time. (A) When only one SNAREpin is involved (left), there remains an activation energy barrier for fusion. Hence, thermal fluctuations will provide the final stroke for fusion pore opening. The mean time for fusion in that case is~1 s. When 6 SNAREpins are acting simultaneously (right), the fusion barrier vanishes and fusion is spontaneous. The only remaining delay is the travel time of the vesicle to a 2 nm distance to the target membrane, typically a few ns. (B) Variation of the fusion time with the number of SNAREpins based on the parameters indicated in Appendix D. An additional membrane merger time may need to be added and become the dominant term for more than 3 SNAREpins. [/fig]
[fig] Figure 7: Role of the various parts of the C-terminal end of the SNAREpin for fusion pore opening and expansion. FEBS Open Bio 12 (2022) 1958-1979 © 2022 The Authors. FEBS Open Bio published by John Wiley & Sons Ltd on behalf of Federation of European Biochemical Societies. [/fig]
[fig] Figure 8: Expansion barrier. Opening the fusion pore is not sufficient to ensure full fusion. An expansion barrier due to curvature energies prevents the growth of the nascent fusion pore. [/fig]
[fig] 100: μs when 3 or more SNAREpins are acting together. The linkers transmit the zipping force to optimize the action of the SNARE domains and provide the final energy stroke to open the fusion pore. The C-terminal hydrophilic residues of VAMP2 reinforce this force transmission by facilitating the deformation of the vesicle membrane. The zipping of part of the linker domains and of the transmembrane domains might be in charge of pore expansion. Experimental results and models suggest that the expansion barrier disappears when 3 SNAREpins or more are simultaneously zipping. DM has worked on all models and the part related to the initiation of the SNAREpin. LB focused on the pore opening and the nature of the pore. PH was involved in the transient pore studies, pore expansion, and the figure design. FP worked on the model and supervised the research. All authors wrote the manuscript. [/fig]
[table] C8: 95% of the molecules (membrane-bound or encapsulated) are released after three characteristic times. TheFig. 9. Mean normalized fusion rates, p v and p t , defined in the text, are presented against the lipid to protein ratio. The curves display a plateau above approximately 500 lipids per SNARE for both v-and t-SNARE, suggesting that the SNARE exhibit no cooperativity and that their contributions are additive. [/table]
|
The Role of Anterior Nuclei of the Thalamus: A Subcortical Gate in Memory Processing: An Intracerebral Recording Study
ObjectiveTo study the involvement of the anterior nuclei of the thalamus (ANT) as compared to the involvement of the hippocampus in the processes of encoding and recognition during visual and verbal memory tasks.MethodsWe studied intracerebral recordings in patients with pharmacoresistent epilepsy who underwent deep brain stimulation (DBS) of the ANT with depth electrodes implanted bilaterally in the ANT and compared the results with epilepsy surgery candidates with depth electrodes implanted bilaterally in the hippocampus. We recorded the event-related potentials (ERPs) elicited by the visual and verbal memory encoding and recognition tasks.ResultsP300-like potentials were recorded in the hippocampus by visual and verbal memory encoding and recognition tasks and in the ANT by the visual encoding and visual and verbal recognition tasks. No significant ERPs were recorded during the verbal encoding task in the ANT. In the visual and verbal recognition tasks, the P300-like potentials in the ANT preceded the P300-like potentials in the hippocampus.ConclusionsThe ANT is a structure in the memory pathway that processes memory information before the hippocampus. We suggest that the ANT has a specific role in memory processes, especially memory recognition, and that memory disturbance should be considered in patients with ANT-DBS and in patients with ANT lesions.ANT is well positioned to serve as a subcortical gate for memory processing in cortical structures.
# Introduction
Following a series of experimental and clinical studies including the large randomized controlled trial SANTE [bib_ref] Electrical stimulation of the anterior nucleus of thalamus for treatment of refractory..., Fisher [/bib_ref] , ANT-DBS was introduced into the therapy of refractory epilepsy. The anterior nuclei of the thalamus (ANT) are the most well established target for DBS in the treatment of epilepsy. In epilepsy pathology the ANT is thought to function as a relay structure to amplify and synchronize epileptic activities in the circuit [bib_ref] Chronic anterior thalamus stimulation for intractable epilepsy, Hodaie [/bib_ref] , [bib_ref] Electrical stimulation of the anterior nucleus of the thalamus for intractable epilepsy:..., Lim [/bib_ref]. The thalamus by its relay function of the many thalamic nuclei does appear to be a station transmitting neural signals primary to the cerebral cortex from a number of cortical and subcortical brain areas.
Interest in the cognitive functions of the ANT in human increased after the SANTE study, in which memory impairment and depression appeared as the most frequent side effects of ANT-DBS. Cognition and mood showed no group differences in objective testing, but participants in the stimulated group were more likely to report memory problems as adverse events than participants on placebo stimulation . High frequency stimulation of the ANT also disrupted the performance of memory paradigms in rats [bib_ref] Anterior thalamus deep brain stimulation at high current impairs memory in rats, Hamani [/bib_ref]. By nature, cognitive behavior such as learning and recall is highly dynamic, and therefore dynamic neuromodulatory devices are likely to alter its function. DBS (of the basal ganglia, temporal lobes, thalamus and limbic system) may also play a potentional role in positively affecting cognitive behavior such as memory formation and recall [bib_ref] Role of deep brain stimulation in modulating memory formation and recall, Hu [/bib_ref].
The impact of ANT-DBS on memory might be explained by the anatomical position of the ANT. The ANT belongs to the associative nuclei of the thalamus getting information from the limbic system (via the fornix and gyrus cinguli). Functionally the thalamus is connected to the limbic system by the hippocampal Papez circuit. The connection is from the hippocampus via the indusium griseum, subiculum and presubiculum to the area entorhinalis and gyrus hippocampi. Through the fornix the Papez circuit continues to the corpus mamillare and as a fasciculus mamillothalamicus to the ANT. Further, the ANT is connected with the posterior cingulate cortex and orbitofrontal cortex and via the cortex of the cingulum (area cingularis posterior) with the hippocampus directly or via the area entorhinalis. Additional circuits connect the ANT to the hypothalamus, nucleus accumbens, habenular nuclei, and the septal nuclei. The connection with the nucleus accumbens links the Papez circuit with the limbic pathway of the basal ganglia. As the ANT is linked with the hippocampus, cingulate limbic structures, and the neocortex and is a key structure in the intrathalamic pathways, it is well positioned to serve the memory circuit as a relay nucleus. The extensive direct and indirect hippocampal connections support the hypothesis that hippocampus and ANT constitute a neuronal network crucial for memory. The important role of the hippocampus in the memory processes is well known [bib_ref] Activity of human hippocampal formation and amygdala neurons during memory testing, Halgren [/bib_ref] [bib_ref] Memory and the hippocampus: a synthesis from findings with rats, monkeys, and..., Squire [/bib_ref] [bib_ref] The hippocampus as a cognitive graph, Muller [/bib_ref] [bib_ref] Changes in hippocampal connectivity in the early stages of Alzhei-mer´s disease: evidence..., Wang [/bib_ref] [bib_ref] Coherent spontaneous activity identifies a hippocampal-parietal memory network, Vincent [/bib_ref] [bib_ref] Intrinsic connectivity between the hippocampus and posteromedial cortex predicts memory performance in..., Wang [/bib_ref] [bib_ref] The influence of age and mild cognitive impairment on associative memory performance..., Oedekoven [/bib_ref] [bib_ref] The role of the hippocampus in flexible cognition and social behavior, Rubin [/bib_ref]. While the role of the hippocampus in the cognitive network is widely studied, less attention has been devoted to the role of the ANT in memory processing. From a clinical point of view, both the hippocampal-anterior thalamic and the perirhinalmedial dorsal thalamic systems are compromised in amnesic cases, leading to severe deficits in both recall and recognition [bib_ref] Episodic memory, amnesia, and the hippocampal-anterior thalamic axis, Aggleton [/bib_ref]. Amnestic syndromes were observed in the pathology of the mammillary bodies, the mammillothalamic tract and the ANT [bib_ref] Pure amnesia after unilateral left polar thalamic infarct: topographic and sequential neuropsychological..., Clarke [/bib_ref] [bib_ref] Degeneration of anterior thalamic nuclei differentiates alcoholics with amnesia, Harding [/bib_ref] [bib_ref] Neuropsychology of infarctions in the thalamus: a review, Van Der Werf [/bib_ref]. The ANT research has focused on animal studies while it is difficult to get the data from human ANT due to the small size and position of this structure. Based on these studies it has been suggested that the ANT serve as a subcortical gate for information used in path integration processes by cortical structures. We used the unique opportunity to record directly from the human ANT through the DBS electrodes as well as to explore the hippocampus by stereotactically placed depth electrodes (SEEG) to explore these two structures in memory processing. We studied and compared local field potentials elicited by memory encoding and recognition tasks recorded in the ANT and in the hippocampus in patients with epilepsy.
We raised three questions:
1. Could local field P300-like potentials be elicited by memory tasks recorded in the ANT?
2. If so, are the ERPs in the ANT identical to or different from the ERPs recorded in the hippocampus?
3. Can the latencies of the P300-like potentials recorded in the hippocampus and in the ANT confirm that the ANT participates in preprocessing memory information?
## Materials and methods patients
Six pharmacoresistent epilepsy patients, all native Czech speakers, were included in our study. Three patients (Patients 1-3, [fig_ref] Table 1: The main characteristics of patients with ANT and hippocampal electrodes [/fig_ref] were implanted bilaterally with ANT-DBS (only), the other three epilepsy surgery candidates had depth electrodes (SEEG) implanted bilaterally in the hippocampi (Patients 4-6, [fig_ref] Table 1: The main characteristics of patients with ANT and hippocampal electrodes [/fig_ref]. No patient had electrodes implanted in the hippocampus and ANT simultaneously, as there was no medical reason for that. Therefore the direct comparision could not be made for the same individuals but only interindividually. All patients underwent complex presurgical evaluation including advanced diagnostic methods (MRI, PET: Positron Emission Tomography, SPECT: Single-Photon Emission Computed Tomography, and SISCOM: subtraction of ictal and interictal SPECT coregistered to MRI) and complex neuropsychological testing (Wechsler Adult Intelligence Scale-III, event. Wechsler Adult Intelligence Scale Revised (short form with subtests: picture completion, arithmetic, similarities, digit symbol-coding, digit span), Rey-Osterrieth Complex , Wechsler Memory Scale-III (Word List subtest), Stroop Test, Verbal Fluency Test, Zung Depression Scale, and Hamilton Anxiety Scale). Neuropsychological examinations conducted prior to the implantation showed slight impairments in various domains with no signs of dementia or psychiatric or behavioral disorders.
We compared the ANT-DBS group with the group of pharmacoresistent epilepsy patients with deep brain electrodes (SEEG) implanted bilaterally in the hippocampi. These patients underwent invasive video EEG exploration before the epilepsy surgery which followed. The ANT-DBS was recommended by the Epilepsy Surgery Commission of the Brno Epilepsy Center and it was considered as the last available option for patients in whom all other therapies failed. None of the patients had responded to long-term vagus nerve stimulation (VNS). The VNS devices were explanted before the DBS implantation. The study and the recordings were approved by the local ethics committee (St. Anne´s Ethics Committee) and the patients gave their written informed consent. The patients were simultaneously treated with the current dosage of their antiepileptic medication [fig_ref] Table 1: The main characteristics of patients with ANT and hippocampal electrodes [/fig_ref].
Surgical methods and procedure DBS procedure. All three patients with ANT-DBS underwent implantation using the ceramic stereotactic Leibinger open frame with the Praezis Plus software and the Talairach and Schaltenbrand Bailey atlases.
The initial coordinates for the ANT as related to the anterior commissure-posterior commissure (AC-PC) line centre were 0-2mm anterior to the midpoint, 5.5mm laterally, and 10-12mm above the AC-PC line. The entry point for the electrode was planned at the proximity of the coronal suture. The final target was modified according to local anatomy, and particular attention was paid to the safe distance of the trajectory from the thalamostriate vein and choroid plexus. All four electrode contacts were planned to be inside thalamic structures. The stimulation leads (Medtronic, Inc.) were implanted bilaterally into the targeted structure using a stereotactic magnetic resonance imaging (MRI)-guided technique and local anesthesia. Intraoperative microrecordings to guide lead placement were employed. Intraoperative stimulation was used to test for possible adverse effects. Once the final target coordinates were defined, a permanent quadripolar DBS electrode (model 3389, with 1.5mm contact length and 0.5mm intercontact distance) was implanted. The electrode position was verified by the intraoperative skiascopy control using the C-arm with postoperative confirmation of electrode position (lateral and anteroposterior radiographs under stereotactic condition and postoperative CT with stereotactic frame and markers). After surgery completion, CT scans under stereotactic conditions covering the entire length of the implanted electrodes were added. The series of images were reimported to the planning workstation and subsequently the coordinates were correlated with the actual positions of the implanted electrodes [fig_ref] Fig 1: Example of the actual positions of the all ANT contacts in patient... [/fig_ref]. Any potential deviation of the final electrode position is readily observed after this correlation without being burdened by the material artifacts. The positions of the electrodes and their contacts in the brain were also later verified using post-placement MRI with electrodes in situ. DBS electrodes were first connected to extracranial extensions (the connection being subcutaneous) enabling recordings from the DBS contacts. After 3-4 days of video-EEG monitoring, the extracranial extensions were removed and the DBS electrodes were connected to the implanted battery.
SEEG procedure. The patients with hippocampal target underwent an exploration of the epileptic focus (seizure onset zone) via stereotactically implanted depth electrodes. The patients were implanted with 6, 9 and 14 orthogonal electrodes respectively using the Talairach and Bancaud methodologyto explore all possible seizure onset zones. Five-to fifteen-contact platinum semiflexible Alcis (Besançon, France) electrodes, each with a diameter of 0.8mm and a contact length of 2mm, and with intercontact intervals of 1.5mm, were used. The exact positions of the electrodes were verified using post-placement MRI with electrodes in situ. The electrodes were implanted bilaterally in the hippocampi (as well as in other target structures that were not included in our study). The study was approved by the local ethics committee. All three patients were fully informed about the character of the study and gave their informed consent. The patient characteristics are shown in the [fig_ref] Table 1: The main characteristics of patients with ANT and hippocampal electrodes [/fig_ref]. The patients later underwent focal cortectomy of the seizure onset zone based on the SEEG findings combined with all noninvasive methods (MRI, PET: Positron Emission Tomography, SPECT: Single-Photon Emission Computed Tomography, SISCOM: subtraction of ictal and interictal SPECT coregistered to MRI), neuropsychological testing, and clinical findings from standard neurological examinations, used in the preoperative phase in order to localize the seizure onset zone. In MRI, as our standard protocol in epilepsy patients, we used T1 and T2 weighted imaging, T2 FLAIR (fluid attenuation inversion recovery) and inverse T2, T1 and T2 coronal and coronal FLAIR scans, DWI (diffusion weighted imaging), T2 FRFSE (with fat suppression), T2 GRE (gradient recalled echo), T1 FSPGR (ultrafast spoiled gradient echo), T2 and PD (proton density) FSE (fast spin echo) on 1.5 Tesla MR.
## Recordings
The episodic memory testing was conducted during the encoding and recognition phases in visual and verbal modalities. Patients were tested in two consecutive days. Visual testing was performed on the first day and verbal (auditory) testing on the second day. The elaborations of the tests for this study were based on protocols used in an intracranial recording study by Jones-Gotman et al. [bib_ref] EEG slow waves and memory performance during the intracarotid amobarbital test, Jones-Gotman [/bib_ref] and on protocols used in an fMRI study by Rabin et al. [bib_ref] Functional MRI predicts post-surgical memory following temporal lobectomy, Rabin [/bib_ref]. The first results obtained in our laboratory were published by Štillová et al. [bib_ref] The posterior medial cortex is involved in visual but not in verbal..., Štillová [/bib_ref].
For the visual task, the patients were seated in front of a computer monitor at a 1.5m distance from the screen. In the encoding phase, 30 emotionally neutral photos were presented for 2 seconds each. The photos showed static scenes, (nature, towns, etc.). During the interstimulus interval, a black screen appeared for 4 seconds. The patients were instructed to remember the presented pictures.
After the encoding phase, a 15-minute break followed, during which the patients watched a well-known fairy tale. They were asked not to remember the story, only to relax. The recognition phase followed: 60 photos were presented, with 30 pictures that had been seen in the encoding phase and 30 new pictures. The pictures appeared on the monitor in random order. The patients held a two-button device in their right hand. They were asked to push the right button when a picture was recognized as having been presented before or the left button when the picture was considered to be a new one.
The verbal memory task was tested the next day. The patients were seated in front of a computer at a 1.5m distance and listened to words through the speakers. In the encoding phase, 30 words were presented with a silent interstimulus interval of 4 seconds. Both abstract and concrete words were intermingled. The words were two or three syllables long, emotionally neutral, and commonly used in the Czech language. The patients were instructed to remember the presented words.
A 15-minute break followed, during which the patients watched a well known fairy tale. The recognition phase followed: 60 words were presented, with 30 words that had been heard in the encoding phase and 30 new words. The old and new words were presented in random order. Using the same device as for the visual task, the patients were asked to push the right button when a word was recognized as having been presented before or the left button when the word was considered to be a new one [bib_ref] The posterior medial cortex is involved in visual but not in verbal..., Štillová [/bib_ref].
The recordings from hippocampal electrodes were performed with the TruScan system (Deymed Diagnostic, Alien Technic) 128-channel EEG machine. The sampling frequency was 1024 Hz with standard anti-aliasing filters.
The recordings from ANT electrodes were performed with the M&I EEG system. The scalp EEG was recorded simultaneously. The recordings were made in a monopolar montage with connected earlobes used as a reference. The signal was filtered in the range from 0.2 to 200 Hz in the time base 2 seconds before the stimulus and 5 seconds after.
## Data evaluation and analysis
The data analysis, segmentation, and evaluation were made by using the ScopeWin and MATLAB software systems. The data were segmented according to the stimulus onset (time 0 in figures), and all the trials were visually controlled. Trials with epileptiform activity or other technical problems were excluded. The trial length was 7s: 2s before and 5s after the onset of the stimuli. The baseline interval was determined 600-100 ms before the stimulus occurred. The mean values from the baseline intervals were subtracted within each trial. After trend elimination in each trial, data were filtered with a 0.2-40 Hz band pass and artifact-free trials were averaged. The statistical significance of the differences between the mean during the baseline interval and the mean computed from the 150ms length sliding window after stimuli is expressed by the probability p value. The same sliding window was used when inter-task differences were computed. We used the non parametric Wilcoxon Rank Sum (Signed Rank) test for paired samples in each trial. The amplitude changes after stimulus were considered significant when the probability value p was lower than 0.05. The statistical significance to baseline is highlighted by a horizontal bar (black and red) in the figures [fig_ref] Fig 2: P300-like potentials in the hippocampus and the thalamus [/fig_ref]. The differences between stimuli were analysed with an unpaired t-test in a time interval 0-1.5s. The level of significance was designated at a p value less than 0.01 two-tailed test was used. The significance between stimuli is shown in figures as delimited by horizontal bars. The local field potential was verified by a bipolar montage for each electrode in every task. The bipolar montage evaluation was used to exclude the volume conduction from other structures, namely from the cortex or transsynaptic propagation along cortical-subcortical pathways [bib_ref] Intracranial volume conduction of cortical spikes and sleep potentials recorded with deep..., Wennberg [/bib_ref] [bib_ref] Restating the importance of bipolar recording in subcortical nuclei, Wennberg [/bib_ref] and confirm the local origin of the potentials. Contacts in the thalamus were placed very close together. Any EEG signal from the common reference was eliminated by a bipolar montage. Even minor bipolar montage activity displays the origins of detected activity in the ANT. The main ERP components in the 200-450 ms following stimuli were identified by visual inspection and quantified by latency and amplitude measures (P300-like potential). The peak latencies were measured from stimulus onset (point 0). The intracerebral potentials occurred with both positive and negative polarities. This was due to variances in the positions of the electrode contact and of the dipole generator. Absolute amplitudes were measured from the baseline. The distance from the electrode to the generator heavily influences the amplitude of intracerebral recorded potentials, and thus the differences of amplitude can only be compared intraindividually.
The bootstrap methodology [bib_ref] Jackknife, bootstrap and other resampling methods in regression analysis, Wu [/bib_ref] with 500 repetitions was used to assess the mean latency and corresponding distribution of P300-like potentials for different tasks. The statistical significance of differences among latencies was tested by the Wilcoxon rank sum test and the Bonferroni correction was used [fig_ref] Table 2: P300-like potentials [/fig_ref].
# Results
We analyzed recordings from the contacts placed in the thalamus, focusing on the anterior nuclei. We compared the recorded data from the thalamus with the data from hippocampal electrodes. The accuracy rate of correct recognition by the patients (characteristics of the patients shown in [fig_ref] Table 1: The main characteristics of patients with ANT and hippocampal electrodes [/fig_ref] , patient 1: 84%, patient 2: 79%, patient 3: 83%, patient 4: 84%, patient 5: 90%, and patient 6: 84.5% of the correct answers.
## Erp analyses
ERPs elicited by the visual encoding task and the visual and verbal recognition tasks were recorded in the ANT. In the visual encoding and recognition tasks an N 115-P 150-N 200-P 250 and N700 complex was detected [fig_ref] Fig 3: Visual recognition task in the ANT [/fig_ref]. The first two waves (N115-and P150) probably represent a visual evoked potential (VEP): such a potential was not recorded in the ERPs detected during the verbal tasks in the ANT. The verbal encoding task elicited only a small late potential peaking at about 1sec; no significant ERP was detected in the period typical for cognitive response.
In the hippocampus the ERPs were elicited by all tasks. An N200-P300-P750 like complex was detected in the hippocampus in both the visual and verbal recognition stimuli tasks. The latencies in the verbal stimuli tasks were longer than in the visual tasks. The early visual potentials with latency about 150 ms that were observed in the ANT were not recorded in the hippocampus. In both recognition tasks (visual and verbal) the latency of the P300-like potentials in the ANT was shorter than in the hippocampus, using the bootstrap for statistical significance [fig_ref] Fig 2: P300-like potentials in the hippocampus and the thalamus [/fig_ref]. The latency of P300-like potential detected in the ANT was significantly shorter (P<0.001) than the latency of P300-like potential in the hippocampus in all cases except visual encoding, where the latency was shorter, but not with statistical significance. Comparing the difference in P300-like potential latencies between the old (repeated) stimulus and the new stimulus during the recognition phase, the ERP latency in the ANT corresponding to the new stimulus was longer (P<0.05) than the latency corresponding to the old stimulus. There were no differences in the hippocampus between old and new stimuli.
Although the seizure onset zones in all patients were localized in different areas, the patients had different structural MR findings and were on a different antiepileptic medications, we believe that the observed ERPs were not fundamentally affected by any pathological processes or by the type of the medication taken at the time of our study.
# Discussion
In recent years, there has been extensive scientific concentration on the cortical structures as primary structures for cognition and behavior, and mainly as the primary structures responsible for regulation. Recent studies have made it increasingly obvious that it is also important to focus on subcortical brain structures including the thalamus. Our results confirm the role of the hippocampus in memory encoding and recognition processes and reveal a selective participation of the ANT in these tasks. The activation of the ANT appears to be mode and task dependent. The ANT participates in the pre-processing of memory tasks, preceding the hippocampal activity. Traditionally, the cognitive networks have been considered to be the cortical network. Far less attention has been paid to the cognitive role of the subcortical structures. Parvizi [bib_ref] Corticocentric myopia: Old bias in new cognitive sciences, Parvizi [/bib_ref] called this situation "corticocentric myopia". The ANT is not only connected with cognition, but is also involved in a complex visuomotor task, as was reported recently in humans [bib_ref] Executive functions are processed in the area of anterior nucleus of thalamus, Bočková [/bib_ref].
The importance of the anterior thalamus in human memory is becoming indubitable [bib_ref] Deficits of memory, executive functioning and attention following infarction in the thalamus;..., Van Der Werf [/bib_ref] [bib_ref] Hippocampal-anterior thalamic pathways for memory: uncovering a network of direct and indirect..., Aggleton [/bib_ref] [bib_ref] Cognitive improvement after long-term electrical stimulation of bilateral anterior thalamic nucleus in..., Oh [/bib_ref]. Damage to the anterior and medial parts of the thalamus, including the ANT, mediodorsal, midline and intralaminar nuclei can contribute to amnesia, although the nature of the memory deficit may vary. This was verified in animal models [bib_ref] Anterior thalamus deep brain stimulation at high current impairs memory in rats, Hamani [/bib_ref] , [bib_ref] Episodic memory, amnesia, and the hippocampal-anterior thalamic axis, Aggleton [/bib_ref] , [bib_ref] Hippocampal-anterior thalamic pathways for memory: uncovering a network of direct and indirect..., Aggleton [/bib_ref] , [bib_ref] The effect of anterior thalamic and cingulate cortex lesions on object-inplace memory..., Parker [/bib_ref] [bib_ref] Disconnecting hippocampal projections to the anterior thalamus produces deficits on tests of..., Warburton [/bib_ref] [bib_ref] Distinct, parallel pathways link the medial mammillary bodies to the anterior thalamus..., Vann [/bib_ref] and in clinical practice in a published study of a group of patients with amnestic syndrome caused by thalamic infarction [bib_ref] Neuropsychology of infarctions in the thalamus: a review, Van Der Werf [/bib_ref] , [bib_ref] What does a comparison of the alcoholic Korsakoff syndrome and thalamic infarction..., Kopelman [/bib_ref].
The ANT was suggested as a critical nodal point in an extended hippocampal system in spatial and non-spatial memory [bib_ref] Episodic memory, amnesia, and the hippocampal-anterior thalamic axis, Aggleton [/bib_ref] ,, [bib_ref] The mammillary bodies: two memory systems in one?, Vann [/bib_ref] , [bib_ref] Beyond spatial memory: the anterior thalamus and memory for the temporal order..., Wolff [/bib_ref]. Several studies focused on the hippocampal-ANT cooperation and on the direct and indirect influence of the ANT on the hippocampus. The involvement of the ANT and the hippocampal-anterior thalamic interconnections in human episodic memory and rodent event memory is known. The hippocampal-ANT axis is considered to be important for memory recall [bib_ref] Hippocampal-anterior thalamic pathways for memory: uncovering a network of direct and indirect..., Aggleton [/bib_ref]. The role of the ANT in episodic memory in terms of learning of visual discrimination (object in place) was found in monkeys by Parker and Gaffan [bib_ref] The effect of anterior thalamic and cingulate cortex lesions on object-inplace memory..., Parker [/bib_ref] and the selective activation of the ANT during the retrieval phase of memory recognition was observed in an fMRI study in humans [bib_ref] The role of the thalamic nuclei in recognition memory accompanied by recall..., Pergola [/bib_ref]. Our results suggest a specific role of the ANT in memory recognition processes and as a recognition "pre-processor" before the information gets further to the hippocampus.
Our previous SEEG recordings with identical tasks displayed similar results, i.e., activity linked with recognition in both modalities and absence of activation with the verbal encoding in the posterior medial cortex, namely the posterior cingulate cortex (PCC) and precuneus [bib_ref] The posterior medial cortex is involved in visual but not in verbal..., Štillová [/bib_ref]. In animal studies the ANT contributes to modulating plasticity in the PCC [bib_ref] Anterior thalamic lesions stop synaptic plasticity in retrosplenial cortex slices: expanding the..., Garden [/bib_ref]. PCC activity decreases following thalamic lesion in humans [bib_ref] FDG-PET findings in the Wernicke-Korsakoff syndrome, Reed [/bib_ref] , [bib_ref] Metabolic impairment in human amnesia: a PET study of memory networks, Fazio [/bib_ref]. The PCC was also more active during retrieval in a study by Pergoli [bib_ref] The role of the thalamic nuclei in recognition memory accompanied by recall..., Pergola [/bib_ref] which supports the interpretation that an ANT-PCC network is critical for the retrieval phase of recognition accompanied by recall. The posterior cingulate gyrus has a bidirectional connectivity via the cingulum with the hippocampus [bib_ref] In vivo human hippocampal cingulate connectivity: A corticocortical evoked potentials (CCEPs) study, Kubota [/bib_ref] , and the ANT is connected via the PCC with the hippocampus. Simple visual, but not verbal evoked potentials were recorded in the ANT. No such potentials were recorded in the hippocampus. It seems that the ANT may participate in the visual system. Further studies are needed in order to understand the nature of this involvement.
# Conclusion
The ANT is a subcortical part of the cortico-subcortical memory network, possibly with particular functional relation to the posterior medial cortex. Our results confirm the importance of the anterior thalamus in cognitive functions, its specific role in memory recognition, and a selective, modality-dependent role in memory encoding. There is a strong implication that the functions of the ANT in memory processing are not only driven by the hippocampus but that actions in the opposite direction may by equally crucial. ANT is well positioned to serve as a subcortical gate for further memory processing in cortical structures.
[fig] Fig 1: Example of the actual positions of the all ANT contacts in patient 2 (right-left orientation of the sagittal scans, fusion of the MR scans and CT correlation). Upper picture: position of the right ANT contacts: the R3 contact (left picture), and position of the R4 contact (right picture). Lower picture: position of the left ANT contacts: the L3 contact (left picture) and position of the L4 contact (right picture). doi:10.1371/journal.pone.0140778.g001 [/fig]
[fig] Fig 2: P300-like potentials in the hippocampus and the thalamus (the mean from all hippocampal and thalamic contacts from all patients in bipolar montage during the encoding and the recognition task). The black curve indicates a new stimulus (new picture/word: NP/NW); red indicates a repeated stimulus during the recognition phase (old picture/word: OP, OW). The arrow shows the beginning of the P300-like potential (red for the repeated stimulus, black for the new one.) Axis x is time (in s); axis y is amplitude in μV. The statistical significance to baseline is highlighted by black and red horizontal lines. The significant difference between tasks in the hippocampus is highlighted by a black rectangle. Simultaneously the difference between tasks is drawn-DIFF.doi:10.1371/journal.pone.0140778.g002 [/fig]
[fig] Fig 3: Visual recognition task in the ANT: N 115-P 150-N 200-P 250 and N700 complex (the mean from all thalamic contacts from all patients in bipolar montage during the recognition task). Arrows mark fast and slow ERP components N115, P150, N200, P250, and N700. The black curve indicates a new stimulus (new picture); red indicates a repeated stimulus during the recognition phase (old picture). On axis x is time (in s). The statistical significance to baseline is highlighted by black and red horizontal lines.doi:10.1371/journal.pone.0140778.g003 [/fig]
[table] Table 1: The main characteristics of patients with ANT and hippocampal electrodes. [/table]
[table] Table 2: P300-like potentials: latency of P300-like potentials recorded from all contacts from the hippocampus versus all contacts from the ANT, mean and standard deviation. [/table]
|
Association of Use of Omega-3 Polyunsaturated Fatty Acids With Changes in Severity of Anxiety Symptoms
et al. Association of use of omega-3 polyunsaturated fatty acids with changes in severity of anxiety symptoms: a systematic review and meta-analysis. JAMA Netw Open.
[fig] 21: Hakimian J, Minasyan A, Zhe-Ying L, et al. Specific behavioral and cellular adaptations induced by chronic morphine are reduced by dietary omega-3 polyunsaturated fatty acids. PloS one. 2017;12(4):e0175090. 22. Hryhorczuk C, Decarie-Spain L, Sharma S, et al. Saturated high-fat feeding independent of obesity alters hypothalamus-pituitary-adrenal axis function but not anxiety-like behaviour. Psychoneuroendocrinology. 2017;83:142-149. 23. Robertson RC, Seira Oriach C, Murphy K, et al. Omega-3 polyunsaturated fatty acids critically regulate behaviour and gut microbiota development in adolescence and adulthood. Brain, behavior, and immunity. 2017;59:21-37. 24. Bandaru SS, Lin K, Roming SL, Vellipuram R, Harney JP. Effects of PI3K inhibition and low docosahexaenoic acid on cognition and behavior. Physiol Behav. 2010;100(3):239-244. Not human clinical trial in the part of anxiety study (n=1) 25. Hamazaki T, Sawazaki S, Nagasawa T, Nagao Y, Kanagawa Y, Yazawa K. Administration of docosahexaenoic acid influences behavior and plasma catecholamine levels at times of psychological stress. Lipids. 1999;34 Suppl:S33-37. Not related to omega 3 supplement effect on anxiety (n=38) 26. Liu JJ, Galfalvy HC, Cooper TB, et al. Omega-3 polyunsaturated fatty acid (PUFA) status in major depressive disorder with comorbid anxiety disorders. The Journal of clinical psychiatry. 2013;74(7):732-738. 27. Jadoon A, Chiu CC, McDermott L, et al. Associations of polyunsaturated fatty acids with residual depression or anxiety in older people with major depression. Journal of affective disorders. 2012;136(3):918-925. 28. Lesperance F, Frasure-Smith N, St-Andre E, Turecki G, Lesperance P, Wisniewski SR. The efficacy of omega-3 supplementation for major depression: a randomized controlled trial. The Journal of clinical psychiatry. 2011;72(8):1054-1062. 29. Sanchez-Villegas A, Henriquez P, Figueiras A, Ortuno F, Lahortiga F, Martinez-Gonzalez MA. Long chain omega-3 fatty acids intake, fish consumption and mental disorders in the SUN cohort study. European journal of nutrition. omega-3 fatty acids are decreased in nondepressed patients with social anxiety disorder. European neuropsychopharmacology : the journal of the European College of Neuropsychopharmacology. 2006;16(2):107-113. 31. Matsuoka Y, Nishi D, Tanima Y, et al. Serum pro-BDNF/BDNF as a treatment biomarker for response to docosahexaenoic acid in traumatized people vulnerable to developing psychological distress: a randomized controlled trial. Transl Psychiatry. 2015;5:e596. 32. Matsumura K, Noguchi H, Nishi D, Hamazaki K, Hamazaki T, Matsuoka YJ. Effects of omega-3 polyunsaturated fatty acids on psychophysiological symptoms of post-traumatic stress disorder in accident survivors: A randomized, double-blind, placebo-controlled trial. Journal of affective disorders. 2016. 33. Emmerich T, Abdullah L, Crynen G, et al. Plasma Lipidomic Profiling in a Military Population of Mild Traumatic Brain Injury and Post-Traumatic Stress Disorder with Apolipoprotein E varepsilon4-Dependent Effect. J Neurotrauma. 2016;33(14):1331-1348. 34. Verly-Miguel MV, Farias DR, Pinto Tde J, Lepsch J, Nardi AE, Kac G. Serum docosahexaenoic acid (DHA) is inversely associated with anxiety disorders in early pregnancy. Journal of anxiety disorders. 2015;30:34-40. 35. Saunders EF, Reider A, Singh G, Gelenberg AJ, Rapoport SI. Low unesterified:esterified eicosapentaenoic acid (EPA) plasma concentration ratio is associated with bipolar disorder episodes, and omega-3 plasma concentrations are altered by treatment. Bipolar disorders. 2015;17(7):729-742. 36. Hamazaki K, Takamori A, Tsuchida A, et al. Dietary intake of fish and n-3 polyunsaturated fatty acids and risks of perinatal depression: The Japan Environment and Children's Study (JECS). Journal of psychiatric research. 2018;98:9-16. 37. Thesing CS, Bot M, Milaneschi Y, Giltay EJ, Penninx B. Omega-3 and omega-6 fatty acid levels in depressive and anxiety disorders. Psychoneuroendocrinology. 2018;87:53-62. 38. Bowler RM, Adams SW, Schwarzer R, et al. Validity of self-reported concentration and memory problems: Relationship with neuropsychological assessment and depression. J Clin Exp Neuropsychol. 2017;39(10):1026-1036. 39. Parellada M, Llorente C, Calvo R, et al. Randomized trial of omega-3 for autism spectrum disorders: Effect on cell membrane composition and behavior. European neuropsychopharmacology : the journal of the European College of Neuropsychopharmacology. 2017;27(12):1319-1330. 40. Pawelczyk T, Piatkowska-Janko E, Bogorodzki P, et al. Omega-3 fatty acid supplementation may prevent loss of gray matter thickness in the left parietooccipital cortex in first episode schizophrenia: A secondary outcome analysis of the OFFER randomized controlled study. Schizophr Res. 2017. 41. Song L, Yoshida S, Tanaka-Mizuno S, Ogawa Y, Furukawa TA, Kawakami K. Association between eicosapentaenoic acid (EPA) medication intake and new onset of depression among Japanese patients with hyperlipidemia: A 3-year follow-up study. J Nutr Intermed Metab. 2017;9:12-16. 42. Wilson PB, Madrigal LA. Associations among Omega-3 Fatty Acid Status, Anxiety, and Mental Toughness in Female Collegiate Athletes. J Am Coll Nutr. 2017;36(8):602-607. 43. Lotrich FE, Sears B, McNamara RK. Polyunsaturated fatty acids moderate the effect of poor sleep on depression risk. Prostaglandins Leukot Essent Fatty Acids. 2016;106:19-25. 44. Chalut-Carpentier A, Pataky Z, Golay A, Bobbioni-Harsch E. Involvement of dietary Fatty acids in multiple biological and psychological functions, in morbidly obese subjects. Obesity surgery. 2015;25(6):1031-1038. 45. Hansen AL, Olson G, Dahl L, et al. Reduced anxiety in forensic inpatients after a long-term intervention with Atlantic salmon. Nutrients. 2014;6(12):5405-5418. 46. Su KP, Lai HC, Yang HT, et al. Omega-3 fatty acids in the prevention of interferon-alpha-induced depression: results from a randomized, controlled trial. Biological psychiatry. 2014;76(7):559-566. 47. Barbadoro P, Annino I, Ponzio E, et al. Fish oil supplementation reduces cortisol basal levels and perceived stress: a randomized, placebo-controlled trial in abstinent alcoholics. Mol Nutr Food Res. 2013;57(6):1110-1114. 48. Parletta N, Cooper P, Gent DN, Petkov J, O'Dea K. Effects of fish oil supplementation on learning and behaviour of children from Australian Indigenous remote community schools: a randomised controlled trial. Prostaglandins Leukot Essent Fatty Acids. 2013;89(2-3):71-79. 49. Souied EH, Delcourt C, Querques G, et al. Oral docosahexaenoic acid in the prevention of exudative age-related macular degeneration: the Nutritional AMD Treatment 2 study. Ophthalmology. 2013;120(8):1619-1631. 50. Young L, Kemper KJ. Integrative care for pediatric patients with pain. Journal of alternative and complementary medicine. 2013;19(7):627-632. 51. Beezhold BL, Johnston CS. Restriction of meat, fish, and poultry in omnivores improves mood: a pilot randomized controlled trial. Nutr J. 2012;11:9. 52. Lok A, Assies J, Koeter MW, et al. Sustained medically unexplained physical symptoms in euthymic patients with recurrent depression: predictive value for recurrence and associations with omega 3-and 6 fatty acids and 5-HTTLPR? Journal of affective disorders. 2012;136(3):604-611. 53. Manor I, Magen A, Keidar D, et al. The effect of phosphatidylserine containing Omega3 fatty-acids on attention-deficit hyperactivity disorder symptoms in children: a double-blind placebo-controlled trial, followed by an open-label extension. European psychiatry : the journal of the Association of European Psychiatrists. 2012;27(5):335-342. 54. Beezhold BL, Johnston CS, Daigle DR. Vegetarian diets are associated with healthy mood states: a cross-sectional study in seventh day adventist adults. Nutr J. 2010;9:26. 55. Nguemeni C, Delplanque B, Rovere C, et al. Dietary supplementation of alphalinolenic acid in an enriched rapeseed oil diet protects from stroke. Pharmacological research. 2010;61(3):226-233. 56. Ramsden C, Gagnon C, Graciosa J, et al. Do omega-6 and trans fatty acids play a role in complex regional pain syndrome? A pilot study. Pain medicine. 2010;11(7):1115-1125. 57. Hibbeln JR, Davis JM. Considerations regarding neuropsychiatric nutritional requirements for intakes of omega-3 highly unsaturated fatty acids. Prostaglandins Leukot Essent Fatty Acids. 2009;81(2-3):179-186. 58. Appleton KM, Gunnell D, Peters TJ, Ness AR, Kessler D, Rogers PJ. No clear evidence of an association between plasma concentrations of n-3 long-chain polyunsaturated fatty acids and depressed mood in a non-clinical population. Prostaglandins Leukot Essent Fatty Acids. 2008;78(6):337-342. 59. Emsley R, Niehaus DJ, Oosthuizen PP, et al. Safety of the omega-3 fatty acid, eicosapentaenoic acid (EPA) in psychiatric patients: results from a randomized, placebo-controlled trial. Psychiatry research. 2008;161(3):284-291. 60. Appleton KM, Peters TJ, Hayward RC, et al. Depressed mood and n-3 polyunsaturated fatty acid intake from fish: non-linear or confounded association? Soc Psychiatry Psychiatr Epidemiol. 2007;42(2):100-104. 61. Nieminen LR, Makino KK, Mehta N, Virkkunen M, Kim HY, Hibbeln JR. Relationship between omega-3 fatty acids and plasma neuroactive steroids in alcoholism, depression and controls. Prostaglandins Leukot Essent Fatty Acids. corticotrophin releasing hormone in perpetrators of domestic violence. Biological psychiatry. 2004;56(11):895-897. 63. Williams LL, Kiecolt-Glaser JK, Horrocks LA, Hillhouse JT, Glaser R. Quantitative association between altered plasma esterified omega-6 fatty acid proportions and psychological stress. Prostaglandins Leukot Essent Fatty Acids. Han X, Eggett DL, Parker TL. Evaluation of the Health Benefits of a Multivitamin, Multimineral, Herbal, Essential Oil-Infused Supplement: A Pilot Trial. J Diet Suppl. 2018;15(2):153-160. 65. Jamilian M, Shojaei A, Samimi M, et al. The effects of omega-3 and vitamin E co-supplementation on parameters of mental health and gene expression related to insulin and inflammation in subjects with polycystic ovary syndrome. Journal of affective disorders. 2018;229:41-47. 66. Barbarich NC, McConaha CW, Halmi KA, et al. Use of nutritional supplements to increase the efficacy of fluoxetine in the treatment of anorexia nervosa. Int J Eat Disord. 2004;35(1):10-15. Protocol but no result of studies (n=2) 67. Matsuoka Y, Nishi D, Nakaya N, et al. Attenuating posttraumatic distress with omega-3 polyunsaturated fatty acids among disaster medical assistance team members after the Great East Japan Earthquake: the APOP randomized controlled trial. BMC psychiatry. 2011;11:132. 68. Matsuoka Y, Nishi D, Yonemoto N, et al. Tachikawa project for prevention of posttraumatic stress disorder with polyunsaturated fatty acid (TPOP): study protocol for a randomized controlled trial. BMC psychiatry. 2013;13:8. Review article (n=7) 69. Bozzatello P, Brignolo E, De Grandi E, Bellino S. Supplementation with Omega- [/fig]
[fig] Funnel: Plot of Standard Error by Hedges's g [/fig]
|
Toll like receptor 7 expressed by malignant cells promotes tumor progression and metastasis through the recruitment of myeloid derived suppressor cells
In non-small cell lung carcinoma (NSCLC), stimulation of toll-like receptor 7 (TLR7), a receptor for single stranded RNA, is linked to tumor progression and resistance to anticancer chemotherapy. However, the mechanism of this effect has been elusive. Here, using a murine model of lung adenocarcinoma, we demonstrate a key role for TLR7 expressed by malignant (rather than by stromal and immune) cells, in the recruitment of myeloid derived suppressor cells (MDSCs), induced after TLR7 stimulation, resulting in accelerated tumor growth and metastasis. In adenocarcinoma patients, high TLR7 expression on malignant cells was associated with poor clinical outcome, as well as with a gene expression signature linked to aggressiveness and metastastic dissemination with high abundance of mRNA encoding intercellular adhesion molecule 1 (ICAM-1), cytokeratins 7 and 19 (KRT-7 and 19), syndecan 4 (SDC4), and p53. In addition, lung tumors expressing high levels of TLR7 have a phenotype of epithelial mesenchymal transition with high expression of vimentin and low abundance of E-cadherin. These data reveal a crucial role for cancer cell-intrinsic TLR7 expression in lung adenocarcinoma progression.ARTICLE HISTORY
# Introduction
Tumors constitute a complex microenvironment, which is composed of genetically altered malignant and non-mutated stromal cells, the latter including fibroblasts, vascular and lymphatic vessels, as well as immune cells [bib_ref] The immune contexture in human tumours: impact on clinical outcome, Fridman [/bib_ref] [bib_ref] Hallmarks of cancer: the next generation, Hanahan [/bib_ref]. Many tumor-infiltrating immune cells including CD4 + and CD8 + T cells, B lymphocytes, or type 1 macrophages (M1) participate in immunosurveillance, meaning that they are commonly associated with good clinical outcome. Other leukocyte subtypes including regulatory T cells (Tregs), myeloid-derived suppressor cells (MDSCs) and type 2 macrophages (M2) have immunosuppressive functions, meaning that their local presence has a negative clinical impact [bib_ref] The immune contexture in human tumours: impact on clinical outcome, Fridman [/bib_ref]. The composition of tumor infiltrating immune cells strongly influences the tumor microenvironment and the behavior of the tumor in terms of progression, metastasis and treatment resistance.
Toll-like receptors (TLRs), which recognize ligands from pathogens (pathogen-associated molecular patterns, PAMPs) or dying cells (damaged-associated molecular patterns, DAMPs), are involved in the immune response against pathogens [bib_ref] Signaling in innate immunity and inflammation, Newton [/bib_ref] , in thus far that they induce the secretion of inflammatory cytokines and chemokines upon their activation [bib_ref] Toll-like receptor modulators: a patent review, Basith [/bib_ref]. These receptors are largely expressed by immune cells and epithelial cells but also by several types of tumors cells [bib_ref] Toll-like receptor stimulation in cancer: A pro-and anti-tumor double-edged sword, Dajon [/bib_ref]. The role of TLRs in tumor microenvironment is complex. On one hand, TLRs expressed by immune cells can favor immunosurveillance, because their stimulation favors the maturation and activation of innate and adaptive immune effectors. On the other hand, TLRs expressed by cancer cells may receive stimuli that favor tumor progression [bib_ref] Dual roles of TLR7 in the lung cancer microenvironment, Dajon [/bib_ref] [bib_ref] Coordinated tumor immunity, Dranoff [/bib_ref].
We have previously demonstrated that TLR7, a single stranded RNA receptor that is usually expressed in endosomes of immune cells including plasmacytoid and conventional dendritic cells (pDCs and cDCs), macrophages and B lymphocytes [bib_ref] Quantitative expression of tolllike receptor 1-10 mRNA in cellular subsets of human..., Hornung [/bib_ref] [bib_ref] TLR-mediated stimulation of APC: distinct cytokine responses of B cells and dendritic..., Barr [/bib_ref] , is expressed by malignant cells from NSCLC patients [bib_ref] Triggering of TLR7 and TLR8 expressed by human lung cancer cells induces..., Cherfils-Vicini [/bib_ref]. High expression by cancer cells was associated with poor prognosis, both in early stages that were treated with surgery alone and more advanced stages treated by neoadjuvant chemotherapy. Moreover, in a mouse model of lung cancer, TLR7 stimulation favored tumor growth [bib_ref] TLR7 promotes tumor progression, chemotherapy resistance, and poor clinical outcomes in non-small..., Chatterjee [/bib_ref]. The present study has been designed to decipher the mechanisms involved in the pro-tumorigenic and possible pro-metastatic effects of TLR7 stimulation.
Here, we report that the stimulation of TLR7 expressed by malignant cells favors tumor progression and metastasis, increases CCL2 and GM-CSF secretion in the tumor microenvironment, and elicits the recruitment of MDSCs into the tumor. In vivo depletion of MDSCs abolished the TLR7dependent pro-tumorigenic and pro-metastatic effect. Finally, we observed in lung adenocarcinoma patients, that high expression of TLR7 by tumor cells is associated to a pro-metastatic gene expression signature as well as to epithelial-mesenchymal transition (EMT).
# Results
## Intratumoral tlr7 agonist injection favors tumor progression through a direct effect on carcinoma cells
Murine lung adenocarcinoma LLC-luc cells express several TLRs including TLR4, TLR7 and TLR9 [fig_ref] Figure 1: Pro-or anti-tumoral effects of different TLRs stimulation [/fig_ref]. Expression of TLR7 was also validated by quantitative PCR (data not shown). We have previously demonstrated that intratumoral injection of TLR7 agonist results in increased tumor progression in immunocompetent as well as in immunodeficient mice [bib_ref] TLR7 promotes tumor progression, chemotherapy resistance, and poor clinical outcomes in non-small..., Chatterjee [/bib_ref]. However, the precise mechanisms involved in such pro-tumorigenic effects are elusive. To determine if this effect was specific, we analyzed the growth of subcutaneous grafted LLC-luc tumors in wild-type (WT) mice locally injected with TLR4, TLR7 or TLR9 agonists. Bacterial lipopolysaccharide (LPS), an agonist of TLR4, failed to affect tumor growth [fig_ref] Figure 1: Pro-or anti-tumoral effects of different TLRs stimulation [/fig_ref] , while CpG, an agonist of TLR9, actually inhibited tumor growth [fig_ref] Figure 1: Pro-or anti-tumoral effects of different TLRs stimulation [/fig_ref]. On the contrary, intratumoral injection of the TLR7 agonist CL264 induced a pro-tumorigenic effect [fig_ref] Figure 1: Pro-or anti-tumoral effects of different TLRs stimulation [/fig_ref] , in accordance to our previous observations [bib_ref] TLR7 promotes tumor progression, chemotherapy resistance, and poor clinical outcomes in non-small..., Chatterjee [/bib_ref].
The pro-tumorigenic effect of TLR7 stimulation could result from a direct stimulation of TLR7 on tumor cells and/ or from stimulation of immune cells that express TLR7. We determined the level of TLR7 expression of immune cells in the tumor, the spleen and the blood of tumor-bearing mice by flow cytometry after optimization of the gating strategy [fig_ref] Figure 1: Pro-or anti-tumoral effects of different TLRs stimulation [/fig_ref]. TLR7 was highly expressed by macrophages, pDCs, and to a less extent by cDCs in all compartments including the tumor [fig_ref] Figure 2: TLR7 stimulation alterates intratumoral immune infiltrates [/fig_ref]. Thus, TLR7 agonists may act both on malignant cells and on tumor-infiltrating myeloid cells to favor tumor progression.
To discriminate the target population of TLR7 agonist, we took advantage of TLR7-deficient mice (TLR7 KO), and generated TLR7-deficient LLC-luc cells (LLC TLR7 KO) by CRISPR/Cas9-mediated gene editing [fig_ref] Figure 1: Pro-or anti-tumoral effects of different TLRs stimulation [/fig_ref]. We injected WT (TLR7 expressing) LLC-luc cells into TLR7 KO mice or alternatively TLR7-deficient LLC-luc cells into WT mice, and then evaluated the effects of the TLR7 agonist CL264 on tumor progression. We observed a growth accelerating effect of CL264 in TLR7 KO mice, similar to that observed in WT mice [fig_ref] Figure 1: Pro-or anti-tumoral effects of different TLRs stimulation [/fig_ref]. In sharp contrast, the TLR7 agonist failed to stimulate the growth of TLR7-deficient LLC-luc tumors [fig_ref] Figure 1: Pro-or anti-tumoral effects of different TLRs stimulation [/fig_ref]. These results clearly demonstrate that the pro-tumorigenic effect of CL264 effect is mediated by an effect on TLR7 expressed by malignant (rather than immune) cells.
## Stimulation of tlr7 alterates the intratumoral immune infiltrates
TLR stimulation is known to modify the inflammatory microenvironment of tumors. We analyzed whether the local injection of TLR7 agonist would impact the composition of the immune infiltrate. Immune phenotyping by flow cytometry (for gating strategy see [fig_ref] Figure 1: Pro-or anti-tumoral effects of different TLRs stimulation [/fig_ref] revealed the percentages of T, B and NK cells, Tregs, macrophages, cDCs, pDCs and MDSCs, including granulocytic and monocytic subpopulations. The immune infiltrate of vehicle-only treated tumors was composed by a majority of myeloid cells including 49% of MDSCs, 15% of cDCs and 5% of macrophages, as well as a by a minority of adaptive immune cells including 10% T cells, 5% NK cells and 3% Tregs [fig_ref] Figure 2: TLR7 stimulation alterates intratumoral immune infiltrates [/fig_ref]. The local injection of CL264 induced a modification of percentages and absolute cells numbers of MDSCs, with a significant (unpaired Student T test, p < 0.05) increase from 49 ± 7% to 66 ± 10% [fig_ref] Figure 2: TLR7 stimulation alterates intratumoral immune infiltrates [/fig_ref]. CL264 also induced a decrease of cDCs (from 15% ± 1% to 9%± 2%) and of pDCs (from 2% to 1.5%) [fig_ref] Figure 2: TLR7 stimulation alterates intratumoral immune infiltrates [/fig_ref]. In control tumors, MDSCs were mostly monocytic, and CL264 treatment resulted in the increase of granulocytic MDSCs [fig_ref] Figure 2: TLR7 stimulation alterates intratumoral immune infiltrates [/fig_ref]. In contrast, CL264 did not affect the frequency of macrophages, NK, T, Tregs, and B cells [fig_ref] Figure 2: TLR7 stimulation alterates intratumoral immune infiltrates [/fig_ref].
## Cl264 stimulates mdsc recruitment by acting on tlr7 expressed by malignant cells
We investigated if the increase of MDSCs within the tumor was a consequence of TLR7 stimulation on cancer or immune cells. The TLR7 agonist CL264 was injected into LLC-luc tumors that were established in WT or TLR7 KO mice.
At baseline levels (without CL264 treatment), WT and TLR7KO mice showed low frequencies of MDSCs, cDCs and macrophages in their blood and spleens [fig_ref] Figure 4: MDSCs are involved in the pro-metastatic effect of TLR7 stimulation [/fig_ref]. TLR7 KO mice exhibited a lower proportion of MDSCs, cDCs, NK and T cells and increased proportion of B cells in the spleen and the blood compared to WT mice [fig_ref] Figure 4: MDSCs are involved in the pro-metastatic effect of TLR7 stimulation [/fig_ref] , B). The tumor immune infiltrate was characterized by high frequencies of MDSCs, both in TLR7 KO and in WT mice. Higher proportions of macrophages and T cells and lower proportions of cDCs were detected in the tumors of TLR7 KO compared to WT mice [fig_ref] Figure 4: MDSCs are involved in the pro-metastatic effect of TLR7 stimulation [/fig_ref].
In tumors growing on TLR7 KO mice [fig_ref] Figure 2: TLR7 stimulation alterates intratumoral immune infiltrates [/fig_ref] , compared to WT mice [fig_ref] Figure 2: TLR7 stimulation alterates intratumoral immune infiltrates [/fig_ref] the injection of TLR7 agonist increased the proportion and absolute number of MDSCs [fig_ref] Figure 2: TLR7 stimulation alterates intratumoral immune infiltrates [/fig_ref] , with a marked shift towards a granulocytic MDSC phenotype [fig_ref] Figure 2: TLR7 stimulation alterates intratumoral immune infiltrates [/fig_ref] , concomitant with a slight decreased proportion of cDCs [fig_ref] Figure 2: TLR7 stimulation alterates intratumoral immune infiltrates [/fig_ref] , but did not modify the frequency of pDCs [fig_ref] Figure 2: TLR7 stimulation alterates intratumoral immune infiltrates [/fig_ref] , macrophages, NK, T, Tregs and B cells [fig_ref] Figure 3: MDSCs are involved in the pro-tumorigenic effect mediated by TLR7 stimulation [/fig_ref].
Collectively, these data indicate that CL264 maintains its local immunomodulatory function resulting in an increased presence of MDSCs in WT and TLR7 KO mice.
Mdscs are involved in the pro-tumorigenic effect mediated by TLR7 stimulation
To decipher the possible role of MDSCs in the tumor growth stimulatory effect of CL264, we experimentally depleted MDSCs from the system. Tumor bearing mice that were treated with CL264 or vehicle only received intraperitoneal injections of anti-Gr1 monoclonal antibody (mAb) or control isotype every 48 hours [fig_ref] Figure 3: MDSCs are involved in the pro-tumorigenic effect mediated by TLR7 stimulation [/fig_ref]. During the experiment, less than 0.4% of MDSCs were detectable in the blood of mice, confirming the efficacy of MDSC depletion [fig_ref] Figure 3: MDSCs are involved in the pro-tumorigenic effect mediated by TLR7 stimulation [/fig_ref]. MDSC depletion resulted in decreased tumor growth, both in WT and in TLR7 KO mice [fig_ref] Figure 3: MDSCs are involved in the pro-tumorigenic effect mediated by TLR7 stimulation [/fig_ref] , G). The TLR7 agonist CL264 stimulated tumor growth [fig_ref] Figure 3: MDSCs are involved in the pro-tumorigenic effect mediated by TLR7 stimulation [/fig_ref] , H), and this pro-tumorigenic effect was completely lost in the absence of MDSCs, both in WT mice [fig_ref] Figure 3: MDSCs are involved in the pro-tumorigenic effect mediated by TLR7 stimulation [/fig_ref] and in TLR7 KO mice [fig_ref] Figure 3: MDSCs are involved in the pro-tumorigenic effect mediated by TLR7 stimulation [/fig_ref].
These results demonstrate that the tumor growth exacerbating effect of TLR7 stimulation is mediated by the increased number of MDSCs within the tumor microenvironment.
TLR7 stimulation induced an increased production of CCL2 and GM-CSF within the tumor microenvironment Given the involvement of MDSCs in the pro-tumorigenic effect of TLR7 stimulation, we investigated the mechanisms that might be involved in the local CL264-induced MDSCs expansion. Several cytokines such as IL-6, GM-CSF or CCL2 stimulate the recruitment or proliferation of MDSCs and may be produced by tumor cells. Local injection of the TLR7 agonist CL264 enhanced the ELISA-detectable concentration of CCL2 and GM-CSF within the tumor. The concentrations of CCL2 and GM-CSF were 10407 ± 1326 and 256 ± 15 pg/ml, respectively, in mice treated with CL264, and 7453 ± 1380 and 186 ± 15 pg/ml, respectively, in control mice receiving PBS [fig_ref] Figure 3: MDSCs are involved in the pro-tumorigenic effect mediated by TLR7 stimulation [/fig_ref] , K). In contrast, CL264 failed to significantly modulate the amount of IL-6 found in the tumor [fig_ref] Figure 3: MDSCs are involved in the pro-tumorigenic effect mediated by TLR7 stimulation [/fig_ref]. The secretion of CCL2 and GM-CSF by tumor cells was confirmed in vitro, which was found increased in supernatants of CL264-stimulated cells [fig_ref] Figure 3: MDSCs are involved in the pro-tumorigenic effect mediated by TLR7 stimulation [/fig_ref].
These results suggest that CL264 may stimulate the recruitment of MDSCs via the induction of CCL2 and GM-CSF.
## Tlr7 agonist does not modify the immunosuppressive capacities of mdscs
To unravel the possible impact of TLR7 stimulation on MDSCs functions, we measured the production of molecules involved in MDSCs-mediated immunosuppressive circuitries. The production of nitrate, H 2 O 2 , and arginase 1 by MDSCs was not affected by CL264 injection. Similarly, the proportion of ROS producing cells was not altered by CL264 [fig_ref] Figure 5: TLR7 stimulation of LLC cells leads to epithelio-mesenchymal signature [/fig_ref]. In addition, the capacity of MDSCs to suppress T cell proliferation was similar for MDSCs purified from PBS-treated and CL264-treated tumors [fig_ref] Figure 5: TLR7 stimulation of LLC cells leads to epithelio-mesenchymal signature [/fig_ref]. These results demonstrate that intratumoral injection of TLR7 agonist does not modify the immunosuppressive properties of intratumoral MDSCs.
## Mdsc are involved in the pro-metastatic effect of tlr7 stimulation
Next, we analyzed the capacity of CL264 to stimulate lung metastasis. NOD/SCID or C57BL/6 mice were subcutaneously grafted with LLC-luc cells, and the apparition of metastasis was imaged over time, by bioluminescence or precisely quantified at necropsy following perfusion with Indina Ink and labelling of the lungs with Feket solution. Either of these methods revealed the capacity of CL264 to stimulate lung metastasis, both in immunodeficient NOD/SCID mice [fig_ref] Figure 4: MDSCs are involved in the pro-metastatic effect of TLR7 stimulation [/fig_ref] and in immunocompetent C57Bl/6 mice [fig_ref] Figure 4: MDSCs are involved in the pro-metastatic effect of TLR7 stimulation [/fig_ref].
The number of metastasis was quantified in PBS-and CL264-treated mice. We observed the presence of lung metastases in both groups of mice, but the mice having received three injections of TLR7 agonist had a significantly higher number of metastases with a mean of 9.5 (± 7.5) metastases per lungs for CL264-treated mice, compared to 4.9 (± 4.5) metastases per lungs for the control group [fig_ref] Figure 4: MDSCs are involved in the pro-metastatic effect of TLR7 stimulation [/fig_ref]. Repeated injections of CL264 (every three days throughout the duration of the experiment) led to even higher number of lung metastasis (19.5 (± 11.5) metastasis per lungs) [fig_ref] Figure 4: MDSCs are involved in the pro-metastatic effect of TLR7 stimulation [/fig_ref]. These results indicate that TLR7 stimulation has a marked pro-metastatic effect.
Given the implication of MDSC in metastatic dissemination 12 , we investigated the potential role of MDSCs in the pro-metastatic effects of TLR7 stimulation. MDSC depletion by injection of anti-Gr1 mAbs led to a significant decrease in the number of metastases (mean of 2.5 (± 2) metastases per lungs) in CL264-treated mice compared to control mice that did not receive depleting antibody (mean of 6.5 (± 4) metastases per lungs) [fig_ref] Figure 4: MDSCs are involved in the pro-metastatic effect of TLR7 stimulation [/fig_ref] , showing that TLR7 stimulation increases the number of lung metastases through a mechanism involving MDSCs.
## High expression of tlr7 in human lung adenocarcinoma is associated with a metastatic gene expression signature
We further characterized the in vitro impact of CL264 on metastatic properties of malignant cells. The stimulation of LLC cells with CL264 resulted in epithelio-mesenchymal transition gene expression signature, with an increased gene expression of N-cadherin, vimentin, snail1, snail 2 and zeb, and a decreased expression of E-cadherin [fig_ref] Figure 5: TLR7 stimulation of LLC cells leads to epithelio-mesenchymal signature [/fig_ref] , B).
Immunohistochemistry was then used to determine the expression of E-cadherin, associated to epithelial cell phenotype, and that of vimentin, expressed by mesenchymal cells, on human malignant cells expressing high or low levels of TLR7. Forty-five adenocarcinoma patients were analyzed for the expression of TLR7, E-cadherin and vimentin (for representative staining patterns see . Patients were divided into two groups that were expressing TLR7 or not on malignant cells. The expression of E-cadherin and vimentin by tumor cells was determined in a semiquantitative fashion. In TLR7 hi patients, the expression of E-cadherin was significantly lower (25.6%± 3.7% of cancer cells expressing E-Cadherin) than in TLR7 low patients (49.3%± 7% of tumor cells expressing E-cadherin) . Concomitantly, in TLR7 hi patients, the expression of vimentin was higher (34.6% ± 6.7% of tumor cells expressed vimentin), than in TLR7 low patients (15.3%± 3.6% of tumor cells expressed vimentin) . In conclusion, in patients with lung adenocarcinoma, high expression of TLR7 on cancer cells is associated with signs of EMT. Data are mean ± SEM (5 mice/group). *: p < 0,05, **: p < 0,01, ***: p < 0,0001. ns = not significant. Each experiment was repeated three times.
We then determined the physiopathology impact of TLR7 expression on the metastasis process in a large cohort of adenocarcinoma patients. We first determined the prognostic value of TLR7 on a cohort of 154 adenocarcinoma patients (sup [fig_ref] Table 1: Univariate and multivariate analysis for 154 patients with lung adenocarcinoma not treated... [/fig_ref] , the malignant cells of whom displayed TLR7 either at a rather low level (< 5%) or were almost completely positive (> 82%, cut off determined in 11 ). We found that high TLR7 expression by tumor cells was associated with poor clinical outcome compared to low TLR7 expression, with a median OS of 45 months, compared to 90 months, respectively [fig_ref] Figure 7: High expression of TLR7 by malignant adenocarcinoma cells is associated with metastatic... [/fig_ref]. Multivariate analysis revealed TLR7 expression as an independent prognostic marker [fig_ref] Table 1: Univariate and multivariate analysis for 154 patients with lung adenocarcinoma not treated... [/fig_ref]. We then selected twelve patients from this cohort, including six patients having low (< 5%, TLR7 low ) and six having high (100%, TLR7 hi ) TLR7 expression, and performed gene expression analysis using nanostring technology and the pan cancer progression panel. Volcano plot analysis revealed that among 770 genes analyzed, six were significantly overexpressed and five were underexpressed in TLR7 hi compared to TLR7 low patients [fig_ref] Figure 7: High expression of TLR7 by malignant adenocarcinoma cells is associated with metastatic... [/fig_ref]. Unsupervised clustering of differentially expressed genes, revealed a small cluster of three genes including ICAM1, keratin 7 (KRT-7) and keratin 19 (KRT-19) that were overexpressed in TLR7 hi patients. Additionally, NME1 (encoding nucleoside diphosphate kinase 1) and syndecan 4 (SDC4) (heparin sulfate proteoglycan molecule) were found overexpressed in the TLR7 hi group of patients [fig_ref] Figure 7: High expression of TLR7 by malignant adenocarcinoma cells is associated with metastatic... [/fig_ref]. More detailed box-plot analysis of these genes confirmed their overexpression in TLR7 hi patients [fig_ref] Figure 7: High expression of TLR7 by malignant adenocarcinoma cells is associated with metastatic... [/fig_ref]. Interestingly, TP53 was also found highly expressed in TLR7 hi patients [fig_ref] Figure 7: High expression of TLR7 by malignant adenocarcinoma cells is associated with metastatic... [/fig_ref]. We then searched for correlation between the level of gene expression and survival of adenocarcinoma lung cancer patients, using the public software KM plotter [bib_ref] Online survival analysis software to assess the prognostic value of biomarkers using..., Győrffy [/bib_ref]. High expression of KRT-7 and KRT-19, as well as that of NME1 and TP53, significantly correlated with poor clinical outcome [fig_ref] Figure 7: High expression of TLR7 by malignant adenocarcinoma cells is associated with metastatic... [/fig_ref].
Collectively, these data indicate the existence of a gene signature that is associated to TLR7 expression by malignant cells as well as with poor prognosis.
# Discussion
Our present data indicate that TLR7 stimulation induces protumorigenic and pro-metastatic effects in a preclinical model of lung cancer while providing two salient insights. First, our data indicate that TLR7 must be expressed by the cancer cells (rather than by immune cells) so that the local presence of a TLR7 agonist can stimulate tumor growth. Second, we reveal the direct involvement of MDSCs in both the tumor growth-promoting and pro-metastatic effects of TLR7 stimulation. The presence of excess amount of a TLR7 agonist within the tumor causes an expansion of the number of MDSCs yet fails to enhance their immunosuppressive capacity. Our results also suggest that MDSC recruitment is mediated via a higher CCL2 and GM-CSF production in the tumor microenvironment after TLR7 stimulation. In addition, we demonstrated that high TLR7 expression by tumor cells from adenocarcinoma patients is associated to signs of EMT, as well as to a pro-metastatic gene expression signature. In particular, we observed a higher expression of KRT-7, KRT-19 and TP53, which are associated to the metastasis process and to a poor prognosis.
Previous studies have reported pro-tumorigenic effects for different TLRs, such as TLR2 [bib_ref] Targeting Toll-like receptor 2 inhibits growth of head and neck squamous cell..., Farnebo [/bib_ref] [bib_ref] Listeria monocytogenes promotes tumor growth via tumor cell toll-like receptor 2 signaling, Huang [/bib_ref] [bib_ref] STAT3-driven upregulation of TLR2 promotes gastric tumorigenesis independent of tumor inflammation, Tye [/bib_ref] [bib_ref] A cell-intrinsic role for TLR2-MYD88 in intestinal and breast epithelia and oncogenesis, Scheeren [/bib_ref] [bib_ref] lipopeptides systemically increase myeloid-derived suppressor cells through TLR2 signaling, Maruyama [/bib_ref] , TLR4 19-23 , TLR5 24 , TLR7 10,11,25 and TLR9 [bib_ref] Pellat-Deceunynck C. Pathogen-associated molecular patterns are growth and survival factors for human..., Jego [/bib_ref] [bib_ref] TLR9 signaling through NF-κB/RELA and STAT3 promotes tumor-propagating potential of prostate cancer..., Moreira [/bib_ref] [bib_ref] TLR9 ligation in pancreatic stellate cells promotes tumorigenesis, Zambirinis [/bib_ref] [bib_ref] Expression of TLR9 in tumor-infiltrating mononuclear cells enhances angiogenesis and is associated..., Belmont [/bib_ref] [bib_ref] Hypoxia induced HMGB1 and mitochondrial DNA interactions mediate tumor growth in hepatocellular..., Liu [/bib_ref]. In prostate and breast cancer, the stimulation of TLR9 induced an over-expression of matrix metalloproteinase 13 as well as an increase of cell invasion [bib_ref] Toll like receptor-9 agonists stimulate prostate cancer invasion in vitro, Ilvesaro [/bib_ref] [bib_ref] Toll-like receptor 9 agonists promote cellular invasion by increasing matrix metalloproteinase activity, Merrell [/bib_ref]. Finally, TLR4 stimulation increased tumor cells invasion in breast cancer [bib_ref] Toll-like receptor 4 prompts human breast cancer cells invasiveness via lipopolysaccharide stimulation..., Yang [/bib_ref] and in lung adenocarcinoma [bib_ref] Gram negative bacteria increase non-small cell lung cancer metastasis via Toll-like receptor..., Chow [/bib_ref].
MDSC, a cell type with marked immunosuppressive properties [bib_ref] Coordinated regulation of myeloid cells by tumours, Gabrilovich [/bib_ref] , has been linked to tumor progression in various murine models of cancer [bib_ref] Aberrant frequency of IL-10-producing B cells and its association with Treg and..., Liu [/bib_ref] [bib_ref] Mesenchymal transition and dissemination of cancer cells is driven by myeloidderived suppressor..., Toh [/bib_ref] [bib_ref] CD14(+)S100A9(+) monocytic myeloid-derived suppressor cells and their clinical relevance in non-small cell..., Feng [/bib_ref]. In cancer patients, the number of circulating MDSCs significantly increases compared to healthy subjects 38 . Our study, however, is the first demonstration in a cancer model that TLR stimulation leads to local MDSCs expansion, accounting for enhanced tumor growth and metastasis.
TLR7 stimulation has previously mostly reported to stimulate anticancer immunosurveillance, for instance through direct stimulation of TLRs on antigens presenting cells, leading to their maturation [bib_ref] Tumor promotion by intratumoral plasmacytoid dendritic cells is reversed by TLR7 ligand..., Mercier [/bib_ref] [bib_ref] Imiquimod inhibits melanoma development by promoting pDC cytotoxic functions and impeding tumor..., Aspord [/bib_ref]. However, in recent years, TLR7 expression by tumor cells has been associated with a pro-tumoral effect in pancreatic adenocarcinoma [bib_ref] Toll-like receptor 7 regulates pancreatic carcinogenesis in mice and humans, Ochi [/bib_ref] and in NSCLC [bib_ref] Triggering of TLR7 and TLR8 expressed by human lung cancer cells induces..., Cherfils-Vicini [/bib_ref] [bib_ref] TLR7 promotes tumor progression, chemotherapy resistance, and poor clinical outcomes in non-small..., Chatterjee [/bib_ref]. We found that stimulation of TLR7 expressed by adenocarcinoma cells modulated the immune infiltrate, leading to a significant expansion in MDSCs, associated with an increased secretion of CCL2 and GM-CSF in the tumor microenvironment. Both these cytokines are known to play a key role in the recruitment of MDSCs [bib_ref] Identification of a new subset of myeloid suppressor cells in peripheral blood..., Filipazzi [/bib_ref] [bib_ref] High-dose granulocyte-macrophage colony-stimulating factorproducing vaccines impair the immune response through the recruitment..., Serafini [/bib_ref] [bib_ref] CCL2 Promotes Colorectal Carcinogenesis by Enhancing Polymorphonuclear Myeloid-Derived Suppressor Cell Population and..., Chun [/bib_ref]. Indeed, the absence of CCL2 production by cancer cells considerably reduces the recruitment of MDSCs into tumors [bib_ref] CCL2 Promotes Colorectal Carcinogenesis by Enhancing Polymorphonuclear Myeloid-Derived Suppressor Cell Population and..., Chun [/bib_ref]. Importantly, TLR7 stimulation had a pro-metastatic effect, while MDSC depletion drastically reduced this number of metastasis. In melanoma, MDSCs are associated with tumor progression, signs of EMT such as decreased E-Cadherin expression and increased migratory capacities of malignant cells [bib_ref] Mesenchymal transition and dissemination of cancer cells is driven by myeloidderived suppressor..., Toh [/bib_ref]. Similarly, MDSCs may contribute to metastatic dissemination of different types of cancer, including NSCLC 43 . Finally, Srivastava's team demonstrated that when MDSCs were depleted, the size of tumors decreased and cell migration was reduced [bib_ref] Myeloid suppressor cell depletion augments antitumor activity in lung cancer, Mk [/bib_ref]. In line with the involvement of EMT in the metastatic process, we found that TLR7 expression in lung adenocarcinoma was associated with EMT. We observed that a high expression of TLR7 was associated with low expression of E-cadherin and overexpression of vimentin. EMT is a process in which cells lose adhesive properties because of down-regulating E-cadherin expression and rearrange their cytoskeleton through enhanced vimentin expression [bib_ref] Epithelialmesenchymal transitions in development and disease, Thiery [/bib_ref]. These changes in cancer cells play a critical role in metastasis development.
In adenocarcinoma patients, high TLR7 expression was associated with gene expression signature possibly linked to metastasis, with overexpression of KRT-7, KRT-19 and SDC4. These three genes have already been ascribed with pro-metastatic properties in different types of cancers. Indeed, KRT-7 is established as a diagnostic marker in a variety of malignancies including biliary duct, bladder, breast, cervix, kidney, liver, lung adenocarcinoma, ovary and pancreas cancer [bib_ref] Keratins in health and cancer: more than mere epithelial cell markers, Karantza [/bib_ref]. In patients with lung adenocarcinoma, strong expression of KRT-7 was associated with a poor overall survival [bib_ref] Keratin 34betaE12/keratin7 expression is a prognostic factor of cancerspecific and overall survival..., Pøhl [/bib_ref]. Keratin 19 is a diagnostic marker in some tumors, such as in biliary duct, bladder, breast, cervix, colon, lung adenocarcinoma, ovary, pancreas, pleura, prostate, stomach and uterus cancer [bib_ref] Keratins in health and cancer: more than mere epithelial cell markers, Karantza [/bib_ref]. In addition, KRT-19-positive hepatocellular carcinomas highly express invasion-related/metastasis-related markers [bib_ref] Keratin 19, a Cancer Stem Cell Marker in Human Hepatocellular Carcinoma, Kawai [/bib_ref]. Finally, SDC4 plays a role in the development and metastatic occurrence in renal cell carcinoma [bib_ref] Up-regulation of TGM2 with ITGB1 and SDC4 is important in the development..., Erdem [/bib_ref].
Strong expression of TLR7 in patients with lung cancer is associated with poor survival prognosis but also with a prometastatic gene signature. As a result, it would be interesting to explore the molecular etiology of this pathogenic TLR7 overexpression. Ligation of TLR7 by agonists can stimulate TLR7 expression in a feedforward loop [bib_ref] IFNs activate toll-like receptor gene expression in viral infections, Miettinen [/bib_ref] , which implies that high TLR7 abundance might be a sign of the natural stimulation of this receptor by ligands present in the tumor microenvironment. The molecular nature and source of such natural TLR7 agonists remains to be clarified. Moreover, it will be important to develop and characterize TLR7 antagonists or inhibitors of TLR7-elicited signals that might have immunostimulatory and anticancer effects.
# Materials and methods
## Cell lines
LLC-Luc cells (gift from Vincenzo di Bartolo, Institut Pasteur, Paris) were obtained from LLC cells (Lewis Lung Carcinoma, ATCC CRL-1642) that were stably transfected with firefly luciferase. LLC-Luc cells were cultured in DMEM+ GlutaMax (Gibco) with 10% heat-inactivated FCS, 1% penicilin/streptomycin mixture, and 0.5mg/ml hygromycin B. The cumulative culture length of the cells was fewer than 6 months after resuscitation. Early passage cells were used for all experiments and they were not reauthenticated.
## Mice
Wild-type C57Bl/6J mice, mice deficient in TLR7 (B6.129S1-Tlr7tm1Flv/J strain), and NOD/SCID mice were purchased from Charles River Laboratories. All the animals were house kept at CEF (Centre d'Explorations Fonctionnelles of Centre de Recherche des Cordeliers, Agreement no. A75-06-12). They were maintained in a constant temperature and humidity in light controlled room with a 12 hours light cycle. They had free access to food (SAFE Laboratory) and tap water. All experiments were conducted in 6-8 week-old female mice in accordance with the institutional guidelines and the recommendations for the care and use of laboratory animals put forward by the Directive 2010/63/ EU revising Directive 86/609/EEC on the protection of animals used for scientific purposes (project has been approved by a user establishment's ethics committee and the Project Authorization: number Ce5/2010/057 and 6629).
## Tumor growth and metastasis
Cancer cells (1 × 10 6 in 100μl PBS/injection) were injected into the flank of recipient mice. Mice then received intra-tumor injections of LPS (40 μg/injection), CL264 (40 μg/injection) or CpG (30 μg/ injection) (Invivogen) at day 0, 3 and 6 after tumor grafting. Tumor volume (length (l), breath (b), depth (d)) was measured every 2-3 days using caliper, and was calculated taking into account the three values according to the ellipsoid volume calculationformula:tumorvolume=4/3×π×l/2×b/2=0.5236×l×b×d. For metastasis detection, mice were injected (i.p.) with 225mg/kg body weight D-luciférine K+ (Interchim), then were anaesthetized using isofluran, 10 minutes before imaging for bioluminescence using an IVIS imaging system (Lumina II). Mice were sacrificed between days 22 and 24. The rib cage was excised and 1 ml of India ink (7.5% solution) was injected into the trachea. The lungs were harvested and washed with 1X PBS. Separation of lobes and connectives tissues elimination was performed. The lungs were immersed and stored in 1ml of Feket solution (300ml of 70% ethanol, 30ml of 37% formaldehyde and 5 ml of glacial acetic acid). The number of metastatic foci was counted.
## Mdscs depletion
When tumors were palpable, mice were injected (i.p.) with anti-Gr1 (clone RB6-8C5, BioXCell) or control antibody (clone LTF-2, BioXCell) (200µg/injection, in 100µl PBS) every 48 hours. MDSC depletion was confirmed in the peripheral blood of recipient mice. Blood sample was recovered in heparin containing tubes. After red blood cell lysis by ACK, nucleated cells were incubated for 20 minutes at 4°C with blocking antibody (24G2, 1mg/ml, 1/100). Cells were stained for 30 minutes at 4°C with the following antibodies: PECy7 anti-CD45 (Clone 30-F11, BD Pharmingen, 1/800), eF450 anti-CD11b (Clone M1/70, eBioscience, 1/100), PerCPCy5.5 anti-Gr1 (Clone RB6-8C5, BD Pharmingen, 1/100), PE anti-CD11c (Clone N418, eBioscience, 1/100) and LIVE/DEAD® Fixable Yellow Dead Cell Stain Kit (Life Technologies, L34959, 1/100). Staining was assessed with a FACS Fortessa cytometer, and flow cytometry data were analyzed using FACSDiva (BD Biosciences) or FlowJO (Tree Star) software.
## Preparation of single cell suspensions and flow cytometry analyzes
Mice were sacrificed 10 days after cancer cell injection, and tumors and spleens were excised, mechanically dissociated, and incubated at 37°C for 30 minutes with DNase I (1mg/ml, Roche) and liberase TM (2,5mg/ml, Roche). Resulting single-cell suspensions were filtered through 70µm cell strainers (BD), and were incubated for 20 minutes with blocking antibody (24G2, 1mg/ml, 1/100). Cells were then stained for 30 minutes with LIVE/DEAD® Fixable Yellow Dead Cell Stain Kit (Life Technologies, L34959) and the following antibodies: PECy7 anti-CD45 (clone 30-F11, BD Pharmingen, 1/800), APC-eF780 anti-CD3 (clone UCHT1, eBioscience, 1/100), PETR anti-CD4 (clone GK1.5, Caltag, 1/100), AF700 anti-CD8 (clone 53-6.7, eBioscience1/100), PE anti-CD19 (clone 1D3, BD Pharmingen, 1/100), BV605 anti-NK1.1 (clone PK136, Biolegend, 1/50), FITC anti-CD11b (clone M1/70, BD Pharmingen, 1/100), eF450 anti-CD11b (clone M1/70, eBioscience, 1/100), BV786 anti-CD11c (clone HL3, BD Pharmingen, 1/50), BV605 anti-Gr1 (clone RB6-5C8, BD Pharmingen, 1/100), AF700 anti-F4/80 (clone CI:A3-1, AbD serotec). For the FoxP3 and TLR7 labeling, the cells were first incubated with fixation/permabilisation reagent (eBiosciences), . Strong TLR7 expression by malignant cells is link to signs of EMT. Representative immunohistochemistry staining of TLR7, E-cadherin and vimentin on malignant cells of adenocarcinoma patients. Images were taken at x20 magnification (A). Percentage of malignant cells that express E-cadherin (B) and vimentin (C) in TLR7 hi and TLR7 low patients (n = 43). *: p < 0,05, **: p < 0,01, ***: p < 0,001. Data are mean ± SEM.
## E1505174-10
as recommended by the manufacturer and then stained during 30 minutes with AF647 anti-FoxP3 antibody (Biolegend) or PE anti-TLR7 antibody (BD Pharmingen).
Staining was assessed with a FACS Fortessa cytometer, and flow cytometry data were analyzed using FACSDiva or FlowJO software. Kaplan-Meier survival curve for overall survival (OS) for the 154 lung adenocarcinoma patients, divided in two groups according to the stratification of TLR7 expression. Optimal cutoff of 82% was previously determined [bib_ref] TLR7 promotes tumor progression, chemotherapy resistance, and poor clinical outcomes in non-small..., Chatterjee [/bib_ref]. Patients are divided into TLR7 hi (> 82% expression of TLR7) and TLR7 low group (< 5% TLR7 expression). p-value was determined by log-rank test (A). Gene expression analysis was performed in 12 adenocarcinoma patients including 6 TLR7 hi and 6 TLR7 low , using nanostring technology. Volcano plot comparison of differentially expressed genes between TLR7 hi and TLR7 low group of patients. The horizontal dimension is the variation of median gene expression (log2 (ratio)) between the 2 groups, and the vertical axis represents the Mann-Whitney test p-value between TLR7 hi and TLR7 low patients. In red are represented genes that are over-or under-expressed in TLR7 hi patients, with a fold-change superior to 2 (11 genes among 770 studied) (B). Hierarchical clustering of the 11 differentially expressed genes in TLR7 hi compared to TLR7 low groups. Individual patients are oriented in columns and expression level for each gene is oriented in rows (C). Box plot representation of ICAM1, KRT-19, KRT-7, SCD4, NME1, and P53 gene expression in TLR7 hi and in TLR7 low patients (D). Kaplan-Meier survival curve for overall survival (OS) for adenocarcinoma patients divided in two groups according to the stratification of ICAM1, KRT-19, KRT-7, SCD4, NME1, and P53 gene expression, using online survival analysis software KM plotter. p value was determined by log-rank test (E).
## Cytokine detection by elisa
Mice were sacrificed 10 days after cancer cell injection and tumors were excised, cut into pieces of 1 mm 3 and placed in RPMI medium supplemented with 1% penicilin/streptomycin and 1% Glutamine for 24h at 37°C. Supernatants were then collected and stored at −80°C.
100000 LL/2 cells were cultured in a 12 well plate for 12 or 24 hours along with TLR7 agonist (2.5 µg/mL CL264), supernatants were collected and stored at −80°C.
Cytokine production was quantified by using GM-CSF ELISA ready-SET-Go! (88-7334-22), mouse CCL2 (MCP-1) ELISA ready-SET-Go! (88-7391-22) and mouse IL-6 ELISA ready-SET-Go! (88-7064-22) from eBiosciences, using the protocols provided by the manufacturer.
## Mdscs sorting and biological activity
Mice were sacrificed 10 days after cancer cell injection and tumors were excised. Single cell suspension was prepared and incubated in blocking solution (24G2, 1mg/ml, 1/100), during 20 minutes, then stained for 30 minutes at 4°C with PECy7 anti-CD45, eF450 anti-CD11b, PerCPCy5.5 anti-Gr1, and LIVE/DEAD® Fixable Yellow Dead Cell Stain Kit. After cell sorting using a FACS Aria cytometer, sorted cells were centrifuged and cultured overnight in RPMI medium with 10% FBS, 2mM L-glutanime, 50U/ml penicilin, 50µg/ml streptomycine, 5% NaPyruvate, 0.1% β-mercapto-ethanol and 10% HEPES. Supernatants were used to perform Nitric oxide assay, (kit KGE001, R&D Systems), H 2 O 2 assay (Kit A22188, Invitrogen, Molecular Probes), and ROS assay (Cell ROX Oxidative Stress Reagent, ThermoFisher Scientific), using protocols provided by the manufacturers.
T cells were purified from spleens using the Pan T cell Isolation Kit II (Miltenyi Biotech). Following centrifugation, purified T cells were labeled with CFSE (10 [bib_ref] Dual roles of TLR7 in the lung cancer microenvironment, Dajon [/bib_ref]
## Immunohistochemistry
Human tumor samples were fixed in neutral-buffered 10% formalin solution and paraffin embedded. TLR7 staining and quantification was performed as previously described 11 using TLR7-specific polyclonal antibody (ENZO Lifesciences) at 10 mg/mL.
The expression of E-cadherin (Dako M3613, 1/100 dilution) and vimentin (Cells Signaling 5741, 1/100) by tumor cells was performed as follows. Serial 5-μm tissue sections were deparaffinized, rehydrated, and pretreated in 10 mM citrate buffer, pH 6, for antigen retrieval. Sections were incubated with hydrogen peroxide for 15 minutes, then blocked in 5% human serum for 30 minutes at room temperature. The slides were then incubated with a primary antibody (diluted in Dako real solution, antibody diluent) anti-E-cadherin (Dako M3613, 1/100 dilution) for one hour at room temperature or anti-vimentin (Cells Signaling 5741, 1/100) overnight. Slides were then incubated for 30 minutes at room temperature with the secondary antibody (anti-rabbit coupled to biotin, 1/200, JIR 715-066-150), followed by 30 minutes incubation with Twenty µm of FFPE sections were obtained from paraffin blocks. Deparaffination was performed by adding 1 mL 100% clearene for 5 min at room temperature. After centrifugation the pellet was washed twice with 1 mL 100% ethanol. Protease digestion was performed by adding digestion buffer and protease form RecoverAll™ Total Nucleic Acid Isolation Kit for FFPE (AM1975), then samples were incubated 3h at 50°C and for 15 min at 80°C for RNA isolation. Gene expression analysis was performed on 100ng RNA, using nCounter XT gene expression assays from nanosting (pancancer progression panel), following the recommendations of the manufacturer.
CRISPR/CAS9 mediated tlr7 gene editing LLC-luc cells were transiently transfected with a plasmid coding for Tlr7-targeting gRNA and Cas9 protein fused to a red fluorescent protein (RFP) (Sigma Aldrich, St. Louis, MO, US) by means of lipofectamin 2000 (Thermo Fisher scientific, Waltham, MA, US) according to the manufacturer recommendation. The following day, RFP expressing cells were selected by single cell sort using a BD FACS Aria Fusion (Becton Dickinson, Franklin Lakes, NJ, US). Following TLR7-deficient clones were then selected.
# Statistical analysis
Statistical analysis for in vivo experiments was performed using GraphPad Prism 5 (GraphPad Software) and using the R software (RFoundation for Statistical Computing; www.r-project. org) and the package survival. Significance threshold for p values was set to 0.05. All mice experiments were analyzed using twoway ANOVA method. Post hoc analyses with Bonferroni corrections were applied for pair-wise comparisons of tumor growth over time in different groups of mice. Tumor volumes at the endpoint of each experiment were compared using a nonparametric Mann-Whitney test with Bonferroni corrections for pairwise comparisons. Mann-Whitney tests were also used to compare percentages and numbers of immune cells infiltrating the tumors in mice. Overall survival (OS) curves were estimated by the Kaplan -Meier method and compared by the log-rank test. The OS was defined from the date of the surgery until the date of death or the last day of the patient's visit to the hospital. Unsupervised hierarchical clustering was performed with Euclidian distance and Ward linkage criterion. Comparisons of the expression of vimentin and E-cadherin in TLR7 hi and TLR7 low patients were performed using a Mann-Whitney test.
## Abbreviations
[fig] Figure 1: Pro-or anti-tumoral effects of different TLRs stimulation. Left panel: TLR4 (A), TLR9 (B) and TLR7 (C) expression by LLC-luc cells and TLR7 expression by LLC-luc cells deficient for TLR7 (obtained by CRISPR/Cas9 technology) (D). Control isotype is shown in blue and stained cells in orange.Right panel: Tumor progression in WT C57Bl/6 mice grafted with LLC-luc cells, after LPS (A) or CpG injection (B). Tumor progression in WT C57Bl/6 or in TLR7 KO mice grafted with LLC-luc cells after CL264 injection (C). Tumor progression in WT C57Bl/6 mice grafted with TLR7-deficient LLC-luc cells (LLC TLR7 KO) after CL264 injection (D).Data are mean ± SEM (5 mice/group). *: p < 0,05, **: p < 0,01, ***: p < 0,001. ns = not significant. Each experiment was repeated three times. [/fig]
[fig] Figure 2: TLR7 stimulation alterates intratumoral immune infiltrates. WT (A-G) or TLR7 KO (H-N) C57BL/6 mice received subcutaneous injection of LLC-luc cells, followed by 4 injections of CL264 (red) or PBS (blue), on days 0, 3, 6 and 9. The percentages and absolute numbers of immune cells were analyzed 10 days after. Radar plot representation for immune cell subpopulations is shown in (A) and (H). Representative flow cytometry image of Gr1 + CD11b + MDSC cells in WT (B) and in TLR7 KO mice (I), having received PBS or CL264. Percentages and absolute numbers of each cell population in WT (C-G) and in TLR7 KO mice (J-N). [/fig]
[fig] Figure 3: MDSCs are involved in the pro-tumorigenic effect mediated by TLR7 stimulation. WT C57BL/6 and TLR7 KO mice received subcutaneous injection of LLC-luc cells, followed by intratumoral injections of CL264 or PBS, and intraperitoneal injections of anti-Gr1 or control antibody every 48 hours (A). Gating strategy for MDSC staining and representative FACS image showing the efficacy of MDSCs depletion in the blood (B). Percentage of MDSCs among total CD45 + cells in the blood is shown for all treated mice (C).Tumor progression in WT C57Bl/6 (D) or in TLR7 KO (G) mice grafted with LLC-luc cells, after isotype or anti-Gr1 antibody injections. Tumor progression in WT C57Bl/6 (E) or in TLR7 KO (H) mice grafted with LLC-luc cells, after CL264 or PBS injection and control isotype antibody injections. Tumor progression in WT C57Bl/6 (F) or in TLR7 KO (I) mice grafted with LLC-luc cells, after CL264 or PBS injection and anti-Gr1 antibody injections. CCL2 (J), GM-CSF (K) and IL-6 (L) titration by ELISA in supernatants from tumor cells from PBS or CL264-treated mice. CCL2 (M) and GM-CSF (N) titration by ELISA in supernatants from in vitro PBS or CL264-stimulated tumor cells.*: p < 0,05, **: p < 0,01, ***: p < 0,0001. Data are mean ± SEM (5 mice/group). ns = not significant. Each experiment was repeated three times. [/fig]
[fig] Figure 4: MDSCs are involved in the pro-metastatic effect of TLR7 stimulation. NOD/SCID (A, C) or C57Bl/6 mice (B, D) received subcutaneous injection of LLC-luc cells, followed by 4 injections of CL264 or PBS, on days 0, 3, 6 and 9. Tumor progression was analyzed by the measurement of D-luciferine bioluminescence (A) or by staining with indian ink and Feket solution at day 25 (B). The number of metastasis per lungs was quantified in NOD/SCID mice treated by PBS or CL264 (3 injections or continuous injections every 72h) (C) and in C57Bl/6 mice having also received control isotype or anti-Gr1 antibody (D). *: p < 0,05, **: p < 0,01, ***: p < 0,0001. Data are mean ± SEM (10-25 mice/group). [/fig]
[fig] Figure 5: TLR7 stimulation of LLC cells leads to epithelio-mesenchymal signature. Gene expression of membrane molecules (A) and transcription factors (B) involved in epithelio-mesenchymal transition, in LLC cells that were stimulated or not with CL264, determined 12 and 24h after stimulation. [/fig]
[fig] Figure 7: High expression of TLR7 by malignant adenocarcinoma cells is associated with metastatic gene expression signature. [/fig]
[table] Table 1: Univariate and multivariate analysis for 154 patients with lung adenocarcinoma not treated with neoadjuvant chemotherapy. minutes with 1x TBS+ 0.04% Tween. Revelation was performed with the DAB kit (Dako, K3468) and stopped by placing slides in 1X TBS and distillated water. Counter coloration was performed with hematoxylin.RNA extraction was performed on LCC-Luc cells stimulated or not with 2.5 µg/mL CL264 using the Maxwell®16 LEV simplyRNA cells kit manufacturer procedure. 200 µg of RNA were used for RT-PCR with High-Capacity cDNA Reverse Transcription Kit with RNase Inhibitor kit from (Applied Biosystems) and the pPCR was performed using Taqman probes (FAM-Applied [/table]
|
Strategies for Interfering With Bacterial Early Stage Biofilms
Biofilm-related bacteria show high resistance to antimicrobial treatments, posing a remarkable challenge to human health. Given bacterial dormancy and high expression of efflux pumps, persistent infections caused by mature biofilms are not easy to treat, thereby driving researchers toward the discovery of many anti-biofilm molecules that can intervene in early stage biofilms formation to inhibit further development and maturity. Compared with mature biofilms, early stage biofilms have fragile structures, vigorous metabolisms, and early attached bacteria are higher susceptibility to antimicrobials. Thus, removing biofilms at the early stage has evident advantages. Many reviews on anti-biofilm compounds that prevent biofilms formation have already been done, but most of them are based on compound classifications to introduce anti-biofilm effects. This review discusses the inhibitory effects of anti-biofilm compounds on early stage biofilms formation from the perspective of the mechanisms of action, including hindering reversible adhesion, reducing extracellular polymeric substances production, interfering in the quorum sensing, and modifying cyclic di-GMP. This information can be exploited further to help researchers in designing new molecules with anti-biofilm activity.
approach and attach to the surface by means of surface appendages such as flagella and pili. However, the mobility of bacteria can easily overcome the bacteria-surface interaction and return to the planktonic state. In the irreversible adhesion stage, initially attached bacteria secrete EPS to bind to the surface covalently, gradually completing the firm attachment. In the early stage, attached bacteria can also aggregate through type IV pili-mediated twitching motilities and use EPSs to adhere surface-associated cells. Middle stage biofilms include the formation of microcolonies and colonies. With the accumulation of bacteria, secreted EPSs increase, gradually forming a layer of water film on the surface of bacteria, and microcolonies are formed. Microcolonies further develop into colonies. The late stage is the maturation and detachment of biofilms. Mature biofilms appear as a 3D network structure with compact structure and coordinated functions. After maturity, biofilms rupture, and bacteria are dispersed to planktonic forms to start a new life cycle of biofilms.
## Regulation systems in biofilms formation
Given that biofilms are the main state of bacteria, the quorum sensing (QS) system and cyclic di-GMP (c-di-GMP) signaling pathway have been shown to play a pivotal role in biofilms formation for various bacterial species, such as Pseudomonas aeruginosa (P. aeruginosa), Escherichia coli (E. coli), Staphylococcus aureus (S. aureus).
QS is a cell-to-cell communication in which bacteria secrete and detect signal molecules. Once a threshold level is reached, specific signals are activated to coordinate pathogenic behaviors, such as biofilms formation, bioluminescence and antibiotic resistance. The connection between QS and biofilms formation has been widely described. In the process of biofilms formation, the QS system regulates bacterial movement, EPSs production, and virulence factor secretion, whereas motility and EPSs play an important role in the initial adhesion stage. Another important regulation system is c-di-GMP signaling pathway. The c-di-GMP is an important second messenger molecule that exists widely in bacteria, since its initial discovery as an allosteric factor regulating cellulose biosynthesis in Gluconacetobacter xylinus, bacterial movement and biofilms formation regulated by the c-di-GMP have grown. Diguanylate cyclases (DGCs) containing the GGDEF domain and phosphodiesterases (PDEs) containing the EAL or HD-GYP domain are, respectively, responsible for the synthesis and degradation of c-di-GMP. Generally, a high c-di-GMP content is favorable for biofilms formation. In the reversible adhesion stage, c-di-GMP inhibits the activity of ATPase. Thus, FleQ cannot regulate downstream flagellar genes, thereby preventing the synthesis of flagella. This inhibitory effect can eliminate surface dynamics to make bacteria stick to the surface stably, and promote irreversible transformation. In the irreversible adhesion stage, c-di-GMP promotes EPS production. For example, in P. aeruginosa, the combination of c-di-GMP with FleQ relieves the inhibitory effect of FleQ on the pel promoter and activates transcription to produce pel polysaccharides. In E. coli, c-di-GMP binds to the PgaCD complex, which has polysaccharide synthesis and transport activity, thereby activating glycosyltransferase activity and promoting the N-acetyl glucosamine (GlcNAc) production. This facilitation can enhance bacterial adhesion and further contribute to early stage biofilms formation.
## Rationales for targeting the early stage biofilms
During biofilms formation, bacteria exhibit a dynamic change in drug sensitivity. In the early stage of biofilm formation , bacteria switch from the free-swimming phase to the surface-attached phase. The cell-surface interaction triggers some mechanically sensitive channels on cell wall opening, making the bacteria susceptible to antimicrobials. At the same time, attached cells get close to each other to initiate active cell-cell interaction. During this process, cells divide, grow, and become increasingly sensitive to antimicrobials. However, in the middle and late stages, high expression of efflux pumps, horizontal gene transfer, and EPS matrix protection can significantly reduce the antimicrobial susceptibility. Surface-attached cells can be more sensitive to antibiotics than planktonic cells during early stage, and the 3D structures of early stage biofilms have not been formed and are characterized by fragile structures. Removing biofilms at this time has evident advantages compared with removing at the mature stage. Inhibiting the formation of early stage biofilms can prevent their development into mature biofilms, thereby reducing biofilm-related infections.
Reversible adhesion, EPSs production, QS system, and c-di-GMP levels are important factors affecting early stage biofilms development. Therefore, reducing motility and EPSs production, and inhibiting QS system and c-di-GMP pathways are effective methods to inhibit early stage biofilms formation. Biofilms prevention strategies have been studied, however, many of these strategies are based on compounds classifications to discuss the anti-biofilm effects, such as antibiotics, phytochemicals, and antimicrobial peptides. Here, we make an effort to review the inhibitory effects of anti-biofilm molecules on early stage biofilms formation from the perspective of the mechanisms of action, providing help to researchers in designing new molecules with anti-biofilm activity. At the same time, we state the disadvantages of each solution in.
## Designing antibacterial surfaces
Applying hydrophilic coatings on the surface is a common method to reduce bacterial adhesion in the initial stage . Based on the repulsive effect between different charges, the surface of bacteria has hydrophobic properties, and applying hydrophilic coatings on the surface can reduce bacterial adhesion and prevent early stage biofilms formation. Commonly used hydrophilic coatings include hyaluronic acid, polyethylene glycol (PEG), and 2-methacryloyloxy ethyl phosphorylcholine (MPC). MPC contains a phospholipid polar group similar to the cell membrane structure, which can effectively reduce protein adsorption and bacterial adhesion, thereby preventing related infections. Researchers used MPC polymers for oral bacterial biofilm infections and found that MPC polymers reduced the adhesion of streptococci on oral epithelial cells by 15-30%.later confirmed through in vivo and in vitro experiments that applying the MPC-coated polymer to the surface of bioabsorbable suture could significantly inhibit the adhesion of S. aureus and reduce surgical site infections. The anti-adhesion activity of PEG is related to its molecular weight. The amount of bacterial adhesion decreases with increased molecular weight in the range of 300-20,000, and when the molecular weight is 5,000, the adhesion amount reaches the minimum. In addition to hydrophilic coatings, applying antibacterial agents on the surface to kill attached bacteria directly is an effective method to inhibit reversible adhesion. Silver ion has broad-spectrum bactericidal activity and is a commonly used antibacterial coating.found that a room temperature processed silver doped perfluoropolyetherurethane coating effectively reduced the adhesion of S. aureus on the catheter surface and prevented the formation of biofilms. Silver nanoparticles (Ag NPs) have the unique properties of nanomaterials and the antibacterial properties of silver, making them an ideal material. Researchers revealed that biosynthesized Ag NPs could reduce the adhesion of Klebsiella pneumoniae by 88%. The antimicrobial peptide is also considered as an effective anti-biofilm coating, which shows better effect than Ag NPs in inhibiting S. aureus adhesion .
## Affecting pili
Pili are important surface appendages that enable bacteria to realize movement and adhere to the surfaces. Thus, pili are one of the targets for removing early stage biofilms. Three types of pili are studied: (1) Type I pili, which are best studied in E. coli, are a functional amyloid primarily responsible for cells' attachment to the surface. (2) Type IV pili are the motility appendage of bacteria that mediate a kind of movement independent of flagella (i.e., twitching motility). The twitching movement plays a critical role in initial adhesion and the development of microcolonies. (3) Curli are the major extracellular surface amyloid fibers that contribute to the initial attachment to a surface and cell-to-cell contact. The drugs that act on type I pili, type IV pili, and curli are introduced in the following section.
The disperse red 15 is a hydroxyl anthraquinone compound that can interfere with FimH, an adhesion factor at the end of the type I pili of E. coli. After the action of disperse red 15, swimming and swarming motilities were inhibited, and the initial adhesion was reduced, resulting in a reduction in biofilm biomass by more than 50%. This result indicated that disperse red 15 could better be used as surface coating to reduce bacterial adhesion and prevent biofilms formation.
Magnesium ions are essential nutrients for the formation of biofilms, which can affect bacterial adhesion and biofilm formation through direct and indirect means. However, the promotion of magnesium ions on adhesion is related to its concentration. Researchers found that 0.1 and 0.5 M magnesium ions could inhibit the expression of type IV pili synthesis genes pilV and pilW, thereby reducing type IV pili formation, weakening the initial adhesion of bacteria and inhibiting early stage biofilms formation.
Ethylene Diamine Tetraacetic Acid (EDTA) is a metal chelating agent. Researchers found that EDTA has the strongest inhibitory effect in the early stage of biofilms formation, but has no biofilm inhibitory effect in the middle and late stages FIGURE 1 | An alternative approach to reduced reversible adhesion: applying hydrophilic coatings. (A) Without applying hydrophilic coatings, most of the planktonic bacteria can adhere to the surface, and only a small part remains planktonic. (B) After using hydrophilic coatings, due to the repulsion of different charges, only a small part of planktonic bacteria can complete adhesion, while most bacteria are still in planktonic states.. EDTA plays an inhibitory role in the early stage by reducing the production of curli to inhibit cell-surface and cell-cell interactions. When combined with ciprofloxacin, gentamicin, and ampicillin, EDTA could reduce the minimal inhibitory concentration of antibiotics and EDTA plus ciprofloxacin treatment had a better therapeutic effect on P. aeruginosa biofilms compared with the single treatment.
Some bacteria also have type II and type III pili in addition to the above-mentioned pili. Type II pili are related to the pathogenic heterogeneity of Porphyromonas gingivalis. Type III pili are an important factor in biofilms formation by Klebsiella pneumoniae, which may be because the pili make the surface of K. pneumoniae hydrophobic, thereby promoting bacterial adhesion. The inhibitors of these two pili are still under investigation.
## Affecting flagella
The role of flagellar motility and surface adhesion in biofilms formation has been reported in many studies. Flagella are believed to promote bacterial adhesion:
(1) Flagella-mediated motility makes planktonic bacteria move toward target sites, promoting subsequent adherence. (2) Flagella-mediated motility allows bacteria to overcome surface tension and reach the surface. (3) Flagella act as bacterial adhesins to help bacteria attach to the surface. Therefore, the strategy with flagella as target is also an effective method to inhibit early stage biofilms formation.
Agaric acid is a fatty acid naturally produced by certain fungi, such as Polyporus, and can be used as anhidrotic to treat tuberculosis patients with extreme sweating. Studies on Salmonella showed that the formation of biofilms was significantly inhibited after the action of agaric acid because it significantly reduced the expression of flagellar rotation genes. These phenomena inhibit the swimming ability and make bacteria stay in the planktonic stage, thereby preventing initial attachment and reducing early stage biofilms formation.
## Inhibiting eps production
Extracellular polymeric substances mediate the irreversible adhesion stage. EPSs allow the long-term attachment of bacteria to surfaces and promote cohesion among bacterial cells, leading to the development of early stage biofilms. Therefore, inhibiting EPS production to reduce irreversible adhesion has become an important target. According to matrix components, extracellular anti-EPS molecules fall into three categories: (1) molecules acting on exopolysaccharides, (2) molecules acting on proteins, and (3) molecules acting on eDNA. These molecules destroy the structural integrity of EPS, facilitating the passage of attached bacteria into their vulnerable, planktonic state, and reduce the early stage biofilms formation. The corresponding drugs are described below.
## Reducing exopolysaccharides production
Among the components of EPS, the exopolysaccharides are the most abundant and the most thoroughly studied Frontiers in Microbiology | www.frontiersin.org component. Different bacteria have different compositions of exopolysaccharides, which are all considered as fundamental matrix for the adhesion and architectural stability of EPS. As molecular glue, the exopolysaccharides allow the initial adhesion of planktonic bacteria on the surface and can fix the bacterial population to increase cell density, becoming an important target in the early stage biofilms removal strategy. At present, most drugs exert their early inhibitory effects by affecting exopolysaccharides and can be divided into chemical drugs, natural components, and probiotics.
## Classic antibiotics
Ambroxol (ABX) is one of the most widely used expectorant drugs. Previous studies showed that ABX could inhibit biofilms at various stages when used aloneand could enhance the killing effect of Staphylococcus epidermidis biofilms in combination with vancomycin . It is presumed that ABX reduces the synthesis of exopolysaccharides and prevents bacterial adhesion to achieve its effect, which was later confirmed in P. aeruginosa. After the addition of ABX, the content of alginate decreased. The expression of algD gene encoding the ratelimiting enzyme GMD for alginate synthesis was down-regulated. algR and algU genes, which promote synthesis, were also down-regulated whereas the mucA gene, which inhibits algD expression, was increased. These results showed that ABX influences early stage biofilms by inhibiting the synthesis of exopolysaccharides.
Antibiotics are the most used drug to treat biofilm-related infections. In this article, we take ceftazidime as an example. Ceftazidime is a β-lactam antibiotic that binds to penicillinbinding proteins to affect cell wall synthesis and reduce bacterial growth. Ceftazidime could inhibit the formation of biofilms of Proteus mirabilis and E. coli. In studies about P. aeruginosa, ceftazidime significantly down-regulated the expression of exopolysaccharides Pel-and Psl-related genes pelA and pslA to impair the structural integrity of EPS, leading to decreased adhesion and inhibited biofilms formation. In addition, the twitching motility was suppressed. The twitching motility is mediated by type IV pili and is associated with adhesion. Ceftazidime decreased the mRNA level of lecB gene, which is necessary for motility, but had no effect on the transcription of the pil that encodes pilus components, thus weakening the expression of protein PilHIJK in twitching-related regulatory pathways, restricting twitching capacity, and affecting adhesion.
## Plant-derived components
Biofilm-growing bacteria are known to show 10 to 1,000time-increased resistance to antibiotics than planktonic microorganisms. The low efficiency of conventional doses of antibiotics and in vivo toxicity of high doses drive researchers toward the discovery of plantderived anti-biofilm molecules. Compared with antibiotics, plant-derived anti-biofilm molecules have the advantages of wide sources, high safety, and low drug resistance, having great potential in controlling biofilms formation. Extracts of Hymenocallis littoralis leaves and tannic acid, the component of Rhus chinensis Mill, could inhibit the formation of S. aureus and P. aeruginosa biofilms, respectively. Some plant components have been found to inhibit biofilms by inhibiting the initial adhesion of bacteria.
Here are some examples.
Thymol is naturally found in the seeds of thyme and oregano, and has a broad-spectrum antimicrobial effect. Studies confirmed that thymol can inhibit the biofilms formation of Candida tropicalis and Enterococcus faecalis. In studies about P. aeruginosa and Candida albicans, thymol is speculated to affect early stage biofilms by inhibiting adhesion. Thymol can significantly reduce the expression of the pga gene that codes the PGA exopolysaccharide , the main component of the Actinobacillus pleuropneumoniae biofilm matrix. This phenomenon causes changes in EPS structural components and weakens the irreversible adhesion of bacteria, which is easy to fall of the surface, thereby reducing the early stage biofilms formation. A study about Enterobacter cloacae showed that thymol could also inhibit the curli fimbriae csgABCEFG to affect initial adhesion in addition to affecting exopolysaccharide synthesis genes.
Artesunate (AS), a water-soluble derivative of Artemisinin, is used in the treatment of malaria and possesses anti-cancer and anti-inflammatory activities. In recent years, the combination of AS and tested antibiotics can significantly enhance the antibacterial activity of antibiotics. In the study with P. aeruginosa,found that using AS alone could reduce the production of exopolysaccharides, thus playing an inhibitory role in the early stage of biofilms formation. In the experiment, a concentration that had no effect on planktonic bacteria was used. In the early stage biofilms formation process, the Fenton reaction between AS and Fe(II) produces reactive oxygen species and carbon-centered radicals. No effect on bacteria is observed, but the corresponding decrease in Fe(II) increases the synthesis of rhamnolipids, competitively inhibits the precursors of polysaccharides, and reduces the production of Psl. The reduction in Psl has a negative effect on irreversible adhesion, thus inhibiting the formation of early stage biofilms. Twitching motility can promote the formation of microcolonies, which is beneficial to biofilms formation. However, excessive twitching movement has a negative effect on the formation of biofilms. Experimental results showed that AS promoted the twitching motility of P. aeruginosa PAO1 in a concentration-dependent manner, resulting in the difficult aggregation and attachment of bacteria and inhibited the formation of biofilms.
Phenyllactic acid (PLA) is an organic acid that exists in many plants and food products, such as honey and lactic acid bacteriafermented food, and has a broad-spectrum antibacterial effect. The researched area found that PLA can also inhibit the formation of bacterial biofilms, such as Listeria monocytogenes (L. monocytogenes) and E. cloacae . In the study about E. faecalis, it was found that the number of bacteria adhering to polyethylene plate and stainless-steel plates decreased significantly after PLA treatment, and that the ability of adhesive bacteria to produce EPSs was also inhibited. Reverse-transcription polymerase chain reaction confirmed that PLA significantly inhibited the expression of EPA polysaccharide synthesis genes, suggesting that PLA could reduce the bacterial adhesion rate and inhibit early stage biofilms development by reducing extracellular polysaccharide synthesis.
## Probiotics
Lactobacillus is one of the probiotic species and can play a potent role in promoting body development, stimulating the immune system, and has an inhibitory effect on the formation of P. aeruginosa and Salmonella biofilms. In the study with S. aureus, it was observed that treatment with Lactobacillus significantly increased the expression of icaR and decreased the expression of icaA, leading to the inhibition of the synthesis of PIA and affecting the adhesion between bacteria and early stage biofilms formation. Lactobacillus salivarius and Lactobacillus kefiranofaciens can inhibit the expression levels of gtfBCD and ftf, respectively, thereby reducing the synthesis of exopolysaccharides, hindering initial adhesion, and inhibiting early stage biofilms formation of Streptococcus mutans (S. mutans).
## Reducing extracellular proteins production
In general, only proteins that play a structural role have been clearly characterized, and matrix proteins are poorly defined and understood. However, matrix proteins have been suggested to promote bacterial adhesion and early stage biofilms accumulation (Matsumoto-Nakano, 2018). For example, in S. aureus, the biofilm-associated protein (BAP) promotes initial attachment and cell-to-cell interaction through mechanisms independent of extracellular polysaccharides. When BAP is inhibited, the abilities of primary attachment and biofilms formation are significantly reduced. Therefore, reducing the production of extracellular proteins is one of the effective methods to control early stage biofilms formation, and D-amino acids and rhamnolipid are believed to exert biofilms inhibition by affecting extracellular proteins.
## D-amino acids
D-amino acids, also called unnatural amino acids, are enantiomers with L-amino acids and naturally exist in organisms. Researchers first demonstrated that 6 mg/L D-tyr could remove biofilms from nylon membrane surfaces. Subsequent studies showed that D-amino acids have an inhibitory effect on the formation of biofilms. For example, a mixture of D-leu and D-try can inhibit the formation of S. aureus and P. aeruginosa biofilms, and a mixture of D-pro and D-ala can inhibit the formation of S. epidermidis biofilms.
At experimental concentrations, the addition of exogenous D-amino acids has no effect on bacterial growth, but can significantly inhibit the formation of biofilms, suggesting that the inhibition of bacterial growth is not the mechanism by which D-amino acids inhibit the formation of biofilms. D-amino acids inhibit biofilms development by reducing initial adhesion.
In P. aeruginosa and S. aureus, the production of extracellular proteins was significantly reduced after D-amino acids treatment, but had no effect on exopolysaccharides. D-amino acids are substituted for L-amino acids in a protein, and the protein undergoes changes in structure, making the protein minimally secreted. This finding suggests that the change in extracellular proteins may be the cause of D-amino acids inhibition of adhesion. Thermodynamic research reported that with the addition of D-tyr, Lewis acid-base interactions are more affected than the overall non-specific interactions. The Lewis acid-base force is closely related to hydrophilicity, and the change in EPS has a negative effect on the water contact angle. The decrease in the protein/polysaccharide ratio is likely to be the cause of increased hydrophilicity, which further indicates that D-amino acids inhibit adhesion and biofilms formation by reducing extracellular protein synthesis.
Peptidoglycan (PG), the major component of the cell wall, is composed of repeating units of GlcNAc and N-acetylmuramic acid (MurNAc). A short peptide chain terminating in D-alanine is attached to the lactoyl group of MurNAc, and researchers revealed that exogenous D-amino acids could replace D-alanine to integrate into PG, changing the composition and architecture of PG. The modification of PG affects the cell wall, and the cell wall is one of the important factors mediating initial adhesion, indicating that D-amino acids may inhibit initial adhesion by affecting the cell wall.
## Rhamnolipids
In recent years, rhamnolipids derived from P. aeruginosa have emerged as an important group of biosurfactants with several applications, and are considered a substitute for traditional antibacterial agents. Current studies found that rhamnolipids disrupted many bacterial biofilms, and the exogenous addition of rhamnolipids could impede initial adhesion. The study on P. aeruginosa biofilm showed that after rhamnolipids treatment, the reduction in protein concentrations was more than twice the reduction in carbohydrate concentrations. This result was similar to the later finding that strains with high protein content in the biofilm matrix were more susceptible to surfactants than strains with carbohydrate-based matrix, indicating that rhamnolipids can have a selective interaction of protein to reduce protein contents and affect the early adhesion of bacteria. Rhamnolipids can affect the swarming motility in addition to reducing protein. The swarming motility is a cell density-dependent, flagella-driven process that enables bacteria to move toward favorable environments and adhere. Studies demonstrated that when rhamnolipids were absent, bacteria lost their ability to swarm, which can be supplemented by providing exogenous rhamnolipids. Therefore, reducing the rhamnolipids content in the bacterial population to hinder swarming motility is also an effective method to inhibit early adhesion. Adding rhamnolipids reduces the protein content but promotes swarming motility. Thus, an in-depth study is needed.
## Others
Proteases, which specifically hydrolyze extracellular proteins, may have great ability to control initial adhesion. When bappositive S. aureus were treated with proteinase K, the number of bacteria attached to the microtiter plate at the primary stage was significantly reduced (Kumar, indicating that exogenous proteases destroyed the physical stability of the EPS and hampered early adhesion. Later studies showed that protease K can hydrolyze extracellular proteins and reduce the eDNA content. Papain and trypsin can digest the fimbrial proteins FimP and FimA. eDNA and fimbriae also play an important role in initial adhesion, which further suggests that proteases can be an antiadhesive agent to affect the early stage biofilms formation.
Curcumin is a hydrophobic polyphenolic substance extracted from the dried rhizomes of Turmeric, a plant of the genus Curcuma, and is recognized as a safe food additive. In addition to being a natural food coloring, curcumin is found as an inhibitor of sortase A activity in S. aureus. Curcumin can inhibit surface protein production or separate protein from cell walls, thereby reducing the initial adhesion and inhibiting early stage biofilms formation.
## Reducing edna production
The content of eDNA is significantly less than those of proteins and exopolysaccharides, and related studies are few. However, eDNA is generally believed to promote initial bacterial adhesion and surface aggregation. eDNA promotes the adhesion of bacteria to biotic and abiotic surfaces, and its essence may have the following points. First, irreversible adhesion is mediated by Lewis acid-base forces, and thermodynamic analysis showed that the presence of eDNA promotes the acid-base interaction, which is more conducive to the attachment of bacteria on the surface, and has a remarkable promotion effect on the formation of biofilms. Second, eDNA does not exist independently in EPSs and interacts with other biomolecules in the biofilm matrix . Maintaining the integrity of the EPS structure and facilitating the adhesion between bacteria and surface are important. Researchers found that the removal of eDNA also results in the reduction of exopolysaccharides and proteins. Third, the presence of eDNA widens the adhesion distance, facilitating bacteria to begin irreversible adhesion at a large distance scale , which remarkably promotes the formation of early stage biofilms. Therefore, reducing eDNA production is also an effective method to inhibit initial adhesion, and current studies confirmed that DNase and reserpine exert early stage biofilms inhibition by affecting eDNA.
## Dnase
At present, DNase is the most in-depth studied strategy of removing eDNA, and can be used to decompose eDNA to reduce bacterial adhesion and control the formation of early stage biofilms. Studies showed that after treating L. monocytogenes, Burkholderia pseudomalleiand S. aureuswith DNase, the bacteria attached to the surface were significantly reduced, and the formation of biofilms were inhibited. The first functional DNase I coating has reduced bacterial adhesion by 95-99%. However, DNase can easily remove young biofilms and partially inhibit middle-stage biofilms, but has no effect on mature biofilms.
## Others
Reserpine, a drug used to treat hypertension and psychosis, has been recently found to inhibit the formation of S. aureus biofilms because it has the same active site as the AtlE protein ligand. Reserpine competitively inhibits the binding of ligands to AtlE protein, thereby reducing the synthesis and secretion of eDNA, resulting in decreased adhesion and inhibiting biofilms formation.
Totarol, a diterpene compound isolated from the totarol tree, has remarkable antibacterial and antioxidant activities. In the study with S. aureus biofilms, the expression levels of the icaA gene encoding exopolysaccharide synthesis and the cidA gene encoding eDNA synthesis decreased significantly after the action of totarol, and the expression amount of cidA decreased more, indicating that totarol reduces the synthesis of eDNA to inhibit early adhesion.
Researchers found that a biosurfactant can significantly downregulate the expression of dltB and cidA genes in S. aureus and that the dltB operon promotes bacterial adhesion . Among the two adhesion-related genes, cidA expression is more significantly affected than dltB expression, suggesting that this biosurfactant reduces adhesion by inhibiting eDNA.
## Inhibition of the qs system
The QS system confers bacteria as a community to approach the surface in a consistent manner and increase the secretion of EPS, making the attachment firmly, which is beneficial to the formation of early stage biofilms. According to the chemical types of signal molecules, the QS system can be roughly divided into three types: (1) N-acyl-homoserines (AHLs)mediated QS system, which exists in Gram-negative bacteria;
(2) auto-inducing peptide (AIP)-mediated QS system, which exists in Gram-positive bacteria; and (3) autoinducer 2 (AI-2)-mediated QS system, which exists in Gram-negative and Gram-positive bacteria. Studies found a QS system mediated by the Pseudomonas Quinolone Signal (PQS) in P. aeruginosa .
Given the important role of QS in early stage biofilms development, the mode targeting at QS has also become an alternative approach to combat undesirable microorganisms. By acting on different steps of the signaling cascade, QS inhibitors can be divided into three categories, (1) blocking the generation of signal molecules, (2) preventing the signal molecules from binding to the corresponding receptors, and (3) degrading the generated signal molecules . Different QS inhibitors are described as follows.
## Blocking the generation of signal molecules
Coumarins, a group of heterocyclic compounds present in many natural plants, have a wide range of biological activities such as antiviral and antioxidation properties, and can be used as QS inhibitors to affect the formation of biofilms such as E. coli, P. aeruginosa, and S. aureus. The study about P. aeruginosa found that the expression levels of lasI, rhlI, and pqsBCH that encoding signal molecules synthesis was significantly downregulated after the action of 2 mM coumarin, whereas those of receptor genes lasR and pqsR remained unchanged. As a result, the syntheses of AHL molecules and PQS was hindered, affecting the QS system and reducing the formation of biofilms. Eugenol has similar mechanism of action with coumarins. After treatment with eugenol and its nanoemulsion at a concentration of 0.2 mg/mL, the expression levels of QS synthase genes lasI and rhlI were inhibited. The inhibition effect led to a decrease in the production of 3-oxo-C12-HSL and C4-HSL, which affected swarming motility and biofilms formation of P. aeruginosa.
Salicylic acid, the bulk drug used to synthesize aspirin, is also a strong inhibitor of AHL and interferes with the production of AHL by binding to the substrates needed for its production. Trans-cinnamaldehyde interacts with the substrate binding site of LasI to inhibit AHL production. Triclosan reduces the production of AHL and inhibits the QS system by inhibiting the enoyl-acyl carrier protein reductase which is an essential intermediate in AHL biosynthesis.
## Preventing signaling molecules from binding to corresponding receptors
Furan compounds are originally discovered in the red marine alga Delisea pulchra. Given their similarity in structure with AHLs, they can replace AHLs to bind with corresponding LasR receptors and inhibit the QS system. This competitive inhibitory effect is dependent on dose. Thus, the side effects of high doses should be considered when developing furan compounds as QS inhibitors. (5Z)-4-bromo-5-(bromomethylene)-3-butyl-2(5H)-furanone has been shown to act as antagonist of AHL and AI-2 QS signals in many organisms. At concentrations of 2.0 and 4.0 µg/mL, furanone C-30 could inhibit exopolysaccharide genes ftf and gtfB and extracellular protein gene gbpB in the biofilm of S. mutans without affecting the growth of planktonic bacteria, thereby reducing the initial adhesion and inhibiting the formation of biofilms. Chlorogenic acid (CA), flavonoids, and cinnamaldehyde have similar mechanisms of action with furan compounds. After the action of CA, the production of rhamnolipid was inhibited , further affecting the early stage biofilms formation of P. aeruginosa.
In addition, the apolipoprotein B can prevent AIP1 from binding to the receptor and inhibit the QS system of methicillinresistant S. aureus.
## Degrading generated signal molecules
In the QS system, it is critical for AHL molecules to reach a threshold concentration. Therefore, if AHLs can be degraded FIGURE 3 | An overview of QS system (A) and the three main strategies to combat it. (B) Blocking the generation of AI. The inhibition of some of synthase can affect the production of more than one signal. (C) Preventing AI from binding to the corresponding receptors. The use of AI analogs can competitively inhibit the binding of AI to the receptor, and the analogs themselves have no pharmacological activity, so they cannot activate the corresponding gene expression and the QS mechanism is ceased. (D) Degrading the generated AI. Degrading the AI with the corresponding enzyme makes it unable to reach the threshold concentration and the degradation products cannot bind to the receptor, inhibiting the QS system. to make it unable to reach the threshold concentration, the formation of biofilms is remarkably weakened. According to the chemical structure of AHLs, three action sites, namely, lactonase site, acyltransferase site, and oxidoreductase site, can decompose AHLs. The degradation of AHL is largely regulated by AHL acyltransferase and AHL lactonase, which hydrolyze the bond and HSL ring, respectively, and degrade AHLs to inhibit the QS system.
## Targeting c-di-gmp signaling
The c-di-GMP signaling plays critical roles during the biofilms formation. Flagellar motility and EPS secretion, which are partly regulated by c-di-GMP signaling and facilitate early stage biofilms development. Thus, blocking c-di-GMP signal transduction is important to inhibit the early stage biofilms. The c-di-GMP-mediated signaling pathway consists of four parts: (1) DGCs that promote the synthesis of c-di-GMP, (2) effector molecules combined with c-di-GMP, (3) downstream targets regulated by effector molecules, and (4) PDEs that degrade c-di-GMP. Among them, the synthesis and degradation of c-di-GMP by DGCs and PDEs are the core of the whole signaling pathway . The following describes the corresponding methods based on these two targets.
## Targeting dgcs
Controlling endogenous synthesis of c-di-GMP is more efficient than directly acting on the downstream targets of c-di-GMP. A glycosylated triterpenoid saponin purified from extracts of garden pea is identified as the first effective inhibitor of benzamide are subsequently demonstrated as inhibitors. Considering that the catalytic domain of DGCs is similar to the catalytic domain of mammalian adenylyl cyclases, and 2 (3 )-O-(N-methylanthraniloyl) (MANT)-and 2 ,3 -O-(2,4,6-trinitrophenyl) (TNP)-substituted nucleotides are potent adenylyl and guanylyl cyclase inhibitors, researchers tested various MANT-and TNP-substituted nucleotides and found that MANT-GTP, MANT-GTPγS, and TNP-GTP acted as potent inhibitors against E. coli DGCs.
## Targeting pdes
Stimulating the activity of PDE keeps the intracellular c-di-GMP at a low concentration, which has a negative effect on biofilms formation. Endogenously produced nitric oxide (NO) is a signaling molecule that reduces the concentration of c-di-GMP in the population by stimulating the activity of PDE. In clinical trials, the inhalation of a low dose of NO (10 ppm) significantly reduced P. aeruginosa biofilm aggregation in cystic fibrosis patients. The application of NO-releasing chitosan dressings to FIGURE 4 | C-di-GMP signaling pathway and (A), (B) are the main inhibitory modes. After the environmental signals are combined with the receptor on the membrane, GTP is produced in the cell. GTP is catalyzed by the GGDEF domain in DGC to form c-di-GMP, and produced c-di-GMP binds to the intracellular receptor to activate the corresponding gene expression and regulate flagellar movement and EPS production; PDE also exists in the cell, and the EAL domain and HD-GYP domain can, respectively, hydrolyze c-di-GMP into pGpG and GMP to inactivate it. (A) Inhibiting the activity of DGC prevents GTP from producing c-di-GMP, so that downstream genes cannot be regulated. (B) Stimulating the activity of PDE to decrease the intracellular c-di-GMP levels that is negatively affects the formation of early stage biofilms.
S. aureus-infected wounds in mice could enhance biofilm dispersal. Here, we emphasize that the low concentration of NO can trigger biofilm dispersal because a high concentration can promote the biofilm formation of some bacteria, such as P. aeruginosa and Nitrosomonas europaea. This phenomenon may be because a high concentration of NO causes anaerobic environments, and forming biofilms in anaerobic environments can be used as a stress defense mechanism. Therefore, the critical concentration should be determined before using NO as an antibacterial agent.
# Conclusion
Researches about biofilms has basically begun in the 1970s. Once the biofilms mature, they are difficult to eradicate due to the EPS barrier function and the efflux pump system, causing serious harm. Thus, intervening in the early stage of biofilms formation can be a solution. The inhibition of early stage biofilms formation can be accomplished by affecting the adhesion of bacteria. Reducing bacterial adhesion can be divided into two points. One is to prevent bacteria from approaching the surface, which is mainly by changing the surface properties or reducing the movement of bacteria, and the other is to reduce the production of EPSs.
Researches on the inhibition of biofilm formation based on hydrophilic coatings changing the properties of the surface and antimicrobial coatings killing the attached bacteria have been relatively mature, but these methods are suitable for biofilms formation of related equipment and implants. Using molecular drugs to inhibit early stage biofilms formation from internal mechanisms is a research hotspot. The movement ability and EPS secretion of bacteria are important factors that affect the formation of early stage biofilms. At the same time, the QS system and c-di-GMP signaling pathway regulate these two behaviors, so early stage biofilms inhibitors can be divided into: (1) reducing bacterial movement ability, (2) inhibiting EPS production, (3) interfering with QS system, and (4) destroying c-di-GMP signal pathway. Inhibitors of the same type can also exert corresponding effects through different pathways, for example, with the same extracellular protein inhibitors, D-amino acids can reduce protein production, rhamnolipids can bind to the protein, and proteases can hydrolyze the protein. Early stage biofilms are fragile, have no 3D structures yet, and are relatively easy to be removed. Early stage biofilms treatments combined with antibiotics have the potential to achieve better therapeutic effects compared to monotherapy, as exemplified byand. Thus, early intervention is necessary.
To date, efforts have already been made in the biofilms field, but current researches on the early inhibition of biofilms have not achieved effective breakthroughs due to two main reasons. First, current techniques for detecting early stage biofilms infections are immature. When infections caused by biofilms are diagnosed, they often develop to the mature biofilm, and this presents a substantial problem in having the opportunity to treat infections at the early stage biofilm progression in the first place. Second, clinical studies are lacking. The differences between in vitro and in vivo experiments are large, and current in vitro assays cannot effectively guarantee the outcomes in clinical practices.
In vitro experiments can only preliminarily prove that drugs have therapeutic effects. When the body is infected with biofilms, it will undergo complex changes, which cannot be effectively simulated by in vitro experiments. These phenomena hinder the progress toward the further application for some types of molecules. Therefore, future researches should focus on the establishment of early stage biofilms infection models, in vivo experiments, and improvement of detection technology.
# Author contributions
JF drafted the manuscript. YZ and SL were responsible for preparing the tables and figures in the manuscript. WZ, JL, HL, FX, HT, GP, LZ, and SC reviewed the manuscript. HF conceptualized the content of the manuscript. All authors contributed to the article and approved the submitted version. |
Subsets and Splits